technologies for detecting heritable mutations in human beings

8/14/2019 Technologies for Detecting Heritable Mutations in Human Beings

http://slidepdf.com/reader/full/technologies-for-detecting-heritable-mutations-in-human-beings 1/148

Technologies for Detecting Heritable Mutations in Human Beings

September 1986

NTIS order #PB87-140158



Recommended Citation:

U.S. Congress, Office of Technology Assessment, Technologies for Detecting Heritable Mutations in Human Beings, OTA-H-298 (Washington, DC: U.S. Government PrintingOffice, September 1986).

Library of Congress Catalog Card Number 86-600523

For sale by the Sup erintendent of Docum entsU.S. Government Printing Office, Washington, DC 20402



Foreword

Ensuring the health of futur e generations of children is of obvious importance toAmerican society. Heritable m utations, perman ent changes in the genetic material thatcan be pa ssed on to succeeding generations, are the cause of a large but currently un-qua ntifiable share of embryonic and fetal loss, disease, disability, and early death inthe United States today. The methods now available to study heritable mutations,how ever, offer relatively little information abou t the kind s of mutations that can oc-cur, their frequency, or their cau ses. Recent adva nces in molecular genetics have open edthe door to n ew a nd innovative technologies that m ay offer a great d eal more informa-tion about DNA. It may soon be possible to characterize mutations precisely, to meas-ure th eir frequency, and perhap s also to associate particular m utations with exposuresto specific mutagenic influences. While some of the new technologies are still on th edra wing board , they are developing quickly and several of them m ay become availablefor wid e-scale u se in the next 5 to 10 years.

The Senate Committee on Veterans’ Affairs, the House Committee on Science andTechnology, and the Hou se Committee on Energy and Comm erce requested that O TA

assess the av ailable information abou t current an d prop osed m eans for d etecting heritablemu tations and on the likelihood and potential imp act of such technological advances.These comm ittees have wr estled w ith the problems of determining wh ether past exposuresto poten tial mutagens have affected th e health of Americans, and of framing reason-able public health laws, given current know ledge and technologies. This report sum -marizes OTA’s findings as they relate to th ese issues.

An ad visory panel, chaired by Arno G. Motulsky, provid ed gu idance and assis-tance during the assessment. The OTA Health Program Advisory Committee, OTAstaff, and scientific and policy experts from the private sector, academia, and the Fed-eral Government provided information during the assessment and reviewed drafts of the report . We thank all who assisted u s. As with all OTA reports, the content of theassessmen t is the sole responsibility of OTA and does n ot necessarily constitute th e con-

sensus or end orsement of the ad visory panel or the Technology Assessment Board. KeyOTA staff involved in the assessm ent w ere Michael Gough, Julie Ostrow sky, and H ellenGelband,

JOHN H. GIBBONS Director

I l l



Advisory Panel for Technologies for DetectingHeritable Mutations in Human Beings

Arno G. Motulsky, Panel Chair

Center for Inherited DiseasesUniversity of Washington School of Medicine

Richard J. AlbertiniDepartment of MedicineUniversity of Vermon tCollege of Medicine

Michael S. BaramSchool of Medicine and Public HealthBoston University

Charles R. CantorDepartment of Human Genetics and

DevelopmentColumbia University College of Physicians

and Surgeons

Dale Hattis

Center for Policy AlternativesMassachusetts Institute of Technology

Ernest B. Hook Bureau of Maternal and Child HealthNew York State Department of Health

Alfred G. Knudson, Jr.Institute for Cancer ResearchFox Chase Cancer Center

Nan M. LairdDepartm ent of Biological StatisticsHarvard University

School of Public HealthMortimer L. Mend elssohnLawrence Livermore National Laboratory

Jeffrey H. MillerDepartment of Biology

University of California at Los Angeles

James V. NeelDepartment of Hu man GeneticsUniversity of Michigan School of Medicine

Norton NelsonDepartment of Environmental MedicineNew York University School of Medicine

Mark L. PearsonE.I. du Pent de Nemours & Co.

Richard K. RiegelmanDepartment of MedicineGeorge Washington University School of

Medicine

Liane B. RussellOak Ridge National Laboratory

Richard B. SetlowBrookhaven Nationa l Laboratory

William J. SchullThe University of Texas Health Science Center

William G. ThillyDepartm ent of Nutr ition and Food Science

Massachusetts Institute of TechnologyRichard M. Myers, Special ConsultantUniversity of California at San Francisco

NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the advisory panel members. Thepanel d oes not, however, necessarily approve, d isapprove, or endorse th is report. OTA assumes full responsibility for the rep ortand the accuracy of its contents.

iv



OTA Project Staff—Technologies for DetectingHeritable Mutations in Human Beings

Roger C. Herdman, Assistant Director, OTA

Health and Life Sciences Division

Clyde J. Behney, Health Program Manager

Michael Gou gh, Project Director

Julia T. Ostrowsky, An a ly st

Hellen Gelband , An a ly st

Other Contributing Staff

Cheryl M. Corsaro, Analys t l

Virginia Cwalina, Administrative Assistant

Carol Ann Gun tow, Secretary/ Word Proces sor Sp ecialist

Diann G. Hohenthaner, Word Proces s or/ P. C. Spe cialis t

Eric Passaglia, Senior Distribution Specialist

Contractors

Elbert Branscomb, Lawrence Livermore National Laboratory

Neal Cariello, Massachusetts Institute of Technology

Leonard Lerman, Genetics Institute

Harvey Moh renw eiser, University of Michigan Med ical School

Richard M. Myers, University of California at San Francisco

Janice A. Nicklas, University of Vermont School of Medicine

Maynard Olson, Washington Un iversity Med ical School

Cassandra Smith, Colum bia University College of Physicians and Surgeons

L.H.T. Van der Ploeg, Columbia University College of Physicians and Surgeons

Diane K. Wagener, Washington, DC

‘On detail from the National Institutes of HealthrApril-June 1985.



Contents

1.

2.

3.

4.

5.

6.

7.

8.

9.

Summary and Options. . . . . . . . . . . . .

Genetic Inheritance and Mutations . .

Kind s and Rates of Hu man H eritable

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

New Technologies for Detecting H eritable Mu tat ions . . . . . . . . . . . . . . . . . . . . . . . . . .

New Meth ods for Measu ring Somatic Mutat ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Laboratory Determination of Heritable Mutation Rates . . . . . . . . . . . . . . . . . . . . . . . .

Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

23

37

55

73

81

91

Mutation Epidemiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..................103

Mutagens: Regulatory Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......113

Appendixes Page

A. Federal Spending formulation Research . . . . . . . . . . ...........................121

B. Acknowledgments and Health Program Advisory Committee . .................123

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ....129

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .139



Glossary of Acronyms and Terms

Glossary of Acronyms

AAECATSDR

BEIR

c

CERCLA

DN ADO EEBVEDBEPAENUEtO

FDAFRCGHbAH bMhprt

HPRT

HTTICPEMC

ICRP

mRNAMSHA

NCRP

NIHNRCNRC

N TPOSHA

OTA

PBLPFGEPKUR

—Adenine—Atomic Energy Commission—Agency for Toxic Substances and Dis-

ease Registry (Centers for Disease Control)–Committee on the Biological Effects of Ionizing Radiations (NRC)

—Cytosine—Comprehensive Environmental Response,

C om pensat ion , and L iab il it y A ct(“Superfund”)

—Deoxyribonucleic Acid—Department of Energy—Epstein-Barr Virus—Ethylene dibrom ide—Environmental Protection Agency—Ethylnitrosourea—Ethylene oxide

–Food an d Drug Adm inistration (DHH S)—Federal Rad iation Coun cil—Guanine—Hemoglobin A—Hemoglobin M—Hypoxanthineguanine phosphoribosyl

transferase (gene)—Hypoxanthineguanine phosphoribosyl

transferase (enzyme)—Heritable Translocation Test—Interna tional Comm ission for Pr otection

Against Environmental Mutagens andCarcinogens

—Interna tional Comm ission on Rad iolog-

ical Protection–Messenger RNA—Mine Safety and H ealth Adm inistration

(Department of Labor)—National Council on Radiation Protec-

tion and Measurements—National Institutes of Health—Nuclear Regulatory Commission—National Research Council (National

Academy of Sciences)–National Toxicology Program (DHHS)–Occupational Safety and Health Admin-

istration (Departm ent of Labor)–Office of Technology Assessmen t (U.S.

Congress)–Peripheral Blood Lymphocyte—Pulsed Field Gel Electrophoresis—Phenylketonuria—Roentgen

RFLP —Restriction Fragment Length Poly-morphism

RN A —Ribonucleic AcidSCE —Sister-chromatid ExchangeSLT —Specific Locus TestT —ThymineT Gr —Thioguan ine resistantTSCA —Toxic Substances Control ActUNSCEAR—United Nations Scientific Com mittee on

the Effects of Atom ic Rad iation2DDGGE —Two Dimensional Denaturing Gel Elec-

trophoresis2DPAGE —Two Dimensional Polyacrylamide Gel

Electrophoresis6TG —6Thioguanine6TG r —6Thioguanine resistant

Glossary of Terms

Achondroplasia: A disease marked by a defect in theformation of cartilage at the en ds of long bones (fe-mur, humerus) that leads to a type of dwarfism.There are a num ber of hereditary forms, the mostcommon of which is autosomal dominant.

Allele: An alternative form of a gen e, or a grou p of functionally related genes, located at the corre-spond ing site on a hom ologous chromosome. Eachallele is inherited separately from each parent.Alleles can be d ominan t, recessive, or co-dominan tfor a pa rticular trait.

Alpha thalassemia: A genetic defect caused by analteration in a p ortion of the gene coding for thealpha globin molecule. The resu lt is an insu fficient

nu mber of alpha globin m olecules and a deficiencyof adult hemoglobin.Am ino acid: One of a group of 20 molecules that bind

together in variou s sequences to form all proteinmolecules. The specific sequence of amino acidsdeterm ines the structure and fun ction of a protein.

Amniocentesis: A procedure that involves withdraw-ing a sampIe (usually 2 to 8 miIliIiters) of the am niotic

fluid su rroun ding th e fetus in u tero. This fluid con-tains cells shed by th e developin g fetus. The cellscan be grown in cell culture and analyzed either bio-chemically or cytogenetically to d etect a variety of genetic abnormalities in the fetus, including geneticdiseases such as Down syndrome and Tay Sachs

disease.Autoradiography: A technique for identifying radio-

actively-labeled molecules or fragments of mole-cules, useful for analyzing DNA for the presenceof mutations.

vii



Autosom e: A chromosom e not involved in sex-deter-mination. In a complete set of human chromo-somes, there are 44 autosomes (22 pairs of homologouschromosom es) and a pair of sex-determining chro-mosomes.

Base pair (of nucleic adds): A pair of hydrogen-bondednitrogenous bases (one purine and one pyrimidine)that join the tw o strand s of the DNA dou ble helix.Adenine pairs with thymine, and cytosine pairs w ithguanine.

Beta globin: A constituent of hem oglobin, which is amolecule consisting of four globin subunits and aheme group (two alpha an d two beta globins).

Carcinogen: A chemical or physical agent that causescancer.

Cell: The smallest membrane-bound protoplasmicbody, consisting of a nucleus and its surroundingcytoplasm, capable of independent reproduction.

Cell culture: Growth in the laborator y of cells isolatedfrom mu lticellular organ isms. Each cultu re is usu allyof one cell type (e.g., lymphocytes, fibroblasts,etc.).

Cell line: A samp le of cells, having u nd ergone the p roc-ess of adap tation to ar tificial laboratory cultivation,that is now cap able of sustaining continuou s, long-term growth in culture.

Chromosome: A threadlike structure that carriesgenetic information ar ranged in a linear sequence.It consists of a complex of nucleic acids and proteins.

Chromosome abnormalities: A group of pathologicalconditions associated with abnormalities in thenumber or structure (e.g., insertions, deletions,rearrangements) of chromosomes.

Chrom osome band ing: The chemical process of stain-ing chromosomes to iden tify each pair of homo-logous chromosom es. Staining by various techniquesproduces patterns of light and dark bands (visibleun der the m icroscope) that are characteristic of each

chromosome pair. Chromosome band ing is particu-larly useful in d etecting stru ctural chromosom e ab-norma lities such as inversions and deletions.

Codon: A sequence of 3 adjacent nucleotides in mRNAthat specifies 1 of the 20 amino acids or th e initia-tion or termination of the p eptide chain. The lin-ear sequence of codons determines the order inwhich amino acids are ad ded to a p olypeptide chaindu ring translation.

Complementary DNA (cDNA): Single-stranded DNAthat is synthesized from a messenger RNA template.It is often u sed as a p robe to help locate a sp ecificgene in an organism.

Congenital: Refers to a cond ition th at is p resent at

birth.Congenital abnorm ality: Any a bnorm ality, genetic ornongen etic, that is present at birth.

Cytoplasm: The contents of a cell exclusive of the nu -cleus. It consists of an aqueous solution and the

organelles susp ended in it and is the site of mostchemical activities of the cell, including proteinsynthesis.

Denaturation: The separation of double-strandedDNA into its complementary, separate strands bytreatment with chemicals or heat. The narrow rangeof temperatu re or chemical concentration at w hichDNA denatu ration occurs is characteristic of thenu cleotide sequence of the p articular m olecule.

Diploid: The chromosome state in which eachhomologous chrom osome is present in p airs. Allhu man somatic cells are diploid (i.e., they ha ve 46 chromosom es), wh ereas reprod uctive cells, with 23chromosomes, are haploid.

DNA: Deoxyribonucleic acid. The nucleic acid in chro-mosomes that contains the genetic information. Themolecule is dou ble-strand ed, with an external “back-bone” formed by a chain of alternating phosphateand su gar (deoxyribose) un its and an internal ladder-like structur e formed b y nu cleotide base-pairs heldtogether by hydrogen bonds. The nucleotide basepair s consist of the bases ad enine (A), cytosine (C),guan ine (G), and th ymine (T), whose structur es aresuch that A can hyd rogen bond w ith T, and C w ithG. The sequen ce of each individu al strand can bededuced by knowing that of its partner. This com-plementarily is the key to the information-trans-mitting capabilities of DNA. (See also nucleotideand gene. )

DNA probe: A segment of complementary DNA thatis used to detect the presence of a particular nucleo-tide base sequen ce.

DNA sequence: The order of nucleotide bases in DNA.Dominant: A term u sed to refer to a genetic trait that

is expressed in an individual who is heterozygousfor a p articular gene. Compar e recessive.

Down syndrome: A genetic abnorm ality characterizedby mental retardation, congenital heart defects,

immune system abnormalities, various morphologi-cal abnormalities and a r edu ced life expectancy.Down syndrome is caused by either an extra copyof chromosome 21 (called trisomy 21) or by twocopies of chrom osome 21 and another chrom osome21 translocated to a different chrom osome (usu allyto chromosome 14). The latter results in “trans-location Down synd rome.” Trisomy 21 has beenshown to increase in frequency with advancedmaternal age.

Electrophoresis: A technique used to separate m ole-cules (such as DNA fragments or proteins) from amixture of similar molecules. By p assing an electriccurrent through a medium containing the m ixture,

each type of molecule travels through th e med iumat a rate th at correspond s to its electric charge andsize. Separation is based on differences in net elec-trical charge and in size or arrangement of themolecule.

Vlli



Enzyme: A p rotein that catalyzes a chemical reactionwithout being permanently altered or consumed bythe reaction, so that it can be used repeatedly.

Enzyme d eficiency variants: Abnormal proteins char-acterized by r edu ction or elimination of enzyma ticactivity. These changes m ay be caused by the ab -sence of a gene p rodu ct, or the p resence of proteinsthat are non-functional or abnormally unstable.

Epidemiologic studies: Stud ies concerned w ith the r e-lationships of various factors determining the fre-

quency and d istribution of diseases in a hum an p op-ulation.Escherichia coli (E. coli): A species of rod-shap ed, gram

negative bacteria that inhabit the nor mal intestinaltract of vertebrates. Many non path ogenic strainsof E. coli are hosts in recombinant DNA tech-nologies.

Eukaryote: Cells or organisms with membrane-bound,structurally d iscrete nuclei and w ell-developed cellorganelles. Eukaryotes include all organisms exceptviruses, bacteria, and blue-green algae. Comp are

prokaryote.Exons: DNA sequences that determine the sequence

of amino acids in proteins. Exons are separated on

the DN A by introns, or intervening sequences, thatare transcribed and later removed, or spliced out,du ring the produ ction of mature messenger RNA.

Gametes: Mature m ale or female reprod uctive cells(germ cells: sperm or ova) with a hap loid chromo-some content (23 chromosomes in humans).

Gene: A linear sequence of nucleotides in DNA thatis needed to synthesize a protein and / or regulatecell functions. A mutation in one or more of thenu cleotides in a gene may lead to abn ormalities inthe structure of the gene prod uct or in the am ountof gene product synthesized.

Genetic code: The sequence of nucleotide triplets alongthe DNA that determines the sequence of amino

acids in protein syn thesis. This code is common tonearly all living or ganisms.

Genome: A term used to refer to all the genetic mate-rial carried by a single gamete.

Genotype: The genetic constitution of an organism,as distinguished from its physical appearance (itsphenotype). For example, an individual may havea h eterozygous genotyp e for eye color consistingof an allele for brown eyes (which is dominant) andan allele for blue eyes (which is recessive) or ahom ozygous genotyp e, with tw o alleles (both dom-inant) for brown eyes. In either case, the phenotyp eis the same: brown eyes.

Germinal mu tations: Mutations in the DNA of repro-

du ctive cells—egg or sp erm. Germinal mu tationscan be transmitted to the offspring only if one of those p articular germ cells is involved in fertili-zation.

Germ line: The tissue or cell line that p rodu ces gametesand is used for reprodu ctive purp oses, as opposed

to tissue or cell line from som atic cells. Also know nas “germinal tissue. ”

Haploid: Half of the full set of genetic material, or onechromosome of each homologous pair. Gameteshave a haploid set of DNA. Fertilization of an ovumby a sperm p roduces a diploid n um ber of chromo-somes in the zygote.

Hemoglobin: The oxygen-carrying molecule found inred blood cells. Adult hemoglobin is a proteincomposed of a single heme group bound to two

alpha globin chains and two beta globin chains.Hemoglobinopathies: A collection of h ereditary dis-orders of hemoglobin structure and/ or function.Examp les are thalassemia an d sickle cell anem ia.

Hemophilia: A genetic disorder d istingu ished by adeficiency of one or more coagulation factors—e.g.,Factor VIII (hemophilia A) or Factor IX (hemophiliaB). The und erlying mu tations are located on th e Xchromosome. Hemophilia occurs most often inmales who have only one X chromosome, and istransm itted to offspring by asym ptom atic females,who have two X chromosomes, one of which carriesthe hemop hilia m utation.

Heritable mutation: A mu tation that is passed from

parent to offspring and was present in a germ cellof one of the parents.Heteroduplex:A dou ble-strand ed nu cleic acid m ole-

cule composed of individual strands of differentorigin, e.g., a parent’s DNA strand hybridized toa child’s DNA strand, or RNA hybridized to DNA.

Heterozygote: An individual w ho has tw o differentalleles of any one particular gen e, e.g., an individ-ual w ho has on e copy of the gene for thalassemiaat the locus for beta globin is heterozygous for th istrait.

Homoduplex: A dou ble-strand ed nu cleic acid mole-cule comp osed of two strand s of the same origin,e.g., genomic DNA isolated from human ceils

(DNA hybridized w ith DNA from the same ind i-vidual).

Homozygote: An individual with the same allele atboth gen es responsible for a p articular trait.

Homozygous: Having identical alleles at a given locus.Compare heterozygous.

Hybridization: The process of joining two single-strands of RNA or DNA together so that theybecome a double-stranded molecule. For hybridiza-tion to occur, the two strands must be nearly orperfectly complementary in the sequence of thenucleotide base pairs.

Induced mutation: A change in the structure of DNA

In

or the number of chromosomes, caused by exposure

of the DNA to a mutagenic agent.vitro: Literally, “in glass, ” p ertaining to a biologi-cal process or reaction taking place in an artificialenvironm ent, usually a laboratory. Sometimes u sedto include th e grow th of cells from m ulticellularorganisms under cell culture conditions.

ix



In vivo: Literally, “in th e living, ” perta ining to a bio-logical process or reaction taking place in a livingcell or organism.

Karyotype: A photomicrograph of an individual’schromosomes arranged in a standard format,showing th e num ber, size, and shap e of each chro-mosome.

Kilobase: A unit of measurement for the length of nu-cleic acids along a chromosome. One kilobase(abbreviat ed kb ) consists of 1,000 nu cleotide bases.

Genes can be several kilobases in length.Lambda: A bacterial virus that infects E. coli, produc-ing ma ny copies of itself within th e bacterium.

Lesch-Nyh an syndrom e: An X-linked recessive dis-order characterized by compu lsive self-mutilationand other mental and behavioral abnormalities. Itis caused by a defect in the gene that produces aparticular enzym e (hypoxanthine-guanine phos-ph oribosyl transferase) imp ortant in metabolism.The causal relationship to the behavioral aspects of the disorder is not yet understood.

Locus: The position of a gene or of a group of functionally related genes on a chromosome.

Locus Test: A measurement of genes examined in a

study of mu tant proteins, based on the n um ber of proteins examined , on the num ber of gene loci rep-resented by each such protein, and on the nu mberof individual samples obtained.

Lymphocytes: Specialized white blood cells involvedin the body’s immune response. B-lymphocytesoriginate in the bone marrow and when stimu latedby an antigen produce circulating antibodies (hu-moral immunity). T-lymphocytes are produced inthe bone marrow and mature in the thymu s glandand engage in a type of defense that does not d ependdirectly on antibody attack (cell-mediated im-munity).

Marfan’s syndrom e: A genetic disorder characterized

by abnorm ally long fingers and toes and by abnor-malities of the eye lenses and heart. (AbrahamLincoln is thought by some to have suffered fromthis d isease). Also called arachnod actyly (“spiderfingeredness”).

Meiosis: The process of cell division in r eprod uctivecells that redu ces the nu mber of chromosomes tohalf of the full set so that each paren t contributeshalf of the genetic material to each offspring. Wh enthe egg an d sp erm fuse at conception, the full setof chromosomes (a total of 46 chromosomes inhu man s) is restored in the offspring.

Mendelian: A term used to refer to a trait that followsMend el’s laws of inheritance and is controlled by

a single gene, and wh ich therefore shows a simp lepattern of inheritance (dominan t or recessive). Sonamed because traits of this sort were first recog-nized by Gregor Mendel, the Austrian monk w hoseearly research laid th e basis for mod ern gen etics.Mendel’s laws include the Law of Segregation,

wh ich d escribes how each pair of alleles separatesinto different gametes and the Law of Ind epend entAssortment, which describes how different allelesare assorted ind epend ently of the other alleles ingametes and how the subsequent pairing of maleand female gametes occurs rand omly.

Messenger RNA (mRNA): Ribonucleic acid that serves

as the temp late for pr otein synthesis. It is prod ucedby transcribing a DNA sequence into a complemen-tary RNA sequence. Messenger RNA molecules

carry the instructions for assembling proteins on theribosomes.Mitosis: The process of division involving DNA rep-

lication, and resulting in two d augh ter cells withthe same nu mber of chromosomes and cytoplasmicmaterial as the p arent cell.

Multifactorial traits: Traits that are not determinedsolely by a single gene. Multifactorial, or polygenic,traits (e.g., height) have variable phenotypic effectsthat d epend for their expression on the interactionof many genes and environmental influences.

Mutagen: A chemical or physical agent that causeschanges in the structur e of DNA.

Mutation rate: The nu mber of mutations p er unit of

DNA (e.g., per gene, per nucleotide, per genome,etc. ) occurring p er un it of time (usually per g en-eration).

Mutations: Changes in the composition of DNA, gen-erally divided according to size into “gene muta-tions” (changes w ithin a single gene, such as nu cleo-tide substitutions) and “chromosome mutations”(affecting larger portions of the chromosome, or theloss or add ition of an entire chrom osome). A “her-itable mutation” is a mutation that is passed fromparent to offspring and therefore was present in thegerm cell of one of the parents. See also induced

mutation an d spontaneous mutation.Nucleic acids: Macromolecules comp osed of sequences

of nucleotides that carry genetic informa tion, Thereare tw o kinds of nu cleic acids: DNA , containing thesugar deoxyribose and RNA, containing the sugarribose.

Nucleotide: A subu nit of DNA or RNA , consisting of a nitrogenous base (adenine, guanine, thymine,cytosine, or uridine), a phosphate molecule, and asugar molecule (deoxyribose in DNA or ribose inRNA). The linkage of thousand s of these subunitsforms the DNA or RNA molecule.

Nucleus: The intracellular structure of eukaryotes thatcontains DNA.

Oligonucleotide: A polymer mad e up of a few (gener-ally fewer th an 10 or 20) nucleotides; a short se-

quence of DNA or RNA.Peptide: A compoun d consisting of two or m ore amino

acids linked together, formed th rough a chemicalprocess that produces one molecule of water foreach joining of one amino acid to an other. Peptidesare the bu ilding blocks of proteins.

x



Phenotype: The appearance of an individual or theobservable properties of an organism that resultfrom the interaction of genes and th e environment.

PhenyIketonuria (PKU): An autosomal recessive ge-netic disorder of amino acid metabolism, caused bythe inability to metabolize ph enylalanine to tyro-sine. The resulting accumu lation of phenylalanineand derived p roducts causes mental retardation,which can be avoided by dietary restriction of phenylalanine beginning soon after birth.

Polymorphism: A single gene t rait (e.g., a red bloodcell surface antigen) that exists in tw o or m ore alter-native forms (such as types A, B, AB, and O blood).Genetic variants are considered polymorphisms if their frequency exceeds 1 percent each, but are con-sidered “rar e mutations” if they are foun d in lessthan 1 percent of the population.

Polypeptide: A sequence of amino acids joined in achain.

Probe: A molecule that h as been tagged or labeled insome way with a tracer substance (a radioactiveisotope or specific dye-absorbing compound) thatis used to locate or identify a specific gene or geneproduct. For example, a radioactive mRNA probe

for a DNA gene, or a monoclinal antibody p robefor a sp ecific protein . See also DNA probes.

Prokaryote: A cell or organism lacking membr ane-bound, structurally discrete nuclei. Compareeu k ary ote.

Protein: A molecule comp osed of hund reds of linkedamino a cids in a sp ecific sequence, which is, in turn,determined by the sequence of nucleotides in DNA.Proteins are required for the structure, function,and regulation of cells, tissues, and organs in th ebody.

Recessive: A genetic trait that is manifested pheno-typically only when both alleles for the trait arepresent at a locus. E.g., sickle cell anemia is mani-

fest only when both copies of the gene for betaglobin contain the beta

smutation, whereas when

individu als have only one copy of the betasgene,

they d o not d evelop the disease, but h ave sickle-cell trait, a generally benign condition. X-linkedtraits act as if they were recessive in females anddom inant in males. Compare dominant .

Recombinant DNA (rDNA) technology: Techniquesinvolving the incorporation of DNA fragments,generated w ith the use of restriction enzymes, intoa su itable host organism’s DNA (a vector). The h ostis then grown in culture to produce clones withmultiple copies of the incorporated DNA fragment.The clones containing this particular DNA fragment

can then be selected an d har vested.Replication: The synthesis of new DNA from existing

DNA.Restriction enzym e: An enz yme th at has th e ability to

recognize a sp ecific nu cleotide sequ ence in a nu cleicacid (ranging from 4 to 12 base pairs in length ) and

cut or cleave the nucleic acid at that sequ ence. Theyare termed “restriction” enzymes because, occurrin g

naturally in bacteria, they recognize foreign nucleicacid (e. g., the DN A of a bacterial virus as it beginsto infect and d estroy its host) and destroy it, thusrestricting the ability of the virus to prey uponcertain potential host strains. Over 400 different re-striction e n z y m e s are known, recognizing a greatvariety of d ifferent nu cleotide base sequ ences. Useof restriction enzymes has made possible the cutting

and splicing together of nucleic acid within andbetween different organisms and species.

Restriction Fragment Length Polymorph ism (RFLP):Fragments of DNA that may v ary in length fromindividual to individual, depending on the presenceof polymorphisms or rare mutations.

Ribosome: A cellular organelle that is the site of protein synthesis, the process of reading theinstructions in m RNA and using it as the guide toconstructing the specified protein.

RNA (ribonucleic acid): The nucleic acid found mainlyin the nucleus and ribosomes and is involved in thecontrol of cellular activities. There are severalclasses of RNA that serve d ifferent pu rposes, includ -

ing messenger RNA (mRNA), transfer RNA (tRNA),and ribosomal RNA (rRNA). The functions of vari-ous kind s of RNA includ e “translating” selectedgene sequences coded in the cell’s DNA , and tr ans-ferring th at information outside of the nu cleus tostructures that then synthesize the proteins indicatedby th e RNA veh icle.

RNaseA: An enzym e that cleaves RNA w here certainsequences of nucleotide bases occur. RNaseA h asbeen used to cut fragments of RNA/ DNA hetero-duplexes where particular mismatches in the hetero-du plexes occur.

Sentinel phenotypes: A group of autosom al dominan tor X-linked cond itions that can occur sp orad ically.

They are thought to result from new germinalmu tations in the p arents’ reprod uctive cells.

Sex chromosomes: The X and the Y chrom osomes, 2of the 46 chromosomes in human cells, that deter-mine the sex of the individu al. Females have tw oX chromosomes, while males have on e X and on eY chrom osome.

Sickle-cell disease: A genetic disorder of hem oglobinfunction caused by the presence of an abnormalbeta globin chain. Patients with sickle-cell diseasehave red blood cells that tend to deform into asickle-like shap e. The specific defect is caused byan abnormal gene, resulting in the replacement of the u sual amino acid, glutam ic acid, with valine,

in the sixth amino acid position in the beta-hemo-globin molecule. This increases its propensity tocrystallize, thus ru ptu ring the red blood cell andcausing the cells to lodge in small blood vessels.Also called “sickle-cell anemia.” See sickle-cezl trait.

Sickle-cell trait: The generally benign condition show n

xi



by individuals carrying the variant beta s gene aswell as the norm al beta

agene. Such ind ividuals are

heterozygous for the sickle-cell gene, and a re healthy(i.e., are u sually asymp tomatic), but two hetero-zygous parents have a 25 percent risk with eachpregnancy of having a child with hom ozygoussickle-cell disea se.

Single gene d isorder: A genetic disease caused by a sin-gle gene and show ing a simple pattern of inherit-ance (e.g., dominan t or recessive, autosomal or X-

Iinked). Also called “Mendelian disorder.”Somatic: A term used to refer to body tissues, as op-

posed to reproductive (germinal) tissues.Somatic cell: Any cell in the body except reproduc-

tive cells or their precursors.Spontaneous mutation: In the absence of any known

causative agent, a change in th e structure of DNAor in th e nu mber of chromosomes. Also called a“background” mutation. Also see mutat ions .

Stem cells: Und ifferentiated cells in the bone ma rrowthat have the ability to replicate and to differenti-ate into blood cells.

Tay-Sachs d isease: An autosom al recessive gen etic dis-ease resulting in developm ental retardation, paral-

ysis, dem entia and blindness, usually fatal in earlychildh ood. The d efective gene codes for h exosamini-dase A, an enzyme that is involved in certain chem-ical pathw ays in the brain. Symp toms are causedby an accum ulation of cerebral gangliosides, fattyacid and sugar molecules found in the brain andnervou s tissue. The gene is foun d in h ighest fre-quen cy among Jews of Eastern Europ ean origin.

Teratogen: A p hysical or chemical agent (e.g., thalido-mid e, radiation , alcohol, etc. ) that can cause con-genital abnormalities as a result of exposure inutero.

Tetramer: A comp lex molecule consisting of four m a- jor portions joined together (e.g., hemoglobin, inwhich two alpha chains and two beta chains are joined to a central heme grou p. )

Thalassemia: A group of autosomal recessive geneticdisorders characterized by abnorm alities in synthe-sis of the globin p olypeptid es of hemoglobin. Thetwo m ost common forms are alpha-thalassemia andbeta-thalassemia, disorders of the alph a- and beta-globin polypep tides, respectively, which cause im-

balances in the production of the globins and leadto an overall deficiency of adult h emoglobin. Thethalassemias are most comm on in peop le of Medi-terranean, Midd le Eastern, and Asian d escent.

T-lymphocytes: See Lymphocytes,Transcription: The synthesis of messenger RNA (mRNA)

on a DNA template. The resulting RNA sequenceis comp lementary to the DNA sequ ence.

Transfer RNA (tRNA): Specialized RNA m oleculesthat fun ction to br ing specific amino acids to ribo-somes that translate messenger RNA (mRN A) intoproteins.

Translation: The process in wh ich th e genetic codecontained in the n ucleotide base sequen ce of mes-

senger RNA directs the synthesis of a specific or-der of amino acids to produce a protein.

Trisomy: The presence of an extra chromosome, re-sulting in three homologous chromosom es insteadof two, e.g., Down syndrome can result fromTrisomy 21, or the p resence of an extra chromo-some nu mber 21 in each bod y cell.

tRNA: See trans fer RNA.X-linked mutation: A mu tation that occurs in a region

of the X-chromosome.Zygote: A fertilized egg that results from the fusion

of sperm an d egg.

xii



Chapter 1

Summary and Options



Contents

Introduction .Request for

. . .

the. . . . . . . . . . . . . . . . . . . . . . . .

Assessment . . . . . . . . . . . . .Scope of the Report.. . . . . . . . . . . . . . . . . . .

Background ... ..~.. . . . . . . . . ........+ . . . .Kind s and Effects of Mutations in Hu man

Current Method s for Studying Mu tations . . .An imal Studies . . . . . . . . . . . . . . . . . . . . . . . .Stud ies in H um an Beings . . . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

Beings

. . . . . .

. . . . . .

. . . . . .

New Technologies for Detecting Hum an Mu tations

Page

. . . . . . . . . . . . . . . . . . . . . . . . . . 3

. . . . . . . . . . . . . . . . . . . . . . . .,, 4

. . . . . . . . . . . . . . . . . . . . . . . . . . 4

. . . . . . . . . . . . . . . . . . . . . . . . . . 5

. . . . . . . . . . . . . . . . . . . . . . . . . . 5

. . . . . . . . . . . . . . . . . . . . . . . . . . 6

. . . . . . . . . . . . . . . . . . . . . . . . . . 6

. . . . . . . . . . . . . . . . . . . . . . . . . . 7

. . . . . . . . . . . . . . . . . . . . . . . . . . 1 0

Detection an d Measurement of Heritable Mutations. . . . . . . . . . . . . . . . . . . . . .. .11Detection and Measurement of Somatic Cell Mutations . . . . . . . . . . . . . . . . . . . . 13

Use of New Technologies in Research and Public Policy. . . . . . . . . . . . . . . . . . . . . .14Feasibility and Validity Testing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14Epidemiologic Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......14

Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......14Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Federal Spending formulation Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Development of New Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...16Feasibility and Validity Testing . . .Field Studies . . . . . . . .Integration

Figure No.

. . . . . . . . . . .

and Human

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

FIGURE

Page

1. Organizational Hierarchy of DNA, the Carrier of Genetic Informationin Human Cells.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6



Chapter 1

Summary and Options

INTRODUCTION

Mutations, lasting changes in the genetic in-formation carried in th e d eoxyribonu cleic acid(DNA) of cells, can cause severe diseases and dis-abilities, none of which is curable and relativelyfew of which can be treated effectively. Suchgenetic diseases represent a significant fraction of chronic disease and mortality in infancy and child-hood; they generally impose heavy burdens ex-pressed in prem ature m ortality, morbidity, infer-tility, and p hysical and mental han dicap. Someof the most common of the 3,000 or more differ-

ent disorders known to result from mutations in-

clude Down syndrome, Duchenne muscular dys-

trophy, and hemophilia. In addition, mutationshave been associated with increased susceptibili-ties to certain chronic diseases, including someforms of diabetes, heart disease, and cancer. Mostmu tations that are expressed as genetic disease al-ready exist in the pop ulation an d are carried fromgeneration to generation. A smaller prop ortionof mutations arises anew, “sporadically,” in eachgeneration, and the specific causes of these mu-tations are unknown .

The public and the government have expressedconcern about th e p ossibility that environm ental

exposures are contributing to or increasing the fre-quency of mutations. Mutations are among thechronic health effects singled out in the Toxic Sub-stances Control Act of 1976 (TSCA) and in the

Comprehensive Environmental Response, Com-

pensation, and Liability Act of 1980 (CERCLA or

“Superfund”). In those laws, Congress specified

public protection from exposures that can cause

mutations. The other major environmental stat-utes contain language broad enough to includeprotection from mutagens.

Unfortunately, little is currently known aboutthe kinds and rates of mutations that occur inhum an beings. Available method s to study suchmu tations are inadequate to p rovide sufficient in-formation for evaluating mu tagenic risks.

Much of our kn owledge of genetic risks to hu-man health from exposures to environmentalagents has been d erived from the stud y of the ef-fects of mutagens on experimental animals. Theseexperiments are useful in manipulating variousaspects of the mu tagenic process, for examp le, toexamine how mutagens act on DNA and to studyeffects of varying doses and rates of exposure tomutagenic agents administered either singly or incombination. Experimentation with animals is es-sential for assessing potential hazards of newchemical and p hysical agents before hum an p op-ulations have been exposed to them . How ever,

at p resent, technical problems in d etecting andmeasuring mutations limit animal experiments asthey limit huma n stud ies, so the results from ani-mal experiments, using current methods, detectonly a small proportion of the kind s and num -bers of mutations that can occur.

Recent adva nces in molecular biology have ledto the d evelopm ent of new technologies for exam-ining DNA that may p rovide insight into the kindsand rates of mu tations that occur in hum an be-ings. This report assesses these developments anddiscusses their potential for predicting risks of mu-

tation from particular exposures.

At present, these new technologies propose rea-sonable and verifiable ways of detecting herita-ble mutations in human DNA and proteins, butnone is efficient enou gh to be u sed on a large scale.However, there is considerable optimism in thescientific comm un ity that these new technologiescan prov ide, for the first time, the means to ob-tain basic know ledge about the pr imary causesof mutation and the means to assess the kind s andrates of mutations that occur in human beings.

Data derived from stud ies in hu man beings,along with verifiable methods to extrapolate fromcorrespond ing animal d ata, will permit a m ore in-formed assessment of the m edical and biological

3



4 q Technologies for Detecting Heritable Mutations in Human Beings

consequences of mutagenic exposures. At present,withou t such comp arative data, it is difficult toknow whether general extrapolations from ani-mal data would lead to underestimates or over-estimates of the genetic risks for human s. Con-tinuing to rely on inadequ ate data may incur both

human and financial costs, since conclusionsdrawn from this information contribute to deci-sions abou t acceptable levels of exposure an d thelevel of society’s resources that ar e devo ted to p ro-viding p rotection from such exposu res.

A combination of factors—concern that env i-ronmental exposures may be contributing to hu-man mutations, questions about the fundamen-tal nature of mu tations, and increasing kn owledgeof the structure an d function of DN A—increasethe likelihood tha t new technologies will be d e-veloped and field tested. How ever, stud ies usingthese technologies may be expensive and willprobably require the collaboration of a large num-ber of scientists; their continued development, pi-lot testing, and large-scale application may requiresufficient interest and financial supp ort ou tsidethe scientific community. With such support, andwith continued d evelopment of the techniqu es,some of these techniques could be ready for large-scale use in the next 5 to 10 years.

Congressional interest in supporting basic re-search on hu man mu tations and in the continueddev elopm ent of these techn ologies is necessary if the regulatory agencies are eventually to have the

tools to evaluate risks from most occupational orenvironmental exposures. The current lack of in-formation on kinds and rates of human mutationsis largely due to the inad equacy of present meth-ods to study heritable mutations. Efforts to com-ply with the agencies’ mandates to protect peo-ple from m utagens ma y be imp eded u nless basicknowledge of causes, kinds, and rates of hum anmutations is obtained.

Request for the Assessment

This assessment w as requested by th e SenateCommittee on Veterans’ Affairs, the House Com-mittee on Science and Technology, and the HouseComm ittee on Energy an d Comm erce. Interest inthe assessment w as also expressed by the SenateCommittee on Public Works and the Environ-

ment, the Senate Committee on Labor and Hu -man Resources, and the H ouse Comm ittee on Vet-erans’ Affairs.

These comm ittees have w restled with p roblemsof determining w hether past exposures to poten-tial mutagens have affected the health of veterans

and civilians and of framing r easonable publichealth laws that can be implemented , given cur-rent know ledge and technologies. OTA w as askedto assess the available information about curr entmeans for detecting mu tations as they relate tothese issues and on the likelihood and potentialimpact of technological developments.

Scope of the Report

This chapter sum marizes current knowledgeabout the kind s and rates of hu man m utations andthe methods that have been used to detect herita-

ble mutations in h uman beings and in experi-mental animals. New technologies assessed in thisreport for detecting and measuring human herita-ble mutations are briefly described. Methods formeasuring human somatic mutations are discussedas tools for evaluating the risks of heritable mu-tations. The final section of this chapter p resentsoptions for congressional action.

Chapter 2 provides background informationabout hu man genetics and DN A, and d iscussesthe types of mutations th at can occur and theirpotential health effects. Chapter 3 reviews the

literature on current m ethods for stud ying muta-tions and summarizes current knowledge aboutthe frequency of heritable mutations in hu manpopulations.

The new technologies for examining humanDNA for heritable mutations are described inchapter 4, followed by d escriptions of new so-matic mutation tests in chapter 5.

Chapter 6 summarizes data from experimental

animals on spontaneous and induced mutations,

and discusses the possible use of such data for

identifying human mutagens and determining

their potency. Chapter 7 focuses on the problems

of extrapolating from the results of animal exper-

iments to human risks.

Chapter 8 discusses epidemiologic considera-

tions in the application of the new technologies,



Ch. l—Summary and Options 5

such as validation of the new m ethods and selec- A sum mary of current Federal expend itures intion of at-risk popu lations to stu dy. Chap ter 9 dis- pu tation research and potential costs of stud iescusses Federal involvement in p rotecting against to detect mutations using the new technologies isgenetic risks and the regulatory mechanisms avail- presented in append ix A.able to control exposur es to mu tagenic agents.

BACKGROUND

Kinds and Effectsin Human Beings

of Mutations

Mutations can occur “spontaneously,” that is,

in the apparent absence of any unusual stimuli,

or they can be “induced” by particular agents. It

is likely that many or most “spontaneous” muta-

tions are caused by external forces, possibly in -clud ing ionizing rad iation, ultraviolet radiation,

viruses, and certain chemicals, but the app ropri-ate links have not been made. Some mutagenspresent aroun d us m ay also be necessary for sus-taining life, for example, oxygen, components of our food, and som e of the body’s own m etabo-lizes. Experiments in an imals have show n thatmany substances present in agricultural, indus-trial, and pha rmaceutical chemicals in u se todayare mutagenic in some test systems. Which of these cause mutations in human beings is still amatter of speculation. Precise causes for essen-tially all mutations th at have been id entified inhuman beings are unknown.

At present, more than 3000 different genetic

diseases have been identified, including d isordersresulting from mu tations in DNA, and d isordersresulting from the interaction of genetic and envi-ronmental components. Approximately 10 in1,000 liveborn infants are born with a single genedisorder and an ad ditional 6 in 1,000 liveborn in-fants are born with a major chromosome abnor-mality. It is estimated that ap proximately 80 per-cent of the single gene disorders are the direct

result of mutations that occurred in germ cells of distant ancestors and were passed along to suc-

ceeding generations. The remaining 20 percent of these cases (0.2 percent of all livebirths) and the

majority of chromosome abnormalities are be-

lieved to be due to sporadic mutations in the re-

productive cells of one of the parents of the in-fant. An additional 10 in 1,000 liveborn infants

ma nifest a serious genetic disease sometim e afterbirth, and a far higher proportion of newbornswill show ind irect effects of one or several p arentalor ancestral mu tations in later life as, for exam-ple, in increased su sceptibilities to some forms of heart disease, diabetes, or cancer.

Mutations are changes in the composition of the genetic material, DNA (see fig. 1), and are gen-erally d ivided according to size into gene m uta-tions and chromosome m utations. Gene mu ta-tions refer to changes within a single gene, forexample, substitutions of single component nu-cleotides, or small losses or additions of geneticmaterial in expressed or nonexpressed regions of the gene. Chromosome mutations affect largerportions of the chromosom e (e.g., structu ral rear-rangements of genetic material in the chromo-somes) or result in the loss or addition of an en-tire chromosome. Since DNA directs the synthesisand regulation of molecules in the body, eithergroup of mutations can influence a wide range of biological and ph ysiological fun ctions, includ ingreproduction, longevity, intelligence, and physi-cal development. Individual differences in suscep-tibility to disease may result from the effects of one gene, several genes, or combinations of genesand environmental factors.

Depending on the nature an d location of themutations and on the function of the genes inwh ich they occur, mutations m ay, in theory, bebeneficial, neutral, or harmful to the individual.

l

The kinds and effects of known m utational eventsrange from single nu cleotide substitutions (thesmallest unit of change in DNA) w ith no clini-

cally observ able effects, to single nu cleotide sub-

ICurrent theory maintains that most newly arising mutations inregions of DNA that directly determine the structure and regula-tion of proteins are more likely to be detrimental than beneficial.Little is known of effects of mutations in other regions of DNA.



6 . Technologies for Detecting Heritable Mutations in Human Beings

Figure I.–Organizational Hierarchy of DNA, theCarrier of Genetic Information in Human Cells

Nucleotides

A G C T

AAA CGC GAC CGA

-ACGAAAATCCGCGCTTCAGATACCTTA –

SOURCE: Office of Technology Assessment.

Adenine, guanine,cytosine, andthymine, the basicbuilding blocks ofDNA.

CodonsNucleotides arrang-ed in a triplet code,each correspondingto an amino acid(components of pro-teins) or to aregulatory signal.

Genes

Functional units ofDNA needed to syn-thesize proteins orregulate cell func-tion.

Chromosome

Thousands of genes

arranged in a linearsequence, consist-ing of a complex ofDNA and proteins.

GenomeThe complete set ofgenetic information;each human repro-ductive cell contains23 chromosomes,and all other cells inthe body contain afull set of 46chromosomes.

stitutions resulting in severe diseases; from majorstructural and numerical chromosome abnormal-ities (the largest observed unit of change in DNA)leading to various abnorm alities and impairments,to those resulting in embryonic, fetal, or neona-tal death.

A child’s entire genetic endowment comes fromthe DN A of tw o single reprodu ctive cells (or ga-metes), one egg and one sperm, from his parents.

2

A mutation occurring in the DNA of either of these ger m cells, a germinal mutation, is passedon to the child, who is born with a “new ” herita-

Me m utation. Mutations in germ cells that are notinvolved in fertilization are not p assed on to theoffspring. If a mutation arises in the DNA of theparents’ nonrep rod uctive cells (collectively termedsomatic cells), such somatic mutations are nottransferred to the reproductive cells. Somatic mu-tations may, however, affect the parents’ health,and indirectly, their ability to bear a healthy child.

‘Each of the gametes (the male spermatozoan and the femaleovum) is the product of a series of developmental stages of repro-ductive (or germ) cells.

CURRENT METHODS FOR STUDYING MUTATIONS

Current emp irical method s to stud y mu tationsin hu man beings focus on p hysiological and bio-chemical effects of mutations because, until re-cently, it was not possible to examine changes inDNA d irectly. Each of the curr ent methods d e-tects only a limited portion of the spectrum of mutational changes. These methods have beenused to derive estimates of baseline frequenciesof some kinds of hum an m utations.

Animal Studies

Much of our knowledge about how substancesinteract with DNA and how they may cause mu-tations is derived from studies with experimentalanimals. In ad dition, some estimates of hu man

mutation rates have been derived, by extrapola-tion, from animal stu dies. Beginning soon afterWorld War II and still continu ing, spontaneou sand induced gene mutation rates have been stud-ied in laboratory m ice. The rate of spontan eousheritable gene m utations, as detected in these ex-periments, is roughly tw o to eight mu tations per1 million genes p er gener ation of mice. Radiationand app roximately two-thirds of about 20 chem-icals

3tested so far increase the frequency of de-

— —30n the basis of animal experiments using the specific locus test

and the heritable translocation test, in wh ich they were found tobe mutagenic in all germ cell stages, these chemicals are stronglysuspected to be mutagenic in humans (154). They included both com-mon environmental agents and chemicals available in the labora-tory that are not normally found in the environmen t.



Ch. l—Summary and Options . 7

tectable heritable gene mutations in these experi-mental mice. Generally, the chemical substancesthat indu ce heritable mutations in mature andmatu ring germ cells also indu ce somatic muta-tions, and vice versa. How ever, substances thatindu ce somatic mutations do not n ecessarily in-

du ce heritable mutations in immatu re germ cells(stem cells). Unlike ma tur e germ cells or somaticcells, these germinal stem cells may have efficientsystems for repair of mutational or premutationaldamage. There is, in fact, indirect evidence fromdose-response and dose-rate experiments that ger-minal stem cells have good repair systems.

It may be u seful to qu antify relationships be-tween som atic and germinal cells with regard tomutagenic potency of different chemicals, and tostud y mu tations in equ ivalent sets of genes in bothtypes of cells. Animal experimentation is useful

for determining the feasibility of using human so-matic mutation rates for predicting the risk of hu-man heritable mutations. This work in anim alsmay demonstrate whether it is possible to extrap-olate from the occur rence of somatic to that of heritable mutations at all, and may help deter-mine whether it is possible to generalize from ani-mal data to human beings. In addition, animalstud ies on heritable mu tations are useful not onlyfor determining whether a given agent is muta-genic but also for more general explorations of the factors that may influence the occurrence of mutations.

Studies in Human Beings

Spontaneous heritable mutations in human be-ings have been stu died by examining: 1) the inci-dence of certain genetic diseases (“sentinel pheno-types”), 2) gross changes in chromosome structureor number, and 3) changes in the structure orfunction of blood proteins. Epidemiologic studiesof specific populations, in particular, the survivorsof atomic bombs in Japan, pr ovide some infor-mation about indu ced heritable mu tations in hu -man beings.

Sentinel Phenotypes

The classic method for identifying human herit-able mu tations is the emp irical observation of in-fants and childr en for the presence of certain raregenetic diseases known as sentinel phenotypes. Ex-

amples include achondroplasia (dwarfism), aniridia,and some childh ood cancers, such as retinoblas-toma and Wilms’ tumor. By recording the occur-rence of sentinel phenotyp es as a prop ortion of the total num ber of livebirths in a d efined pop u-lation over time, the frequency of each disorder

(and of its corresponding mutation) can be esti-mated (165).

The characteristic of sentinel phenotypes thatis most useful for mu tation stud ies is that theseconditions are “sporadic” in most or all cases; theyalmost alwa ys result from a n ew germ inal muta-tion in one of the paren ts of the afflicted individ -ual.4 Each different sentinel phenotype is thoughtto result from a different, single, mu tant gen e, al-though precise genetic information to confirm thesingle gene hyp othesis is lacking in most cases.

Despite the d istinctive characteristics of sentinel

phen otypes, the relevance of existing d ata on thefrequency of the various sentinel phenotypes tothe study of kind s and rates of hum an m utationsis limited by a lack of knowledge of the geneticbases of the ph enotypes, and by the sm all frac-tion of DN A that accounts for these ph enotypes.Of the several thousand know n genetic diseases,only 40 are thought to satisfy the criteria for in-

clusion as “sentinel phenotypes. ” Roughly 40

genes are involved in the 40 sentinel phenotypes,

among a total of an estimated 50,000 to 100,000

expressed genes in an individual’s DNA.

Sentinel phenotypes are severely debilitatingconditions that require accurate diagnosis and

long-term medical care. However, practical dif-

ficulties arise in gathering and maintaining data

on the incidence of sentinel phenotypes for the

purpose of tracking mutation rates. Infants with

sentinel phenotypes are rare, numbering approx-

imately 1 in 10,000 to 1 in 10 million Iiveborn

infants, depending on the particular disease. Con-

sequently, a huge number of infants must be ob-

served in order to find even a few infants withsentinel phenotypes. Diagnosis of individual phe-notypes is complicated by the genetic heterogeneity

of these disorders, so that highly trained special-ists in various pediatric subdiscipline, which are

‘Various tests are done to exclude nonsporadic cases, which couldresult from X-linked recessive inheritance, mistaken p arentage, andthe occurrence of other genetic or nongenetic conditions that mimicthe appearance of sentinel phenotypes.




few in number, would be needed to make theseobservations. Millions of consecutive newborn in-fants would have to be monitored thoroughly andaccurately for many years, and registries, muchlarger than those currently in use, would beneeded to collect and maintain the necessary d ata.

Chromosome Abnormalities

Another m ethod for identifying a certain classof hum an h eritable mu tations is the examinationof chromosomes und er a light m icroscope for thepresence of chromosom e abnor malities (“cyto-genetic analysis”). Norm al hum an DN A is orga-nized into 46 chromosomes (44 autosomes and 2sex chromosomes), distinguishable by size, pro-portional shape, and staining pattern, or “band-ing.” Chromosome abnormalities are defined aseither numerical (extra or missing whole chromo-some[s]) or structural (deletions, insertions, trans-locations, inversions, etc., of sections of chromo-somes). In total, chromosome abnormalities areestimated to occur in at least 5 percent of all hu-man conceptions. The majority of such concep-tuses are spontaneously aborted, but the few thatsurvive comprise approximately 0.6 percent of allliveborn infants. The incidence of Down syn-dr ome is 1 in 650 at birth, making it the m ost com-

mon chromosomal disorder in newborns.

With the most advanced chromosome staining

methods currently available, approximately 1,000

bands are distinguishable in one set of human

chromosomes, although far fewer bands are pro-duced with routine methods. Mutations in DNA

sometimes cause a change in the banding pattern,

particularly if such mutations involve large sec-

tions of a chromosome. With routine banding

methods, there may be several hundred genes

present in each visible band, and with higher reso-lution banding methods, a single band may con-

tain about 100 genes. However, smaller muta-

tions, from single nucleotide changes within genes

on up to some deletions and insertions of entire

genes, generally are not visible by any banding

method.

Measurement of Mutant Proteins

Most of the available information on rates of

spontaneous human mutations has been derived

from studies of sentinel phenotypes and chromo-

some abnormalities, but data on different kindsof mutations in hu mans is now emerging fromstudies of mutant proteins. In general, mutantproteins ar e more p recise indicators of geneticdam age than are clinical and cytogenetic obser-vations. Certain m utations in genes that d etermine

the structure of proteins alter the chemical char-acteristics of the proteins, causing them to behavedifferently in separation and purification proce-dures. These differences suggest that a mutationhas occurred because proteins are constructedaccording to blueprints in DN A, and changes inDNA can lead to the prod uction of altered pro-teins. If the protein un der stud y is an enzym e, amutation within the gene that codes for it canalter, diminish, or eliminate the enzyme’s bio-chemical activity.

Operationally, mutant proteins are identifiedby taking samples of blood from each memberof a “triad,” includ ing both p arents and the child.Proteins are extracted from blood compon ents,and the proteins are separated by electrophore-sis. Putative mutations are identified when a pro-tein from a child behaves differently from the cor-responding p rotein from both p arents.

One-Dimensional Separation of Proteins.—Thetechnique most commonly used to study mutantproteins is electrophoresis, a method of separat-ing p roteins on the basis of their electrical charges.The term “electrophoretic variant,” or “electro-morp h, ” is used to describe a protein that behaves

differently in electrophoresis from the correspond-ing protein found in the parents.

Although one-dimensional electrophoretic anal-ysis of proteins is well established and can beimproved by includ ing functional assays for ad -ditional enzymes, it is limited to detecting: 1) mu-tations that do not eliminate the functional abil-ity of the proteins, 2) nucleotide substitutions onlyin coding regions of genes for the proteins exam-ined, and 3) only those nu cleotide substitutionsthat alter the electrical charge on these proteins;such substitutions are thought to account for

about one-third of all nucleotide substitutions incoding regions, wh ich, in turn , account for a frac-tion of all the kinds of mutations that can occur.Electrophoresis does not detect many other typesof mutations, including small duplications, re-arrangements, or mutations that result in the



Ch. l—Summary and Options q 9

absence of gene products; these mutations arethought to constitute the majority of the muta-tions ind uced by certain mu tagens, including ra-diation. Moreover, electrophoresis does not de-tect m utations that occur anywh ere outside thecoding regions of a certain set of genes, includ-

ing mu tations in other coding genes and in non-expressed regions of the DNA.

Data from several studies using one-dimen-sional electrophoresis have been used to estimatethe rate at wh ich m utations produ ce electropho-retic variants, and from this estimate, to infer thetotal rate of amino acid substitutions in proteins,and the correspond ing mutation rate per nu cleo-tide in hum an DNA.

Two-Dimen sional Separ ation of Proteins.—Anextension of one-dimensional electrophoresis in-volves separation of proteins in a second d imen-sion. With two-dimensional electrophoresis, about300 proteins from each person can be separated

and examined, compared with about 100 proteins

per sample that can be separated in one-dimen-

sional electrophoresis. Further improvements may

be possible with the use of computer algorithms

to assist in interpreting the complex two-dimen-

sional gels. This technique is currently feasible,

and although it detects the same types of muta-

tions as one-dimensional electrophoresis, it can

examine more proteins per samp le.

Epidemiologic StudiesAn extensive body of data from experimental

animals dem onstrates that exposure to radiationand to certain chemicals can ind uce mu tations inmammalian germ cells. In humans, exposure toionizing radiation is known to cause somatic mu-tations, and it is suspected to enh ance the prob-ability of heritable mutations. To date, however,the av ailable m ethods hav e provided no direct evi-

den ce for the ind uction (by chemicals or by radi-

ation) of mutations in human germ cells.

The single largest population studied for in-

du ced mu tations is the group of survivors of theatomic bombs detonated in Hiroshima and Naga-saki in 1945. Many survivors of the bombs re-

ceived doses of radiation that could have caused

germinal mutations; in experiments with muta-

tion induction by radiation in mammals, similar

kinds and doses of radiation were sufficient tocause observable mutations in offspring. There-fore, it was assumed that germinal mutationscould have been indu ced in p eople exposed to therad iation from th e blasts.

Medical examinations of the survivors soon af-ter the b lasts revealed the imm ediate effects of whole body irradiation: loss of hair; reduction inbone m arrow activity; and redu ction in circulat-ing white blood cells, associated with a reductionin the body’s resistance to infection. Among thosew ho recovered from the imm ediate effects of theradiation, there was a significant excess of can-cer deaths later in life. Certain types of leukemiawere the first cancers to appear in excess, but con-tinuing followu p has r evealed later increases inother cancers, such as multiple myeloma andcancers of the breast, thyroid, colon, esophagus,

stomach, lung, ovaries, and p ossibly of the spi-nal cord an d nerves (24,55).

Exposure of pregnant women to radiation was

found to be associated with an increased incidence

in their liveborn infants of small head circumfer-

ence, mental retardation, and an increased inci-

dence of childhood cancers. The critical time for

fetal brain damage from radiation exposure wasidentified as the period of 8 to 15 weeks of gesta-

tion (99).

Analysis of the chromosomes prepared from

peripheral blood lymphocytes of survivors ex-

posed to the radiation has indicated an excess of chromosome aberrations (7). Certain types of

chromosome aberrations (mainly balanced struc-

tural rearrangements, such as reciprocal translo-cations and inversions) have been found to per-sist in circulating lymphocytes long after exposureto radiation, whereas other types of chromosomeaberrations (e.g., unbalanced rearrangements) inlymph ocytes declined in nu mber soon after ex-posure. Overall, the frequency of chromosomallyaberrant cells in the survivors’ blood w as foundto be prop ortional to the estimated d ose of radi-ation received at the time of the bombing. It hasnot yet known whether these somatic mutationsare correlated with specific cancers or other dis-eases in the survivors.

Survivors’ children who were conceived afterthe acute radiation exposure were examined to




study mu tagenic effects on the p arents’ reprodu c-tive cells. Using various methods available from1945 to the p resent, survivor s’ offspring were stu d-ied for “untoward pregnancy outcomes, ”

5for cer-

tain chromosome abnormalities, or most recently,for abnormal blood proteins. The offspring of par-

ents exposed to atom ic radiation were comp aredwith the offspring of parents who were beyondthe zone of radiation (greater than 2,500 metersfrom the h ypocenter at the time of the bombings).Observation and analysis of some 70,000 off-spring has rev ealed n o statistically significant ex-cess in the incidence of stillbirths, congenital mal-formations, neonatal deaths, or chromosomeabnormalities. These findings suggest that the fre-quency of radiation-indu ced germinal mu tationsthat led to certain gross abnormalities in newborninfants was not high en ough to be detectable in

a population of that size and genetic heter-ogeneity. However, they do n ot rule out the pos-sibility of other manifestations of genetic damagein these children , or of latent expressions of suchdam age, since the methods u sed to study th is pop-ulation could examine only a small subset of DNAand only a limited nu mber of genetic endp oints.

Analysis of the children’s blood proteins forelectrophoretic variants was later done to detectrecessive mutations, that is, mutations th at arenot expressed as disease (unless present in both

5These were defined as major congenital defects and/ or stillbirths

and/ or death in the survivors’ offspring du ring the first postnatalweek. These abnormalities can be caused by exposure to radiation,as well as to other environm ental agents, and by socioeconomicfactors.

copies of a particular g ene). This ana lysis, begu nin 1976, foun d few mu tations in either the exposed

or control groups, making interpretation problem-

atic. While the results indicate no significant ex-

cess of mutant proteins in the children of exposed

parents, they do not exclude the possibility that

an excess exists undetected. Unfortunately, elec-trophoresis is inefficient at detecting deletions, one

of the most likely types of radiation-induced mu-

tations. Overall, these findings do not rule out

the possibility of genetic damage to the offspring

of survivors of the atomic bombs, but they put

upper limits of the frequency of occurrence of cer-

tain types of mutations that the current methods

are able to detect.

Taking these findings at face value, and cog-nizant of an enormous body of data on the genetic

effects of radiation on experimental animals, the

investigators suggest that the dose of radiationnecessary to double the human mutation rate (the“doubling dose”) was between 139 and 258 rem,

6

but they caution that there is a possibility of large

error attached to that estimate since genes other

than the ones sampled may demonstrate differ-

ent sensitivity to radiation, and some types of mu-

tations may be repaired more efficiently than

others. Their estimate of the doubling dose, if cor-

rect, how ever, ind icates that m an could be con-siderably less sensitive to radiation than labora-tory mice.

6A rem (Roentgen-Equivalent-Man) is a measure of absorbed ra-diation dose. For comparison, a chest X-ray exposes an individualto about 0.1 rem.

NEW TECHNOLOGIES FOR DETECTING HUMAN MUTATIONS

Recent advances in molecular biology have ledto the development of techniques that allow di-rect examination of DN A for evidence of mu ta-tions. These methods can examine large regionsof human DNA without requiring detailed knowl-

edge of the genes contained in those regions andwithout preparing genetic probes for particularsequences. Unlike current methods—observingsentinel phenotypes, chromosome abnormalities,

and electrophoretic variants—that are limited todetecting a small fraction of all kinds and nu m-bers of mu tations, the new techniqu es have thepoten tial for detecting a wide, unselected spectrumof mutations across the DNA. These techniques

are examples of the state-of-the-art in moleculargenetics and th ey are now p romising to providethe basis for better approaches to studying m uta-genesis.



Ch. I—Summary and Options q 1 1

Applying recent developments in molecular bi-

ology to the problem of detecting sporadic mu-

tations, the new techniques described in this report

propose reasonable and verifiable ways of exam-

ining human DNA for alterations in sequence and

structure. These developments include the abil-

ity to clone specific genes, to cut up DNA intopred ictable fragments, to h ybridize complemen-tary DNA strands, to detect less-than-perfecthybridizations due to single base pair changes, andto separate large fragments of hu man DNA. Someof these methods d etect similar types of mu tationsand some complement each other by d etecting dif-ferent typ es. Some of these m ay m erit further de-velopment and , eventually, pilot testing. Severalnew technologies, representing d ifferent app roaches,are discussed below.

None of these new techniques has been ap pliedto large human popu lations or to experimentalanimals and it is not known h ow w ell they willperform. At present, none of these techniques isapproaching the efficiency needed for examiningthe kinds and rates of mu tations in a pop ulationor for determining whether mutation rates are in-creasing. With technical improvements in effi-ciency, some of the techniques, or derivatives of them, could be available in the next 5 to 10 yearsfor large-scale u se. Since these techn ologies pro-vide new information about DN A, the health im-plications of any newfound mu tations may notbe immed iately know n. Add itional research and

method s wou ld be needed to examine biologicaland physiological implications of the identifiedmu tations for the popu lations studied an d for theirdescendants.

Detection and Measurement ofHeritable Mutations

Restriction Fragment Length Polymorphisms

Restriction endonucleases are enzymes, isolatedfrom bacteria, that can be used experim entally tocleave isolated DNA molecules into fragments at

specific sites in the DNA sequence that they rec-ognize. If any of these sites have been altered bymu tation, the resulting pattern of fragment sizeswou ld also be altered. Restriction enzym es canbe used to detect mutations that either: 1) createa new restriction site, 2) eliminate an old one, or

3) change the distance between existing restrictionsites. (These mutations may include single nucleo-tide substitutions, or multiple nucleotide deletionsor insertions. )

To use this method to detect mutations, DNAwould be treated with a set of restriction en-

donu cleases and the resulting DNA fragments sep-arated by electrophoresis and examined for differ-ences between those present in the child’s DNAand those in either pa rent’s DNA. Restriction siteanalysis d oes not allow examination of every nu -cleotide. How ever, the u se of a set of combinedrestriction enzymes increases the number of re-striction sites identified, allowing a larger portionof the DNA to be examined , includ ing both ex-pressed and nonexpressed regions.

Genomic Sequencing

The most straightforward app roach to lookingfor mutations is by determ ining the sequence of every nucleotide in a child’s genome and thencomparing th is with the DNA sequences of thechild’s parents. To determine its sequence, humanDNA is cut w ith restriction end onu cleases intofragments, and then each fragment is analyzed forits sequence of nucleotides. Genomic sequencingwou ld detect mutations regardless of where theyoccur — in regions that code for specific pr oteinsand regulatory functions as well as in regionswithout kn own functions—and is therefore poten-tially very informative. While it is technically p os-

sible at present, sequencing is currently feasibleonly for very limited sections of DNA, such asthe length of DNA comprising only a few genes.Because of current technical inefficiencies, it wouldbe an enormous task, involving many labora-tories, a large nu mber of scientists, and at leastseveral decades to sequence even one entire ge-nome, the complete set of DNA in an individual’sgerm cell. At present, it is not feasible to usegenomic sequencing to examine several peoples’genomes for m utations, although sequencing canbe used in conjunction w ith other techniques toexamine small sections of DN A.

One-Dimensional Denaturing GradientGel Electrophoresis

A m odification of the standar d electrophoreticgel procedure, d enaturing gel electrophoresis al-



12 qTechnologies for Detecting Heritable Mutations in Human Beings

lows DNA fragmen ts to be separated not only onthe ba sis of size, bu t also on th e basis of sequenceof nucleotides. Double-strand ed DNA separates(“denatures”) into its constituent strands w hen itis heated or when it is exposed to denaturingchemicals. A grad ient of increasing strength of such chemicals can be produced in an electro-

phoretic gel so that DNA samp les will travel inthe d irection of the electric current, separa te bysize, and begin to d issociate as th ey reach th eirparticular critical concentration of denaturingchemical.

Every unique strand of DNA d issociates at adifferent concentration of den aturant. In fact, asequence difference of only one nucleotide be-tween tw o otherw ise identical strands of dou ble-stranded DNA is enough to cause the strands todissociate at different concentrations of denatur-ant chem ical, and to stop traveling at d ifferent

locations in th e gel. Using this techniqu e, the par-ents’ and child’s DNA would be cut into fragmentswith restriction enzymes, dissociated into singlestranded DNA, and reannealed with r adioactivelylabeled probe DNA. The resulting heteroduplexfragments would then be separated on the basisof their DNA composition in a den aturing gra -dient gel. Mismatches between the sequences of probe and child’s DNA and not between probeand parents’ DNA would appear as differentband ing patterns on the gel. Again, a comp ari-son between the band ing pattern of parents’ andchild’s DNA analyzed in th is way m ay identify

a wide range of mutations in all DNA regions.

Two-Dimensional Denaturing GradientGel Electrophoresis

Another approach to detecting mutations is atechnique whereby sizing and denaturing gels areused to differentiate among DNA sequences com-mon to parents’ and child’s DNA, polymorp hismsin either p arent’s DNA wh ich are tran smitted tothe child, and any new mu tations in the child’sDNA. Like the two-dimensional polyacrylamidegel procedure for protein separation d escribedearlier, this approach compares locations of spotson a gel (in this case, DNA spots) for evidenceof new mutations. In this method, various com-binations of parents’ and child’s DNA are pro-du ced and compared on the basis of the dena-

turant concentration at which they dissociate.Parents’ and child’s DNAs are compared to eachother, rather th an to relatively small probes. Thisapp roach, wh ich w ould allow d etection of mu ta-tions in the complete genome, would detect differ-ences (“mismatches”) between a child’s DNA andhis or her parents’ DNA. Such mismatches would

represent various types of mu tations in the nu cleo-tide sequence in expressed and nonexpressed DNAregions.

DNA-RNA Heteroduplex Analysis

This technique hinges on the production of DNA boun d to comp lementary strand s of ribo-nucleic acid (RNA) or “DNA-RNA heteroduplexes,”and on the use of specific enzymes, such as“RNaseA,” that cleave the DNA at particular se-quences where the DNA and RNA are n ot per-fectly boun d at every nu cleotide. This is similar

to the u se of restriction enzym es, which cleaveDNA at particular normal sequen ces, except thatRNaseA cleaves the RNA strand in RNA/ DNAhybrid molecules where there are mismatchednucleotides, indicating mutations. The resultingfragments are then separated electrophoreticallyto detect differences between parents’ and child’sDNA. The efficiency of this approach depends onthe number of different-mismatches that can berecognized and cleaved. This method would de-tect nucleotide substitutions over a large portionof the DNA.

Subtractive Hybridization

Detecting mu tations would be much easier if it were p ossible to ignore the m ajority of DNAsequences that are the sam e in parents and childand, instead, focus only on the few sequences thatmay be different. Subtractive hybridization is anidea for selecting and characterizing sequences ina child’s DNA that are different from either par-ent’s DNA.

First, the double-stranded DNA of both par-ents is cut into fragm ents with restriction enzym es,dissociated into single-stranded DNA, and thenmixed together with a set of single-stranded refer-ence DNA sequences. These reference sequencesrepresent all possible sequences of 18 nucleotides(analogous to a dictionary of 18-letter words using




only 4 d ifferent letters. ) Each r eference sequen cebinds to its comp lementary sequence in the p aren-tal DNA and can be removed from the mixture.Any reference sequences left over, not bound toparental DN A, represent sequences not found inthe parents’ DNA. If the child’s DNA binds with

any of these left-over sequences, such hybridswould indicate that the child’s DNA containsdifferent sequences from those in the parents’DN A. These hybrids could be separated and ana-lyzed for mu tations.

This approach is the least well developed of allthe ones discussed in this report, and its feasibil-ity is unknown. If it does prove feasible, how-ever, this app roach wou ld identify short sequencescontaining mu tations in any p art of the DNA, al-lowing further d etailed stu dy (e.g., by DNA se-quencing) of the kinds of mutations that mayoccur.

Pulsed Field Gel Electrophoresis

If hu man DNA were short an d simp le, it couldbe cutup w ith restriction enzymes and separatedelectrophoretically into discrete bands, each rep-resenting a particular segment of the total DNA.How ever, hum an DN A is so long that when re-striction-digested DNA is electrophoretically sep-arated, the resulting fragments of the wh ole setof DNA form a continuou s smear of band s. Evenif the DN A could be cut into 100 or 200 fragments,the pieces would be too big to pass individually

through the pores of a standard electrophoreticgel. A new technique, pulsed field gel electropho-

resis is being developed to separate large frag-

ments of human DNA and to examine such frag-

ments for evidence of mutations. The proceduremay detect submicroscopic chromosome muta-tions, including rearrangements, deletions, breaks,and transpositions. At the present time, however,the method cannot handle whole human chromo-somes, though it works well with fragments of human chromosomes and with smaller wholechromosomes from lower organisms. This tech-nique may be useful in d etecting chromosome mu -

tations that are intermed iate in size between m a- jor rearrangements (observable by cytogeneticmethods) and single base pair changes, potentiallya large prop ortion of all possible mutat ions.

Detection and Measurement ofSomatic Cell Mutations

The method s for detecting h eritable mu tationsrely on comparing the DNA of parents with theDNA of their children to infer the kinds and rates

of mutations that p reviously occurred in parents’reproductive cells. While this information is val-uable in learning about heritable mutations, itmay come month s or years after the mu tationshave actually occurred, and this temporal sepa-ration of events m akes it difficult to d raw asso-ciations between mutations and their causes. Teststo detect somat ic mu tations m aybe useful in sig-naling the probability of heritable mutations. Suchtests may be useful in relating exposures to spe-cific mu tagens w ith p articular genetic events inthe cell, and they m ay help to identify ind ivid-uals and popu lations at high risk for mu tations.7

It is thought that the mu tation p rocess is fun-dam entally similar in germ inal and somatic cells.If this is true, then it may be possible to pred ictthe risk of germinal and heritable mutations onthe basis of measuremen ts of somatic mutations,wh ich are inferred from the frequency of mutantcells. Several investigators ar e curren tly work ingon m ethods to relate the frequency of mu tant cellsto the number and kinds of underlying mutations.

Several new techniques for detecting and meas-uring somatic mutations are described in thisreport. Mutant somatic cells may occur during

growth and development and may appear asrarely as one in a million normal cells. Detectionof somatic mu tants requires method s for scanningthrou gh a m illion or so nonm utan t cells to finda single variant cell. Two general ap proaches areused: 1) screening, which uses high-speed machin-ery to look at the total population of cells andeither count or sort ou t the varian ts; and 2) selec-tion, in which a population of cells is cultured inthe laboratory un der conditions that permit thegrowth of variant cells and that restrict the growthof the majority of cells that are nonvariants. The

7Even without knowing the exact relationship between rates and

kinds of somatic and heritable mu tations, it is reasonable to pre-dict that people with high somatic mutation rates might beat higherrisk for heritable mutations, either because of a particular environ-mental exposure or a genetic susceptibility to mutations.



14 Technologies for Detecting Heritable Mutations in Human Beings

new DNA technologies could be u sed to charac-terize the mutations in any such somatic variantsfound.

Different mu tagens have been shown to pro-du ce distinguishable types of mutations (“muta-tional spectra”) in hum an cells grown in culture.

Determining mutational spectra may be useful inassociating specific mutational changes in somaticcells with particular mutagenic agents to whichindividuals may be exposed. Such informationcould be useful in understanding the causes of

USE OF NEW TECHNOLOGIES IN

Feasibility and Validity Testing

The new technologies now range from ideas onpap er to being in various stages of laboratory de-velopmen t, but none is yet ready for use in thefield. A critical step before a technology is usedin an investigation of a pop ulation thought to beat risk for mutations is that the technology betested for validity and feasibility. To assure thata technology is a “valid” method , that is, that itdetects the types of mutations that it theoreticallyis capable of detecting, and to characterize the de-gree to w hich it gives the “right” answ er, it w illbe necessary to test a technology in a series of vali-dation stu dies. A first step might be to test thetechnologies against pieces of DNA with know n

mutations, and at a later stage, in offspring of ani-mals exposed to know n mu tagens. Feasibility test-ing will be required to m ake sure the technologycan be efficiently scaled up for analyzing largenumbers of samples.

Epidemiologic Activities

If the value of new technologies for detectingmutations is to be realized, it will be as tools fordetermining rates and patterns of mutations in epi-demiologic studies of human beings. Once a tech-nology has successfully passed through validation

and feasibility tests, it will become a candidatefor use in three major types of epidemiologicactivities: surveillance, monitoring, and ad hocstudies.

mu tational changes as w ell as in m onitoring at-risk populations.

Data on kinds and rates of somatic mutationsmay provide a monitor for exposure to mutagensand carcinogens, and are relevant to the study of carcinogenesis and of aging. How ever, measure-

ments of somatic mutation rates per se have lit-tle direct applicability to intergenerational (ortransmitted) effects, without corresponding infor-mation on heritable mutations.

RESEARCH AND PUBLIC POLICY

Surveillance is a rou tine activity wh ose aim, inthe context of this report, wou ld be to measu rethe “baseline” rate of mutations in a d efined p op-

ulation over the course of time and to facilitaterapid recognition of changes in those rates. Mon-

itoring consists of observations over time in a pop-ulation thought to be at increased risk of, in thiscase, heritable mutations, because of exposure toa known or suspected mutagen, for the purposeof helping the specific at-risk population in what-ever way is possible. People living around h az-ardou s waste d isposal sites have been m onitoredfor endpoints other than mutations, e.g., cancerand birth defects, and they w ould be likely can-didates for mutation monitoring when technol-ogies become available to do so efficiently. Ad

hoc studies of a variety of designs are carried outto test hypotheses about suspected causes of mu-tations. Ideally, the results of ad hoc studies canbe generalized to populations other than thosespecifically studied.

Extrapolation

Making predictions from observations of causeand effect in one system to probable effects inanother is one form of extrapolation. The p roc-ess involves a set of assump tions in moving fromone system to another. The p ractical imp ortance

of extrapolation for mutagenicity is, ideally, tobe able to predict mutagenic effects on human be-ings from the response in laboratory animals orlower test systems. The ability to extrapolate to




human responses addresses one of the major goals

of public health protection, the ability to iden-

tify substances harmful to human beings before

anyone is exposed, thereby providing a rational

basis for controlling exposure.

Extrapolation can be qualitative or quantita-

tive. Qualitative, also called biologic, extrapola-

tion involves predicting the direction of a result,for example, if a chemical causes mutations in alaboratory test, can we also expect mutations inhuman beings? Quantitative extrapolation in-volves translating a quan titative result in an ani-mal test into a qu antitative estimate of mutagenicrisk in h um ans. Going a step further in extrapo-lation, can an estimate of mu tagenic damage betranslated into a measu re of genetic disease?

A n um ber of theoretical mod els for extrapolat-ing mu tagenic effects have been prop osed, basedon various parallel relationships. For instance, itmight be tru e that if the relationship betw een so-matic and heritable effects in animals were knownafter exposu re to a sp ecific mu tagen, and if onecould measu re a somatic effect in hu man beingswho had been exposed to the same substance, aheritable effect in human beings could be pre-dicted, assum ing the relationship between som aticand heritable effects is parallel in animals and hu-man beings. Because of the paucity of data, par-ticularly from hum an stud ies, it has been imp os-sible to validate such an extrapolation model. Thenew technologies should allow a m ajor increasein the database wh ich, in turn, should allow re-searchers to more fully explore relationshipsamong various typ es of test results.

Regulation

Congress has mandated public protection frommu tagens in certain environmental health laws

(e.g., TSCA and CERCLA), and other laws pr o-vide mandates broad enough to emp ower agen-cies to take action against mutagens in virtuallyany appropriate situation. Except for radiation,however, very little regulatory evaluation has

taken place on the subject of heritable mutations.

This is directly related to the lack of sensitivemethods to detect heritable mutations in humanbeings, and the related difficulty in extrapolatingfrom results in nonhuman test systems to prob-able human responses.

The Environmental Protection Agency (EPA)has recently issued “Guidelines for MutagenicityRisk Assessment. ” EPA’s app roach is relativelysimple and pragm atic. It requires only the typesof information that can be acquired w ith currenttechnologies, but allows for information from newtechnologies, as they become available. Theguidelines require evidence of: 1) mutagenic activ-ity from any of a variety of test systems, and 2)chemical interactions of the mu tagen in th e mam -malian gonad . Using a “w eight-of-evidence” d e-term ination , the evidence is classified as “su ffi-cient, ” “ suggestive, ” or “limited” for predictingmutagenic effects in human beings.

Although chemicals have not generally beenregulated as mutagens, it is probable that ex-posures to mutagens have been redu ced by regu-lations for carcinogenicity. Strong evidence sup-ports th e idea that a first stage in many cancersinvolves mu tation in a som atic cell, and one of the most w idely used screening tests for poten-tial carcinogens, the Ames test, is actually a testof mutagenicity. The extent to which people areprotected against heritable mutations if their can-cer risk from a specific agent is minimized is atpresent unknown. The new technologies shouldgreatly improve ou r ability to make that jud gment.

Federal Spending forMutation Research

OTA queried Federal research and regulatoryagencies about their supp ort of research directedat un derstand ing hum an m utations. For fiscal year1985, they repor ted about $14.3 million spent on

development or applications of methods for de-

tecting and/ or counting hu man somatic or herita-

ble mutations. An estimated $207 million was

spent in the broader category of related genetic

research.




OPTIONS

Research related to the new technologies de-scribed in this report has the dual aims of increas-ing the knowledge base in hum an genetics aboutthe causes and effects of mutations, and p rodu c-ing information that could be used to estimate

mu tagenic risks for the p urp ose of protecting pu b-lic health. The pace and direction of researchtoward developing these methods an d the qual-ity and efficiency of preliminary testing of meth-ods could be influenced by congressional and ex-ecutive branch actions and priorities. Integrationof research in different test systems and progressin developing extrapolation models also can beinfluenced by actions now and in the near future.

Continued progress in the development and ap-plication of new technologies will depend not onlyon support of the individuals and laboratories

directly involved in this research, but also onwork in other areas. Although not directly ad-dressed in this assessment, support of research inmedical, human, mammalian, and moleculargenetics will be essential to a full understandingof the mutation p rocess.

The options that follow are group ed in four sec-tions: 1) options to influence the development of new technologies, 2) options to ad dress v ariousaspects of feasibility and valid ity testing, 3) op-tions concerning the use of new technologies infield stud ies, and 4) options to encour age coordi-nation of research and validation of extrapolationmodels.

Development of New Technologies

The Department of Energy (DOE), severalagencies of the Department of Health and HumanServices, and EPA currently provide funding toresearchers in indepen dent and government lab-oratories for various research and developmentactivities pertaining to hu man mu tation researchand the d evelopm ent of new technologies. Eachagency, approp riately, proceeds dow n a slightly

different path. OTA estimates that a total of about$14.3 million w as spent on hum an mutation re-search in fiscal year 1985. Progress in this researchcould be speeded u p by increased fun ding, thoughit is difficult to quantify an expected gain. It is

clear also that less money sp ent on hum an m uta-tion research will slow progress in laboratories al-ready engaged in this research, and could deterother scientists from pursuing research in thisfield.

Option I : Congress could assure that funding

levels for human mutation research and closely

related stu dies d o not decline w ithout the re-

sponsible agencies assessing the impact of fund-

ing cuts on research progress. This requirement

could be expressed in appropriation, authori-

zation, or oversight activities.

Of the several funding agencies, DOE has takenthe lead in funding m uch of the research on newmethod s for mutation d etection described in thisreport, and researchers at some of DOE’s NationalLaboratories are among the leaders in the re-

search. DOE also is the agency responsible forfunding U.S. participation in the Radiation EffectsResearch Foundation, the joint United States-Japan body that continues to study the h ealth of Japan ese atomic bomb su rvivors.

Option Z : A lead agency for research related to

detecting and characterizing mutations could

be des ignated. The lead agency w ould be re-

sp onsible for tracking the d evelopm ent of new

techn ologies , facilitating th e interchan ge of in-

formation among scientists developing the tech-

nologies and those in related fields, and en-

couraging and facilitating coordinated studiesinvolving different subdiscipline. The lead

agency a lso w ould k eep Congres s informe d

about activities in this area. DOE may be the

logical choice for a lead agency.

The types of activities that a lead agency mightengage in are described below. These activities areimportant w hether or not a lead agen cy is desig-nated, and Congress should consider d irectly en-couraging them if there is no lead agency.

Tracking the Developm ent of

New TechnologiesAll the new technologies require improvements

in efficiency before th ey become useful tools forstudying human beings. As research proceeds,some techniques w ill develop m ore quickly than




others, some will be dropped, and some maychange in character, altering th e kinds of muta-tions they can d etect. The lead a gency wou ld beresponsible for keeping track of these devel-opments.

It would be useful if the “tracking” responsi-bility could lead to actions on the pa rt of the leadagency that would promote the rap id and efficientdevelopmen t of the technologies. At some p oint,the techn ologies will be read y for feasibility test-ing and eventually, field testing. It would help re-searchers to know the stages of development of various methods, and it could h elp agencies makedecisions about funding studies using certainmethod s. In ad dition, the lead agency could as-sist researchers by anticipa ting need s that w ill becommon to all research programs an d by encour-aging efficient u se of resources.

Facilitating Information ExchangeAmong Researchers

In 1984, DOE organized a nd fund ed a m eetingthat brou ght together for the first time man y of the researchers involved in laboratory-based mu-tation research. This meeting is acknowledgedamong th ose who attend ed as a m ilestone for in-formation exchange and the generation of newideas concerning detection of heritable mutations.In fact, the ideas for some of the new method sdescribed in this report w ere born at that meet-ing. There is a continuing n eed for this type of information exchange.

Keeping Congress Informed

Congress has already directed regulatory agen-cies to redu ce exposures to environmental agentsthat m ay cause mu tations. As part of its oversightof both regu latory and research activities, Con-gress could benefit from up-to-date informationabout the development of various technologies.The lead agency could report in a specified m an-ner, e.g., a brief annual report describing the levelof current research, its goals, results of completedor ongoing work, and expected near-term d evel-

opments. This information could also be the ba-sis for informing the p ublic about mu tation re-search. This activity will continue to be valuableto Congress through later phases of developmentand app lication of new technologies.

Feasibility and Validity Testing

After technologies pass throu gh a phase of de-velopment to improve their efficiency and to work out technical details, a period of feasibility andvalidity testing w ill be necessary before a tech-nology can rationally be u sed as a tool in a large-

scale study of heritable mutations in human be-ings. There are some w ays to make validity test-ing an efficient process. As an example, a “DNAlibrary, ” a collection of known DNA sequences,particularly sequences carrying known mu tations,could be established and maintained by on e lab-oratory, which could make DNA sequences avail-able to all researchers developing mutation detec-tion technologies. This material could be used todetermine w hether a particular technology d etectsthose mu tations that it is designed to detect, anal-ogous to testing chemical procedur es and equip-ment against know n chemical “standar ds. ” At a

more ad vanced stage, new technologies might alsobe tested in animal experiments with a selectionof mutagens kn own to cause specific types of mu-tations. If a lead agency is designated, some of these options would logically be among thatagency’s responsibilities. If there is n o lead agency,these functions could still be encour aged by Con -gress through oversight activities.

Banking Biological Samples

Biological samples, especially blood samples,are often collected during the course of medical

examinations for people thought to be exposedto environmen tal or occup ational agents. Imp le-mentation of a plan to bank those samp les wouldfacilitate hum an mu tation stud ies and other re-lated research.

Examining stored samp les is not n early so dis-rup tive as collecting samp les from a p opu lation.The very act of specifically collecting samples fora stud y of mutations wou ld raise expectations thatdefinitive information about risks would be forth-coming. Examining stored samp les would a voidthat human cost and could, at the same time, pro-

vide a realistic test of new techn ologies before theyare app lied to people w ho are anxious about theeffects of environmental exposures on their genes.

Although stored samp les offer many advan-tages, prepa ring hum an sam ples for storage and




maintaining the stored sam ples is a significanttask. Currently the high cost of storing sam plesinhibits establishment of sample banks. There hasbeen little research d irectly aimed at imp rovedmethod s for storing biological samp les.

Option 3: Congress could encourage banking of

biological material obtained from at-risk indi-

viduals and their spouses and offspring, with

the objective of studying somatic and herita-

ble mutations from these stored samples before

technologies are used directly in a study of an

exposed population. A centralized samples

ban k a vailable to all resea rchers w ould be the

most efficient means of establishing a reposi-

tory of sufficient size and scope for validity

test ing.

Option 4: Congress could encourage the appro-

priate agencies (through the lead agency , i f one

is d es ignated ) to inves t igate the potential for im-

provng techn ologies for storing biological s am - ples. Where possible, funding should be en-

couraged for such improvements. This option

is relevant to a broad sp ectrum of hu m an h ealth

research, and collaboration among disciplines

should be encouraged for determining the spe-

cific storage needs of different lines of research.

Field Studies

Sometime in the n ext 5 to 10 years, it is likelythat one or more of the technologies for stud y-

ing human mutations will be brought to the pointof readiness for ep idemiologic stud ies of humanbeings thought to be at risk for mutations. Sucha stud y could be initiated by an ind epend ent sci-entist or a group of scientists involved in the de-velopment of a new technology who have iden-tified an at-risk population; a new technology ortechnologies could be used in an environmentalhealth investigation triggered by the discovery of a potentially at-risk population; or Congress couldman date a stud y of mutations in a specific pop u-lation. Depending on the way a stud y is initiated,

different agencies will be draw n in, and differentfund ing mechanisms w ill come into p lay.

Should a stud y be initiated by on e or a groupof investigators w ho subm it a grant p roposal forfunding, the current system for review of researchawar ds is probably app ropr iate for making sci-

entific jud gments abou t wh ether a method is readyfor large-scale testing. From that point on, how-ever, there are significant differences between aprop osal to use a new technology in a human pop -ulation and most other p roposals. The major dif-ferences are size and money. Ap plication of anypromising technology will require that state-of-

the-art methods in molecular biology be expandedfrom sm all-scale laborator y u se to large-scale ex-amination of blood samp les collected from hu n-dreds or thousands of people.

A study such as that described above wou ld re-quire a significant comm itment of fund s over aperiod of years, and could account for a large per-centage of a granting agency’s funds. The amountof money necessary can only be estimated, butit could am oun t to tens of millions of dollars, nota large amou nt in government spend ing, but largein comparison to most biomedical research grants,which usually range from less than $100,000 to

abou t $500,000 ann ua lly. The necessary comm it-ment of funds from any single agency, consider-ing current spending for this type of research, islikely to be impossible, no matter h ow worthw hilethe proposed study.

Option 5: Congress could consider providing spe-

cific add-on funding to finance an investigator-

initiated realistic test of a promising method to

study human heritable mutations.

A study mandated by Congress or an agency-initiated stu dy u sing new technologies will not

necessarily undergo the rigorous review and crit-icism that a proposal submitted to a grantingagency would. Congress already has some experi-ence in m and ating epidem iologic stud ies. Con-cerns about cancer and birth d efects convincedCongress to mand ate studies of military veteranswho w ere exposed to Agent Orange or w ho par-ticipated in atomic bomb tests. In addition, Con-gress has vested the Public Health Service withthe auth ority to carry out a w ide range of healthinvestigations of exposures from toxic wastedu mp s through “Superfund .” The role of the man-dated “Agency for Toxic Substances and Disease

Registry” (ATSDR) is to “effectuate and imp le-ment th e health related au thorities” of Sup erfund .ATSDR, located within the Centers for DiseaseControl, is a logical place for new technologiesto be used as exposed pop ulations are identifiedthrough other provisions of Superfund.



Ch. l—Surmmary and Options q19

An un usua l array of health problems or an ap -parent excess of disease among people livingaround toxic waste sites could trigger a stud y byATSDR. Such problems could also prompt localresidents to press Congress for studies to d eter-mine w hether, among other things, the disordershad been caused by mutations, and whether these

mu tations could have resulted from chemicals inthe w aste site.

Some scientific societies, such as th e Am ericanSociety of Human Genetics, could be asked topa rticipa te in a review of the feasibility of detect-ing possible health effects from environmental ex-posu res. If the stud y is deter min ed to be feasible,Congress and the p ublic can be reassured that thestudy’s findings are likely to be useful in decision-making. Alternatively, a decision that no stu dyis currently feasible would und erline the imp or-tance of developing and testing n ew m ethods. Re-

viewing th e feasibility of a stud y may be perceivedas d elaying a n investigation u nnecessarily. How-ever, performing a study that has little or no likeli-hood of answering questions about exposure andmu tations has marginal value at best and w ouldnot serve the people wh o requested it.

Option 6: If Congress mandates a study of heri ta-

ble mu tations in an a t-risk population us ing the

new technologies, the mandate should include

a feas ibility a ss ess men t by a pa nel of experts

before the s tudy begins .

Option 7: Congress could require agencies to plan

for a rigorous ou tside review by a pa ne l of ex-

perts of any agency-init iated study using a new

technology, before such a study can proceed.

Integration of Animal andHuman Studies

Much of our current ability to estimate the ef-fects of various external agents on hum an beingsis derived from animal stud ies. In th e future, ani-mals w ill continue to be u sed to test for mu tage-nicity and, ideally, to predict effects in human be-ings. Right now , the available methods a nd bodyof data provide an inadequate basis for makingpredictions from animal results to human effects.The new techn ologies described in this repor t fordetecting h eritable and somatic mutations can beapplied in both animal systems and in human be-ings. This inform ation and information from cur -

rently available animal systems could be u sed inan integrated fashion to study relationships be-tween somatic, germinal, and heritable m utations,and to pursue the development of extrapolationmod els for p redicting effects in hu man beings.

The kind of integrated research necessary w ill

not occur spontaneously if the current patternholds. There are few researchers engaged in studiesthat integrate comparable information from differ-ent systems for the purp ose of elucidat ing rela-tionships am ong such systems. Imp rovements inextrapolation from animal to hum an d ata andfrom somatic to heritable mutations await effortsto encourage and coordinate the appropriate re-search. Single laboratories or centers a re u nlikelyto be able to perform all the different tests n eces-sary to develop and test extrapolation models,making coordination among laboratories essential.

The Na tional Toxicology Program (NTP) is thecenter of Federal mu tation research using experi-mental animals, funding research grants and con-tracts nationally and internationally. The NationalCenter for Toxicological Research, a laboratoryof the Food and Drug A dm inistration that is partof NTP, has facilities and experience to carry ou tlarge-scale animal tests. EPA also has a genetictoxicology pr ogram, and it is actively pu rsuingdevelopment and application of methods forstudying human somatic mutations. DOE’s Na-tional Laboratories have d irected or carried outthe majority of large-scale animal studies of mu-

tation rates and mechanisms, and this experience,as well as the ad vanced technology that these lab-oratories have developed for sorting and study-ing chromosomes and cells, will be useful in im-proving methods for extrapolation. These threeagencies could encourage the development of methods for linking information from animalstudies to somatic and heritable mu tation risks inhumans.

Option 8: Congress could en courage s tudies of so-

matic, germinal, and heritable mutations in ex-

perimental animals using both currently avail-

able an d new technologies . Further, rese arch

funding agencies should encourage animal

s tud ies d irec ted a t ident ify ing the m echanisms

of mutagenesis and elucidating relationships be-

tw een m utagen ic potencies in animals an d in

h u ma n b e in g s .



Chapter 2

Genetic Inheritanceand Mutations



Contents

Page

Function and Organization of DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......23

Protein Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Inheritance of Genetic Traits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......27Kinds of Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Gene Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Chromosome Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......29

Health Effects of Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......30

Early Mortality and Morbidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......31

Diseases During Adult Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......31

Effects of Somatic Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......31

Persistence of New Mutations in the Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

FIGURES

Figure No. Page

2. The Structure of DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3. Replication of DNA... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4. Diagrammatic Representation of Protein Synthesis According to the

Genetic Instructions in DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265. The Genetic Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

6. Chromosomal Translocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30



Chapter 2

Genetic Inheritance and Mutations

“Like begets like” is an expr ession that d escribesrepeated observations in families. Except in rareinstances, most children are norm al and gener-ally resemble their biological pa rents, brothers,and sisters. In these cases, the genetic informa-tion has been passed inta ct from the parents totheir children.

This report is about the instances when some-thing goes wrong, wh en a genetic mistake is made;a mu tation in the DNA of the parents’ reprod uc-tive cells, both of w hom are p hysiologically nor-

FUNCTION AND ORGANIZATION

Deoxyribonu cleic acid, DN A, is the carrier an dtransmitter of genetic information. Its functionsrequire that it is relatively stable and that it is ableto produce identical copies of itself. Certainagents, such as some forms of radiation, viruses,and certain chemicals, alter DNA’s ability tomaintain these characteristics. Mutations, changesin the composition of DNA, can occur as a resultof these agents acting on DNA or on the systemsin place to repair DNA damage.

Human DNA can be thought of as an immenseencycloped ia of genetic information; each person ’sDNA contains a un ique compilation of roughly3 billion nu cleotides in each of its two chains, theexact sequence shared by no other p erson exceptan identical twin. The genetic information en-coded in DNA is contained in the nucleus of cells,and is packaged in u nits, or chromosomes, thatconsist of long tw isted d ouble strands of DNA sur-round ed by a comp lex of proteins. Each chromo-some contains thousands of genes, the functionalunits of DNA, which are “read,” or transcribed,by the cell so that genetic information can be usedto make proteins or to regulate cell functions. It

is thought that only a sm all proportion, app rox-imately 1 to 10 percent, of human DNA is trans-lated into sp ecific proteins. Functions of the non-translated majority of DNA are largely unknown.There are 23 pairs of chromosomes in all nucleated

real, is passed to a child wh o may or m ay not ap-pear norm al, depend ing on the nature and severityof the mutation. All types of mutations, from thebenign to the severe, are considered in this report.As an introduction to human genetics for non-specialists, this chapter summarizes some basic in-formation about normal functions of DNA, thevarious kinds of mutations that can occur, andhealth effects of mu tations. More comp lete infor-mation can be found in a general reference suchas Vogel and Motulsky (165).

OF DNA

human cells except human reproductive cells(germ cells). The latter contain on e of each p airof chromosomes, since two germ cells (one eggand one sperm ) fuse at conception and re-createa full set of DNA in the offspring.

DN A is composed of two chains of nucleotidesbound together in a d ouble helical structure (fig.2). The backbones of the tw o chains, formed bysugar (deoxyribose) and phosp hate m olecules, areheld to each other by hydrogen-bonded nitroge-

n ou s b a s e s : adenine, thymine, guanine, and cyto-

sine, abbreviated A, T, G, and C, respectively.Each of the chains’ units, or nucleotides, consistof a sugar, phospha te, and nitrogenous base. Thelinear sequence of nu cleotides repeated in vari-ous combinations thousand s of times along theDNA d etermines the particular genetic instruc-tions for the pr odu ction of proteins and for theregulation of cell functions.

A chang e in even a single nu cleotide, e.g., sub-stituting G for T, among th e 3 billion n ucleotidesin the DN A sequence, constitutes one type of mu -tation. Depending on wh ere it occurs, such a m u-

tation could be sufficient to cause significantpa thological changes an d d ysfunction, or it couldcause no imp airment at all. In thalassemia, forexample, a genetic disorder of hemoglobin syn-thesis, appr oximately 40 different “spelling er-

23




Figure 2.—The Structure of DNA

AdenlneGuanlne 0

0-0’

3

5

0 .

~

22a, The pairing of the four nitrogenous bases of DNA: o

Adenine (A) pairs with Thymine (T)Guanine (G) pairs with Cytosine (C)

o7

-O” p

Y

d

2 H2

, o–

‘ 0

; H2

,o– \

CH2

o -

‘ 0

; H2

)sugar-phosphate

backbone

2b. The four bases form the four letters in the alphabet ofthe genetic code. The sequence of the bases along thesugar-phosphate backbone encodes the genetic in-formation.

e pairs

phosphateone

A schematic diagram of the DNA double helix. A three-dimensional representation of the DNA double helix.

2c. The DbJA molecule is a double heli x composed of two chains. The sugar-phosphate backbones twist around the out-side, with the paired bases on the inside serving to hold the chains together.




Ch. 2—Genetic Inheritance and Mutations q 2 5

rors, ” or (single or mu ltiple) nu cleotide cha nges,have been identified in a par ticular gene tha t speci-fies the p rod uction of globin, an essential constit-uent of hemoglobin. In complex diseases such as

heart disease, susceptibility is thought to be in-fluen ced by the intera ction of environm ental fac-tors such as d iet, and genetic factors such as m u-tations in one or several genes.

During DN A replication, each chain is used asa template to synthesize copies of the originalDNA (fig. 3). New mu tations are tran smissibleto daughter cells. To replicate, the two chains of the DN A d ouble helix separate between nu cleo-tides, and each is copied by a series of enzymesthat insert a complementary nitrogenous base op-posite each base in the original strand , creatingtwo identical copies of the original one. As a re-sult of specific pairing between nitrogenous bases,each chain builds its complementary strand using

the single chain as a guide.1

Protein Synthesis

The normal functioning of the hu man body—includ ing the d igestion of food; the p rodu ctionof muscle, hair, bone, and skin; and the fun ction-ing of the brain and nervous system—is depen-dent on a well-ordered series of chemical re-actions. Proteins, especially the enzymes, promotethe chemical reactions on which these processesare based. Each enzyme has a specific function;it recognizes a chem ical, attaches to it, reacts w ith

it, alters it, and leaves, ready to promote the samereaction again w ith another chemical.

To synthesize a prot ein, the genetic informa-tion contained in one or more genes along theDN A is transcribed to sm all pieces of ribonu cleicacid (RNA), which are faithful replicas of onestrand of DNA (see fig. 4). Each stran d of RNAmoves out of the nucleus to the cytoplasm of thecell, where it serves as a template for proteinsynthesis. Proteins are comp osed of hun dred s of linked subun its called amino acids. At the ribo-somes, RNA is used to gather and link amino

acids in a specific sequence and length accordingto the d esign sp ecified in the DN A to form thedifferent proteins.

‘The molecular structure of the nitrogenous bases in DNA requiresthat A pairs only with T (or with U in RNA) and G pairs only with C.

Figure 3.–Replication of DNA

Old Old

Old New New OldWhen DNA replicates, the original strands unwind and

serve as templates for the building of new complementary

strands. The daughter molecules are exact copies of theparent, with each having one of the parent strands.

SOURCE: Office of Technology Assessment

Genetic information in DN A is organized in atriplet code, or sequ ence of three nucleotides. Eachtriplet specifies a complementary codon in “mes-senger RNA” (mRNA) which, in turn, specifiesa particular amino acid. The sequence of such

triplets in a gene ultimately d etermines the aminoacid sequence of the corresponding protein.Codons exist for each of the 20 amino acids thatmake up the m yriad of different pr oteins in thebody. Additional codons exist to signal “start” and




Figure 4.—Diagrammatic Representation of Protein Synthesis According to the Genetic Instructions in DNA

1. TRANSCRIPTION

3’ DNA 5’

A A A T T C T C G C A C A C C C T T C G

5’ 1Nuclear ~ ~ ~

RNAU U U A A G A G C G U G U G G G A A G C

IRemove introns,

2. RNA PROCESSINGprocess ends

5’ 3,

U u u G A G G U G A A G

Transferto I Nucleus

Nuclear membrane

Maturemessenger

RNA(mRNA)

5,i 3’

U U U G A G G U G A A G

3. TRANSLATION mRNA 5’ I

transfer RNA

.. / ”

TRibosome

Protein polypeptide chaincomposed of linked amino acids

SOURCE: Adapted from A.E, H, Emery, An /nfroduction fO Recombinant DNA (Chichester: John Wiley & Sons, 1984)



Ch. 2—Genetic Inheritance and Mutations • 27

Figure 5.—The Genetic Code

Codon Amino Acid Codon Amino Acid Codon Amino Acid Codon Amino Acid

UuuEach codon, or triplet of nucleotides

Phenylalanine Ucu Serine UAU Tyrosine UGU Cysteine in RNA, codes for an amino acid (AA).UUC Phenylalanine UCC Serine UAC Tyrosine UGC Cysteine Twenty different amino acids are pro-U UA L eu cine UCA Serine UAA atop UGA stopUUG Leucine UC G Serine UAG

duced from a total of 64 differentRNAstop UGG Tryptophan codons, but some amino acids are

Cuu Leucine Ccu Proline CAUHistidine

CGUArginine

specified by more than one codon (e.g.,

CuC Leucine ccc Proline CAC Histidine CGC Arglnine phenylalanine is specified by UUU andCUA

by UUC). In addition, one codon (AUG)Leucine C CA P roline CAA Glutamine CGA Arginine

CUG Leucine CCG Proline CAG GIutamine CGG Argininespecifies the ‘(start” of a protein, and 3codons (UAA, UAG, and UGA) specify

AUU Isoleucine ACU Threonine AAU Asparagine AGU Serinetermination of a protein. Mutations inthe nucleotide sequence can change

AUC isoleucine AC C Threonlne AAC Asparagine AGC Serine the resulting protein structure if theAUA Isoleucine ACA Threonine AAA Lysine AGA ArginineAUG Methionine ACG Threonine AAG Lysine AG G Arginine

mutation alters the amino acid speci-fied by a triplet codon or if it alters the

(start) reading frame by deleting or adding a

GUUnucleotide.

Valine GC U Alanine GAU Aspartic acid GG U GlycineGU C Valine GCC Alanine GAC Aspartic acid GG C Glycine U - uracil (thymine) A - adenineGUA Valine GCA Alanine GAA Glutamic acid GG A Glyclne C - cytoslne G. guanineGUG Valine GC G Alanine GAG Glutamic acid GGG Glycine

SOURCE Off Ice of Technology Assessment and National Institute of General MedicalSciences

“stop” in the construction of a polypeptide (pro-tein) chain. The triplet code p rovides m ore infor-mation, however, than is needed for 20 amino

acids; all except 2 of the amino acids are speci-

fied by more than one codon (see fig. 5).

Inheritance of Genetic Traits

A child’s entire genetic endowment comes fromthe DNA in a single sperm an d egg that fuse atconception. The fertilized egg then divides, be-coming a multicellular embryo, and the cellsdifferentiate into specialized tissues and organsduring embryonic development. In terms of genetic material, there are tw o general types of cells: somatic cells an d germ inal cells. Those tha thave a full set of DNA at some time in their d e-velopment and do not participate in the transmis-sion of genetic material to future generations ares oma tic cells, which include all cells in the bodyexcept the reproductive cells. Germ cells includethe gam etes (egg and sperm, each of which con-tains half of the total set of DNA) and the celltypes from w hich the gametes arise.

A norm al, full set of hu man DNA in som aticcells consists of 46 chromosomes, or 23 pairs of

homologous chromosomes: 22 pairs of “auto-somes” and 1 pair of sex-determ ining chromo-

somes. Females have 22 pairs of autosomes andtwo X chromosomes, while males have 22 pairsof autosomes and an X and a Y chromosome. Re-productive cells, with half the normal set of DNA,are produced by reduction division, or meiosis,from sex-cell progen itor cells in the female ovar yor male testis. These progenitor cells are set asideearly in fetal developm ent and are destined to bethe only source of germ cells. Each ovum or spermhas 23 single, unpaired chromosomes: ova have

22 autosomes and one X chromosome; sperm

have 22 autosomes and one X or one Y chro-mosome.

Most genes are present in two copies, one on

each member of a homologous pair of chromo-

somes. For instance, if gene A is found on chro-mosome 1 it w ill be present at the sam e locationor locus on both copies of chromosome 1. Theexact nature of the DNA sequence of gene A maydiffer between the two chromosomes, and thew ord allele, classified as either normal or mu ta n t ,

is used to refer to different forms of the same gen e.Many different mutant alleles are possible, eachone defined by a particular change in the DNAsequence. A m utation that is not expressed w hen

a norm al gene is present at the sam e locus on th esister chromosom e is called recessive and a mu-tation that is expressed even in the p resence of a normal gene is called dominant.




A Normal Karyotype of 46 Human Chromosomes

7 8 9 )1 12

13

Photo credit: Gail Stettt?n, Johns Hopkins Hospital

Karotypes typically show all the chromosomes of a single cell, and are used mainly to identify abnormalities in the numberor structure of chromosomes. Chromosomes can be isolated from any nucleated cell type in the body, but are mostoften isolated from white blood cells derived from a sample of venous blood or from amniotic fluid cells obtained atamniocentesis. The cells are stimulated to grow and divide in the laboratory, yielding a large number of cells whichcontain a sufficient amount of DNA to isolate and examine. The cells are then blocked in their dividing phase by theaddition of a chemical, such as colcemid, to their growth media. After harvesting the cells, they are treated with a hypotonicsolution to cause them to swell, and then with a chemical fixative to maintain this fragile, swollen state. When smallamounts of the cell suspension are carefully dropped onto microscope slides, the cells break open, spreading out theirchromosomes. To visualize the chromosomes and to distinguish the different homologous pairs, the chromosomes arestained to produce a characteristic banding pattern for each of the 23 pairs of homologous chromosomes. The chromosomescan then be photographed under the microscope, cut out from the final print, and numbered and arranged in order of

decreasing size as shown in this karyotype.



Ch. 2—Genetic Inheritance and Mutations q 2 9

KINDS OF MUTATIONS

Mutations are changes in the composition of

DNA. They may or may not be manifested in out-

ward appearances. Some mutations, depending

on wh ere they occur in the DNA, have no appar-

ent effects at all, and some cause changes that aredetectable, but are without obvious imp airment.Still other mutations lead to profound effects onhum an h ealth and behavior: stillbirths, neonataland early childhood deaths, various diseases,severe physical impairments, and mental defi-ciencies.

If mutations occur in genes that determine thestructure or regulation of essential proteins, theresults are often detrimental to health. Since mostDNA in hum an cells is not expressed or u sed inany know n w ay, mu tations in these regions mayor may not be clinically app arent, although it ispossible that subtle physiological variations mayoccur. Interactions among several such mutationsmay result in unusual responses to environmentalstimuli or they may influence susceptibility tovarious chronic diseases.

Mutations can be generally divided a ccordingto size into two gr oup s: gene mu tations and chro-mosome mutations. Chromosome mutations in-clude numerical and structural abnormalities.

Gene Mutations

Gene mutations occur within or across a sin-gle gene, resulting from the substitution of onenu cleotide for another or from the rearrangements(e.g., deletions) within the gene, or they maydu plicate or d elete the entire gene. The mu tationresponsible for sickle cell anemia, for example,is a point mutation: a single nucleotide in the genecoding for globin, a constituent of hem oglobin,is substituted for another nucleotide, resulting inthe substitution of one amino acid (glutamic acid)for another (valine) at a certain position in theglobin chain. One of the mu tations responsiblefor a form of alpha-thalassemia, another form of

anem ia, involves the d eletion of an entire alpha-globin gene, resulting in the absence of alpha-globin synthesis.

Since pr oteins are prod uced from the instruc-tions in genes, a mutation in a gene that codesfor a specific protein may affect the structure, reg-ulation, or synthesis of the p rotein, A mutation

that results in a d ifferent codon m ay or m ay notchange the resulting protein, since differentcodons can code for the sam e am ino acid. Alter-natively, such a mu tation m ay result in the sub-stitution of one amino a cid for another, and pos-sibly change the charge d istribution and structureof the protein. Another mu tation in the same genemay result in gross redu ction or complete loss of activity of the resulting protein. Detailed analy-sis of the m utations un derlying thalassemia d is-ease has shown that mu tations occurring in th ebordering areas, not within the translated partsof the gen e, can also imp air gen e function (171).

Chromosome Mutations

Major changes affecting more than one gene arecalled chromosom e m utations. These involve theloss, addition, or displacement of major parts of a chromosome or chromosomes. They are oftenlarge enough to be visible und er the light m icro-scope as a change in the shap e, size, or stainingpattern (“banding”) of a chromosome. Majorchromosome a bnormalities are often lethal or re-sult in conditions associated with reduced lifespanand infertility.

Structural Abnormalities

Structural mutations in the chromosomeschange the arran gement of sets of genes along th echromosom e. Sections of one chrom osome, in-cluding many genes, may break and reattach toanother chromosome (e.g., in a balanced trans-location, shown in fig. 6) or another location on

the same chromosome. Alternatively, sections of

chromosomes can be deleted, inserted, duplicated,

inverted.

Numerical AbnormalitiesNumerical chromosomal mutations increase or

decrease the number of whole chromosomes,




Figure 6. —Chromosomal Translocation

Normal Translocationchromosome 8 chromosome 8q –

ISites of

breakage


HEALTH EFFECTS OF MUTATIONS

In Western countries, genetic diseases are morevisible now than in the earlier part of this cen-tury. Advances in medical care, public healthmeasures, and living conditions have contributedto a grad ual d ecline in the contribution of envi-ronmental factors to certain types of diseases,particularly, infectious d iseases and nu tritionaldeficiencies (32). Genetic disorders as a g roup now

represent a significant fraction of chronic diseases

and mortality in infancy and childhood, althougheach disorder is individually rare. Such disorders

generally impose heavy burdens, expressed inmortality, morbidity, infertility, and physical andmental handicap. Approximately 90 percent of them become clinically app arent before pu berty.

The incidence of genetic disease is not knownprecisely, but it is estimated that genetic disordersare manifested at birth in app roximately 1 per-cent of liveborn infants, and are thou ght to ac-count for about 7’ percent of stillbirths andneonatal deaths and for about 8.5 per cent of child-

hood deaths. Nearly 10 percent of admissions topediatric hospitals in North America are reportedfor genetic causes. The majority of these geneticdiseases lack effective treatm ent. A sm all, but in -creasing number are avoidable through prenataldiagnosis and selective abortion (21,32,40).

without changing the structure of the ind ividu alchromosomes. Errors in the production of thegerm cells can lead to abnormal numbers of chro-mosom es in the offspring. For example, Downsyndrom e can be caused by the presence of an ex-tra chromosome (Trisomy 21); three copies of chromosome 21 are present w hile all other chro-

mosomes are p resent in normal pairs, giving a to-tal of 47 chrom osomes in each cell. A similar, butopposite, error in production of the germ cells re-sults in the lack of one chromosome. Turner syn-dr ome (Monosom y X), in w hich affected femaleshave 44 autosomes and only one sex chromosom e,an X chromosome, results in abn ormal d evelop-ment of the ovaries and in sterility. The effectsof these mutations may result not from alterationin the nature of the gene products but from theabnormal amount of gene prod ucts present.

In most o f these cases, the p articular genetic dis-order is passed along in fam ilies, or the asym p-toma tic carrier state for the d isease is passed alongfor many generations, before it maybe expressedin a child with th e disord er. In som e of these cases,however, the disorder may sud denly appear in achild wh en no relatives have had the disorder. Notun commonly, such an event can be explained bymistaken parentage or by one of the parents show -

ing mild or almost unnoticeable signs of the dis-order, so that th is parent has a disease-causinggene that can be passed on to h is or her children.H owever, the sud den, unexpected appearan ce of a new condition in a family can also result froma new m utation in th e reprod uctive cells of theparents. Examples of conditions that result fromnew mu tations in a large percentage of cases in-clud e Duchenne m uscular dystroph y, osteogenesisimp erfect (a cond ition resu lting in brittle bones),and achondroplasia (dwarfism).

The exact causes of new heritable mutations are

unkn own . It is thought that the risk of some newheritable mutations increases with increasing ageof the father. Certain rare disorders du e to mu-tations occur at as m uch as four times the aver-age rate when fathers are 40 years of age or olderat the time of conception of their children. Such



Ch. 2—Genetic Inheritance and Mutations . 31

disord ers include achondrop lasia, Apert’s syn-drom e (a disorder w ith skull malformations andfusion of bones in the hands), and Marfan’s syn-drom e (a comp lex syndrom e includ ing increasedheight). Advanced maternal age is a risk factorfor some newly arising chromosome abnormal-ities, includ ing Down syn drom e.

Very small deletions of a p articular chrom o-some have been found in children having certaincancers such as Wilm’s tum or (chromosome nu m-ber 11) and retinoblastoma (chrom osome nu m-ber 13). Some chrom osome m uta tions are com-patible with normal life, and others may beassociated with increased morbidity and repro-du ctive pr oblems. For example, when a sectionof a chromosome breaks off and reattaches toanother chromosome, both copies of all of thegenes may still be present, bu t they are locatedabnormally. If no genetic material has been lost

in the move, this type of rearrangement is calleda balanced translocation. However, a gene thatwas p reviously “tu rned off” m ay, in its new loca-tion, be sud den ly “turned on”; this explanationhas been invoked to explain how cancer genes canbe activated. Furtherm ore, in the p rodu ction of germ cells, an abnormal complement of DNAcould result, so that a parent with a balancedtranslocation may have offspring with u nbalancedchromosomes, an outcome that u sually has dele-terious consequences for the child.

Early Mortality and MorbidityNumerical and structural chromosome abnor-

malities have been d etected in as m any as 50 to60 percent of spontaneous abortions, 5 to 6 per-

cent of perinatal deaths, and 0.6 percent of live-

born infants (43). Fetuses that survive to term w ith

such abnorm alities are apparent ly only a smallpercentage of the number of fetuses w ith chro-mosome abnormalities that have been conceived.Abnormalities such as trisomy 13 or trisomy 18

are among the more common disorders found inthe liveborn group, whereas many numerical and

structural abnormalties, most of which are notfound in surviving infants, have been found inspontaneous abortuses.

Some of the numerical chromosome abnormal-ities, such as trisomy 13 or trisomy 18, can be

compatible with a full-term pregnancy but maycause severe handicap in infancy and childhoodand may be associated with a reduced lifespan.The health effects of structural chromosome aber-rations (involving only parts of chromosomes) de-pend on the size and location of the aberration.

How ever, features shared by man y of these con-ditions includ e low birthw eight; mental retard a-tion; and abnorm alities of the face, hand s, feet,legs, arms, and internal organs, often necessitat-ing m edical care throu ghout an individual’s life.

Diseases During Adult Life

Some genetic disorders, although present in anindividu al’s DN A since his conception, do n otman ifest themselves un til adult life. Hu ntingtondisease, a degenerative neurological disorder char-acterized by irregular movements of the limbs andfacial muscles, mental deterioration, and deathusually within 20 years of its first physiological

appearance, is an example of a genetic disorder

with adult onset.

Many disorders that develop during adulthoodand that tend to cluster in families do not haveclassical patterns of genetic inheritance. Severalgenes may be involved in the d evelopm ent of these“multifactorial diseases, ” and environmental fac-tors may also influence the manifestation andseverity of these disorders. Susceptibilities to con-ditions such as diabetes mellitus, hypertension,

ischemic heart disease, peptic ulcer, schizophre-nia, bipolar affective (manic depressive) disorder,and cancer are though t to have genetic as well asenvironmental components.

Effects of Somatic Mutations

Muta tions that occur in som atic cells may af-fect the individu al, but they are not p assed on tofuture generations. Somatic mu tations m ay p laya role in the development of some malignanttumors by removing normal inhibition to cell

growth and regulation, producing cells with aselective grow th ad vantage. A w ide range of ex-periments in animals and observations in humansindicate that somatic mutation may be one stepin the dev elopm ent of certain typ es of cancer, in-




eluding leukem ia and cancers of the breast and disease xeroderma pigmentosum, can also lead tothyroid. Mutations that cause deficiencies in the cancer following exposure to a mu tagenic agent.normal DNA repair system, such as in the genetic

PERSISTENCE OF NEW MUTATIONS IN THE POPULATION

The “spontaneous mu tation rate” represents thesum of “natural” error arising d uring the courseof life and r epr odu ction, and external influen ces,both natural and m anmad e. The mu tation ratesthat have prevailed throughout recent human his-tory may be in a state of relative equilibrium, al-thou gh accurate informa tion is lacking. Sup pose,though, that mutagens are suddenly introducedheavily into the environment, and that there isan increase in mutation rates. It is impossible toaccurately pred ict the n um ber of people born w ithgenetic diseases in future generations that will re-

sult from that increase, but some predictions canbe made based on w hat is known abou t the ef-fects of different types of mutations on subsequentgenerations. Two cases are considered below: 1)the effect of a d oubling of the spontan eous m u-tation rate in one generation, and 2) the effect of a permanent doubling of the spontaneous muta-tion rate.

Assume that a pop ulation w as exposed to a“pu lse” of a m utagenic agent, such as rad iationfrom an atom ic bomb or from chemical agentsin a toxic spill, and that the exposure caused a

dou bling of the heritable m utation rate. Overall,the n um ber of extra cases of disease associatedwith mutation in the first generation after thepu lse (the generation in w hich the effect wou ldbe greatest), will be comparatively small. One esti-mate (from a 1977 United Nations Scientific Com-mittee on th e Effects of Atomic Rad iation repor t)is that 10.5 out of every 100 liveborn infants has

some kind of genetic disorder (88). Almost all of

those cases are from mutations already present

in the gene pool, and not from new m utations.The National Research Council (NRC) estimatesthat one mu tational pu lse wou ld add an extra 0.66

cases, so instead of 10.5 per 100, 11.16 caseswould occur. The effects on subsequent genera-tions would be smaller, but even after hund redsof generations, there might be some small incre-ment of the mutational load as a result of the

pu lse. To und erstand th is phenomenon , it is use-ful to consider different kinds of mutations sep-arately.

Chromosome m utations would be the shortestlived in the population, because in most cases theyresult in sterility, so they wou ld not be passed onto the next generation. They w ould p robably beeliminated completely after about 1.25 gener-

ations.

Dominant and X-linked (sex-linked) mutations

are the next shortest lived. Many of these, because

they often cause such severe d isease, interfere witha normal lifespan and reproduction. After aboutfour or five generations, on average, they would

no longer be transmitted to future generations.

New recessive mutations will probably have the

greatest chance of being maintained in the popu-

lation. Virtually none would be eliminated in thefirst generation, because a single individual willhave only one copy of the gene, and two wouldbe needed to cause overt d isease. Recessive m u-tations may t ake generations to sur face as overdisease. There is mu ch less certainty abou t the u lti-

mate fate of recessive mutations than about dom-inant, X-linked, or chrom osome m utations.

Many “common” or multifactorial diseases arepartially influenced by genetic traits carried infamilies or arising anew. These include congeni-tal malformations, and such common diseases asheart disease, diabetes mellitus, bipolar affectivedisorder, immunological disorders, and cancer.Since it is difficult to estimate the p roport ion of disease that is currently influenced by mutations,it is also difficult to estimate how long such mu-tations indu ced from a pu lse of a mutagenic agent

would persist in the p opulation. However, theNRC report (88) estimates that th ese mutat ionswou ld p ersist for an average of 10 generations.

These estimates suggest a surprisingly small in-crease over background rates of genetic disease



Ch. 2—Genetic /inheritance andMutations Ž 33

in the first few generations after intense exposureto mutagenic agents, which would decline slowlyover many generations. The NRC report (88) sug-gests that about half of the total effect would oc-cur in the first six generations. The m ost severemutations would be more easily observable asdom inant genetic diseases. These w ould be elim-inated within one or two generations since theyinterfere with survival and fertility. Mutationsw ith a less severe initial impact w ould persist inthe population for a longer period of time.

A perm anent d oubling of the current m utationrate would, however, after perhaps hundreds of generations, lead to a new equilibrium, w ith theincidence of genetic disease at about twice its cur-rent level. One of the biggest unknowns aboutpred icting futu re effects of mu tations is the im-pact of an accum ulation of recessive mutat ions

in the population. There is also little informationto bring to bear on the qu estion of the impact of an increase in mutations to the prevalence of com-mon diseases.



Chapter 3

Kinds and Rates of HumanHeritable Mutations



Chapter 3

Kinds and Rates of Human

Heritable Mutations

INTRODUCTION

The genetic code specified in DNA directs thetranscription of RNA, which is translated intopolypeptide chains; these chains, in turn, are as-sembled into p roteins, which are the bod y’s toolsfor the regulation of physiological, biochemical,and behavioral functioning. Mutations, transmis-sible alterations in DNA, can change the mes-sages, causing alterations in RNA molecules andproteins. Some of these changes can lead to sub-tle abnormalities, diseases, and disabilities.

Environmental factors and specific interactionsbetween env ironmen tal and genetic factors arethou ght t o accoun t for the ma jority of diseasesor suscep tibilities to disease. Mutation s alone arethought to account for only a fraction of the cur-rent burden of disease. However, there is a groupof disorders for which new mutations are the pri-mary causes. Among these “genetic” diseases isa grou p of rare clinical conditions, “sentinel phe-notypes, ” that occur sporadically in each gener-ation; children w ith sentinel phen otypes are u su-ally born to p arents who d o not have the samedisease, indicating a n ew m utation that arose in

a germ cell of one of the par ents,

There are several other ind icators of new m u-tations in hum an beings that do not rely on de-tecting diseases. The m ost widely used method shave been cytogenetic analysis of chromosomesand biochemical analysis of proteins.

The structure and number of chromosomes vis-ible by light microscopy has been used to detecta certain class of heritable muta tions. Gross chro-mosome abnormalities, often associated with highfetal and neonatal death rates, are identifiablewith increasing p recision by cytogenetic analy-sis. A large portion of these chromosom e abnor-malities arise anew each generation,

Studies of mutationally altered proteins havebeen d one on a limited basis. The p roteins m ost

accessible for analysis are those circulating inblood; other body proteins are not routinely ob-tained for analysis. The method available fordetecting abnorm al proteins is electroph oresis.Heritable mutations have been d etected by iden-tifying electrophoretic variants of blood proteins.In addition, a combination of approaches has beenused to estimate the mutation rate per nucleotidein genes coding for globin polypeptides, constit-uents of hem oglobin. These app roaches to iden-tifying mu tations are discussed in this chapter.

The term “m utation rates” for particular typesof mu tations in sp ecific regions of DN A is u sedfrequently in this report. Mutation rates for differ-ent kinds of mutations, or corresponding to differ-ent overt effects of mutations, can be expressedas mu tations per locus, per gene, per nucleotide,and per gamete. All of these indicate a specifictype of mu tation occurring per generation, reflect-ing mu tations arising an ew from one generationto the next. This chapter summarizes currentknow ledge about such mutation rates as they havebeen measured thus far; it is not yet known

whether these rates apply also to other pop ula-tions, to other gene regions, or to other typ es of mu tations than th e ones examined.

Rather than looking d irectly at germinal mu -tations—mutations in egg and sperm DNA—themethod s described in this report focus on herita-ble mutations in liveborn offspring, a subset of all germinal mutations.

lA larger subset of ger-

minal mu tations includ es new m utations observedin offspring before birth. In recent years, cyto-genetic techniques have been used to identifychromosome abnorm alities in spontan eous abor-

tuses. These investigations have shown that thefrequency of new mutations in spontaneous abor-

lA t present it is not feasible to examine egg or sperm DNA forgenetic damage, although some information on major chromosomeabnormalities can be obtained by cytogenetic examination of sperm.

37




tuses is mu ch higher tha n in liveborn infants; the agents. Nevertheless, the most lasting medical and

observed incidence of new mu tations at birth on ly social consequences of new heritable mutations

partially represents the “true” incidence of such are derived from those mutations that are com-

heritable m utations in all conceptuses, and may patible with life at least through infancy and

give an incomplete picture of the p otential risk childhood.

from environmental exposures to mutagenic

SURVEILLANCE OF SENTINEL PHENOTYPES

The United Nations Scientific Committee on theEffects of Atomic Radiation (144) and the Com-mittee on the Biological Effects of Ionizing Radi-ations (87) estimate that 1 percent of liveborn in-fants carry a gene for an autosomal dominant

disease; in 20 percent of these cases (0.2 percentof livebirths) their disease is du e to a new , or “spo-

rad ic, ” mutation that arose in the reproductivecells of one of their paren ts, Observing and re-cording the birth of infants with such sporadicconditions is the classical app roach to identify-ing heritable mu tations that lead to serious dis-ease (36). 2

A subset of the group of autosomal dominant

and X-linked conditions, indicator conditions

known as “sentinel phenotypes,” maybe particu-larly useful in identifying heritable mutations.

Sentinel phenotypes are clinical disorders that oc-

cur sporadically, probably as the result of a sin-

gle mutant gene. They are m anifested at birth orwithin the first months of life, and usually require

long-term, multidisciplinary medical care (83). Inaddition, sentinel phenotypes are associated with

low fertility, so their appearance suggests muta-

tions that have been passed n ot from affected par-ents to offspring, but instead, have been passedfrom unaffected par ents to offspring as a resultof a n ew ly arising mu tation in a germ cell of oneof the parents of that offspring. Affected infantswou ld have this mu tation in the DN A of all theirsomatic cells and 50 percent of their germ cells.

‘Rates of a wid e variety of new heritable mu tations leading tosevere diseases have also been estimated by an “indirect method. ”This method is a theoretical approach, based on the loss of disease-

causing genes from the population (due to reduced reproductive ca-pabilities of individuals with these genes) and on the sporadic fre-quency of the genes in the population. The mutation rate necessaryto maintain such genes in the population, despite their recurrent loss,

is inferred from these observations. The types of populations andconditions suitable for this method are limited, and its precision isunknown (91,165).

Mulvihill and Czeizel (83) compiled a list,show n in t able 1, of 41 genetic disorders (36 dom-

inant and 5 X-linked) that satisfy the above cri-teria. Each of these phenotypes is individuall y

rare, some of them occurring in only one in a m il-lion liveborn infants.

Since children with sentinel phenotypes ha veserious health problems associated with their con-

ditions, they may first come to the attention of the family practitioner or ped iatrician, and thenbe referred to subspecialists for a complete diag-nosis. Clinical expertise in dysm orp hology,

3med-

ical genetics, pediatric ophth almology, and on-cology may be needed to diagnose accurately thesmall numbers of children with sentinel pheno-types present among the vast majority of un-affected children and among children w ith pheno-typ ically similar, but n ongenetic cond itions (called“phenocopies”) (71,83,165).

Reliable incidence data are available only for

13 dom inant and 5 X-linked sentinel phen otypes(166); only those that are manifested in infancy,rather than in childhood or adolescence, Q are sys-tematically recorded. Diagnostic records of sen-tinel phenotypes m ay be available in pop ulation-based registries of childh ood cancers, birth de-fects, and genetic diseases, with initial entriesbased on birth certificates and later entries addedas the conditions are diagnosed during infancy.After investigators rule out the possibility of a dis-crepancy between stated and biological parent-age by genetic testing (166), sporadic cases of asentinel phenotype a re accepted as evidence for

‘Physicians who specialize in diagnosing syndromes character-ized by unusual physical form and stucture, for example, congeni-tal malformations.

‘lSuch disorders include achondroplasia, acrocephalosyndactyly,osteogenesis imperfect, retinoblastoma, an d Wilms’ tumor (66).



Ch. 3—Kinds and Rates of Human Heritable Mutations 39

Table 1 .—Candidate Sentinel Phenotypes’

Inheritancea

Phenotypes identifiable at birth: Achondroplasia . . . . . . . . . . . . . . . . . . . . . . . . . .Cataract, bilateral, isolated . . . . . . . . . . . . . . . .Ptosis, congenital, hereditary . . . . . . . . . . . . . .Osteogenesis Imperfect type I . . . . . . . . . . . .

Oral-facial-digital (Gorlin-Psaume) syndrometype I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Incontinentia pigmenti, Bloch-Sulzberger

syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Split hand and foot, bilateral atypical . . . . . . .Aniridia, isolated . . . . . . . . . . . . . . . . . . . . . . . . .Crouzon craniofacial dysostosis . . . . . . . . . . .Holt-Oram (heart-hand) syndrome . . . . . . . . . .Van der Woude syndrome (cleft lip and/or

palate with mucous cysts of lower lip). . . .Contractual arachnodactyly . . . . . . . . . . . . . . .Acrocephalosyndactyly type 1, Alpert’s

syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Moebius syndrome, congenital facial

diplegia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Nail-patella syndrome . . . . . . . . . . . . . . . . . . . .Oculodentodigital dysplasia

(ODD syndrome) . . . . . . . . . . . . . . . . . . . . . . .Polysyndactyly, postaxial . . . . . . . . . . . . . . . . .Treacher Collins syndrome, mandibulofacial

dysostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . .Cleidocranial dysplasia . . . . . . . . . . . . . . . . . . .Thanatophoric dwarfism . . . . . . . . . . . . . . . . . .EEC (ectrodactyly, ectodermal dysplasia,

cleft lip and palate) syndrome . . . . . . . . . . .Whistling face (Freeman-Sheldon)

syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Acrocephalosyndactyly type V, Pfeiffer

syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Spondyloepiphyseal dysplasia congenita. . . .

Phenotypes not identifiable at birth: Amelogenesis imperfect . . . . . . . . . . . . . . . . .Exostosis, multiple . . . . . . . . . . . . . . . . . . . . . . .Marfan syndrome . . . . . . . . . . . . . . . . . . . . . . . .

Myotonic dystrophy . . . . . . . . . . . . . . . . . . . . . .Neurofibromatosis . . . . . . . . . . . . . . . . . . . . . . .Polycystic renal disease . . . . . . . . . . . . . . . . . .Polyposis coli and Gardner syndromeRetinoblastoma, hereditary . . . . . . . . . . . . . . . .Tuberous sclerosis . . . . . . . . . . . . . . . . . . . . . . .von Hippel-Lindau syndrome . . . . . . . . . . . . . .Waardenburg syndrome. . . . . . . . . . . . . . . . . . .Wiedemann-Beckwith (EMG) syndrome . . . . .Wilms’ tumor, hereditary . . . . . . . . . . . . . . . . . .Muscular dystrophy, Duchenne type . . . . . . . .Hemophilia A. . . . . . . . . . . . . . . . . . . . . . . . . . . .Hemophilia B. . . . . . . . . . . . . . . . . . . . . . . . . . . .

ADADADAD

XD

XDADADADAD

ADAD

AD

ADAD

ADAD

ADADAD

AD

AD

ADAD

ADADAD

ADADADADADADADADADADXRXRXR

aAD refers to autosomal dominant inheritance, XD refers to X-linked dominant

inheritance, and XR refers to X-1inked recessive inheritance.

SOURCE: J.J Mulvihill and A. Czeizel, “Perspectives in Mutations Epidemiology6: A 1963 View of Sentinel Phenotypes,” Mutat. Res. 123:345-361, 1963.

a new germinal mu tation originating in a germcell of one of the parents (65).

In general, recording the incidence of sentinel

phenotypes is not useful for monitoring a smallpopulation exposed to a suspected mutagen. Sur-veillance of sentinel phenotypes in large popu la-

tions is potentially more useful for estimating thebackground mutation rates leading to dominantgenetic disorders. Given the rarity of the ind ivid-ual sentinel phenotypes, the most reliable data forestimating mutation rates would be derived fromthe largest target pop ulations with, for examp le,international cooperation among study centers(62,66).

Mutation Rates Estimated From

Phenotypic Data

Data are available on the incidence of various

sentinel phenotypes from different time periodsin particular regions worldwide, though represen-

tation is incomplete. Crow and Denniston (22)

summarized the most current data; these data

have not been significantly updated since the1940s and 1950s. The frequencies for any given

phenotype are general ly consis tent between

studies and the frequencies of occurrence of the

different phenotypes vary over a thousandfold

range, from 1 in 10,000 to 1 in 10 million depend-

ing on the particular disorder. The arithmeticmean of the rates for the different disorders is ap-proximately 2 mutations per 100,000 genes per

generation (17,89,166). This average rate is often

cited as the “classical” rate of heritable mutations.

‘Mutation rates corresponding to sentinel phenotypes are expressed

as the frequenc-y of mutations per gene. It assumes one gene per dis-order and one disorder per person and it is designated “per genera-tion” because the mutations are detected in offspring of individualswhose germ celk have incurred mutation. In general, the mutationrate is defined as the incidence of a sporad ic case of some indicator

(e.g., a sentinel phenotype or a protein variant) divided by two timesthe total number of cases examined (e.g., newborn infants or pro-tein determinations). The two in the denominator is to express themutation rate, by convention, as the rate “per fertilized germ cell, ”instead of “per concepts,” which is made from two germ cells (162).



Ch. 3—Kinds and Rates of Human Heritable Mutations • 41

hood (169). The significance to human health of structural rearrangements of the chrom osomes isless well defined, but such defects have been asso-ciated w ith mental retard ation, physical malfor-mations, and a variety of malignant diseases (43).In ad dition, carriers of balanced translocations

may p rodu ce offspring w ith unbalanced comple-ments of DN A; these offspring m ay d ie in u teroand may account for a high spontaneous abor-tion rate among such carrier parents.

It is believed tha t at least 5 percent of all rec-ognized hum an conceptions have a chromosom eabnormality, although this is a crude estimate be-cause it is not corrected for factors w hich couldinfluence the occurr ence of such abn ormalities(such as ma terna l age) or for factors wh ich couldaffect the rate of embryonic or fetal death in someof the disorders (43). Numerical and structuralchromosome abnormalities have been detected inso to 60 percent of recognized spontaneous abor-tions (data su mm arized in ref. 12), 5 to 6 p ercent

of perinatal deaths (ibid.), and 0.6 percent of live-

born infants (42).

Chromosome abnormalities are useful for meas-

uring heritable mutation rates since they are usu-

ally identified as new mutations, either because

the defect causes virtual infertility or sterility (as

in the case of num erical chromosome abnorm al-ities) and therefore could not have been inheritedfrom a par ent w ith the sam e defect, or becausethe parents of an individual with a chromosome

abnormality (e.g., a structural abnormality) havebeen shown n ot to carry the same m utation. Nu -merical chromosome abnormalities and some struc-tural abnormalities are detectable with cytogeneticmethod s using current chromosome staining andbanding techniques.

In general, little is known about the geneticmechanisms of either numerical or structural rear-rangemen ts. They appear to resu lt from comp lexprocesses during gametogenesis in which variousdefects may lead to the same or different outcomes(18,119). Consequently, frequencies of the vari-

ous cytogenetic abnormalities may correspond toa combined rate of several different mutational

events or to a single mutational event.

Individuals with chromosome abnormalitiescan be identified on the basis of karyotypes (see

ch. 2). However, since the vast majority of con-ceptuses with chromosome abnormalities such astrisomies and unbalanced structural rearrange-ments are spontaneously aborted, many beforea pr egnan cy is recognized, the frequency of thesemutations at birth is known to rep resent only a

small fraction of the presumed frequency of thesemutations at conception.

Data on the frequency of chromosome abnor-malities in fetuses at the tim e of amniocentesis(generally d one at 16 to 18 weeks gestation) pro-

vides informat ion not avai lable in newborn

screening. In addition, recent progress in cyto-genetic analysis of chorionic villi cells makes itpossible to identify chromosome abnormalities infetuses at 8 to 10 weeks gestation (116). This tech-nique provides even earlier information on pos-

sible chromosome abnormalities, before fetuses

with certain abnormalities (particularly trisomies

and unbalanced structural rearrangements) would

have aborted spontaneously. The analysis of fe-

tal cells obtained at amniocentesis has provided

valuable data on the frequency of chromosome

abnormalities in the second trimester of pregnancy

(12,45,159.) However, estimates of mutation rates

from those measurements may be inflated unless

certain biases are corrected. Women who elect

amniocentesis—generally older mothers, couples

who have had previous reproductive problems,

and/ or couples who may have been exposed to

mutagenic agents (101) —are at increased risk for

chromosome abnormalities in their offspring.Cytogenetic surveys of consecutive liveborn in-

fants have provided the most extensive data on

chromosome mutation rates, although it is known

that these data underestimate the “true” rate of new chromosome mutations. It should also be

noted that the prevalence of sentinel phenotypes

and biochemical variants in liveborn infants may

also underestimate their “true” rate at conception,

but little is known about selection against these

mutations in the developing fetus (e. g., molecu-

lar repair of the defect, or spontaneous abortion

of the fetus), in contrast to the documented evi-den ce for selection against chromosome mutationsin the fetus (135. ) How ever, the frequen cy at birth

is probably the best indicator of events with the

highest cost in terms of the social and economic

burden of these diseases.




Mutation Rates Derived From

Chromosome Abnormalities

At least 10 major research efforts w orldw idehave attempted to d etermine the p revalence of chromosome abnormalities in liveborn infants. It

is possible to estimate th e rate of new mu tationsunderlying these defects, but difficult to comparewith rates of sentinel phenotypes, since chromo-some mutations, by definition, are multigenicchanges and sentinel phenotypes correspond tosingle gene m utations.

The combined d ata from studies surveying atotal of 67,014 newborn infants for variousperiods between 1974 and 1980 in the UnitedStates, Canada, Scotland, Denmark, the U. S. S. R.,and Japan (120) are shown in table 2. Two of thestudies used newer banding techniques that iden-tify more abnormalities than conventional stain-

ing method s used in the rem aining eight stud ies.

For conditions that are a ssociated w ith virtualinfertility, complete sterility, or prereproductivedea th, it is often assu med that each case is the re-sult of a new mu tation (unless there is evidencefor gonadal mosaicism 6). Such conditions includ ethe autosomal trisomies (e.g., Down syndrome)and the sex chromosome aneuploidies (e.g.,Turner’s syndrome).

For these defects, the apparen t mu tation rateobserved in liveborn infants is calculated as the

nu mber of individuals identified w ith a given ab-normality divided by the total number of indi-viduals examined. This mutation rate is expressedas the num ber of mu tations per generation. Forconditions which could be the result of either newmu tations or inherited mu tations, such as the vari-ous balanced and unbalanced structural rearrange-ments, it is necessary to karyotype the parents of infants found with these chromosome abnorm al-ities, and to exclude inherited mutations from cal-culations of the d e novo mu tation rate. For these

6Mosaicism refers to the presence of two genetically different cellpopulations in the same individual. This can result from a muta-tion during one of the earliest divisions in the zygote, a few daysafter conception, that led to the proliferation of a large number of cells that were derived from the original cell bearing the mutation.

conditions, the observed m utation rate is deter-mined using the following formula:

Number of patients X Proportion who are likelyMutation + with abnormalit y to have de novo mutations

7

rateTotal number examined

The data show that the frequencies of chromo-some abnormalities in liveborn infants range from2 in 100,000 newborns per generation for inver-sions to 121 in 100,000 newborns per generation

for Trisomy 21 (Down syndrome), the most com-mon newly arising numerical chromosome abnor-mality found at birth. Overall, the numericalabnormalities are mu ch more frequent than th estructural abnormalities. However, the rates forbalanced structural rearrangements may be lowin these data, since only two of the surv eys usedmore sensitive banding m ethods w hich are need-ed to iden tify these more subtle aberrations.

Limitations of Cytogenetic Analysis

Screening for chromosome mu tations in n ew-born infants gives only a p artial picture of theirincidence. The prevalence of a particular geneticmu tation at birth can be thou ght of as the resultof the incidence of the mu tation at conception in-teracting with the probability of survival of con-ceptu ses with the mu tation. Fetal survival, in turn,depend s on ind ividu al fetal and maternal attri-butes, as well as on the interaction between thetw o (136). Some mutations in the fetus may af-

fect fetal survival, whereas other mutations mayhave no effect on survival. Major chromosome

aberrations, in particular, numerical abnormal-

ities and most unbalanced structural rearrange-

ments, are often incompatible with fetal survival

and are associated with an increased risk of prena-

tal loss (45).

Balanced chromosome rearrangements, how-ever, ma y not lead to excessive p rena tal loss. Forthis reason, the balanced structural rearrange-ments m ay be a useful indicator of new germinalmutations observed in liveborn offspring. Sincesome of these infants identified as having a bal-

“Such proportions are based on empirical data for each type of chromosome aberration. In this case, data from Jacobs, 1981 wereused and are shown in table z in parentheses in the third column.



Ch. 3—Kinds and Rates of Human Heritable Mutations “ 43

Table 2.— Prevalence of Chromosome Abnormalities at Birth

Number New chromosomeTotal of new abnormalities per 100,000

Chromosome abnormality population mutantsa

newborns per generation

Numerical anomalies: Autosomal trisomies:

b

Trisomy 13 . . . . . . . . . . . . . . . . . . . . .Trisomy 18 . . . . . . . . . . . . . . . . . . . . .Trisomy 21 . . . . . . . . . . . . . . . . . . . . .

Male sex chromosome anomalies:47,XYY . . . . . . . . . . . . . . . . . . . . . . . .47,XXY . . . . . . . . . . . . . . . . . . . . . . . .

Female sex chromosome anomalies:45,x . . . ., , . . . . . . ... , . . . . . ., ., . .47, XXX . . . . . . . . . . . . . . . . . . . . . . . .

Balanced structural rearrangements: Robertsonian translocations:

c

D/D d . . . . . . . . . . . . . . . . . . . . . . . . . .D/G . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . 67,014

. . . . . . . . . . . . . . . . . . 67,014

. . . . . . . . . . . . . . . . . . 67,014

. . . . . . . . . . . . . . . . . . 43,048 males

. . . . . . . . . . . . . . . . . . 43,048 males

. . . . . . . . . . . . . . . . . . 23,966 females

. . . . . . . . . . . . . . . . . . 23,966 females

. . . . . . . . . . . . . . . . . . 67,014

. . . . . . . . . . . . . . . . . . 67,014

38

81

4342

224

48(2/ 29) = 3.314(2/ 1 1) =2.5

512

121

10098

8100

54

Rec ip ro cal tran slocat io ns an d i nser ti ons . . .. .. .. .. .. .. .6 7,014 60(13/43) =18 27Inversions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67,014 12(1/8)= 1.5 2

Unbalanced structural rearrangements: Tr anslocat ions , inv er sions , and d eletions . . ... .. .. .. .. .67. 014 37(7/16) =16 24,asee discussion in text for proportion of new mutants among all those identified with Structural r0arran9ementS.

bTrisomies refer t. conditions inwhich there are 47 chromosomes (instead of the normal 46 chromosomes or 23 pairs of homologous chromosomes, indicated as46,XX for females or 46,XY for males) because of the presence of an extra copy of one chromosome; three copies of this particular chromosome would be presentinstead of the normal two. Autosorna/ trisomies indicate an extra autosome, which includes any chromosome except one of the sex chromosomes

Prearrangements in which the long arms of two chromosomes fuse, resulting in one “hybrid” chromosome and the IOSS of the short WTls Of each chromosome,

dD and G refer to groups of chromosome numbers 13-15 and 21 ’22, resPectivf31Y

SOURCE: K Sankaranarayanan, Genetic Effecfs of /orrizina Radiafiorr in kfu/tice//u/ar Eukarvofes and the Assessment of Genetic Radiafion Hazards in Man (Amster-dam Elsewer Blomedlcal Press, 1982), pp 161-i64

anced chromosomal mutation may have inheritedthe relatively subtle abnormality from one of theirparents, it is necessary to d etermine wh ether themutation observed in the child is actually a new

germinal mutation or not.

Although it is not understood how bands ob-

servable under the microscope correspond to thestructure and composition of DNA, mutations inDNA can cause a visible chan ge in the bandingpattern, p articularly if such m utations involve alarge section of a chromosome. Increased resolu-tion and sensitivity of cytogenetic techniques, spe-cifically the “high-resolution” banding methods(72), makes it possible to detect smaller and moresubtle aberrations tha n w as p reviously p ossible.Currently, routine cytogenetic method s can p ro-du ce 200 to 300 band s (containing several hun -dred genes in each band) in a complete set of DNA, and the high-resolution methods can dis-

tinguish as many as 1,000 bands (containing about100 or fewer genes per band) (59). With moresmaller bands d istingu ishable, fewer genes are

present per band, and a change in one gene has

a greater likelihood of showing up as a change

in the banding pattern.

At present, data available for large populations

rely on only the routine staining and banding

methods. Like the observations of sentinel pheno-

types in newborn infants, cytogenetic methods

focus on a subset of mutations, albeit those with

serious clinical consequences. Even though chro-mosome abnorma lities are not as rare as sentinelphen otypes, these two group s of abnormalities to-gether p robably account for only a sm all fractionof heritable mutations.

Mutations detectable by the methods discussedin the remaind er of this chapter and in the fol-lowing chapter do not necessarily have adverseclinical consequences. Their detection broadensthe spectrum of identifiable m utations by includ-ing those that may not have an imm ediate adverse

effect on the individual, and also offers more pre-cise information about the nature of the geneticchanges.



44 • Technologies for Detecting Heritable Mutations in Human Beings —

DETECTION OF BLOOD PROTEIN VARIANTS

In a single-gene, dominant genetic disease, if the causative mu tant gene occup ies one allele of the pair of genes and the n ormal form of the geneoccupies the other allele, the one mutant gene isdom inant over the norm al gene and is sufficientto cause overt disease. Dominant diseases are saidto be expressed in individuals w ho hav e a singledose, or are heterozygous, for the disease-causinggene. Autosomal recessive diseases, however, areclinically apparen t only when the mu tant gene ispresent at both alleles, or in the hom ozygous state;a single dose, the heterozygous state, can, for in-stance, destroy the a ctivity of an enzym e or p ro-du ce a defective gene produ ct, but the correspond-ing norm al gene sup plies a sufficient amou nt of normal gene product to prevent overt clinicalsymp toms. It is difficult to infer the frequency of

recessive genes by pop ulation studies of pheno-types since individuals with one copy of the reces-sive mutant gene are p henotypically indistinguish-able from ind ividu als with only norm al genes forthe particular protein. Biochemical means of iden-tifying of heterozygotes, who are phenotypicallynormal, is one w ay of identifying new mu tationsthat are expressed as recessive traits.

Biochemical stud ies on asymp tomatic individ-uals with protein variants (indicative of singlegene, recessive mu tations) can be u seful in express-ing the cumulative, “invisible” effect of an in-creased mu tation rate. Any p rotein in the bodycould be studied, but since peripheral blood is eas-ily accessible, pr oteins from blood are m ost com-monly used for these studies. Three major tech-niques for the d etection and characterization of variant proteins have been used in stud ies of hu-man

1.

2.

3.

populations:

electrophoresis of blood p roteins,quantitative and kinetic studies of enzymedeficiency variants from eryth rocytes, andmolecular analysis of Hemoglobin M (HbM)and unstable hemoglobins.

One-Dimensional Electrophoresis

Electrophoresis separates proteins according tonet molecular charge. A mixture of proteins is ap-plied to a starch or acrylamide gel and exposed

to an electric current. Each protein moves accord-ing to its own electric charge and separates fromproteins with different charges. The subsequentadd ition to the gel of stains that bind to the pro-teins makes it possible to v isualize the locationof the proteins. Electrophoresis can be used to de-tect variant proteins in individuals and in theirparents. The identification in an offspring of a var-iant protein that is absent from either parent isevidence for a new mu tation. As in the other ap-proaches, it is necessary to exclude the possibil-ity of a discrepancy between stated and biologi-cal parentage so that only new, not inherited,mutations are considered.

Electrophoretic variants are detectable becausea mu tation results in a change in the net chargeof a protein. On a theoretical basis it is expected

that app roximately one-third of amino acid su b-stitutions prod uce a change in m olecular chargeof a protein, and there is also evidence that elec-trophoresis can detect chan ges in a molecule’s con-figu ration d ue to a nu cleotide su bstitution. All inall, it is estimated that electrophoresis probablydetects about 50 percent of nu cleotide substitu-tions in expressed regions of coding genes (96).

The protein variants detected by electrophoresisare u sually not associated w ith clinically recog-nizable problems. However, this technique has animportant advantage for mutation research over

the ph enotypic approach; it measures changes ata level closer to the level of the mu tational even tsand therefore may provide more accurate esti-ma tes of mu tation rates. It also reflects mor e sub-tle genetic changes, broadening the sp ectrum of detectable gen etic endpoints.

Mutation Rates Derived From Electrophoresis

Several large-scale studies to detect protein var-iants in human populations have been reported(excluding studies in Japanese atomic bomb sur-vivors, which are described below). Neel, et al.

(94), and Mohrenweiser (76) devised a large-scalepilot program to study m utation rates using pla-cental cord blood obtained from new born infantsin Ann Arbor, Michigan (approximatel y 3,500samples). They found no new heritable mutations



Ch. 3—Kinds and Rates of Human Heritable Mutations q 45

in 36 different proteins in a total of 218,376 lo-cus tests in Caucasian and 18,900 such tests in

black infants.gHarris and colleagues (38) summa-

rized data from studies in the United Kingdom inwhich 43 different loci coding for blood proteinswere analyzed, and no new mutations were found

in 133,478 locus tests.

9

Altland and colleagues (5)examined blood from filter p aper su bmitted origi-nally for screening for phenylketonuria in approx-imately 25,000 newborn infants in West Germany.They identified one p utative new mu tation amon gfive different hemoglobin proteins in approxi-mately 225,000 locus tests.

Feasibility of Electrophoresis

The availability of hum an blood for analysisof variant p roteins makes electrophoresis feasi-ble for stud ies of large p opu lations. In a dd ition,it perm its comp arisons of the frequ ency of occur -rence of variant proteins between humans andother species, since comparable proteins are usu-ally associated with homologous gene sequences(93). Perhaps the most important advantage of this approach is that it permits the stud y of a de-fined set of genes expressing functional proteinsand offers more precise information on the num -ber of genes involved in the biochemical obser-vations made and on the effect of mutation onsome pro teins (90).

How ever, electrophoretic analysis can p rovideinformation only on mutations in genes that are

transcribed into RNA and subsequently translatedinto proteins, and only for mutations that alterthe protein’s net molecular charge or configura-tion. It cannot pr ovide information on mu tationsin nontranscribed regions of the DNA, whichcomprise the larger portion of the genome (165).

Since the baseline frequency of new protein var-iants is still poorly d efined , it is difficult to esti-mate the size and scope of a population study nec-essary to detect an increase over a period of timeor to detect differences in rates between twopopulations. One-dimensional electrophoresis is

limited by the number of proteins that can beexamined for variants, approximately the samenumber of different phenotypes included in thelist of sentinel phenotypes. However, one-dimen-sional electroph oresis in a large p opu lation is amore straightforward approach to studying new

heritable mutations than is population surveil-lance for sentinel ph enotypes.

Vogel and Altland (164) estimated that analy-sis of two pop ulations of 10 million ind ividualseach would be needed to have a 95-percent chanceof detecting a 10-percent increase in spon taneou smutation rates. This calculation is based on atheoretical rate of 1 mu tation p er 1 million genesper generation and 50 loci tested per individualin the stud y. The p opu lation size could be redu cedif the proteins that w ere examined m utated m orefrequently or if detection of only large increasesin the mutation rate (e.g., 50 percent) were accept-able, or if many more proteins could be exam-ined from each ind ividual tested.

Quantitative and Kinetic Studies ofEnzyme Deficiency Variants

Mutations that result in variant proteins withgreatly redu ced or no biological activity, whenpresent in the heterozygous state, are not readilydetectable in electroph oretic studies. Such “en-zyme deficiency variants, ” or “nulls,”

10can be

identified through automated biochemical tests

that m easure enzym e activity in red blood cells.These studies increase the detectable range of mutational events.

Loss of function of a protein may result from:

q the absence of a gene product,q the presence of a protein which is non-

functional catalytically, orq abnormally unstable enzymes (75).

The types of mu tations causing these abnorm al-ities includ e chromosome rearrangemen t or loss,deletions, frameshift mu tations, and n ucleotide

*The number of locus tests takes into account the number of pro-teins examined, the number of gene loci represented by each suchprotein, and the number of individual samples obtained.

9The number of individuals tested varies depend ing on the en-

zyme examined , but for some of the en zyme tests, more than 10,000individuals were tested.

l’JThese mutations are characterized by the presence of an enzy -

matically inactive protein, or by an absence of a protein. These canbe caused by gene d eletions, nucleotide substitutions in critical loca-tions in the gene, mutations in the “start” and “stop” codons forthe protein, or even by nucleotide substitutions at splicing junctionsfor the int rons .




substitutions in coding or non coding regions of DNA. The range of genetic events underlying theproduction of enzyme deficiency variants isthou ght to be larger than for electrophoretic mo-bility variants, since mutations in noncoding re-gions (e.g., intervening sequences, flanking re-gions, etc. ) as well as in coding regions could bedetected (74).

Four studies have provided d ata on the fre-quency of enzyme deficiency variants. The fre-quencies of such variants have been measured in:1) placental cord blood samp les from app roxi-mately 2,500 individuals in Ann Arbor, Michi-

gan (74,75); 2) blood samples obtained from par-ticipants of the studies in Japan coordinated by

the Radiation Effects Research Foundation (122);3) blood samples obtained from 3,OO O hospitalizedindividuals in West Germany (3o); and 4) bloodsamples obtained from Amerindians of South and

Central America (78). No new mutations werefound in any of these studies. However, rare butnot new enzyme d eficiency mu tations that havebeen inherited over m any generations have beenmeasured at an overall frequency of 2 variantsper 1,000 determinations in a total of approxi-mately 110,000 locus tests. If new mutationsproducing enzyme deficiency variants occur, aspred icted, no more than tw ice as often as new mu -tations leading to electrophoretic variants, the lack of new m utations in these 110,000 locus tests isnot inconsistent w ith electroph oretic data .

Molecular Analysis of Hemoglobin Mand Unstable Hemoglobins

Stamatoyannopoulos and Nute (131) calculatedmutation rates for certain nucleotides in the globingenes, which, when mutated, produce autosomaldom inant disorders expressed as chronic hemo-lytic anemia or methemoglobinemic cyanosis.Their data provided the first direct measurementsof heritable mutation rates for particular nucleo-

tides in human DNA. Although these data arelimited to the specific characteristics of the globingenes, they suggest the kind of information which

may be possible to obtain for other genes as thegenetic basis for variou s single gene disord ers isdiscovered. Even if the mutation rates derivedfrom these studies of the globin genes are not rep-

resentative of rates of mu tation at oth er loci, thehemoglobin system described below provides amod el for combining ep idem iologic, clinical, andmolecular information to estimate heritable mu-tation rates in hu man popu lations.

Structure and Function of Hem oglobin

Hemoglobin molecules, which exist in a num-ber of slightly different forms, are the universalcarriers of oxygen in the blood from the lun gs toall cells of the h um an body. Different m ixturesof hemoglobins are p resent at all stages of devel-opment from embryonic to adult, although a par-ticular type of hemoglobin tends to pred ominateat any on e stage. In add ition, hund reds of abnor-mal hemoglobins have been identified, ranging inclinical severity from benign through serioushematologic disorders. Hum an h emoglobin is aprotein consisting of four sep arate globin poly-

peptide chains and four iron-containing hemegroup s. In adu lts, the pred ominant form of he-moglobin is hemoglobin A (HbA), composed of two alpha-globin chains and two beta-globinchains each joined to a heme group .

Hemoglobin’s primary function of binding, car-rying, and releasing oxygen is closely regulatedby several factors, including the maintenance of a precise three-dimensional conformation of thehemoglobin molecule, the composition and bal-ance of globin chains, and the distribution of ioniccharges in and aroun d the m olecule. Muta tionswh ich result in amino acid substitutions at par-

ticular sites in the globin polypeptide chain canmarkedly alter these specific functional proper-ties of the hem oglobin molecule.

Two types of abnormal hemoglobin, HbM andunstable hemoglobin, usually occur sporadicallyas a result of a new m utation and are expressedas dominant mutations.

11Stamatoyannopoulos

and Nu te (131) described the genetic changes lead-ing to these particular abnormal h emoglobins andtheir incidence in human populations. From this

11A heterozygous individual, having one mutant globin gene and

one normal allele, would have clinically recognizable symptoms of

the disease. This is in contrast to other, more common hemoglobindiseases, such as sickle cell anemia and thalassemia, where the het-erozygote is asymptomatic and the disease occurs only when an in-dividual is homozygous for the variant gene.



Ch. 3—Kinds and Rates of Human Heritable Mutations q47

information, they were able to calculate the rateof mutation leading to these abnormal hemo-globins.

Unstable Hemoglobins

Various m utations in the globin genes can lead

to an altered sequence of amino acids in the globinchains and to a major change in the th ree-dimen-sional structure of the hemoglobin molecule.Often these chan ges create a less stable hem oglo-bin molecule than its norm al coun terpart. “Un-stable hemoglobin” can result in the hemoglobinprecipitating w ithin the red cell, leading to m em-brane damage and premature destruction of thered cells by th e liver and spleen (165).

The degree of clinical severity of the unstable

hemoglobin depends in part on the site at which

the amino acid replacement occurs and on the

properties of the amino acid that is substituted(37). However, manifestations of the disordervary from mild hemoglobin instability which may

not be clinically significant, to severe hemoglo-

bin instability, which causes chronic hemolytic

anemia.

Unstable hemoglobin is diagnosed by isolating

and characterizing the abnormal hemoglobin, and

by determining its molecular stability. Since theamino acid sequence of normal hemoglobin isknown, and the sequences for some abnormalhemoglobins have been determined, the geneticmu tation und erlying these un stable globins havebeen identified.

Hemoglobin M—The Methemoglobinemias

A blue, cyanotic appearance is the chief clini-cal manifestation of methemoglobinemia, causedby HbM. Cyanosis occurs whenever a significantprop ortion of the hemoglobin in the circulatingred b lood cells is not carrying oxygen , and in thiscase it is because of an inability of the hemoglo-bin to bind and carry oxygen (37).

Five different mutations in the globin gene

wh ich lead to the prod uction of HbM h ave beenidentified, All of these are expressed as dominantphenotypes. The common characteristic of thesemu tations is that each results in substitu tion of an amino acid in the globin peptide w here the

heme groups are attached. (These mutations arespecific substitutions of histidine for tyrosine onthe alpha-globin chain [at positions 58 and 87] oron the beta-globin chain [at positions 63 and 92],

or by substitution of valine for glutamic acid on

the beta-globin chain [at position 67]. ) By alter-

ing the site of attachmen t of heme on the globinchains, these mu tations change the three-dimen-sional organization of the entire hemoglobin mol-ecule and interfere with its ability to combine withoxygen (100).

Measurement of Mutation Rates in theGlobin Genes

Using published d ata on the num ber of casesof hemog lobin disord ers from 10 countries (134),Stamatoyan nopou los and N ute (131) determ inedthe number of children with HbM and unstablehemoglobin among an estimated number of live-

births in the d efined p opu lations. They recordeda total of 55 cases of de nov o hem oglobin mu -tants, each of wh ich d erived from substitutionsof single nu cleotides in th e coding p ortions of theglobin genes: 40 children with unstable hemoglo-

bin (all beta chain mutations) and 15 with HbM

(10 beta chain and 5 alpha chain).12

The requi-

site paternity data, also from published informa-

tion on these mutations, had been collected on

19 of the 55 cases which were used to calculate

mutation rates (131). Their data are shown intable 3.

The mutation rates derived from these data areexpressed in mu tational events per alpha- or beta-gene nucleotide per generation, and range from5.9 X 10-9 to 19X 10-9 for a lpha- and beta-gene

nucleotides, respectively. The rate for all beta-gene variants together is 7.4X 10

-9per nucleo-

tide p er generation. These estimates correspondto mu tation rates per globin chain gene per gen -eration of 2.6X 10

-6to 8.3 X 10

-6. The rate per

gene is obtained by m ultiplying the rate p er nu -

‘2The number of alpha chain mutations detected wa s less than

expected, given the existence of two alpha-globin genes per beta-globin gene in an individual. A possible reason for this may be that

alpha-chain defects would produce physiological abnormalities atan earlier stage of fetal development than would beta-chain abnor-malities, since the alpha gene is activated much earlier, and m aybe causing a greater loss of fetuses early in gestation.



48 q Technologies for Detecting Heritable Mutations in Human Beings —

Table 3.–Rates of Mutation Leading to HbMand Unstable Hb Disorders

Mutation rate/ nucleotide/ Mutation rate/

Type of disorder generation globin geneb

Unstable Hbs . . .A lphaM variants . 4.2x 10-6 /423 nucleotidesBeta

Mvariants . . 18.9x 10

-98.3 x 10

-6 /438 nucleotides

Ail beta-chainvariants . . . . . . 7.4 x 10-9 3.2x 10-6 /438 nucleotides

aThe mutation rate per nucleotide is based on the rate of appearance Of the Par-

ticular disorder (number of mutants observed divided by 2 multiplied by the num-

ber of births in the generation(s) Included). This is the average mutation ratefor the particular disorder expressed per globin gene affected per generation.To calculate the mutation rate per nucleotide per generation, Stamatoyannopou-Ios and Nute took the frequency of each mutant divided by 2 (for two beta genesin a genorne) multiplied by 3 (since a given nucleotide can be substituted byany one of three other nucleotides), then divided that product by the numberof different substituted nucleotides actually observed In their data.

bT’he estimated mutation rate per alpha-globin and beta-globin gene was caku-

Iated as the mutation rate per nucleotide multiplied by the number of nucleo-tides constituting the coding portion of the corresponding gene (423 in the

alpha-globin gene and 438 in the beta-globin gene).

SOURCE: G. Stamatoyannopoulos and P.E. Nute, “De Novo Mutations Produc-ing Unstable Hbs or Hbs M. 11: Direct Estimates of Minimum Nucleo-tide Mutation Rates in Man,” Hum, Genef. 60:181-88, 1982.

cleotide by the n um ber of nucleotides in the cod-ing portion of the gene.

orders. The d iagnosis of these hemoglobin d is-orders requires structural and functional analy-ses of the aberrant proteins, and not all caseswou ld be id entified initially by th eir phenotyp icabnormalities; some degree of underascertainmentis inevitable in these data. This is particularly truein the data for the unstable hemoglobins, where

clinical manifestations of the genotype range frommild to severe. The more severe cases are likelyto be identified and reported in the literatur e.

Overall, the ord er of magn itude of the estimatesfor gene mutation rates, 1 mutation in 1 milliongenes per generation, is consistent with that of theestimates of the sp ontaneous h eritable m utationrate made by Neel and colleagues in the Japanesepop ulations that comp rised th e control group intheir ongoing stud y of the genetic effects of theatomic bombs (see below). This suggests that eventhough the rates derived from these data are based

on mutations observed in only alpha- and beta-globin genes, they are not very different from mu-tation rates in other expressed genes.

The rates calculated from these data are likelyto be minimum rates correspond ing to these dis-

STUDIES OF POPULATIONS EXPOSED TO ATOMIC RADIATION

The first d emonstration that ionizing rad iationcould induce genetic mutation is usually creditedto Muller (81) for his work with Drosophila, and

since then, many scientists have examined the na-ture and frequency of induced mutations in vari-ous sp ecies of animals, including w hole mamm als(120). There is an extensive b ody of experimen taldata th at suggests that exposure to rad iation andto certain chemicals can induce mutations in mam-malian germ cells. In humans, exposure to ioniz-ing radiation is known to cause somatic mu tation,and is suspected to enhance the p robability of ger-minal mutation (165). To date, however, the

avai lab le methods have provided no documented

evidence for the induction (by chemicals or by ra-

diation) of mutations in human germ cells.

The single largest population at risk for inducedheritable mutations is the group of surv ivors of the atomic bombs dropped on Hiroshima andNagasaki in 1945. Survivors of these bombs are

estimated to have received doses of radiation con-sidered to be “biologically effective”; in exper i-ments with mutation induction by acute exposureto radiation in mammals, similar kinds and dosesof radiation were sufficient to cause phenotypi-cally apparen t mu tations in offspring (and p re-suma bly other nonexpressed m utations as well).On the basis of this kind of experimental data,the assum ption was mad e that germinal mutationscould h ave been ind uced in people exposed to theradiation.

Efforts to detect genetic consequences of theatomic bombs detonated in Hiroshima andNagasaki have been in pr ogress continuously since1946, sponsored by the Atomic Bomb Casualty

Commission and by its successor in 1977, the Ra-diation Effects Research Foundation. These studies

have attempted to determine whether germinal

mutations are expressed as heritable mutations in

the survivors’ offspring, particularly as “untoward



Ch. 3—Kinds and Rates of Human Heritable Mutations q 49

pregnancy outcomes,” as single gene disorders,as chromosome abnormalities, or most recently,as abnormal blood proteins.

Early Epidemiologic Studies

The earliest clinical studies, as well as the con-

tinuing mortality surveillance study, sought to

identify a variety of health problems in the off-

spring of exposed atomic bomb survivors. Thesestudies consisted of epidemiologic surveys of 70,082 newborn infants born in Hiroshima and

Nagasaki between 1948 and 1953, including about

38,000 infants for which at least one parent had

been proximally exposed to the radiation (defined

as being within 2,OO O meters of the hypocenter,

wh ile distally exposed w as defined as being greater

than 2,500 meters from the hypocenter or not

exposed at all). Stillbirths, infant mortality, birth-

weight, congenital abnormalities, childhood mor-tality, and sex ratio were examined. An “un-

toward pregnancy outcome” was defined as a

stillbirth, a major congenital defect, death dur-

ing the first postnatal week of life, or a combina-

tion of these events. The investigators found no

statistically significant, radiation-related increase

for any “untoward pregnancy outcomes” (123).

One reason for this finding may be that these

health problems each have some genetic basis, but

that all are influenced by a variety of nongenetic

factors as well, some of which may have obscured

an effect of genetic mutation. In addition, a con-

tinuing study (the F1 Mortality study) of mortal-ity among children born between 1946 and 1980

has not demonstrated a relationship between ra-

diation exposure of the parents and mortality in

the offspring (56).

Cytogenetic Studies

Since 1967, Awa and colleagu es hav e screenedthe children of survivor s of the atom ic bom bs forcytogenetic abnormalities to determine whetherexposure of the p arents’ gametes to the radiationcaused a higher frequency of chromosome abnor-

malities in the offspring. After first reporting ahigher frequency of sex-chromosome abnormal-ities among 2,885 children of exposed parentscompared with 1,090 children of the control groupof parents, Awa (6) concluded that a mutagenic

effect of the radiation w as not d emonstra ted inthese data. The total frequency of chromosomallyabnormal children of exposed parents (0.62 per-cent) was higher than in the control group (0.28percent), although the difference was not statis-tically significant, and was similar to the fre-

quency of such a bnormalities observed in severalstudies of unselected, consecutive new born p op-ulat ions (120).

Results of their more recent analysis of the datashowed that 0.52 percent of the offspring of ex-

posed parents were determined to have an abnor-

mal chromosome constitution, compared to 0.49

percent of the offspring of the control group of

parents, which suggests that there is no signifi-

cant difference between the two groups of off-

spring in frequency of autosomal balanced and

unbalanced rearrangements and of sex-chromo-

some numerical abnormalities (8). This study did

not provide information on chromosome abnor-

malities which are associated with high mortal-

ity rates, such as most of the unbalanced struc-tural rearrangements and numerical chromosomeabnormalities involving the autosomes.

Biochemical Analyses

Beginning in 1976, one-dimensional electropho-resis was used to look for mutations affecting pro-teins found in red blood cells and blood plasma.Rare varian ts of these proteins, differing by a sin-

gle constituent amino acid, are taken to representmutations which occurred in the germ cells of oneof the parents if the variant is detected in a childand is absent from both parents. Classification asa “probable mutation” is made after inquiry intopossible errors of diagnosis and of stated parent-

age (124).

Neel and colleagues (93,96,121) analyzed 28 en-

zyme and serum proteins from blood samples ob-

tained from children who were identified in the

F1 Mortality study and who were born between

1959 and 1975 to survivors of the atomic bombs.

Over 10,000 children were examined in eachgroup of exposed and unexposed parents. These

children were also participants in the cytogenetic

studies, and the biochemical analyses made use

of the same blood samples that were used for the




chromosome analyses. To date, two probable mu-tations have been identified am ong offspring of proximally exposed parents (estimated to have re-ceived an average conjoint gonadal exposure of approximately 600mSv [60 rem 13]) in 419,66610-cus tests, while 3 new mu tations have been foundamong 539,170 locus tests of unexposed individ-uals. Although the number of new mutants issmall, a crude mutation rate can be calculatedfrom these data. The rate of mu tation in the ex-posed group is approximately 5 mutations in 1million genes, and the rate in the u nexposed groupis 6 mu tations in 1 million genes p er generation.Such rates have wide m argins of error and are n otstatistically different (96).

In 1979, the tests of electrophoretic variants

were expanded to include assays for reduced activ-

lsThe doses of radiation that the survivors are estimated to have

received are currently being re-evaluated in light of new informa-tion concerning the qu antity and quality of radiation released bythe two bombs (125). The revised estimate of exposure will prob-ably be lower than the current estimate of 60 rem.

CONCLUSIONS

This chapter has described several currently fea-sible approaches for determining rates of herita-

ble mutations. Each of the method s has its par-ticular problems. The genetic defects underlyingthe sentinel phenotypes are largely u nknow n.These phenotypes may result from mutationswithin genes at one or several genes, and may becomplicated by the presence of phenotyp icallysimilar but nongenetic traits, all serving to biasestimates of their frequency upw ards an d to cre-ate a w ide m argin of error. Even with a collec-tion of sentinel phenotyp es that can be reliablyidentified in newborn infants, population surveysof a sufficient size are difficult to perform. Todetect a small or moderate increase of sentinel

phenotypes, consecutive newborns of entire na-tional populations would have to be monitoredthoroughly for many years, perhaps as p art of anongoing surveillance program for birth defects andchildhood cancers.

ity variants of certain red cell enzymes .14 The p res-ence of such variants su ggests a deletion in theDNA, producing “null” mutations which wouldnot be identified by one-dimensional electroph o-resis. No su ch variants w ere found among 11,852locus tests of 10 erythrocyte enzymes in childrenof exposed and un exposed p arents (122).

Mutation Studies in PopulationsExposed to Radiation in

the Marshall Islands

N eel and h is colleagu es (93) rep orted r esults of electrophoretic studies of blood proteins from off-spring of parents exposed to rad ioactive falloutfrom the Bravo thermonuclear test explosion atBikini in the Marshall Islands in 1954. No vari-ant p roteins indicating new mu tations were foundin 1,897 locus tests of 25 proteins m easured in chil-

dren of unexposed p arents, and in 1,835 such testsin childr en of exposed paren ts.

‘—14An enz—yme level three standard deviations below the mean

and/ or less than or equal to 66 percent of normal was defined asa deficiency variant.

The routine staining and band ing of chromo-somes is the most straightforw ard of the current

methods, but they allow only major chromosomalchanges to be iden tified. More advanced ban d-ing techniques, which are not performed rou-tinely, allow smaller chromosome abnormalitiesto be d etected in some cases.

Among the current m ethods described here, thedetection of electrophoretic protein variantscomes closest to mirroring precise changes inDNA. However, one-dimensional electrophore-sis is limited by the relatively small nu mber of pro-teins it can examine and by the fact that it doesnot read ily d etect the absence of gene prod ucts,

caused by deletions or by specific nucleotide sub-stitutions. Detection of enzyme deficienc

yvari-

ants, although still in early stages of development,could complement electrophoresis by detectingloss-of-activity variants. In addition, large pop-



52 q Technologies for Detecting Heritable Mutations in Human Beings —

The different approaches currently available re-flect the various forms that m utations an d th eirconsequences take. Each of the current approaches

to studying mutations is generally l imited to par-

ticular kinds of mutations or to a particular set

of outcomes . The cytogenetic and biochemical ap -proaches discussed are limited to detecting a small

subset of mu tations that occur in h um an DNA.Identification of sentinel phenotypes and chromo-some abnor malities limits the observations to apart icular set of outcomes, mostly in th e severeend of the spectrum. The available data are basedon m utations observed in a small fraction of thegenes, about 40 to sO, out of a total of about50,000 to 100,000 in the human genome.

Aspects of the current method s that limit theirscope also make th em inefficient. The prop ortionof all mu tations that these method s can identifyis relatively small and the limited number of suit-

able genetic endpoints (acceptable sentinel pheno-types, chromosome abnormalities, and protein

variants) makes it difficult to sample more thana few m utational events per individu al. As a re-sult, huge populations of “exposed” and “un-exposed” people would be needed to find outwh ether the exposure resulted in excess heritablemutations (71). Not surprisingly, perhaps, thereis no documented evidence to date for a meaning-

ful difference in r ates for any typ e of mu tationbetween two populations (or in the same popu-lation at different times).

Sentinel phenotypes and cytogenetic analysesprov ide a highly selective app roach to some of the worst possible consequences of new muta-tions. How ever, they do not p ermit more thana partial view of the broader consequences of mu -tations in human beings. Detailed molecular,physiological, and clinical information on thekinds and rates of new mu tations w ill eventuallybe needed to assess the imm ediate or potential im-pact of new mu tations.



Chapter 4

New Technologies forDetecting Heritable Mutations



Contents

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Short Descrip tions of the New Technologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Two-Dimensional Polyacrylamide Gel Electrophoresis . . . . . . . . . . . . . . . . . . . . . 56DNA Sequencing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Restriction Fragment Length Polym orp hism s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58One-Dimensiona l Denatur ing Grad ient Gel Electroph oresis . . . . . . . . . . . . . . . . . 59

Two-Dimen sional Denatu ring Grad ient Gel Electrophoresis. . . . . . . . . . . . . . . . . 61Ribonuclease Cleavage of Mismatches in RNA/ DNA Heterod up lexes . . . . . . . . 64Subtractive Hybridization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Pulsed Field Gel Electrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

TABLE

Table No. Page

6. Current and New Method s for Detecting Hu man Heritable Mutations . . . . . . . 55

FIGURESFigure No. Page

7. Produ ction of Restriction Fragmen t Length Polymorp hisms UsingRestriction Enzym es and Separa tion of Different DNA FragmentSizes by Agarose Gel Electrop horesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

8. Restr iction Fragm ent Length Polym orp hism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 609. One-Dimensional Denatu ring Grad ient Gel Electroph oresis. . . . . . . . . . . . . . . . 62

10. Two-Dimensional Denatu ring Grad ient Gel Electroph oresis . . . . . . . . . . . . . . . 6311. Ribonuclease Cleavage of Mismatches in RNA/ DNA Heterod up lexes. . . . . . . 6512. Subtractive Hybridization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67



56 . Technologies for Detecting Heritab/e Mutations in Human Beings

SHORT DESCRIPTIONS OF THE NEW TECHNOLOGIES

Two-Dimensional Polyacrylamide GelElectrophoresis

A d evelopm ent following from the experiencewith one-dimensional electrophoresis for protein

varian ts (see ch. 3) is two-dim ensional p olyacryl-am ide gel electrophor esis, “2-D PAGE, ” the on lytechniqu e described in this chapter that is cur-rently in use in stud ies of hu man beings. With thistechnique, proteins are separated first by iso-electric focusing on th e basis of their molecularcharge, and then by electrophoresis on the basisof their molecular weight. From 500 to 1,000 pro-teins may be visible from a single sample on one

gel, and about 100 of these may be sufficiently

clear to detect rare variants in a child that are not

present in the parents (92).

To search for new mutations in proteins using2-D PAGE, blood samples are taken from child/

mother/ father trios. The program that has hadmost experience with this technique for mutationresearch, the University of Michigan MedicalSchool, collects cord blood from newborns andparental blood sam ples at the time of the child’sbirth. Each blood sample is fractionated into sixparts: platelets, plasma, nonpolymorphonuclearcells, polymorphonuclear cells, erythrocyte (redblood cell) membranes, and erythrocyte cytosol(the contents of the cell) (95). Each fraction istreated and analyzed separately.

Separation in the first dimension is by thecharge of the protein molecules, qualitatively sim-ilar to the separation using one-dimensional elec-troph oresis. Each sample is run on a long, thin(1 to 2 mm in d iameter) cylinder of polyacryl-amide in a hollow glass tube using a techniquecalled isoelectric focusing. The separation time canbe speeded u p or slowed d own d epending on thevoltage applied. After the first separation, the gel,wh ich looks like a clear nood le, is “extruded ” fromthe tube.

In the second sep aration, the “noodle” is laid

across the top of a gel abou t 7 inches squ are, thatis held between two glass or plastic plates. A cur-rent is passed through the gel, and the proteinmolecules migrate down through the gel at a speed

that is proportional to their size. A final steprenders the proteins visible by autoradiography(for proteins that can be isotonically labeled) orby stains such as Coomassie Brilliant Blue andsilver- or nickel-based stains.

The trained eye or a computer program detectspattern s of spots on the gel, though m ost of thestained areas are too crowded or indistinct fordrawing conclusions about specific proteins. Onlyspots in relatively clear areas and of a sufficientintensity are chosen for scoring. The ana lysis en-tails comparing th e parent/ child trios for d iffer-ences. No tw o gels are run in precisely the sameway, so it is often d ifficult to sup erimpose a setof three gels over one another and to look fordifferences. While protein spots bear the sam erelationships to each oth er on d ifferent gels, they

routinely appear at slightly different coordinates.For each typ e of samp le mentioned above, a

standard constellation of protein spots is scoredfor genetic variation. A mu tant p rotein can ap-pear on the gel as a spot for w hich th ere is no cor-respond ing spot on parental gels. The gels can beinterpreted visually, which is the traditionalmethod, and currently the most accurate, butcomputer-assisted interpretation is desirable. Twosteps in analysis lend themselves to computeriza-tion. First, and now routinely done, the proteinspots can be “read” by computer for location andintensity. The second step is compu terized com-parison of different gels. Two approaches forcomparison are being d eveloped in different lab-oratories, each of which will probably be usefulon a “pr odu ction scale” in the near futu re (92).

Eventually, computers will be relied on to screen

gels, and to eliminate from further consideration

the overwhelming majority of paren t/ child trios

with no variants.

In comparison to one-dimensional protein sep-aration, a relatively small number of parent/ childtrios have been analyzed w ith 2-D PAGE, and n oheritable mutations have been found. However,

the results of these analyses have r evealed a widearray of largely unknown proteins, and distinctd ifferences in the sp ectrum of proteins from d iffer-ent cell types. The range of electrophoretic vari-



Ch. 4—New Technologies for Detecting Heritable Mutations q 5 7

ants of a nu mber of these proteins have been char-acterized in plasma (104) and in erythrocyte lysate(105). In plasma, the variation is substantially

higher than has been reported in other blood

fractions.

The advantage of 2-D PAGE is that it can be

done now without a major developmental break-

through and without a great initial cost, though

the running expenses are significant. It may besome years before any of the DNA analytic meth-ods catch up to protein analysis in practical ap-plication. The future of protein analysis for d e-tecting mutations after DNA methods becomeroutine is unclear, however.

There are t echnical limitations of the 2-D PA GE

technique. First, the technique itself is technicallydemanding. Second , the nu mber of proteins thatcan be scored on each gel is limited by the degreeof separation possible on plates that are of a prac-

tical size.

As with most of the other techniques, only acertain spectrum of mu tations is d etectable with2-D PAGE. Chromosome rearran gements in gen-eral are not revealed through electrophoretic var-iants. This technique has its greatest strength indetecting p oint mu tations and larger deletions inthe DN A that change the size or charge of a pro-tein molecule. Point m utations that d o not causesuch changes, and small insertions or deletions,are less likely to be detectable with this method.Currently, null mutations cannot be detected relia-

bly, and this is probably the greatest disadvantageof 2-D PAGE. Overall, 2-D PAGE should detectan estimated 25 to 35 percent of all spontaneousgene mutations. Several aspects of 2-D PAGE of blood proteins, in addition to computerized com-parisons of gels, are amenable to improvement,and at least one approach is being developed toimprove detection of null mutations (92).

The identity of most of the proteins th at canbe visualized on two-dimensional gels is un-known, Although it has not yet been done, thetechnology exists to recover proteins from gels,

purify them, and then determine their amino acidsequences (92). Once amino acid sequences areworked out, the nucleotide sequences that codefor them can be ded uced an d n ucleic acid probescan be assembled to correspond to the protein-

coding nucleotide sequences. The p robes couldtheoretically be used to locate the correspondingDNA within the human genome, thereby deter-mining exactly where in the genome the codingsequen ce lies.

DNA Sequencing

Conceptually, the most thorough an d straight-forw ard analysis of the genome w ould consist of lining up the chromosomes and determining thenucleotide sequence from end to end on each one.Once the sequence was kn own , it could be storedin computer memory.

The advantage of determining the completegenomic sequence would be tha t any nu cleotidesequence that was identified in any laboratorycould immediately be mapped to its chromosomallocation. As nucleotide sequences for various

parts of the genom e were obtained in d ifferent lab-oratories, both common polymorphisms and raremutations could be identified by making compar-isons between th e total genomic sequence and p ar-tial sequences. Furthermore, the method woulddetect mutations and p olymorphisms no n-tatterwh ere they occur —in introns, exons, and repeti-tive sequences.

Walter Gilbert, Fred Sanger, and their col-leagues developed the theoretical basis for thetechniques that makes it possible to consider se-quencing the genome (67,118) and in 1984, Ge-orge Church and Walter Gilbert described amethod for sequencing any part of the genome(2o), although the method could be applied to se-quencing the entire genome. The genomic DNAwou ld be cut into pieces small enough to be han-dled in the laboratory, and then chemical meth-ods w ould be u sed to determine the nu cleotide se-quence with in each fragment. The method s arewell understood, are being applied in dozens of laboratories around the world, and are a stand-ard exercise in the ed ucation of molecular biol-ogy graduate students (33). Through February

‘Polymorphisms are “common” mutations that are present in thepopulation at a relatively high frequency. Several alternative ver-sions of the same basic gene sequence, all equally functional, canco-exist. New mutations are defined as occurring in less than 1 per-cent of the population, excluding the more common polymorphicvariants.



58 . Technologies for Defecting Heritable Mutations in Human Beings —

1985, the m ethods h ave been used to examine d is-crete and defined segments of DNA from organ-isms with large genomes and complete genomesfrom smaller organisms, sequencing a total of 4,045,305 base p airs (150)—rough ly 1/ 800 of the

human genome.

Practically, we are far away from being able

to sequence the entire genome, not because thetechniques are una vailable but because they areexpensive and slow when applied to a task of thema gnitu de of sequen cing 3 billion nu cleotides. Ameeting during the summer of 1985 consideredthis enormous task and estimated the cost at about

$1 per nucleotide, or $3 billion overall. The ef-fort can also be estimated in terms of technician-years, and that comes to abou t 200 technician-years to seq uen ce 100 million nu cleotides, w hichwou ld be expected to yield about one new mu -tation.

Although it is conceptually satisfying, the enor-mous resources necessary to sequence the entiregenome m ake it unlikely that the task w ill be un-dertaken soon. However, genomic sequencingtechnology is still progressing, and further auto-ma tion of the p rocess ma y increase its efficiency.Continued encouragement of those efforts andpr ovision of sufficient resour ces to develop tech-niques may make it possible to sequence the en-tire human genome in a reasonable time. Whenit is don e, it w ill probably be und ertaken as p artof an effort to p rovide m ore general informationabout human genetics, rather than to identify hu-man mu tations. If the genom e is sequenced, forwh atever reason, it will be a great boon to mu ta-tion stud ies by provid ing a reference for all tech-niques that involve sequencing or the u se of re-striction enzymes.

Restriction Fragment Length

Polymorphisms

A m ethod th at offers a less exhaustive surveythan genomic sequencing, but which is potentiallymor e efficient, involves the use of restr iction en-

zymes and gel electrophoresis to detect nucleo-tide substitutions or small deletions in the DN A(98).

Genomic DNA is isolated from white bloodcells and treated w ith restriction enzym es, chem-icals that recognize specific sequences in the DNAand “ cut” the DNA w herever such sequences oc-cur (see fig. 7). (The number of fragments pro-du ced is d etermined by the frequency of occur-rence of the particular enz yme recogn ition site in

the DNA sequence. ) Restriction site analysis doesnot, in practice, examine every nucleotide. How-ever, the use of a set of combined restriction en-

zymes increases the number of restriction sitesidentified, allowing examination of a larger por-

tion of the DNA, including both expressed and

nonexpressed regions.

After the DNA is fragmented (see fig. 8), there

are two alternative ways of proceeding, one using

a gene cloning method, and the other using a non-cloning, direct method.

Cloning is the process of “growing” human

DNA in other organisms; it reduces the amountof DNA needed—in most cases corresponding to

the amount of blood that must be drawn—but it

increases the number of laboratory manipulations.

Addit ionally, i t introduces some uncertainty

about the possibility that DNA changes will oc-

cur during cloning. This can be checked, but re-

quires diligence.

In the gene cloning method, the human DNA

fragments are incorporated into the genetic ma-

terial of a virus called “lambda. ” Lambda repro-

duces in the bacterium E.coli, and large quanti-

ties of viral DNA containing the hum an segmentsare produced. Following isolation of the clonedhu man DNA, each clone is treated w ith a set of restriction enzymes that cut it into smaller frag-ments. These collections of fragments are thenseparated by gel electrophoresis, producing d is-tinct bands visible by gel staining methods. Avisual comparison of the location of the bands de-rived from parental DNA and from children’sDNA show s whether a mu tation has occurred ina site for a r estriction enzyme. The p resence of a band in the child’s DNA th at is not pr esent inthe parents’ DNA, or vice versa, indicates a

nu cleotide su bstitution or deletion restriction en-zyme recognition site and suggests that a herita-ble mu tation has occurred.




Figure 7.— Production of Restriction Fragment Length Polymorphisms Using Restriction Enzymes andSeparation of Different DNA Fragment Sizes by Agarose Gel Electrophoresis

Agarose sizing gelelectrophoresis

DNA I

DNA II

I

t t t

I

4 fragments result: c, b, d, and a

,

t t

Base change leads toloss of restrictionenzyme site and

creates 3 fragments c, e, and a

t = restriction enzyme site of cleavage

SOURCE: Off Ice of Technology Assessment

An alternative method bypasses the lambdacloning procedure, and an alyzes DNA fragmentsdirectly. Fragments from the entire human ge-nom e are spread ou t on an electroph oretic gel,forming smears of indistinguishable bands. In or-der to visualize the location of particular segmentsof DNA and to evaluate their position relative tothe equivalent band in the reference samp le, thenaturally dou ble-stranded DNA is dissociated intosingle-stranded DN A w hile still in the gel and a l-lowed to reassociate w ith sma ll pieces of specificgenes which are added to the gel and which areradioactive (probes). Those original DNA se-quences that pair with the probes are now tagged .The DNA is then transferred to a flexible plasticmembrane conserving the spatial arrangement of the DNA (a procedure known as Southern blot-ting) and autoradiographed; bands are detectedw here the tagged sequen ces are present. As withthe cloning m ethod above, the app earance of a

t

( -)

Largerfragments

IDirection

ofcurrent

Smallerfragments

(+)

DNA I DNA II

Positions of differentfragment sizes after

treatment withrestriction enzyme

new band or the disappearance of an old bandsuggests a nucleotide substitution or deletion in

a recognition site of one of the restrictionenzymes.

One= Dimensional Denaturing GradientGel Eiectrophoresis

A m odification of the standard electrophoreticgel procedu re, proposed by Fischer and Lerman(34), allows DNA to be separated not only on the

basis of size, but also on the basis of sequence

of nucleotides even if differences do not occur at

restr ict ion enzyme recognit ion si tes. Double-

stranded DNA dissociates into single-stranded

DN A wh en it is heated or w hen it is exposed todenaturing chemicals (e.g., formamide or urea).A gradient of increasing strength of such chemi-cals can be pr odu ced in a gel so that DNA sam -ples w ill travel in th e direction of the electric cur -



.

60 • Technologies for Detecting Heritable Mutations in Human Beings

1

Figure 8.— Restriction Fragment Length

Mother’s DNA Father’s DNA

Polymorphism

Child’s DNA

L /

GENE CLONING METHOD/ \

DIRECT (NONCLONING) METHOD

ViralDNA

Fragmentsof

viralDNA

o3A ( - )

Electriccurrent

\Separate fragments

restriction enzymes, cut DNA into double-stranded fragments of

2A Clone each fragment into the bacterial virus lambda, infect into E.co/i, grown in petri dishes, and allow the virus toreplicate within the bacteria. Isolate large quantities of the viral DNA, which now contains segments of human DNA, andcleave it with restriction enzymes.

o3A Separate DNA fragments using agarose gel electrophoresis. Visualize bands corresponding to different sizes offragments using fluorescent staining. Fragments found in the child’s DNA samples and not in the parents’ DNA may con-tain heritable mutations. These bands can be removed and the DNA analyzed for specific mutations.

Direct (noncloning) Method:

o2B Apply samples of DNA fragments to an electrophoretic gel, producing smears of indistinguishable bands.

o3B Dissociate DNA fragments into single strands and incubate with radioactive, single-stranded, 32P-labeled DNA probes forspecific human genes. The probes hybridize with complementary sequences in the DNA and label their position in the gel.Visualize the position of bands using autoradiography. Changes in the position of bands in the child’s DNA compared tothe parents’ DNA suggest possible new mutations. Bands can be isolated and DNA analyzed for sequence differences.





rent, separating by size, and will also begin to

dissociate as they reach their particular critical

concentration of denaturing chemical. Dissocia-tion causes the molecule to split into constituent

parts, or unravel, and get stuck in the pores of the gel (84).

Every unique strand of DNA dissociates at a

unique concentration of denaturant. In fact, thedifference of only one n ucleotide betw een tw ootherwise identical strands of double-strandedDNA of 250 nucleotides in length is enough tocause the strand s to d issociate at different con-centrations of denatu rant chemical, and to stoptraveling at different locations in the gel. Again,a comparison between the banding pattern of par-ents’ and child’s DNA analyzed in this way m ayidentify a w ide range of m utations in all regionsof the DNA.

In this method, total genomic DNA is isolatedfrom individuals and is heated to form single-stranded DNA (see fig. 9). It is then mixed withradioactive, single-stranded DNA probes that cor-

respond to a small portion of the genome, Fol-

lowing incubation so that double-stranded hybrid

molecules (“heteroduplexes”) will form, they are

treated w ith restriction enzymes and eIectroph oresedin a denaturing gradient gel. The presence of a

single mismatch between a nucleotide in the “nor-

mal” probe and the corresponding segment from

a person bearing a mutation makes the heter-oduplex slightly more sensitive to denaturing

chemicals so that such a m olecule w ill stop trav -eling in the gel before the point at wh ich a p er-fectly p aired heterodu plex will stop. The gel isdr ied and exposed to X-ray film, and the result-

ing autoradiogram is examined visually for differ-ences in banding patterns between parents’ andchild’s DNA.

Two-Dimensional Denaturing Gradient

Gel Electrophoresis

Leonard Lerman has proposed a technique

wh ereby a two-dimensional separation of parents’and child’s DNA is u sed to d ifferentiate amongDNA sequences common to all three members,polymorph isms in either parent w hich are trans-mitted to the child, and new mutations in thechild’s DNA (57). Like the two-d imensional p oly -

acrylamide gel procedure for pr otein separ ationdescribed above, two-dimensional denaturinggradient gel electrophoresis (2 DGGE) compareslocations of spots on a gel (in this case, DNAspots) representing va rious types of m utations inthe nu cleotide sequ ence in expressed a nd nonex-

pressed regions of the DNA. This method sepa-rates DNA fragments on the basis of differencesin base composition (or sequence), after separat-ing them on the basis of differences in length.2DGGE detects differences between DNA hetero-du plexes by responses of their structure to grad -ual changes in d enaturant concentration in the gel.

Genomic DNA from the parents in one sampleand from the parents and children in anothersamp le would first be treated w ith restriction en-zymes and separated by size on separate agarosegels in a single lane (see fig. 10). The gel strips

are then laid across the top of two denatu rant gra-dient gels and the DNA is electrophoresed throu ghthe increasing gr adient of d enaturan t. Each gel isthen cut into horizontal slices, each of which con-tains pieces of the original homodu plex DNA of all sizes, but including on ly those pieces that dis-sociated at the same denaturing concentration(“isomelting” groups). These gel slices containinghomod up lexes of parents’ DNA or of parents’ andchild’s DNA are physically removed (or collectedon removable membranes) heated, and allowedto reassociate as heteroduplex molecules (see fig.lo).

In order to d ifferentiate between common areasof the DNA among parents and children, anothergradient denaturing gel is used to d enature themixtures of double-strand ed, isomelting DNA of the parents (M 1 / M z + F1 / Fz) and another of theparents’ DNA m ixed w ith the child’s (M,/ Mz

2+

F1 / F2 +C 1 / Cz). Various double stranded combi-nations of parents’ and child’s DNA are produced(e.g., M

2 / C

l, F

1, / C2, etc.). The two sets of mix-

tures are compared in the final two gels.

Sequences that pair up perfectly represent com-mon sequ ences to all three (level “i” in fig. 10).

At the bottom, a narrow region, dense with hetero-duplex fragments, represents the largest fractionof the sample—those that are perfectly matchedbetween base pairs and therefore are located atthe same denaturant concentration as the origi-nal homod up lexes contained in the narrow slice



62 . Technologies for Detecting Heritable Mutations in Human Beings —

Figure 9.—One-Dimensional Denaturing Gradient Gel Electrophoresis

Child’s DNAFather’s DNAMother’s DNA

o1 /- I

\ I i I

1Treat with

1 /

1 I restriction \

/ 1 I enzymes

/ 1 ~

&

, I

i

‘,

ITreat with

/’ \ I \ restriction

Dissociate andreanneal with32P-DNA probe

Dissociate andreanneal with32P-DNA probe

o2 Dissociate andreanneal with‘2P-DNA probe

ISeparate fragments

o3

Probe Father Mother Child

( - )

( + )

Increasingconcentrationof denaturant

Iisolate genomic DNA from white blood cells. Using restriction enzymes, cut DNA into (double-stranded) fragments of- various lengths.

o2 Dissociate double-stranded DNA fragments into single strands, and reanneal in the presence of radioactive 32P-labeled,single-stranded DNA probes. Heteroduplexes form between probe and sample DNA even if the base sequences are not

perfectly complementary; if mutations are present, some of these heteroduplexes will contain mismatches.

o3 Separate heteroduplex fragments in a denaturing gradient gel. Fragments with mismatches denature in a lower

concentration of denaturant than fragments that are perfectly complementary. Visualize the position of the fragmentswith autoradiography. Fragments containing the child’s DNA that denature sooner than the parents’ fragments can beanalyzed for new mutations.




Ch. 4—New Technologies for Detecting Heritable Mutations • 63

01

02

03

Figure 10.—Two-Dimensional Denaturing Gradient Gel Electrophoresis

Mother’s DNA Father’s DNA Child’s DNA

o5

06

Increasing

o7 concentra t ion

of denaturant

Increasing

concentrationof denaturant

(1)lsolate genomic DNA from white blood cells. Using restriction enzymes, cut DNA into double-stranded fragments Of- various lengths.

o2 Mix mother’s and father’s DNA fragments (M 1 /M, + F,/F,) in one set, and mix parents’ and child’s DNA fragments(M,/M, + F,/F, + Cl /C,) in another set.

o3 Separate fragments by length in an agarose sizing gel. Cut out a lengthwise strip of the gel, containing a range of fragmentsizes, and lay the strip across the top of a denaturing gradient gel.

o4 Separate fragments along an increasing concentration of denaturant. Cut out a slice of the gel at a narrow interval of the

temperature gradient, or collect a similar interval on a removable membrane.

o5 While maintaining spatial arrangement of fragments in the narrow slice, dissociate the double-stranded fragments into

single strands, and reanneal them to form heteroduplexes (e. g., M2 /C1, F,/CZ, F2 /M,, etc.) of various combinations of strandsof the original homoduplexes (M,/Mz, F1 /F z, and Cl /C Z).

o6 Lay the thin strip of heteroduplex fragments arranged by size across the top of a new gradient gel, and separate in thedirection of an increasing concentration of denaturants.

o7 Comparing the two final gels, the DNA spots are grouped in three distinct regions (see text for explanation). Analyze spotsin the highest region of the gel for DNA fragments with possible new mutations.




64 c Technologies for Detecting Heritable Mutations in Human Beings

removed from the gel or collected in th e mem-brane.

Polymorphisms that are present in one parentor the other show u p as spots that d enature earlierthan the perfectly matched DNAs (level “ii”). Inthe m idd le region of the gel, a diffuse d istribu-tion of fragments represents heteroduplexes withsingle mismatches between the tw o strand s, caus-ing them to dissociate earlier in the gel (at a lowerdenaturant concentration, occurring above theperfectly m atched heterodu plexes). This regionwould contain inherited polymorphisms as wellas new m utations, but the region would be toodense with DNA spots to distinguish individualspots.

Molecules that contain more than one polymor-phism per fragment denature even earlier (level“iii”) since multiple mismatches in a fragment have

an additive effect on the fragments stability in thedena turing grad ient. The highest area in the gel,where heteroduplexes dissociate in the lowestdenatu rant concentration, is where fragments w ithmore than one mismatch would be located. Thesemismatches could be new mu tations or inheritedpolymorphisms. Since the number of such frag-ments in the sample is likely to be small, this re-gion would show discrete, nonoverlapping spots;DNA spots in this third level of the gel can beremoved from the gel and analyzed to identifypossible new heritable m utations. By comp aringthe two final gels, new spots should be ap parentto a trained eye. These would represent singlebase pair changes or very small deletions and re-arrangem ents in all regions of the DNA.

In theory, it should be possible to cut manystrips of isomelting regions out of the first gel andderive an analysis of each melting region by a ser-ies of second gels. By comp aring th e child’s DN Ato his or her p arents’ DN A, this approach allowsfor large portions of the genome to be examinedsimultaneously.

Ribonuclease Cleavage of Mismatchesin RNA/DNA Heteroduplexes

A technically simpler approach to detecting mu-tations in cloned an d genom ic DNA has been p ro-posed by Richard Myers (85,86). This technique

uses an enzyme, ribonu clease A (RNaseA), thatcleaves double-stranded RNA/ DNA h eterodu-plexes where a specific mismatch of base pairs oc-curs. RNaseA cleaves the RNA / DN A moleculewhere a cytosine (C) in RNA occurs opposite toaden ine (A) in DNA . (C norm ally pairs with G

[guanine] and A nor mally pairs with T [thym i-dine]. ) The idea behind this method is similar tothat of restriction enzymes, each of which cleavesDNA at sp ecific normal sequences, except thatRNaseA cleaves RNA/ DNA hybrid m oleculeswhere mismatches occur. The efficiency of thisapp roach depend s on the num ber of different mis-matches that can be recognized a nd cleaved. Thegreater the num ber of enzymes that can be foundto cleave d ifferent mism atches, the more typ es of mu tations can be detected. This method shoulddetect nucleotide substitutions over a large por-tion of the DNA.

With this m ethod, genom ic DNA is first iso-lated and mixed w ith radioactively labeled RNAprobes (see fig. 11). The mixture is heated so thatthe strand s dissociate and then reassociate ran-domly with complementary strands. When RNAmolecules bind w ith DN A m olecules of homolo-gous sequences, the result is one type of “heter-odu plex.” “Hom ologous” means that the sequenceis close enough between th e two molecules thatthey will form a stable double-stranded molecule.The presence of a single base pair m ismatch doesnot prevent double-stranded heteroduplexes from

forming, however.

The heteroduplexes are treated with RNaseAand separated according to size on a standardagarose gel. Perfectly paired molecules will be un-affected by RNaseA and will form bands on thegel that correspon d to th eir original size. How -ever, imper fectly paired molecules w ill be cut bythe enzyme at the site of the C:A mismatch, andtwo separate fragments will result, showing upas two sep arate band s on the gel. For analysis,paren ts’ and child’s DNA are treated an d elec-trophoresed separately, and the gels comp ared for

different patterns by autoradiography to revealthe occurrence of any possible mismatches. Cur-rently this method is applicable only for (RNA)C:A(DNA) mismatches, which represent 1 of 12 pos-sible types of misma tches between these heter-oduplexes.




Figure 11 .—Ribonuclease Cleavage of Mismatches in RNA/DNA Heteroduplexes

o Child’s DNAMother’s DNA Father’s DNA

Probe

o3

mismatchcut into 2 pieces)

o4 ( - )

Electriccurrent

i

(+)

-

—

—

—




Subtractive Hybridization

Detecting mutations would be much easier if

it were possible to ignore the millions of nucleo-

tide sequences that are identical in both parents

and child and , instead, focus only on the relativelyfew sequences that are different. Currently, thebest estimate for the frequency of human muta-tions is about 1 in 100 million base pairs, or about

30 per genom e, so it is necessary to exam ine 3 bil-lion base pairs from each of the parents and a childand find those 30 that differ. George Church had

conceived of a method to find sequences in achild’s DNA that are n ot pr esent in either parent’sDN A (25).

While there is no experience with Church’s pro-posed method, and therefore, no informationabout its feasibility, it is a promising idea. Th ebasis of Church’s method is to view the hu man

genome as a group of unique nucleotide se-quences. In this context, “unique sequence” meansa sequence that m ay occur on ly once in the en-tire genome. To explain what a unique sequenceis, consider the shortest and longest sequence inthe genome, a single nucleotide and the entire ge-nome. A single nucleotide is not un ique; A, T,G, and C occur repeatedly throughout the ge-nom e. At the other extrem e, the entire genom eis unique. Somewhere in between those extremesis a length of nu cleotides that is long enough tobe uniqu e withou t being too long to hand le ex-perimentally.

It turns out that a sequence of 18 nucleotides

is long enough to be unique; on stat ist ical

grounds, any sequence of that length or greaterthat occurs once in the genome is not expectedto occur a second time. Therefore, if every p os-sible sequence of 18 nucleotides, or “18-mer,” is

synthesized, this collection of 18-mers will include

sequences that together represent every possible

sequence of 18 nucleotides of DNA (there are 418

or 70 billion possibilities). Some proportion of

these are actually present in human sequences,some a re not. It is feasible to make th e entire col-

lection of 70 billion different 18-mers, and a sin-gle synthesis provides sufficient 18-mers for manyexperiments.

To look for mutations, DNA from parents andchild is isolated an d cut w ith restriction enzym es

into relatively small lengths of 40 to 200 base p airs

and dissociated into single strands (see fig. 12).

This parental DNA alone is mixed with the col-

lection of 18-mers under conditions that permit

the formation of perfect hybrids: each nucleotidesequence of an 18-mer hybridizes exactly with

each corresponding genomic sequence. The con-ditions are also such that an 18-mer that inter-acts with a genomic sequence complementary to just 17 of the 18 nucleotides will not form a sta-

ble hybrid. When the hybridization reaction iscomplete, the 18-mers that did not find a perfect

match with any parental DNA sequences are left

behind as single-stranded molecules, unchanged

by their participation in the hybridization re-

action. Because they are single stranded, they canbe separated from all the other 18-mers that arehybridized to parental DNA. This unbound frac-tion of the mixture is retained, while the fragments

of parental DNA that bind to the 18-mers are dis-carded.

This pool of single-stranded 18-mers which didnot bind to any pa rental DNA sequ ence is thenmixed with similarly prepared DN A from thechild, and hybridized under the same conditions.This time, how ever, the m olecules of interest arethe sequences that bind to the child’s DNA (andnot previously to the parents’ DNA). Any 18-merthat hybridizes perfectly with a sequence in thechild’s DNA will identify a sequence that is presentin the child’s DNA an d absent from the parents’

DNA. Such sequences may contain new muta-tions present in the child’s DNA. It should thenbe possible to isolate and characterize this hybridformed b etween the 18-mer and child’s DNA, sothat the m utant sequence can be d etermined. If mu tations occur at a frequ ency of 1 per 100 mil-lion nu cleotides, and if this technique w orks per-fectly, it could detect 30 to 40 nucleotide sequ encesin the child’s DNA that are not p resent in the par-ents’ DNA.

This approach is the least well developed of allthe ones d iscussed in this report, and its feasibil-ity is unknow n. If it does pr ove feasible, this ap-proa ch could identify short sequences containingmu tations in any p art of the DNA, allowing fur-ther d etailed stu dy (e.g., by DNA sequ encing) of the kinds of mutations that are occurring.



Ch. 4—New Technologies for Detecting Heritable Mutations 67

Figure 12.—Subtractive Hybridization

Father’s DNA Child’s DNA

o2 Child’s DNA fragments

/

Separate anddiscard bound

fragments

o3

04

05

Discardunbound

fragments

o1 Isolate genomic DNA from parents’ and child’s white blood cells; mix parents’ DNA samples together, while keepingchild’s DNA separate. Cleave DNA into fragments with restriction enzymes.

o2 Dissociate parents’ DNA fragments into single strands, and reanneal with a set of unique single-stranded 18mersequences. 18mer sequences complementary to sequences in the parents’ DNA will form double-stranded, or bound,fragments. The remainder, the unbound fraction, will remain single stranded.

o3 Separate the bound fraction from the unbound fraction. The bound fraction is discarded.

o4 Unbound 18mers for which no parental complement exists are allowed to hybridize with the child’s DNA fragments underthe same conditions of dissociation and reannealing. Any fragments of the child’s DNA that hybridize with unboundfraction of 18mers represent sequences present in the child’s DNA that are not present in either parents’ DNA.

o5 Isolate these heteroduplexes and analyze them for specific new mutations.

SOURCE Off Ice of Technology Assessment




Pulsed Field Gel Electrophoresis

If hum an DN A w ere short and simp le, it couldbe cutup with restriction enzym es and separatedelectroph oretically into discrete band s, each rep -resenting a particular segment of the total DNA.However, human DNA is so long that electropho-retically separated fragments make a continuoussmear of bands. If the DNA were cut in only afew p laces to prod uce only 100 or 200 band s, thepieces would be too big to pass individuallythrough the pores of a standard electrophoreticgel. A new technique, pulsed field gel electropho-resis (PFGE) is being develop ed to allow sep ara-tion of fragments of hum an DN A larger than canbe separated with other electrophoretic techniquesand to examine such fragments for evidence of mutations. In theory, the procedure will detectsubmicroscopic chromosome mutations, includ-

ing rearrangements, deletions, breaks, and trans-positions. At present, the method cannot hand leintact hu man chromosomes although it worksvery w ell with sm aller chromosomes from lowerorganisms and with small fragments of humanchromosomes (157,158).

The standard process for isolating genomicDNA rand omly breaks the long molecules intopieces. In the PFGE procedu re, genomic DN A istreated so that random breaks are avoided. Wholecells are suspend ed in liquid agarose and after theagarose solidifies into a gel, enzymes are addedto degrade proteins and RNA. This apparently

minimizes random damage to the high molecu-lar weight DNA. DN A is cut into exact and p re-dictable fragments by addition of a “rare cutter, ”a restriction enzyme (or group of enzymes) thatcuts DNA consisting of 6 billion nu cleotides, into

CONCLUSIONS

In a short p eriod of time, enormous ad vanceshave been mad e in the ability to “read” the geneticmaterial and to und erstand wh at it means. The

techniques described in this chapter are examplesof state-of-the-art molecular genetics. Some of thecomponents of these techniques were originallydeveloped for other p urp oses in genetic research,while others were designed specifically for detect-

only 3,OO O pieces averaging 2 million nucleotides

each (about 1/ 60 of an average chromosome)

(126).

Ten different fragments are each labeled witha radioactive probe, so that each lane in the gel

will have 10 visible bands, the most that can be

analyzed per lane. The fragments are then sepa-rated on an agarose gel in wh ich the electrical fieldis applied alternately (“pulsed”) in perpendiculardirections for 24 to 72 hours. The pulse time is

adjusted, according to the particular size of the

fragments, to maximize separation of the frag-

ments. The optimal pulse time is one in which thefragments are constantly untangling and reorient-

ing themselves. Autoradiography is the final stepin the proced ure. Each of the 10 fragments boundto a labeled p robe app ears as a visible band onthe gel on exposure to X-rays; chrom osomal mu -tations wou ld ap pear as a shift in the position of the fragment containing the mutation.

Recent evid ence suggests tha t fairly large p iecesof human chromosomes can be separated, al-though most of the experience with the techniqueis limited to certain lower organisms such as yeastsand unicellular parasites. Results of experimentswith DN A d erived from th ese organisms, whichhave much smaller genomes than humans, sug-gest that chromosome mutations would be appar-ent if they were at least 5 percent as large as thefragment itself. The procedure would have to bemodified for hum an chromosomes due to their

larger size and complexity (127). This techniqu emay be useful in detecting chromosome mutations

that are intermediate in size between major rear-rangem ents (observable by cytogenetic meth ods)and single base pair changes.

ing new heritable mu tations. They all are successesin the sense that they prop ose reasonable and ver-ifiable ways of examining human DNA for alter-

ations in the nucleotide sequence. However, noneof the DNA-based techniques is approaching theefficiency that would be needed for any of themto be appropriate for field use. As these technol-ogies develop and information is shared am ong



Ch. 4—New Technologies for Detecting Heritable Mutations q 69

investigators, it may be p ossible to d esign com-binations of methods that complement each otherand that select for those samples in which thereis a greater likelihood of find ing new mu tations.

In general, despite impressive achievements thathave been ma de, the technology remains too un-

wieldy for large-scale applications. Olson (98)likens the current situation in m utation researchto the status of comp uter technology in the late1950s, The spectacular improvement in computerscame not from scale-up of 1950s technology butfrom developm ent of the integrated circuit. Hesuggests that big imp rovements in technology willprovide the appropriate methods; doing more of what we have already done is unlikely to besufficient.

Analyzing comp lete sets of hu man DN A fromlarge numbers of people using these methods

wou ld, at present, require an enormou s comm it-ment of resources. In considering the w isdom of dev oting major resources to such research, Olson(98) emphasizes that the program’s ultimate goal

would have to be far more ambitious than the

detection of a few bona-fide examples of new mu-

tations. A program large enough to have a legiti-

mate impact on public policy about environ-

mental exposures would have to be much larger

than a 5 or 10 person laboratory and be severalorders of magnitude more efficient than existingrestriction fragment length polymorphism tech-nology.

One of the outstanding characteristics of cur-rent research in m olecular genetics is the rap id

pace of new developments and ideas; assumingthis continues or accelerates, we can count on bet-ter methods to replace the proposed ones de-scribed in this report. The better ones should beable to examine at least the major portions of thegenome and be practical in studying pop ulationsat higher risk for heritable mutations. They shouldalso be able to detect a wide, if not the entire,spectrum of mutationa l events.

Until the next generation of techniqu es is de-veloped, the best cour se for the n ext few yearsmay be to m aintain a balance between the sup -

port of current m ethods and the supp ort of basicresearch which is the source of better ones. Main-taining a balance implies that resource-intensiveefforts should not pull unwarranted amounts of resources away from developing new techniques,although no special scientific effort seem s to beneeded to encourage this highly innovative andrapid ly paced area of research.



Chapter 5

New Methods for MeasuringSomatic Mutations



Contents

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73Detection of Somatic Mutations in White Blood Cells . . . . . . . . . . . . . . . . . . . . . . . 74

Selection for HPRT Mutants in T-Lymphocytes. . . . . . . . . . . . . . . . . . . . . . . . . . . 74Mutational Spectra From Human T-Lymphocyte HPRT Mutants,.. . . . . . . . . . 75The hprt Gene. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Results of Studies of hprt- T-lymphocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Detection of Somatic Mutations in Red Blood Cells . . . . . . . . . . . . . . . . . . . . . . . . . 76

Glycophorin Somatic Cell Mutation Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Hemoglobin Somatic Cell Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

TABLE

T able N o. Page

7. Frequencies of hprt- T-Lymphocytes in Human Beings . . . . . . . . . . . . . . . . . . . . 76



Chapter 5

New Methods for Measuring

Somatic Mutations

INTRODUCTION

All the techniques for learning about heritablemutations described in chapter 4 rely on compar-ing DNA or proteins of parents with DNA or p ro-teins of their childr en to infer the kind s and r atesof mutations that occurred in parental germ cells.While that information is of great value in learn-ing abou t heritable mu tations, it comes well af-ter the m utations have actually occurred . Herita-ble mutation tests are not useful for d eterminingrapidly that an individual or a population has

been exposed to a mutagen, so that people whohave been exposed can be appropriately moni-tored and others, who have not been exposed, canbe warned and spared exposure to that samemutagen. For this reason, it is desirable to developtests for somatic cell mutations that might allowinferences to be made about possible heritable ef-fects of some mutagenic exposure. Instead of pro-viding an exhaustive catalog of somatic cell mu-tation assays, this chapter provides examp les of developing technologies that m ay soon be avail-able to measure somatic mutations in human pop-ulations.

It is not known w hether the ra tes and k inds o f

somatic mutations in blood or any other non-

germinal t issue have any predictable relationship

w ith rates a nd kind s of germinal mu tations; it is

only w ith further dev elopm ent an d tes ting of

those m ethods , together w ith me thods for detect-

ing germinal mutations, that such relationships

might be elucidated. However, even withoutknowing the p recise relationships between somaticand germinal mutations, somatic tests may pro-vide an early warning that people are being ex-posed to environmental mu tagens. Furthermore,identifying and characterizing somatic mutantswill provide important gu idance for redu cin

gex-

posures to m utagens and carcinogens, and is rele-vant to th e stud y of aging and of carcinogenesis.

Exposure to mu tagens may increase the risk of developing som e kinds of cancer. In add ition, in-creases in the frequencies of second cancers occur-ring years after treatm ent of a first cancer withradiation therap y h ave been noted (see e.g., ref.11). In a stud y of su rvivors of cervical cancer, therisk of developing a second cancer was found tobe highest for organs close to where radiationtherapy was delivered, strongly suggesting thatthe development of second cancers maybe related

to the mutagenic effects of radiation (11).

Measurement of somatic mu tants offers the p os-sibility of identifying differences in somatic mu-tation rates over time w ithin ind ividu als, amon gdifferent individuals, and am ong d ifferent pop u-lations. Such stu dies can be used to exam ine pos-sible associations between exposures to particu-lar environmental or occup ational agents and thefrequency of mutations. Ultimately, they maybeused to identify individuals or populations whoare exposed to m utagens. This information couldbe useful to the r egulatory agencies in identify-

ing genotoxic compou nd s and th eir potencies.To be useful for studying hu man pop ulations,

somatic tests must be based on reasonably acces-sible cells (e.g., blood cells) and on recognizablemarkers for m utation (e. g., variant p roteins orenzymes). The most easily sampled human tissueis blood, and thu s far, all the somatic tests thathave been developed for studying mutations haveused red and white blood cells. Somatic cellmutation assays focus on two types of cellularchanges tha t result from m uta tions: 1) alterationsin ph enotyp ic pr operties of cells, such as chan ges

in cells’ resistance to drugs added to their growthmed ia; l and 2) alteration in a gene prod uct, such

‘Studying drug resistance in humans is inherently limited to study-

ing single copy genes, such as genes on the X chromosome (includ -ing the hprt gene) or dominant selectable markers.

73




as a p rotein. The first type of change has been bers of m ostly identical norm al cells of a sp ecificmost extensively studied using the mutation as- type. Although somatic cell mutation assays d e-say based on 6-thioguanine (6-TG) resistance, and tect phenotypic events that represent, as closelythe second type has been studied using two types as possible, actual m utations in DN A, the dataof proteins found in circulating red blood cells. are expressed as numbers of mu ta n t s , rather than

Instead of examining DNA directly, current and

in num bers of mutations. Various analytic tech-

developing somatic tests detect rare m utant cells niques have been used to estimate the mutation

based on marker phenotypes among large num-frequency from the observed mutant frequency (97).

DETECTION OF SOMATIC MUTATIONS IN WHITE BLOOD CELLS

Albertini and his colleagues (3) and Morley and

his colleagues (79) have developed assays usinghum an p eripheral blood lymphocytes to d etectmu tations that result in a phenotyp ic change inthese cells. The assay s are based on th e ability of mu tant T-lymphocytes, wh ich are present in theperipheral blood at the time the sample is drawn,to replicate DNA and to grow under conditionsthat d o not allow grow th of normal cells. The m u-tant cells are identified by their resistance to6-TG, a drug that is normally toxic to humanT-lymphocytes.

Norm al cells produ ce an enzyme, hypoxanthine-guanine phosphoribosyl transferase (HPRT), thatmet abolizes 6-TG to a toxic chem ical, killing th ecells. Mutations in the hprt gene (the gene thatdirects the p roduction of HPRT), can resu lt in cellsthat do not prod uce HPRT. The mu tation from

the normal gene (hprt +

) to the mutant gene(hprt-) makes the cell resistant to 6-TG (6-TG’)because they cannot metabolize 6-TG into its toxic

form.

This particular selective system has been shownto meet all the necessary criteria for a valid geneticassay: 1) the 6-TG

rmu tants are genetic variants

that “breed true”: whether grown with or with-out 6-TG, they remain resistant to 6-TG whenchallenged with the drug (3,79); 2) exposure of laboratory cultures of T-lymphocytes to knownmutagenic agents such as X-rays or ultraviolet ra-

diation increases the frequ ency of 6-TGr

mutants,indicating an increase in mutation in these cells(17,160); 3) there is a demonstrable change in the

gene p rodu ct—a d eficiency in the enzym e HPRT;and 4> there is a dem onstrable change in the DNA

sequence of the hprt gene—a direct demonstra-tion that the hprt gene has mutated (4,143).

Selection for HPRT Mutants inT= Lymphocytes

Nicklas (97) described two methods for iden-tifying mutant T-lymphocytes in human bloodsamples. The first is the autoradiographic assay.After blood is dr awn , the T-lymphocytes are pu -rified and frozen. On thawing, the cells are splitinto two cultures, both of w hich are treated w ithphytohemagglutinin, a chemical that stimulatesDNA synthesis. One of the cultures is also treatedwith 6-TG. The other culture is an untreatedcontrol.

A label, radioactive thymidine, is added to bothcultur es, and a ll cells capable of DN A rep licationand growth incorporate the labeled thym idine intotheir DNA. In the the control culture not exposedto 6-TG, essentially all the cells d ivide an d incor-porate thymidine, In the culture exposed to 6-TG,norm al cells are killed an d therefore do not in-corporate the radioactive thymidine into theirDN A. The cells from both cultures ar e then fixedon m icroscope slides, covered w ith an X-ray filmor emulsion, and th en developed using standar dradiographic techniques. Cells that have incorpo-rated radioactive thym idine emit rad iation, caus-ing a black, exposed spot on the d eveloped film.

The developed films are then viewed under amicroscope. This technique is quite sensitive;rad ioactive cells can be counted at a rate of 2 per10,000 to 1 per 10 million cells. The procedurecan also be automated to make it relatively effi-



Ch. 5—New Methods for Measuring Somatic Mutations q 7 5

cient for identifying rar e mu tants. Dividing thefraction of radioactive cells in the 6-TG-exposedculture by the fraction in the culture not exposedto the dru g prod uces an estimate of the frequencyof 6-TG-resistant mutants.

The second method, called the clonal assay,selects for cells that are resistant to 6-TG. Whiteblood cells are d ispensed in small test tubes orwells in a small plastic device resembling a tinymu ffin tin, each test tu be or w ell containing 6-

TG in the growth medium . Only the mutant 6-

TG-resistant cells grow and divide, forming visi-

ble colonies within 10 to 14 days. These colonies

can then be analyzed further by isolating their

DNA and analyzing it for mutations. The clonalassay allows characterization of mu tations foun din these cells, whereas the autoradiographicmethod permits only determination of their fre-

quency.

Mutational Spectra From HumanT= Lymphocyte HPRT Mutants

Cariello (16) described a techniqu e for an alysis

of the hprt gene that may correlate mutationalpatterns w ith particular mu tagenic exposures. Inthis method, all of the mutant hprt- T-lympho-cytes in a small blood sample are isolated andgrown in culture for 2 weeks to prod uce enoughDN A for ana lysis. DN A of these clones of differ-en t hprt- mu tants is then examined by first cut-

ting it into fragmen ts with restriction enzym es,adding a group of different DNA probes and sep-arating DNA bound to the probes by gradient gelelectrophoresis. DNA with different mutationswill bind to different probes. The pattern of m u-tations in the original blood sam ple app ears asthe pattern of DNA probe positions on a gel.Different mutagens are known to produce clearlydifferent patterns of mutation in human as wellas bacterial cells, so the aim of this techn ique isto identify the p robable causes of mu tations inhu man blood cell samples by using these patterns,or “mutational spectra” (139). This approach in-

clud es studies of the pattern of spontaneous m u-tation in hum an T-lymphocyte samples as wellas studies of the mu tational spectra produ ced bypotentially mutagenic chemicals to which humanbeings may be exposed. The goal of this approach

is to “fingerpr int” for the presen ce of a pa rticularmu tagen in an individu al’s life history.

The hprt Gene

One of the advantages of stud ying the hprt gene

is that much has been learned abou t it from stud iesof a particular genetic disease. Heritable mutationsin the hprt gene located on the X chromosome leadto Lesch-Nyhan syndr ome, a severely d ebilitat-ing disorder characterized by m ental retardationand self-mu tilation. The hprt gene has been iso-lated from blood samp les taken from patients withLesch-Nyhan syndrome, and the mutations in thehprt gene have been studied. Comparisons havebeen mad e between the types of mu tations presentin cells of Lesch-Nyhan patients and those presentin rare, mutant hprt- T-lymphocytes selectedfrom blood samp les of people w ithout the d isease.DNA from 5 of 28 Lesch-Nyhan patients were

found to have major hprt gene alterations—deletions of DN A, amp lification of DN A, and de-tectable changes in DN A sequence (172). Studies

of mutant T-lymphocytes selected from normal

blood samples have shown similar gene alterations

in several cases (4,143). These results show thatth e hprt- T-lymphocyte assay detects spon tane-ous mu tations in the hprt gene in somatic cellsthat are qualitatively similar to clinically appar-ent heritable mutations in humans, and also dem-onstrates that the clonal assay in T-lymphocytes

detects true mutations, not phenotyp ic changesthat mimic mutations.

Results of Studies ofhprt

- T-lymphocytes

A number of studies have measured frequen-cies of T-lymphocytes resistant to 6-TG in nor-mal ind ividu als. Average m utant frequencies forthe groups of people tested fall within a fifteen-fold range, between abou t 1 and 15 mutations permillion cells (table 7). However, mutant frequen-

cies for different individuals vary widely. For in-

stance, Albertini reports a range of frequenciesfrom 0.4 to 42 mutants per million T-lymphocytes

from 23 individuals, representing a hundredfoldrange in somatic mutant cell frequencies among

a small number of people (2). This wide range in




Table 7.—Frequencies of hprt -T-Lymphocytesin Human Beings

mutant frequencies may be due to different ages

and other characteristics of the T-lymphocyte

donors. Two investigators have shown that there

is a linear increase of hprt- mutants in T-lym-phocytes with increasing age (141,160). Donor

characteristics, such as smoking behavior, have

been shown to increase somatic mutant frequency;genetic variation in sensitivity to m utagens mayalso exist among individuals.

The degree to w hich the observed frequency of mutants reflects the actual frequency of somaticmu tations in this assay is not fully und erstood.

Descendants of the original mutant cells increasethe observed frequency, so growth of mutants invivo could falsely lead to a conclusion of highermutation rates than is actually the case, unless thedata are corrected for this problem. In add ition,hprt- cells maybe shor ter-lived than hprt

+cells.Before the techn ique can be used to estimate pos-sible increases in somatic mutation rates fromenvironmental exposures, the wide range of back-ground rates of these mutants has to be investi-gated, includ ing discrimination between true ge-netic events and any possible phenocopies. Whileimpr ecision of the assay may contribu te to thiswide range of mutant frequencies, such a rangeis not un expected for genetically and geographi-cally diverse human populations.

DETECTION OF SOMATIC MUTATIONS IN RED BLOOD CELLS

Unlike white blood cells, mature red blood cellshave no n uclei and contain no DNA. They can-not be grow n in the laboratory since they do n othave the capacity to rep licate. Red b lood cells can,

however, be used to detect mutations expressedin altered p roteins on their surface. Red blood cellsare also easily obtained from a blood sample.There are about 1 billion red blood cells per mil-liliter of whole blood. Branscomb (13) describedtwo methods to detect variant red cells, usinghigh-speed automated microscopy and flow cy-tometry to analyze and sort hund reds of red bloodcells per second to d etect rare red cell mutan ts.These two m ethods are described below.

Glycophorin Somatic Cell

Mutation AssayThe glycophorin assay detects gene-loss muta-

tions expressed by the absence of cellular proteinscorresponding to those genes. Such mutations canbe detected w here two or more variants of a par-

ticular protein normally exist, where the genes forthe alternative forms are said to be codominant(both variants are expressed when a gene for eachis present), and wh ere different variants are ex-

pressed in th e same cell. A mu tation causing theloss of function of the gene for on e of the varian tforms results in a cell that expresses only the otherform of the protein (9,10,5o).

One such protein is glycophorin A, a glycopro-

tein present on the surface of red blood cells. Two

variants of glycophorin A normally exist, called

“M” and “N,” which differ from each other by2 out of a total of 131 amino acids. The M and

N serotypes have no known biological function,

and each allele functions independently of the

other. The M and N variants of glycophorin Acan be labeled ind epend ently using mon oclinalantibodies carrying different fluorescing mole-cules. As a resu lt of the labeling, the M variantapp ears green and the N app ears red wh en viewedunder a fluorescent microscope. A flow cytome-ter, a machine through w hich cells are p assed at



Ch. 5—New Methods for Measuring Somatic Mutations q 77

very high speed, can sort and count the numberof cells that a re green, containing on ly the M var -iant, and the nu mber that are red, containing onlythe N variant.

Individuals w hose red cells display both vari-ants of glycoph orin are heterozygou s for the M

and N alleles (i. e., they inherited an M gene fromone parent and an N gene from the other). As usedcurrently, this assay detects gene loss mutationsat glycophorin genes in individu als who a re het-erozygous for the M and N variants. Red bloodcells from heterozygotes show both red and greenfluorescence; if a mu tation inactivates one of theglycophorin A alleles, only the other variantwould be present on the cell surface, hence onlyone color would appear. Single-color cells are rec-ognized and distinguished from the majority of dou ble-colored cells, and are counted as somaticmutant cells.

The glycophorin A assay is currently being usedat the Lawrence Livermore Laboratory to studymutations in two populations: cancer patients’blood cells examined before and after treatm entwith know n d oses of chemotherap eutic dru gs, andJapanese atomic bomb survivors whose dose of rad iation can be estimated (69).

Hemoglobin Somatic Cell Assay

The frequency of particular hemoglobin vari-ants that exist at low frequencies in red blood cells

of normal individuals can be measured. Cells con-taining these variants presumably arise in bonemarr ow stem cells, the p recursors of red bloodcells. Methods have been developed using mono-clonal antibodies, prep ared for each typ e of var-iant, to label particular mu tant hem oglobin m ole-cules. This method detects small mutations, suchas single nucleotide changes resulting in am inoacid substitutions, frame shift mutations, etc., inthe gene for beta-globin, a constituent of hemo-globin.

Investigators at the University of Washington

(132,133) and at Lawrence Livermore NationalLaboratory (50) have developed an approach to

detect hemoglobin mutations in red blood cells.

In this technique, an antibody to the mu tant he-moglobin is bound to a chemical called a fluoro-

chrome, which is then mixed with red cells in sus-pension. The fluorochrome emits fluorescent lightof a specific color when light of a particular wave-length is shone on it. The cells are screened undera microscope that allows the investigator to d i-rect light of a particular w avelength on the sam -ple. If any of the cells contain the mutant hemo-globin bound to fluorescing antibody, they canbe viewed an d counted.

Stamatoyannopoulos and his colleagues haveused this technique on blood samp les from 15 in-dividu als using antibodies specific for H emoglo-bin S (the mutant form of hemoglobin present insickle cell anem ia) and Hem oglobin C. An aver-age of 50 million cells per ind ividual w as screened,

and an average mutant frequency of 1.1 in 10 mil-

lion cells per subject was obtained, with a range

of 4 in 100 million to 3 in 10 million. These in-

vestigators also screened the red blood cells of 10

individuals who had been exposed to knownmutagens—X-rays or mutagenic drugs used to

treat cancer. The frequencies of mutant he m o-globins obtained from these samples were withinthe range found for normal ind ividuals.

This techn ique is labor-intensive because it re-lies on visual screening of the cells: it takes onetechnician nearly 1 month to screen 100 millionred cells. The investigators have since collaboratedwith Dr. Mend elssohn an d h is associates at theLawrence Livermore Laboratory to use their auto-ma ted fluorescence-activated cell sorter to screen

the cells. Although som e technical difficulties wereencountered w ith this instrument, results obtainedon red cells from six normal individuals (onscreening 500 million cells per individual) yieldedan average mutant frequency of 1 in 10 million,about th e same as the average frequency deter-mined from visually counted cells.

With an array of different mutant hemoglobinsto use for pr odu ction of antibodies, a variety of mu tations can be d etected in hem oglobins foundin rare cells present in norm al individuals: singlenu cleotide substitutions, deletions, frame-shift

mu tations, or virtually any mu tation that causesthe prod uction of mu tant hemoglobins.

There are currently two technical problems withthis techniqu e. First, a significant nu mb er of falsepositives are generated. Better m ethods of fixa-




tion and antibody staining of the hemoglobins are reported that with present capabilities using flow

necessary, along with verification of each vari- sorting, 88 antibody-labeled red cells can be iden-

ant found. Second, the method is not yet fully tified in a mixture that contains a known quan-autom ated, but research in several laboratories tit y of 100 such labeled r ed cells and 1 billion u n-is progressing toward full automation. Significant labeled red cells, a relatively good retrieval rate.progress has already been m ade; Mendelson (68)

CONCLUSIONS

The methods to detect and quantify somaticmu tations suggest some ad ditional approaches forstudying the nature and mechanisms of mutationin hu man beings. Somatic tests offer the possi-bility of drawing associations between exposuresto specific mutagens and particular genetic eventsin the cell, and they m ay help to iden tify individ-uals and populations at high risk for mutations.They may lead to understanding the relationships

between the frequency and nature of somatic andgerminal mutations, and if those relationships canbe identified, somatic tests may contribu te signif-icantly to und erstanding w hy hu man heritablemu tations occur and m ay help pred ict their fu-ture occurrence. At present, however, it is notpossible to directly extrapolate from m utat ionsin som atic cells to risks of heritable mu tations inoffspring.

Animal studies can be used to measure quan-

titatively the correlations between exposure to

mutagens and rates of somatic mutation, as well

as to investigate possible associations between so-matic and heritable mutations. Coordinated tests

of somatic and germinal mutations in experi-

mental animals may help guide such research inhum ans. Such data, combined with d ata on hu-man somatic mutation rates, may suggest specificrisks of heritable mutations in hum an p opu lations.

In this regard, the tests for somatic mutation

described in this report appear promising. Noneof these methods is currently ready for applica-tion to human population studies, however, eventhou gh m any of the technical details have been

worked out. Since the assays do n ot detect DNAmutations directly, it is particularly important toverify the results from these methods by isolat-ing variant cells and confirming that the alteredproteins do ind eed reflect primary changes in thegenes coding for these proteins. It is also neces-sary, although m ore d ifficult, to d etermine howoften these method s miss the mu tations they aredesigned to d etect,

In general, the most useful somatic cell muta-tion assay m ay be one that d etects a spectrum of mu tations (single am ino acid substitutions to largedeletions and translocations) and that can d etectmu tations in d ifferent cell types. Different tissuesmay demonstrate different frequencies of somaticmu tants, and equally as important, some typesof mutan t cells may have longer lifespans in thebody than others. Consideration of such charac-teristics could be useful in choosing the mostapp ropr iate assay for particular circum stances.Short-lived m utants could be useful in studying

the effects of short-term exposures to poten tialmu tagens, wh ile longer lived mu tants could bebetter ind icators of mu tagenic effects of long-term,chronic exposure to environmental mutagens.

The genetic analysis of white blood cell DNAoffers the possibility of identifying genetic “fin-gerprints, ” or particular patterns of mu tations in-duced by specific mutagenic agents. Such patternsare an essential first step toward understandingthe molecular mechanisms of genetic change inhuman DNA,



Chapter 6

Laboratory Determination ofHeritable Mutation Rates



Introduction , . . . . . . . . . . . . . . . . . ., . :. ..,..... . . . . . . . . . . . . . . . . . . . . . . . . . . .

Roden t Tests for Heritab le Mu tat ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .The Dominant Lethal Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .The Heritable Translocation Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Specific Locus Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Other Studies in Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Factors That Influence Induced Mutation Rates in Mice . . . . . . . . . . . . . . . . . . . . . .Effects in Males . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Effects in Females . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Estimating Effects in Human Beings From Anim al Data . . . . . . . . . . . . . . . . . . . . . .

818181828283

84848586

TABLE

Table No. page

8. Spontaneous Mutation Rates in Mice. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

FIGURE

Figure No. Page

13 . Sperm atogonial Mu tation Rates in Mice After Low-Dose Rate andHigh-Dose Rate Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85




in the developm ent of the em bryo (after fertili-zation), the lethal events are detectable by dissect-ing female uteri a few days after implantation of embryos (implantation occurs about nine days af-ter fertilization). If there are more dead embryosthan expected (compared with matings of unex-posed m ice) and / or more implantation sites than

expected in th e uterus where there is no longeran embryo, the effect is assumed to be the resultof a mu tation in th e male germ cell. Chromosomalstudies in mice have shown that almost all “dom-inant lethal” events are associated with majorchromosome abnormalities.

Human analogs of the postimplantation fetallosses counted in dom inant lethal tests are spon-taneous abortions (miscarriages). Fifty to 60 per-cent of spontaneous abortions are associated withnumerical and structural chromosome abnormal-

ities, the same basic types of abnormalities de-

tected by dominant lethal tests (12). These simi-larities between mice and men, and the general

pragmatic acceptance of higher animals as predic-

tors for human risk, underlie the idea that agents

that are p ositive in a m ouse dom inant lethal testmay also cause spontaneous abortions or chromo-some abnormalities in liveborn infants.

The Heritable Translocation Test

The heritable translocation test detects breakageand rejoining of fragments from d ifferent chro-mosomes. When a translocation involves no loss

of genetic information (a “balanced transloca-tion”), often it has n o ad verse effect on th e func-tioning of the carrier individual. The unusualchromosomes produced by the breaking and re-

joining are present in all cells, and they can pairwith each other and divide and function like nor-mal chrom osomes. How ever, such translocationscause reproductive problems because the germcells may have incomplete sets of genetic mate-rial and may lead to unbalanced chromosome ab-normalities in offspring, often resulting in deathduring embryonic or fetal development. Herita-ble translocations are thus detectable in mice be-

cause they result in reduced litter size in the sec-ond generation after exposure to the susp ect agent,or in sterility in the first generation. In addition,cytogenetic studies can detect abnormal chrom o-

somes and can be used to study th e associationbetween redu ced fertility and chromosome abnor-malities.

Specific Locus Tests

Since 1927, when Herman Muller first demon-

strated that radiation causes mu tations in fruit flies(8 1 ), there has been concern about possible geneticeffects of radiation in human beings. FollowingWorld War II, in an ticipation of widespread useof nuclear power, the Atomic Energy Commission(AEC) contracted for research on the mutageniceffects of radiation in mice to provide an experi-mental underpinning for estimating effects inhuman beings, Specific locus tests that detect mu-tations in a few specific genes in m ice were d e-veloped for that p urp ose. This work and use of specific locus tests to measu re chemically-indu cedmu tations have provided the animal data mostuseful for estimating human risk.

Through research sup ported by the AEC, Wil-liam L. Russell (109) bred a strain of “tester” micethat differs from wild-type mice in 7 morpho-logical features—coat colors, pa tterns of p igmen-tation, eye color, and ear shape—that are reces-sive traits. Each of the features is thou ght to becontrolled by a d ifferent genetic locus. Tester m iceare homozygous for the recessive alleles of thesegenes, while wild-type m ice are h omozygous forthe dominant alleles. When the tester strain isma ted to a strain of wild-typ e mice, the offspr ing

should all be heterozygous at the seven test loci,and therefore they should appear wild-type be-cause the the wild-type alleles are associated withthe dom inant phenotyp e. How ever, if a germinalmu tation occur s in a gene for one of the sevenloci in the w ild-type strain, causing the d ominantallele not to be expressed , the offsprin g from tha tmu tation w ill have a phen otype characteristic of the tester strain.

In a sp ecific locus test, the w ild-typ e m ouse isexposed to radiation or to a know n or susp ectedchemical mutagen and then m ated to the tester

strain. In most cases, only male w ild-type miceare exposed, and the tester strain mice are female.The reason for this is largely a pr actical one: oneexposed male can sire a large num ber of progeny




mutation, would be inherited from one of the par- in the mou se derived from these experimental dataents (51,52,103,161). is in the same range as the estimate of the spon-

Data from animal experiments using electro-taneous rate derived from studies in human be-

phoretic and quantitative enzyme assays are lim-ings (77), recognizing, however, that these esti-

ited, but at th is stage a general, cautious conclu-mates are based on sm all num bers and m ay beheavily influenced by chance fluctuations.

sion can be drawn. The background mutation rate

FACTORS THAT INFLUENCE INDUCED MUTATION RATES IN MICE

When r adiation or a chemical is tested for mu-tagenicity, the experiment does n ot necessarilyproduce definitive answers to the questions: “Isthis agent mu tagenic’?” And “if so, how mu tagenicis it?” The answ ers differ dep ending on severalfactors, two important ones being whether theanimal is male or female and the stage of devel-opm ent of the exposed germ cell. There is ampleevidence from radiation studies that male and fe-

male germ cells differ in their sensitivity to muta-gens. In males, germ cells appear to be more likelyto sustain a mu tation the further along the d evel-opmental path they are. Overall, female germ cellsapp ear to be more resistant to mutagens than aremale germ cells.

Effects in Males

The male testis, from p uberty on, contains amixture of germ cells at different developmentalstages, including spermatogonial stem cells, dif-ferentiating spermatogonia, and intermediatestage cells, the spermatocytes and spermatids,which are maturing into spermatozoa. The mech-anisms of sperm prod uction are qualitatively verysimilar in mice and m en, differing largely in tim-ing: in mice, sperm production from spermatogo-nia to spermatozoa takes 35 days; in men, it takes74 days (60). When a male is irradiated, germ cells

at all stages of development are exposed. By vary -ing the times of mating, the effects of radiation

on different stages can be examined. For instance,

if male mice are irradiated an d m ated immedi-ately, any mutants that ap pear in the progenymu st result from m utations that were caused infully mature sp erm. In a m ating two w eeks afterexposure, eggs are fertilized by spermatozoa thatwere sperm atids at the time of irradiation. If sevenor more w eeks are allowed to p ass, sperm in the

ejaculate would h ave been irradiated at the sper-matogonial stem-cell stage. Experimental evidencesshow s that cells at each stage have va rying sensi-tivities to indu ction of m utations by r adiation, thelater, postspermatogonial stages being most sen-sitive.

While spermatogonial stem cells are more resis-tant to the effects of mutagens than are germ cellsin later stages of development, mutations in stem

cells are of greater concern. Stem cells persistthroughout the lives of males, constantly givingrise to new genera tions of sperm cells. A mu ta-tion in stem cell DNA results in every cell thatbud s off from th e stem cell bearing a muta tion.The lifespan of later germ -cell stages constitutesonly a small fraction of the total reproductivelifespan (especially in hu man s). Thus, in th e caseof acute exposur e to these later germ-cell stages,only conceptions occurring during a short postex-posure interval are at risk. For these reasons, theeffects of radiation and some chemicals have beenespecially closely studied in spermatogonial stemcells in anim als.

The most common way for people to be ex-posed to ionizing radiation is through acute ex-posure to medical X-rays, or long-term, low-levelexposure to natural background radiation. Cer-tain groups are exposed occupationally (e.g., ura-nium miners, nuclear pow erplant workers, somemed ical personnel); and some m ay ha ve speciallow-level environm ental exposures from livingnear nuclear installations, nuclear waste sites, ura-nium mine tailings, or areas of high natu ral ra-don concentrations.

in m ice, exposu re of m ales to X-rays or oth erforms of ionizing radiation at low dose ratescauses a measurable increase in the number of mutations, while such exposures do not measura-




birth, app ear resistant to the lethal effects of ra-diation.

The differing sensitivities of oocytes of differ-ent species to lethal effects of ionizing radiationcomplicate efforts to extrapolate from mutationrates in female mice to the expected rates in hu-

ma n females (28). In pa rticular, radiation m ay re-sult in more mu tations in hu man females thanwou ld be p redicted from female mice because hu -man arrested oocytes are less likely to be killedby radiation. That migh t allow a hu man oocyteto survive a mu tagenic dose of radiation and tobe fertilized.

Using experimental data from female mice, Rus-sell (1977) made several calculations relating ef-

fects of radiation in mice to potential effects inhu man females. In each calculation, he mad e the“w orst case” assum ption that all stages of femalegerm cells were as sensitive to the mutagenic ef-fects of radiation as are th e most sensitive stages—matu ring and matu re oocytes-in m ice. Becausethere are different method s of hand ling the d ata,

he m ade four d ifferent calculations. Three pro-duced estimates indicating that radiation wouldnot increase the mu tation rate in hu man arrestedoocytes. The fourth estimated that human ar-rested oocytes would be som ewha t less than half as sensitive (from 17 to 44 percent as sensitive)

to radiation as are human spermatogonia.

ESTIMATING EFFECTS IN HUMAN BEINGS FROM ANIMAL DATA

Most extrapolations from animal data to hu-man mutagenic risks are based on studies in malemice because: 1) there are m ore data for male micethan for female mice, 2) there appear to be fewerbiological differences in germ cell developmen t be-tween m ales of the two species than th ere are be-tween females (see above), and 3) male germ cellsapp ear to be more sensitive to radiation than fe-male germ cells and may provide a more sensitiveindicator of genetic risk.

The dose-response curves in figure 13 are de-

rived from experiments on m ale mice and th eyprovide information for estimating hu man risks.The absence of a detectable threshold for geneticeffects in male mice suggests that humans are atsome increased risk for mu tations at any level of radiation exposure. This finding increases the im-portance of making quan titative estimates of theeffects of radiation; since risk d oes not g o to zeroexcept at zero dose, wh at is it at levels of hu manexposure? The “d oubling dose, ” wh ich is the d oseof radiation that induces an additional numberof mutations equal to the num ber that occur sp on-taneou sly (resulting from all mutagenic influences,

known and u nknown ), can be derived from theinformation in figure 13. The doubling dose differsdepend ing on the radiation dose rate. The dou -bling dose in male mice for high d ose rates is abou t40 R; for low dose rates, about 110 R. This is of

practical importance for human beings, sincesome exposu res are at high dose rates—e.g. thepopulations of Nagasaki and Hiroshima and peo-ple receiving rad iation therap y for some typ es of cancer—but most population exposure to radia-tion is delivered at a low d ose rate—e.g., peoplewho live near radioactive mines. About threetimes as many m utations are caused by radiationdelivered at a high d ose rate as opp osed to anequal total radiation d ose delivered at a low d oserate (114).

In chapter 3, the effects on mu tation rates of acute irradiation of the populations at Nagasakiand Hiroshima are discussed. The exact dou blingdose for those populations is still being debatedbecause of uncertainties about the total dose andbecause of insensitivity of the method s used todetect mutations. Using the best informationavailable about those pop ulations, how ever, hu-mans app ear to be less sensitive than m ice to highdose rate exposures.

Curren t rad iation exposu res of United Statescitizens average about 200 millirem (mrem) per

year. On a population basis, this exposure is gen-erally divided about equally between high dose

rate medical exposures and low dose rate expo-

sures to natural sources of radiation. These ex-posures do not approach the doubling dose for



Ch. 6—Laboratory Determination of Heritable Mutation Rates • 87

mutations. If the average rate of human mutations

is now on the order of one mutation per 1 mil-

lion genes per generation, we can estimate the ef-

fect of a hypothetical increase in human radia-

t ion exposure to twice the current level . If

exposures increased to 400 mrem and if that ex-

posure were as mutagenic as the high dose ratedelivered to mice, it would increase the average

human rate to 1.005 mutations per 1 million genes

per generation. Such an increase would be un-detectable by any current method.

Animal stud ies have proven u seful for learn-ing about th e various relationships of dose, par-ticularly for radiation, and germ-cell and herita-ble mutation rates. They are limited, however,in providing information directly applicable to hu-man beings. First and most important, human be-ings differ from animals in ways th at we kn owabout and in more ways that are not und erstood.Second, the endpoints measured in animal testsdo not generally correspond directly to endpointsmeasurable in hu man beings un der non experi-

mental conditions. Third, the available animaltests, like the currently available methods forstudying h um an beings, detect mu tations arisingin specific, selected , loci which m ay not be rep -resenta tive of other loci. Evidence from Russell’sspecific locus test, cited above, shows a 35-fold

difference between the least sensitive and mostsensitive of the loci tested. A fourth considera-tion, again similar to the limitations of stud ies inhumans, is that each test does not detect morethan a fraction of the mu tations that can occur,and the exact nature of the mutations that are de-tectable cannot be characterized with thesemethods.

When th e new technologies d iscussed in this re-port are further developed , they presum ably willbe app lied to experimental anim als as well as tohum an beings. Stud ies that examine correspond -ing genes with th e same techniques inanimals may provide some of thecomparisons that now are lacking.

hum ans andinterspecies



Chapter 7

Extrapolation



Contents

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .The Aims of Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Extrapolat ion Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Attemp ts a t Qualitat ive Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Attemp ts at Qu antitat ive Extrap olation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Dose-Response Relat ionships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Gener alizations Abou t Dose-Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Limits of Extrapolat ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

91

91

93

94

97

97

99

99

TABLE

Table No. Page

9. Weighting of Test Results for Presumption of Germ-Line Mutagenicity . . . . . . 95

FIGURES

Figure No. Page

14. Biological Test Systems for Studying Mutations . . . . . . . . . . . . . . . . . . . . . . . . . 9215. An Example of a-Parallelogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93



Chapter 7

Extrapolation

INTRODUCTION

Currently, predictions of possible risk of herita-ble mutations in hu man beings are based on infer-ences, or “extrap olating” resu lts, of muta genicitytests in other organisms or in laboratory cell cul-tures. One of the key problems in genetic extrap-olation is that, while there is no shortage of muta-genicity tests using a variety of organisms and celltypes, researchers have just begun the p ainstak-ing work of draw ing out relationships among re-sults of different tests, and of eventually validat-ing models for extrapolating to heritable effectsin hu man beings (see fig. 14). This chap ter reviews

the basic constructs that have been d evised forframing genetic extrapolations and then p resentssome efforts that have been mad e to carry out sp e-cific extrap olations.

The Aims of Extrapolation

For m any years, the emp hasis in mu tagenicitytesting h as been on developing individual tests andlearning about their prop erties. Test developmenthas not been targeted exclusively toward devel-oping p redictive m odels for heritable mu tationsin hum an beings and in fact the qu est for tests thatcould be u sed to pred ict carcinogenicity throu ghsomatic mutation has predominated. This beingthe case, the test systems v ary tremend ously andinclud e, for examp le, tests in bacteria, insects, ani-mal somatic and germ cells in culture, whole mam -mals, and hu man somatic cells. Efforts to relateresults from one test system to another or fromvarious tests to human beings, even on a qu alita-tive level, have been relatively recent and havenot progressed very far to date. Bridges (15) sum-marized this situation:

Despite the extremely large num ber of screen-

ing tests that have been performed, remarkablylittle rational though t has been devoted to theuse that shou ld be mad e of the results of suchtests. Mutagenicity tests in lower organism s

lar e

“’Lower organisms” refers to bacteria, yeast, Drosophila, etc.,an d not to whole mammals.

essentially screening tests w arning of a p ossiblehum an hazard. They do not give any quantita-tive indication of the level of risk to man, notso mu ch because of differences in the organiza-tion of DNA in man an d lower organisms, butbecause of the complex m etabolic capabilities of man which may greatly enhance or diminish themu tagenic effectiveness of an agent.

In recent years, scientists w ho u se different ex-perimental approaches to study mutations havebegun to d iscuss m ethods to correlate results from

various test systems. The first consideration iswhether results from one system are qualitativelypred ictive of results in anoth er, i.e., are app ro-priate endp oints being considered and do the re-sults agree in w hether they are p ositive or nega-tive. This is called “ biologic extrap olation. ” Thesecond type of extrapolation is quantitative, whichdeals with the relationship betw een the quan tita-tive response in a test to a quan titative estimateof the likely effect in human beings.

Examples of questions posed in extrapolationare the following: If experiments demonstrate that

a single exposure to a h igh d ose of a chem ical in-du ces heritable mutations in mice, wh at wou ldbe the result of a single exposure to a lower realis-tic dose or of a long-term exposure to a lower d oseof that chemical in humans? How do results of mu tagenesis experiments on rodent somatic cellsin a test tube cell culture (e. g., Chinese hamsterovary cells “in vitro”) relate to the issue of mu ta-tion rates inhuman somatic cells, or mutation ratesin hum an germ cells? What are approp riate doseconversions betw een various experimental systemsand hum an beings? If a worker in a chemical planthas a tenfold increase in mutation frequencies in

his or h er somatic cells, what are the long-termhealth imp lications? Is that worker at increasedrisk of having a child w ith a new mu tation? Thesetypes of questions illustrate the comp lex issues thatarise in u sing results from on e system to make p re-dictions of risks to hum an health.

91



Ch. 7—Extrapolation q 9 3

In translating resu lts from on e system to another(using similar genetic end point s), a nu mber of sep-arate extrapolations may have to be made: 1) fromspecies to species; 2) from exper imenta l doses toactual environ men tal doses; 3) from o ne cell typ eto anoth er; 4) from in vitro to in vivo ph ysiologi-cal conditions; and 5) the biggest and most u n-certain leap, from estimates of mutation frequen-cies to estimates of genetic disease in humans. The

EXTRAPOLATION MODELS

Several researchers have begun developing mod-els for extrapolating from one test system toanother (14,15,46,49,61,64,106, 129,130,138). oneof the key features common to all the extrapola-tion mod els developed is that a result from a sin-

gle test system w ould n ot be used alone to predicta result in another test system. Instead, results fromseveral related test systems are correlated and usedtogether.

Several of the methods described are extensionsor rearrangem ents of the first extrapolation m eth-od developed by Sobels (1290)—the parallelogram(130):

The und erlying principle is to obtain informa-tion on genetic damage that is hard to measuredirectly, for examp le mutation in m ouse germcells, by compar ison with end points that can be

determined experimentally, e.g., alkylation pernu cleotide in mam malian cells in vitro and inmouse germ cells, and mutation induction inmam malian cells in vitro.

Sobels’ parallelogram is illustra ted in figure 15.It is relatively easy to determine the mutation fre-quency in mouse somatic cells (a quantity called“A”) up on exposure to a particular chemical muta-

gen in culture. With certain types of chemicals.

(alkylating a gents), it is also possible to d erive ameasu re of the interaction of the mutagen withthe DN A of those cells, which is quan tified as“alkylations p er nu cleotide” (a quantity called

“A”). Alkylations per n ucleotide can also be meas-ured in mouse germ cells after exposure to the samemu tagen (B’). The relationship of these d ifferentvalues is then used to calculate the expected m u-tation frequ ency in mouse germ cells (B) on ex-

kind of information that would give the biggestboost to the ability to pred ict effects in hu manswith information from other test systems is know-ing exactly wh at kind s of mu tations (e.g., pointmutations, chromosomal rearrangements, etc. )each of the tests detects. The new technologies dis-cussed in this report may provide this type of in-formation.

Figure 15.—An Example of a Parallelogram

A ’

A

B ’

B

Estimated mutations

(Note that the same test system is represented in the verticaldimension and the same genetic endpoint in the horizontaldimension.)

SOURCE: F.H, Sobels, “The Parallelogram: An Indirect Approach for the Assess.

ment of Genetic Risks from Chemical Mutagens, ” pp. 323-327 in K.C.Born et al. (eds.), Progress in Mutation Research, vol. 3 (Amsterdam:Elsevier, 1982).

posure to the same mutagen. A major, untestedassumption is that the ratio of A to A‘ is propor-tional to th e ratio of B to B‘, i.e., A/ A = B/ BIf that is true, it is then a matter of simple algebra

to predict the mutation frequency in mouse germcells (B) by solving the equa tion for B, wh ich isthe only unknow n quantity.

A similar parallelogram can be used to extrap-olate results from mouse germ cells to human germcells. Mutation frequencies are m easured in so-matic cells of both humans and mice. Germ-cellmu tation frequencies are measured in the mouseand compared to somatic-cell mu tation frequen-cies in the m ouse. Assuming that the ratio of germ-cell to somatic-cell mutation frequencies is thesame in mice and human beings, germ-cell muta-

tion frequencies in human beings can be predictedfrom the m easured hum an soma tic-cell mu tationfrequencies (64).

Streisinger (138) has proposed a more complex

extrapolation method using two sequential paral-



94 • Technologies for Detecting Heritable Mutations in Human Beings

lelograms. In the first parallelogram, a measureof chemical interaction with hum an germ cell DNAis estimated from measurements of the effects of the chemical in human and animal (e.g., monkey)somatic cells, and a measure of interaction of thesame chemical with germ-cell DNA in the sam eanimal. In th e second p arallelogram, the ratio of a measure of chemical interaction with mousegerm-cell DNA to mouse germ-cell mutation rates(from the specific locus test) is used to predict thehuman germ-cell mutation rate using the estimatedvalue of chemical interaction w ith hum an germ -cell DNA from the first parallelogram. This con-struct embodies several untested assumptions.

Bridges (1980) also deve loped a m ore comp lexextrapolation m odel based on the original par al-lelogram model of Sobels. Based on his approach,Bridges outlined the types of results needed to fillinformation gaps, ultimately to assess the impact

of mu tagen s at th e: 1) molecular; 2) cellular; and3) wh ole organism levels in both animals and man .He su ggested stud ies to determine: 1) the presenceof an effective dose of mutagen at the molecularlevel by m easuring the concentration of mu tagenin the gonads or blood or t he extent of reactionwith DN A; 2) whether there ap pears to be a rela-tionship betw een the presence of the mu tagen anda biological response at the cellular level by meas-uring somatic mutation frequencies or chromo-somal changes in lymp hocytes; and 3) wh etherthere is an effect at the whole organism level by

measuring the frequency of heritable geneticdefects, congenital malformations, or fetal loss.

The values Bridges specifies are obtainable inanimal systems. To obtain such values for man ,Bridges suggests that use be made of certain other-wise normal human populations that are exposed

to large doses of mu tagens. Examples of such pop -ulations are pa tients treated for diseases, such ascancer, with dru gs that are known to be muta-genic, and certain occupational cohorts in whichthere are known excesses of cancer (1,155). Simul-taneous studies using the same mutagens couldbe carried out in experimental m ice to determ inethe relative sensitivities of mouse and m an to thesemutagens.

Parallelogram mod els are attractive for theirsimplicity and inherent logic. They are ap prop ri-ate starting points for exploring relationships

among test results when sufficient data becomeavailable to do so. However, the assump tions em-bodied in parallelogram models—consistent, pre-dictable relationships amon g var ious cell types,translatable amon g species—are almost entirely

untested. The great differences among speciesmake it u nlikely that these parallelogram mod elswill survive validation studies intact. While theymay continue to be useful research tools for pos-ing logical questions, they may or may not p rovepractical for pred icting risks of heritable mu ta-tions in h um an beings.

ATTEMPTS AT QUALITATIVE EXTRAPOLATION

L. Russell and colleagu es (106) compared theresults of muta genicity tests carried out in a var i-ety of systems other than whole mammals withresults from specific locus tests (SLTs) and herita-ble translocation tests (HTTs) in mice (see ch. 6for descriptions of these two tests). The purposeof the comparison w as to find ou t how w ell re-sults of the nongerm-cell tests corresponded to thequalitative results (positive or negative) of the twogerm-cell tests. The an alysis was limited by th erelatively small number of chemicals that h avebeen tested in either the SLT or the HLT. About35 chemicals have been tested in one or both of the germ -cell assays, out of a tota l of about 2,000

chemicals for which som e test results are avail-able from any system.

2

The comparison tests were grouped into 18 cat-egories and the categories given r elative w eightsaccording to their biological relationship to oneof the germ-cell tests. The categories and theirweights are given in table 9. The lowest weightedcategory includes tests u sing pr okaryotes, such

as bacteria, directly treated by the suspected muta-gen. Higher scores signify m oving toward higher

‘These test systems are not the focus of this report and are n otdescribed here in d etail. Descriptions of these tests can be foundin (88).



Ch. 7—Extrapolation q 9 5

Table 9.—Weighting of Test Results for Presumption of Germ-Line Mutagenicity

Exposure Exposurenot within withinmammalian mammalianbody Weight body Weight Germ cells Weight

Prokaryotes, all endpoints SAL 1 BFT 2W P – HMA

Lower eukaryotes, all endpoints YEAYEP 2ASPNEU

Higher eukaryotes, chromosome aberrations PYC 3

Higher eukaryotes, gene mutations PGM 3

Mammals, genetically nondefined endpoints SC1SC2 4UDHUDP

CYC 5Mammals, chromosome aberrations

Mammals, gene mutations CHOV79 5

SC3SC4

DANDHT

6

MNT 7 DLTCYE

}

CY9CYB 10 CYOCY5CY8

8

8

4

8

15

MST 10 SPF 15L51

NOTE’ In general, the weights increase from top to bottom and from left to right in the table, From top to bottom, the tests progress from lower to higher organisms

and from more general endpoints to endpoints of direct relevance to human heritable mutations. From left to right, the categories progress from in vitro testsIn both somatic and germ cells, to in vivo germ-cell tests

SOURCE L.B Russell, C S Aaron, F de Serres, et al , “Evaluation of Mutagenicity Assays for Purposes of Genetic Risk Assessment,” Mutation Research 134:143-157, 1984

mammals, toward germ cells, and toward treat-ment with the chemical in a whole mammal.

A single composite score was calculated for eachchemical tested, adding together scores from eachcategory in which there were test results. Thereis only one score per category regard less of thenu mber of tests. Positive results are scored as posi-tive num bers; negative results as negative nu m-

bers, e.g., an in-vitro somatic-cell chromosomeaberration test with a positive result yields a scoreof +5, one with a negative result, a score of —5.

Russell and colleagues found that nearly allchemicals that tested positive in either or both the

SLT and H TT had h igh comp osite scores fromother tests. A number of chemicals with negativeSLT and HTT results also had high, positive com-posite scores, representing “false positives” in thecomparison tests.

Similar analyses looked separately at the SLTand HTT and th e comp arison tests that specifi-cally detect gene mutations or chromosome aber-rations, respectively. The results are similar tothose matching the results in all comparison testsagainst both mammalian germ-cell assays: highscores for most chemicals positive in th e germ-cell tests, and a nu mber of false positives.



96Technologies for Detecting Heritable Mutations in Human Beings

In an add itional analysis, the compa rison testsare ranked according to how well each p redictsthe results of the two germ-cell tests. In general,the tests in higher numbered categories in theearlier an alyses, i.e., those that are closer biolog-ically to whole mammal germ-cell tests, had bet-ter correlations w ith the SLT and HTT. For theSLT, the best predictors overall were the mousespot test, unscheduled DNA synthesis in mousetestis, and the m icronu cleus test. For the H TT,unscheduled DNA synthesis in testis, the domi-nan t-lethal test, and one lower ra nked test, sister-chromatid exchange in cultured animal cells, werethe strongest predictors.

While it ap pear s that the resu lts of some of thecomparison tests correlate relatively well with themam malian germ-cell tests, in fact, not on e of thesecorrelations reaches conventional statistical sig-nificance, meaning that the tests do not p redictreliability better than chance. The lack of signifi-cant results is du e, in large part, to the sm all num-ber of comparisons for many tests, and in partbecause of th e p rocess for selecting chem icals forSLT and HTT. From a practical, public policystandpoint, this is an important finding. The lack of statistically significant results does not meanthat these comparisons are without value. Thestud y provides a status report on the quality andquan tity of existing d ata.

Since the two whole mammal tests (the SLT and

HTT) are relatively expensive and time-consum -ing, they are usually reserved for testing chemi-cals highly suspected of causing heritable muta-tions. The suspicion is based on resu lts of othertests, specifically the comparison tests examinedin Russell and colleagues’ analysis. It is hard ly sur-prising, therefore, that the comparison test resultsare largely p ositive for chemicals eventu ally testedin the mammalian germ-cell assays. Russell andcolleagues took the preponderance of positive re-sults into accoun t in their analyses.

Many chemicals have been tested in m utage-

nicity assays because, for reasons of chemicalstructure or other properties, there is a high likeli-hood that they will be mutagenic. While thesechemicals have proved useful as laboratory tools,they are n ot necessarily u seful for dra wing con-clusions about w hat p eople are actually exposedto. W. Russell (111), using the same database used

by L. Russell and colleagues (106), looked exclu-

sively at the 11 “environmental chemicals” (those

found in the home or workplace) that have beentested in the SLT and examined th e results. All11 are positive in the Drosophila sex-linked lethaltest, the 10 that have been tested are positive inmammalian somatic-cell tests, and there are a va-riety of positive results in other test systems. Noneof the 11, each of wh ich w as tested at very highdoses, is positive in the SLT, suggesting no increasein mu tations in spermatogon ia (the pre-meioticmale germ-cell stage) although several have posi-tive results in tests of later sp erm developmen talstages.

What conclusions can be drawn about the va-lidity of qualitative extrapolation from variousmutagenicity tests to a risk of heritable mutationsin human beings? The available data give no directinformation ab out m uta genic effects of chem icals

on hu man germ cells in vivo. Application to hu-mans rests on a series of assumptions about theresponse of human germ cells in relation to re-spon ses in other types of cells and in oth er species.

The analysis of the these test results does allowsome generalizations about biologic extrapolationand about the nature of the available test data-base. On the first point, there is evidence suggest-ing that chem icals that test positive in some com-parison tests for gene m utations or chromosomalaberr ations hav e a likelihood of being positive inthe SLT or HTT, respectively, but are not invari-

ably so. To da te, chem icals testing negative in th ecomparison tests have not prod uced clear posi-tives in w hole mam mal germ -cell tests, but thedatabase supporting this comparison is limited.Biologically, it seems un likely th at chem icals tha tare consistently negative in comparison testswou ld, in fact, be germ-cell mu tagens in w holeanimals. But it is unlikely also that very manychemicals with negative results in one or two testswill have been tested in enou gh systems to allowan empirical test of that hypothesis.

Both L. Russell’s and W. Russell’s analyses de-

scribed above suggest that a number of chemicalsthat test positive in the simpler comparison testswill be not be show n to cause mu tations in sper-matogonial germ cells. At present, however, it isimpossible to know based on comp arison test re-sults, which chemicals are true hum an germ-cellmu tagens and wh ich are not.



Ch. 7—Extrapolation . 97

A major limitation of the database is that nearly The new technologies d escribed in th is report,all tests have been in exposed m ale animals. Two wh ich could be used to p rovide information onstrong animal mutagens—rad iation and a chemi- the types of mutations detected by the variouscal agent, ethylnitrosourea—do not appear to tests, may improve our ability to apply the re-cause heritable mu tations in the imm ature germ suits of simpler tests to pred icting hu man risk.cells of fema le mice, but the d ata for females ar e

not su fficient to d raw firm conclusions.

ATTEMPTS AT QUANTITATIVE EXTRAPOLATION

Not surp risingly, quantitative extrapolation iseven less advan ced than is qualitative extrapola-tion. However, there are some data bearing onquan titative relationships that may eventually beuseful in predicting effects in human beings. First,there is some information abou t dose-responserelationships. Second, some preliminary attempts

are being made to determine relationships amongcertain corners of the parallelogram models de-scribed earlier in this chapter. One such effort isdescribed below.

In an ongoing effort, a group of investigatorsis using a parallelogram approach to evaluate theeffects of gam ma ra diation, cyclophosp ham ide (amedical drug) and benzo[a]pyrene (an environ-mental chemical) on mouse and human cells byassaying chromosomal aberrations and sister chro-matid exchanges in mitogen-stimu lated peripheralblood lymp hocytes (PBLs). While these investi-gators have as yet little da ta, a pap er presentingsome preliminary results (170) lays ou t the r igor-

ous procedures necessary for proper extrapola-

tion of results from different studies of just one

substance to p redictions in un tested systems. Forinstance, in the case of radiation, the authors re-late an experimental result in irradiated culturedhuman PBLs to reports from the literature of thesam e chromosom al endp oint (dicentrics) in PBLsfrom patients wh o have received th erapeutic ra-diation. Thu s far, these auth ors have n ot pr esentedany conclusions about the relationships they arestudying.

Dose= Response Relationships

Extrapolation from high to low dose, and fromhigh to low d ose rate, requires knowledge of dose-response relationships. For heritable gene m uta-

tions, adequ ate data on d ose-response relation-ships is limited to the effects of ionizing rad iationand one chemical, ethylnitrosourea (ENU), in malemice. The available information comes from re-sults of SLTs in mice, most of the wor k being d onein a small number of laboratories in different partsof the world. Radiation and ENU ha ve also been

assayed in many short-term in vivo and in vitrotests in human somatic cells, so some generaliza-tions may be drawn about the nature of dose-response relationships for these two agents. In bothcases, there are independent effects of dose rateand of total dose. This means th at a fixed dosemay cause a d ifferent rate of mutations dep endingon the intensity of the dose, i.e., a more protractedadm inistration may result in fewer mu tations thanif the total dose is administered at one time.

Radiation

Specific-Locus Test. —In a series of experim entsfrom th e m id-195@ to the p resent, a ra nge of ra-diation doses, delivered at a ran ge of dose rates,has been tested in male mice. Results are avail-able for both spermatogonial and postspermato-gonial stages. In spermatids and spermatozoa(postspermatogonial stages), the dose-responserelationship is generally linear, and there is no ef-fect of dose rate. This means that approximatelythe same m utation rate results from a short, high-dose exposure an d from a chronic, low-dose ex-posure when the same total dose is given.

In spermatogonia, the early, pre-meiotic stage,for equal total exposure, radiation given at highdose rates causes more m utations per u nit of dosethan rad iation at lower d ose rates. At a high d oserate of 90 Roentgens per minu te (R/ rein) or inter-

med iate dose rate of 8 R/ rein (work cited in 111),




the mutation rate decreases with decreasing dosefaster than w ould be pred icted by a linear rela-tionship. At dose rates of 0.8 R/ rein or below,how ever, the rate of mu tations per u nit dose ap-pears to be constant, and w ithout a threshold. Atlow d ose rates, the d ose-response relationship islinear, and above abou t 0.8 R/ rein, the mu tationrate per u nit of dose rises more qu ickly than w ouldbe pred icted by a linear relationship. Above a to-tal dose of between 600 and 1,000 rem, the mu ta-

tion rate begins to decline rapidly.

Investigators using the SLT offer an explana-

tion for the observed dose-response patterns (111,

114,115). They postulate that the difference indose-rate response between spermatogonia andpostspermatogonial stages is a function of an ac-tive repair mechanism in m etabolically active sper-ma togonial cells, w hich does not fun ction at laterstages. In the earlier, spermatogon ial stages, a

larger percentage of changes can be repaired be-fore the spermatogonia comp lete meiosis w hen ex-posure is at a lower d ose rate than is the case withan acute, high-dose-rate exposure. In post-sperma-togonial stages when capacity for repair is low,the total rad iation dose, irresp ective of dose rate,is the determinant of the mutation rate.

Heritable Translocation Test.—Generoso andco-workers (135) have investigated heritable trans-locations induced by high-dose rate (96 R/ rein)irradiation of spermatogonial stem cells of mice.They report a linear d ose-response relation be-

tween O and 600 R of total irradiation, and repeti-tion of doses in this range gave ad ditive effectsup to 2,OOO R. From these data, the expected in-crease in heritable translocations at h igh-dose-ratesis calculated to be about 0.00004 per R.

Cytogenetics.—Waters an d colleagues (170) de-

scribe a dose-response relationship for gamma ra-

diation after both in vitro and in vivo irradiation,

using as an endpoint a certain type of chromo-

somal mutation, dicentrics, in PBLs. The in vivodose-response, which was derived from reportsin the literature, is linear at low er dose ra tes, and

quadratic at higher dose rates, meaning that theincrease in the number of dicentrics rises fasterthan it wou ld if the relation continued to be lin-ear. The qu adra tic component m ay be explainedby an interaction of mutational events causing

some d icentrics, and the interactions being increas-ingly mor e likely as the dose rate increases.

Ethylnitrosourea (ENU)

Specific Locus Test . —As is the case with radia-tion, both total dose of ENU an d dose-rate inde-pendently affect the mutation rate in mouse sper-matogon ia. Experimen tal data ind icate that theresponse at doses below 100 mg/ kg of body weightis “infralinear,” meaning that as the dose is lowered ,the mutation rate drops faster than would be pre-dicted by a linear relationship. At doses between100 and 400 mg/ kg, the response app ears to belinear (4 I). No threshold was d etected over therange of doses tested, but the possibility of athreshold at a dose lower than 25 mg/ kg (the low-est single dose tested) is not excluded by the data.

In an experiment examining mutational re-

sponses at d ifferent d ose rates, the m utation ratewa s measured in mice that w ere given 10 weeklydoses of 10 mg/ kg of ENU, and compared w iththe mu tation rate for a single dose of 100 mg/ kgof ENU. The m utat ion rate for the fractionateddose w as only about 15 percent of the rate for asingle dose (112).

Russell (111) notes th at, in light of informationindicating th at ENU reaches germ cells in dosesproportional to injected amounts (see next section),these results cannot be explained by differencesin metabolic processes. He interprets the infralinear

portion of the dose-response curve and lower mu-tation rate that follows d ose fractionation to bethe result of effective mutational repair systemsin sperm atogonia, the same reasoning as in thecase of radiation.

Unscheduled DN A Synthesis in Mouse Sperma-tids.—Carricarte and Sega (cited in 111) found thedose-response of “unschedu led DN A synthesis”in mouse spermatids to be linear over the rangefrom 10 to 100 mg/ kg, the same range over w hichW. Russell (111) found an infralinear relationshipin the SLT. Sega (cited in 111) also measured “ad-

du ct formation” after injections of ENU, and founda linear response in the range from 5 to 100 mg/ kg.One conclusion from these observations is thatchemical interaction with DNA may not alwaysbe directly related to th e rate of mutation, and



Ch. 7—Extrapolation . 99

this presents a major difficultyparallelogram, described earlier in

Sister-Chromatid Exchange and

for Sobels’this chapter.

ThioguanineResistance.–Jones and colleagues (54) investigatedthe dose-response relationship for two somatic-cell endpoints in whole mice exposed to ENU. The

frequency of sister-chromatid exchange (SCE) andthe frequency of thioguanine resistant (TG r) cellswere measured in white blood cells in the spleen.Corresponding end points exist in hu man beings,which makes them potentially useful for extrapo-lating to hu man responses. The investigators meas-ured both the response over time at several doselevels, and the response to a range of doses at sev-eral points in time after exposure.

Linear relationships w ere discovered betw eendose and reponse for both SCEs and TG

rcells,

but the timing of these responses was q uite differ-

ent. For all dose levels, the highest SCE levels weremeasured on the first day after exposure, and SCE

LIMITS OF EXTRAPOLATION

The main limitation to validating extrapolationmod els is a virtual lack of data to complete theparallelogram m odels. Until there is enough em-pirical information to determine whether qualita-tive and quan titative generalizations can be mad e,it is impossible to kn ow how useful extrapolation

could be. At present, qualitative extrapolation isat least a tentative guid e for iden tification of mu ta-genic agents.

Even when d ata are available, man y questionsbedevil the use of experimental animal data to ex-trapolate risks to hum ans. In part icular, little isknow n abou t the comparab ility of species withrespect to activation, detoxification, and tissue d is-tribution of specific chemicals, as well as otherinterspecies differences in metabolism. In add ition,mu tational responses of germ cells at differentstages of developm ent can d iffer for d ifferent typesof mutagens. Many of these gaps in information

could be filled by studies that are now technicallyfeasible.

levels d ecreased back to baseline after about 70days, whereas the TG

rresponse rose linearly

over time to a peak after about 80 days for eachdose level.

Generalizations About Dose-Response

While dose-response relationships cannot begenerally and simply described some generaliza-tions can be mad e. It appears that m utations insomatic cells are more predictable from the vari-ous experimental test systems available than aremutations germ-cell. Repair of mutations in germcells may explain this difference. Both a betterun derstand ing of the mechanisms of mu tation andrepair at th e molecular level, and emp irical com-parisons of test data on more su bstances will con-tribute to a better und erstanding of dose-responserelationships. Once again, the new technologiesmay contribute to this und erstanding.

In using test resu lts from somatic cell systems,there are differences in the sensitivities of varioustypes of som atic cells and g erm cells to d ifferentmutagens. Even within a single cell type, differ-ent gene loci can respond very differently to a spe-cific mutagen. In making comparisons, for in-

stance, between somatic and germ cells, if the sam elocus (or set of loci) cannot be tested, an y d iffer-ences in outcome cannot necessarily be attributedsolely to a difference in cell type. These are a fewof the many uncertainties in extrapolations fromsomatic to germ cells.

An organizational p roblem that affects extrap-olation in a practical way is that collaborationshave been rare amon g researchers working w ithdifferent systems, and are particularly rare amongscientists working in d ifferent laboratories. In addi-tion, the test systems are not necessarily stand -ardized am ong laboratories working with the same

systems, so results from different laboratories can-not always be combined or directly compared.



.


Even relatively simple aspects of experimen tal pro-cedu res, for instance, the conditions un der w hichcells or organisms are exposed to radiation orchemicals, vary enough among laboratories to ren-der results incomparable to varying degrees.

Overall, the development of methods to extrap-

olate from different experimental test systems tothe risk of heritable mutations in human beings

is still in its infancy. While the va rious p arallelo-gram mod els are appealing because of their poten-tial usefulness, some of the data already availablesuggest that they m ay never broad ly applicable.Both qualitative and quantitative improvementsin the database of experimental results will be nec-

essary before the usefulness of extrapolation canbe reasonably assessed.



Chapter 8

Mutation Epidemiology



Contents

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........................103Valida tion of Measurem ent Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...104Surveillance and Disease Registries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......104Monitoring and Exposure Registries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........106Ad Hoc Epidemiologic Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..107Population to Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ...108

Radiation-Exposed Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............108Cancer Patients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......................108Other Populations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................109

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .............................110



Chapter 8

Mutation Epidemiology

INTRODUCTION

This report is about biochemical and genetictechniques for studying heritable mutations. These

new techniques will ultimately be used to study

people who are suspected of being at high risk for

excess heritable mu tations. The types of epidem io-logic activities in which mutation detection tech-niques may eventually be used are: surveillance,

monitoring, and ad hoc studies. Before any tech-

nique can be used for those purposes, a series of

validation studies will be needed, calling for dif-

ferent populations that are appropriate for study

at different stages of development of the tech-

nologies.

Sur veillance is a rout ine activity who se aim, inthe context of this report, wou ld be to measu rethe “baseline” rate of mu tations in a defined pop -ulation over the course of time, and to facilitaterapid recognition of changes in these rates. Fol-lowing the distinction ma de by H ook and Cross(44), the term mon itoring is reserved for obser-vations over time in pop ulations thou ght to beat increased risk for heritable mu tations becauseof a know n or suspected exposure to a known orsuspected mu tagen, for the pu rpose of helping thespecific popu lation in wh atever w ay p ossible. Ad

hoc studies of a variety of designs are carried outto test h yp o th eses about suspected causes of mu -tations.

The different aims of surv eillance, monitoring ,and ad hoc studies require that different criteriabe app lied for deciding w hen and whether to carryout one or more of those activities. Surveillanceand monitoring are not designed as hypothesis-testing activities, thou gh th ey ma y be sources forhypothesis development.

The reasons a population is chosen for surveil-

lance may be largely opportunistic. It is unlikely

that an entirely new system of data collectionw ould be p ut in p lace for m utation surveillance.It is more probable that mutation surveillancewou ld be add ed to an existing program that isestablished for another purpose, for instance,

birth defects surveillance. The population coveredmust be large enough to generate reliable rates formutational events that may be relatively rare, butthere is no fixed requirement for size. Of the threetypes of activities, surveillance generally wouldinvolve the sm allest effort an d resource expendi-ture per individual, but because large numbers of people w ould be routinely subject to the surveil-lance test, the total cost could be large. There is,therefore, a great need to consider the costs andbenefits of such a program before embarking onone, and for choosing the detection techniqu e ac-cordingly. The threshold for initiating mutation

surveillance w ould be relatively h igh. Informa-tion about trends som etimes can be obtained bymeans other than full-scale surveillance, and somesuch studies for that p urp ose have been d one.

The main reason for instituting a mutationmonitoring program is concern about the poten-tial effects of a mutagenic exposure in a specificgroup of people. If there is enough concern, thereun dou btedly will be a greater expenditure of re-sources and effort per person than is the case forsurveillance, meaning that m ore extensive con-tact with the p opulation and testing would be

justified. Given today’s knowledge, it is hard toconceive of a situation in w hich there wou ld bea concern only about h eritable m utations, so anymu tation m onitoring effort w ould m ost likely bepart of a larger program. The obvious concurrentconcerns for exposures thought to be mu tagenicare cancer and birth defects. Strict criteria basedon tests of the statistical power to detect certainlevels of effects are not appropriate for monitoredpopu lations, since the concern is about that p ar-ticular grou p of people. Any finding is of inter-est in that situation. The information maybe usedincidentally to calculate upper limits of risk which

could be generalized to other pop ulations withsimilar exposures. The most imp ortant consider-ation in making a decision to mon itor is that thereis reasonable evidence suggesting that th e pop u-lation may be at a substan tially increased risk.

103




This decision m ay w ell be influenced by p oliticalpressures to act, but ideally there should be a rec-ognition that the best scientific jud gment eitherdoes or does not support a monitoring effort.

The purp ose of ad hoc studies is to test hyp oth-eses about exposures and effects. This is the oneplace where it is imp erative that stud ies be de-

signed to achieve a high probability of detectingan effect if it is present. Such stu dies are valua blenot only for the sake of the pop ulations involvedin the studies, but for their generalizability toother populations. Results of these studies canform the basis for public health actions, if levels

of risk can be established.

VALIDATION OF MEASUREMENT TECHNIQUES

For the most part, the new technologies de-

scribed in this report have n ot been “validated. ”They are new, and have not been applied to largenumbers of people. Surveillance, monitoring, andad hoc stud ies all requ ire tools—in this case tech-niques for measuring mutations—with an acceptabledegree of validity. That is, the tests mu st meas-

ure w hat they are d esigned to m easure, withinsome definable limits. Generally, it is not possibleto gather reliable information about a populationand concurrently gather validating informationabout a technique used to measu re outcome, un-less another technique, w ith know n v alidity, andknown relationship to the new technique, is alsoapp lied in the stud y. Even thou gh that is techni-cally feasible, it is probably not an efficient wayto gather validating d ata.

A first step in the v alidation p rocess w ould beto use laboratory-prepared samples with known

DNA sequences to confirm that the types of mu -tations tha t shou ld theoretically be d etected witha new technique actually are reliably detected. Be-yond that stage, the need to move to clinical sam-ples can be met u sing stored blood from individ-uals stud ied p reviously for other reasons. These

stored samp les need not be from p arent/ childtriads initially, but triads w ill be needed at a laterstage.

A nu mber of research organ izations are stor-ing samples that would be appropriate for studiesof mutations using new DNA techniques. The Na-

tional Cancer Institute for example, is storingblood samples from cancer patients who havebeen treated with d rugs and radiation, and theRadiation Effects Research Foundation has storedblood from Japan ese citizens who were exposedto atomic bombs. In both of these cases, DNAis stored according to an established techniquethat u ses Epstein-Barr virus to transform lymp h-ocytes, thereby “immortalizing” the cells so theycan be grown indefinitely. The transformed cellscan be frozen for the long term in liquid nitro-gen. Both sample preparation and long-term stor-age costs are substantial. With currently available

technologies, it is unlikely that large numbers of samples will be stored. This limits the number andvariety of samples available for validation studies,and also for later studies of people exposed to p o-tential mutagens.

SURVEILLANCE AND DISEASE REGISTRIES

The method s and aims of disease surveillance track the incidence of cancer. Although the firsthave developed based on experience in infectious national cancer surveillance system, which cov-

disease control. Reporting of vital statistics, in ers about 12 percent of the pop ulation, was pu tparticular births and deaths for calculating birth in place as recently as 1972, New York State in -and death rates, is also a form of surveillance. Sur- stituted a reporting system for cancer cases inveillance of non infectious diseases is a relatively 1940, and Connecticut followed the next year.recent development, with roots in the desire to There are now dozens of cancer surveillance sys-



Ch. 8—Mutation Epidemiology • 105

terns operating around the country and interna-tionally, covering a range of popu lations fromcounties to whole countries (145).

Information about individual patients in can-cer surveillance systems forms the basis for “can-cer registries. ” The routine output of registries

consists of cancer rates by sex, age, and race(where applicable) for each cancer site. Registriesalso are an imp ortant sou rce for researchers in-vestigating hypotheses about cancer causation. Inthis sense, cancer surveillance, with informationabout ind ividu als recorded on registry forms, issimilar to m ortality statistics, with informationabout ind ividu als recorded on d eath certificates.

It should come as no su rprise that there are no“heritable m utation su rveillance systems” n ow inplace. There are, in various places around theworld , registries of birth d efects, wh ich includ erecords of at least some sentinel phenotypes. Be-yond that, as this report shows, there are atpresent no techniques for detecting m utations thatare suitable for use in a large-scale population sur-veillance program. Because developments havebeen so rapid, however, there maybe one or sev-eral good candidate techniques within 5 or 10years.

Surveillance trad itionally has involved report-ing, to a central place, information already col-

lected by some segm ent of the health care systemfor reasons directly related to the health of indi-viduals. Even for infectious diseases, only cases

that come to the attention of physicians are re-ported . Active “case-find ing” in the pop ulationis not a usu al feature of surveillance. For chronicdisease, the same is true. Cancers diagnosed byphysicians are entered into registries. Case-findingprog rams, such as breast cancer screening p ro-gram s, are instituted on th e basis of their effec-tiveness in identifying cases early in the courseof disease, for the benefit of the individual withthe d isease.

Infants are examined for birth defects becauseof the potential imp act on the children and their

parents’ lives, and not m ainly for the p urp ose of compu ting the rates of birth d efects in a popu la-tion. Testing programs for newborns, includingbiochemical tests for metabolic diseases (not nec-essarily a r esult of a m utation), also have been

instituted because of their importance to the healthof the individu al. Nearly all States now requiretesting newborns for phenylketonuria (PKU), andsome require additional tests. For instance, NewYork r equ ires testing for PKU, sickle-cell anemia,and congenital hypothyroidism, which are m oder-ately rare, and also for very rare cond itions in-

cluding maple syrup urine disease, homocystin-uria, histidinemia, galactosemia, and adenosinedeaminase deficiency (102). The tests do not im-pose an added burden on the newborn, since allare carried out using the same blood sample.

Cytogenetic techniques h ave not been used forpop ulation-based surveillance of chromosome ab-normalities, but some large hospital-born seriesof newborns have been tested (102). Most of therecorded cases of chromosome abnormalities foundin this way m ight eventually have been detectedlater in life because of health an d r eprod uctive

problems, but some others might otherw ise havegone undetected.

There is no formula for deciding whether to in-stitute a surveillance program, but there are char-acteristics of the disease, of the population, andof the particular test to be used that contributeto the d ecision: 1) the seriousness of the d isease(if the measu red en dp oint is known to be associ-ated w ith a d isease); 2) the a bility to alter its clin-ical course after diagnosis; 3) the prevalence of the disease in the population; 4) the reliability andvalidity of the test; 5) the acceptab ility of the test

to the population; 6) the cost of the screening pro-gram; and 7) the cost of not screening (i.e., thecost of treatment and social support). It is worthconsidering these factors in thinking about screen-ing and surveillance for heritable mu tations.

The idea of surveillance for heritable mutationsrepresents a departure from the traditional appli-cations of surveillance. It appears to be the casethat most heritable mutations are not related todisease over the course of an individual’s lifetime,and no predictions useful to the individual cancurrently be made about the effect of a heritable

mutation in the absence of recognizable disease,beyond those that are associated with sentinelphenotypes and major chromosome abnormal-ities. As mutation detection techniques becomemore and more sensitive, in fact, a greater per-



106 Ž Technologies for Detecting Heritable Mutations in Human Beings

centage of the mu tations detected may not be re-lated to a known effect on health.

Heritable mutation surveillance beyond report-ing sentinel phenotypes w ill require more th an justa rep orting of events already d etected. It will re-quire imposing a test burden on a p opulation for

the sole purpose of collecting information aboutmutations that may never affect an individual’slife. This argues against instituting surveillance.A reason in favor of surv eillance is that it is clearthat increased m utation rates w ill be looked for

in special popu lations, those being monitored be-cause of worries that they have been exposed toa mu tagen. Surveillance systems can p rovide arange of estimates of “baseline” or “background”rates, even though they may be from differentpopu lations. In a more general sense, one of theorigina l aims of surveillance is relevant: to sub-

stantiate long-term trend s and patterns in healthevents and to detect changes that may be ad-dressed by p ublic health action.

MONITORING AND EXPOSURE REGISTRIES

Monitoring is the “long-range observation of individuals wh o are at presum ptive high risk foradverse ou tcomes because of sp ecific life events, ”

(44) in particular, exposures to suspected muta-gens. The event may be catastrophic, such as ex-posure at the time of detonation of an atomicbomb, or a chemical plant explosion. Or the“event” may be long term, such as an occupationalor an environm ental exposu re. There are abouttwo dozen populations around the world cur-rently monitored for long-term health effects, andsome of those programs include various studiesof heritable mu tations.

The most extensive pop ulation mon itoring, in-cluding monitoring for mutations, is of the Japa-nese residents of Hiroshima and N agasaki, man yof whom were exposed to substantial amounts of radiation during World War II when atomicbombs w ere detonated in those cities. The popu -lation around a chemical plant that exploded nearSeveso, Italy in 1976, releasing several pounds of

dioxin, is the subject of health monitoring activi-ties, including monitoring for birth defects. Thepeople exposed to meth yl isocyanate in Bhopal,India, will undou btedly be followed for years tocome. Because these group s were exposed , andbecause it is conceivable that someth ing could bedon e to alleviate health p roblems if they are d e-

tected early, or if wa rning signals are picked u p,

they are being mon itored; the scientific knowl-edge gained as a result is a secondary benefit.

The most p rominen t examples of chron ic ex-

posures are from occupational activities and toxicchemicals in the environm ent. The popu lationsexposed to hazar ds often are not geograp hicallydeterm ined, but m ay be a collection of workersfrom around the country. Workers exposed toradiation in the nuclear power industry are an ex-ample of this. There are several “exposure regis-tries” in existence worldw ide, though non e spe-cifically because of a perceived increased risk of mu tations. One such registry has the n ames of allworkers who were exposed to dioxin during themanufacture of various chemicals in th is coun -try. There also is an international d ioxin r egis-try, with names of workers from all around theworld . The registry does not, how ever, have in-formation abou t the health status of those work-ers. A similar registry for workers exposed toberyllium exists in this country. A report preparedfor the Nuclear Regulatory Commission in 1980recommended that a registry be started for work-ers exposed to low-level ionizing radiation in cer-tain typ es of workp laces, because of a possibleincreased cancer risk (29). These registries could

be used for monitoring and as a potential popu-

lation to include in ad hoc studies.



Ch. 8—Mutation Epidemiology q 10 7

AD HOC EPIDEMIOLOGIC STUDIES

Surveillance, monitoring, medical case reports,and laboratory research can all lead to hypothe-ses about possible causes of heritable mutations.An investigator wishing to test a hypothesis mustfind suitable subjects to study, in contrast to a

monitoring activity, where the existence of the ex-posed p opu lation is the reason for acting. A stud yshould be undertaken only if there is a good chanceof answering the question of interest. Disease andexposure registries are common sources of indi-viduals to study, depending on the question.

A cohort design will probably prove the m ostuseful app roach for studies of heritable mu tations,though case-control studies of sentinel phenotyp esmay also prove valuable. A cohort stud y involvesidentifying a grou p of ind ividu als, some exposedto the suspected mutagen and some not exposed.

The health outcomes, i.e., the presence or absenceof mutations in offspring, of the two sub-cohortsare compared. A higher rate of mutations in theexposed group wou ld signify an “association” be-tween the exposure an d heritable m utations. Sta-tistical tests are applied to the results to estimatethe likelihood of the result occurring if in fact therewa s no real d ifference in mu tation rates betw eenthe two groups.

In a case-control d esign, a group of “cases,” in-dividuals with conditions of interest, e.g., sentinelphenotypes, is comp ared to a group of individ-uals wh o do not h ave the condition of interest,but w ho are otherw ise similar d emograp hically.The cases and controls are comp ared accordingto their past histories of exposures or other char-acteristics that might be associated with the mu-tation and an assessment made as to wh ether theirhistories differ in imp ortant w ays.

The important question for all studies is not justwh ether the exposure is “associated with” m uta-tions, but w hether it ca u ses them . That is a d iffi-cult if nearly impossible judgment to make in mostinstances, but there are some generally acceptedguidelines for evaluating the likelihood of an asso-

ciation being causal based on epidemiologic evi-dence. These are:

ent times and in different pop ulations, andin studies of different designs.

2. Strength: The size of the effect of an exposu reis the measure of strength of association.This is usu ally measured as an estimate of

relative risk (a ratio of the rat e of muta tionsin an exposed group to the rate in an u nex-posed group). The presence of a dose-re-sponse r elationship , that is, the size of theeffect changes in a logical way with the levelof exposure and in at least some cases, withthe dose rate.

3. Specificity: Specificity refers to the degreeto w hich the exposu re is associated exclu-sively w ith the ou tcome of interest, in thiscase a mutation, and the degree to which amutation is associated exclusively with theexposu re. The concept of specificity derives

from study of infectious disease and is rele-vant to the stud y of mutations (and chronicdiseases gen erally) only in special cases, forexample, a specific mutation that almostnever occurs in the general population butapp ears to be exclusively related to a par -ticular exposur e. While a h ighly specific rela-tionship can p rovide p ositive evidence fora causal relationship , a lesser d egree of speci-ficity does not necessarily argue stronglyagainst causality.

4. Temporal Relationship: The exposure mustoccur before the effect. In the case of h erita-ble mutations, the p icture is mor e comp li-cated. See chapter 6 for a d iscussion of thetiming of exposure for males and females fora plausible effect on germ cells.

5. Coherence: All available information frommedical and biological science, and from epi-dem iologic observa tions and stu dies, fits to-gether in a way that supports the hypothe-sis. The greater the variety of information,and types of stud y designs, the stronger thefinding of coherence.

These criteria are quite stringent, and even in

the best of cases, often cann ot be m et, but th eyare useful as standards.

1. Consistency: The association is observed instudies by different investigators, at differ-




POPULATIONS TO STUDY

There are elements in the environment thatdamage human health under certain conditionsof exposure. Biologic, chemical, and physicalagents cause acute and chronic diseases in hu-mans. At present, there are no exposures une-

quivocally known to cause heritable mutationsin hu man beings. A combination of factors, in-cluding the rather insensitive methods for detect-ing heritable mutations that h ave been available,and the possibility that hu man germ cells may notbe very su sceptible to some mu tagens, probablycontribute to this situation. As a consequ ence, in-vestigators looking for the effects of mutagensmu st do so in p eople who have been highly ex-posed to agents that are very likely to be muta -genic in germ cells. There are not very many largegroup s of people fitting that description, a factthat many m ight find surp rising.

Radiation= Exposed Groups

Radiation causes heritable mutations in labora-tory mice and is the most likely potential germ-cell mutagen to which large numbers of humanbeings have been exposed, either intentionally oraccidentally. The largest population with a knownhigh radiation exposure, the Japanese atomicbomb survivors, continue to be followed for ef-fects on cancer incidence, birth outcomes, andheritable mutations. Heritable mutations havebeen studied by clinical observations, cytogenetic

techniqu es, one-dimensional electroph oresis of blood p roteins, and m ore recently with the m ostsensitive technique of two-dimensional gel elec-trophoresis of blood proteins (see ch. 3).

A report was prepared in 1980, under contractto the N uclear Regulatory Com mission (NRC),that evaluated opportunities for studying thehealth effects of low-level ionizing radiation (29).The report is focused on cancer, but the evalua-tion m ethods apply equally to stud ying muta-tions. The authors initially identified 100 candi-date popu lations. Abou t 30 remained after two

broad criteria were ap plied: 1) that there be d ataidentifying exposed individuals, and 2) that therew ere at least 10,000 people in a single p opu lationgroup or one comprising several similar grou ps.

Those 30 pop ulations were evaluated further, and

recommendations made that if additional studies

were to be undertaken, three occupational groups

an d on e group with environmental exposure held

ou t the greatest promise of yielding a reliable re-

sult. Even the best of these, however, has a rela-tively low power: less than a 50 percent chance

of find ing an excess of cancer if it exists. In gen-eral, this level of power w ould be unacceptablein an epidemiologic study. Although political con-siderations might influence a decision to go aheadwith a study, they do nothing to increase thepow er of the method.

The power figures for these studies refer to can-cer detection, and the probability of detectingheritable mutations is undoubtedly far lower,making it u nlikely at best that anything could belearned about radiation and heritable mutationsby studying an y of these group s with currently-available methods.

The report to the NRC contained one other rec-ommendation, that a registry for radiation w ork-ers be initiated. The registry would maintaininformation about radiation doses and some in-formation abou t other exposures. This recomm en-dation has not been acted on. There are exam-ples of radiation-exposure registries, but these aremainly for people acutely exposed accidentally,and not for the more u sual long-term chronic ex-posures of workers.

Cancer Patients

Treatment for many cancers includes chemo-therapy with cytotoxic drugs, some of w hich arecarcinogenic in laboratory animals and mutagenicin vitro, and treatment w ith high doses of radia-tion. There is a growing body of evidence thatcancer patients are at a severalfold increased risk of developing second cancers, and some of thesesecond cancers ma y be attributable to treatmentof the first cancer with d rugs an d radiation (see,

e.g., ref. 149). Cancer is m ostly a d isease of oldage, but certain cancers have their peak incidencein younger people. Hodgkin’s disease, for in-stance, occurs with greatest frequency in young



Ch. 8—Mutation Epidemiology . 109

men. Childhood leukemias, some brain cancers,and tum ors with strong genetic comp onents, e.g.Wilms’ tumor, retinoblastoma, and neuroblastoma,occur in the first few years of life. As treatmen tfor these early cancers has improved over the lasttwo d ecades, larger num bers of peop le are sur-

viving, and it is these survivors wh o are at an in-creased risk of a second cancer, and possibly of heritable mutations.

Results from four stu dies of the offspring of childh ood cancer survivors, and nine stud ies of offspring of ad ult cancer p atients have been pu b-lished as of mid-1985. Several other studies arein progress (82).

The combined published studies represent morethan 700 cancer p atients (both m ale and female)and more than 1,500 pregnancies, about 1,200 of which resulted in live births. Four percent of the

liveborn babies had major birth defects, which issimilar to the incidence in the general population.Two of the liveborn children had cancer. One hada hered itary bilateral retinoblastoma, as h is fa-ther had . The other, the dau ghter of a brain can-cer survivor, had acute myelocytic leukemia. Onechild had a condition that could have been theresult of a new mu tation, the Marfan synd rome,which fits the definition of a sentinel phenotype.Several other children had defects that might havehad genetic components, but none of these rep-resented sentinel phenotypes.

The largest study of offspring of childhood can-cer sur vivors, including abou t 2,300 individuals

from five population-based cancer registries, isnearing completion. Preliminary results indicateno increased risk of cancer in offspring comparedwith a control group, but th e analysis is not yetfinal (82). Another long-term followup study,

with more than 3,3oo cases enrolled to date, isunder way in the United Kingdom . No results areyet available from that study (82).

The findings of a large international coopera-tive study of second tum ors in children treatedfor cancers are provocative (142). Overall, 12 per-cent of children wh o surv ive at least 2 years af-ter a first cancer develop a second cancer some-time dur ing the 25 years following the first cancer.

Most of the patients in the stud y were treated w ithhigh-dose rad iation th erapy. The risk of secondcancers was highest among children w ith cancersknow n to be strongly genetically influenced. Inthat group, there may well be a genetic defect thatpredisposes to mutations, e.g., a faulty repairmechanism, which could also be related to ahigher risk of heritable mutations in that grou p.

Cancer registries are the most nu merous regis-tries of any type, and cohorts of treated p atientsand patients with second tumors are relativelyeasy to identify, comp ared with identifying otherpopulations potentially exposed to mutagens.These group s should be considered w hen stud iesof heritable mu tations u sing the n ew technologiesbecome feasible.

Other Populations

A study of birth outcomes in people who hadattempted suicide by self-poisoning in Hungaryis an examp le of opp ortun istic us e of availableinformation (23), A cohort of about 1,300 indi-viduals w ho took large doses of drugs in suicideattempts has been studied since 1976. Early on ,the investigators looked for short-term effects onsomatic cells, using cytogenetic and biochemicaltesting. Long-term followup of birth outcomes ex-amined spontaneous a bortions, ectopic pregnan-cies, stillbirths, low birthweight, and congenitalanomalies. The study suffered a large loss of fol-lowup of stud y subjects, but in those evaluated,no imp ortant excesses in an y of these endp ointswere discovered.




CONCLUSIONS

A very important question is answered by sim- greatly increase the pow er to identify mu tationsply observing birth outcomes in people thought in studies such as those described above, addingto be at high risk, nam ely whether those individ- another dimension to knowledge about the rela-uals are at risk of having childr en w ith serious tionship between exposure to mutagens, the pres-diseases and d isabilities. The new m utation d e- ence of detectable mutations in DN A, and the ex-tection technologies discussed in this report may istence of observable health effects.



Chapter 9

Mutagens: RegulatoryConsiderations



Chapter 9

Mutagens: Regulatory Considerations

INTRODUCTION

It is clear from the information gathered in thisreport that w e are some years distant from beingable to muster convincing direct evidence of anybut very large increases in th e rates of heritablemutations in human beings. The ability to pre-dict from experimental data which agents arelikely to increase the mutation rate if human be-ings were exposed h as not been pu t to the test.Even for the most likely mutagens, the ability tomake quantitative extrapolations is relatively un-developed, N onetheless, it is reasonable and p ru-dent to accept that the environment m ay containhuman germ-cell mutagens and that, to the ex-

tent possible, huma n beings should be protectedfrom them at levels that m ight cause heritable mu -tations.

Ionizing radiation was recognized as a cause of heritable m uta tions in fruitflies in the 1920s, andthe Federal Government has since made effortsto protect workers and the public from excessive

radiation exposure. Widespread concern abou t themutagenic potential of chemicals is more recent,an issue brought into focus by the environmentalmovem ent that took shape in the late 1960s. Wh ilethe potential cancer-causing properties of m a n-mad e chemicals have been the d riving force be-hind environmental health laws, two more recentlaws specifically mention m utation as an endp ointagainst which the public should be protected.About a dozen other statutes includ e languagebroad enou gh to charge the Federal Governmentwith the responsibility to protect against herita-ble mutations. Evidence from current methods for

measuring m utation rates suggests that basing reg-ulation of environmental agents on carcinogenic-ity will likely assure protection against heritablemutations, but new, more sensitive detection tech-nologies, such as those d iscussed in this report,may necessitate a reexamination of that con-clusion.

FEDERAL INVOLVEMENT IN PROTECTING AGAINST GENETIC RISK

Radiation Protection

In 1928, the n ew ly created Interna tional X-Rayand Radium Protection Commission was chargedby the Second International Congress on Radiol-ogy with developing recommendations for pro-tection against radiation (156). The followingyear, the Ad visory Committee on X-Ray and Ra-dium Protection was formed to represent the U.S.viewpoint to the international commission. Thesetwo bod ies were the forerunn ers of the currentInternational Comm ission on Rad iological Pro-tection (ICRP) and the N ationa l Council on Ra-diation Protection and Measuremen ts (NCRP),

the latter char tered by the U .S. Congress in 1964.The ICRP and NCRP have, since their first rec-ommendations in the 1930s based their accept-

able radiation exposure limits on both heritable

an d somatic effects. The limits recommended have

been lowered over the years, reflecting increasedknowledge about radiation effects, and particu-larly about th e effects on th e pop ulation of lowlevels of radiation.

Neither the ICRP nor N CRP recomm endationshave the force of law, but by an d large, they haveformed the basis for the radiation protection limitsadop ted by U.S. regulatory agencies. The firstFederal entity officially charged with providingthe agencies with gu idance for d eveloping rad ia-tion protection standard s was th e Federal Radia-tion Council (FRC), established in 1959. In 1960,FRC issued recommendations for both occupa-

tional exposure and exposure of members of thepublic, which drew on ICRP and NCRP work (156). Over the years, the National Academy of

Sciences Comm ittee on the Biological Effects of Ionizing Radiation and the Un ited N ations Sci-

113




entific Comm ittee on th e Effects of Atomic Radi-ation have also been influential in providing anal-yses that undergird exposure limits.

The N ational Environm ental Protection Act of 1970 transferred the responsibilities of FRC to thenew Environmental Protection Agency (EPA).

EPA administers several environmental healthstatutes under which that Agency is responsiblefor setting standard s for radiation exposure in spe-cific conditions. Under its broader responsibili-ties, EPA has provided guidance for exposurefrom d iagnostic X-rays, which is the regu latoryresponsibility of the Federal Food and DrugAdm inistration (FDA), and for exposure of ura-nium miners, who are the responsibility of theMine Safety and Health Administration (MSHA).

Agencies other than EPA with responsibility forsome aspects of radiation protection include: the

Nuclear Regulatory Commission, the Departmentof Energy, the Department of Defense, FDA,MSHA, the Occupational Safety and HealthAdministration (OSHA), and the Department of Transportation. The States have responsibilitiesas well. Each entity, depending on its specificcharge, is required to protect workers, the pub-lic, or both in accordan ce with the guidan ce pro-vided by EPA.

The basic occupational and population ex-posure guidelines have not been revised since1960. In 1981, EPA proposed new occupational

guidelines (153), in line with 1977 ICRP recom-mend ations (47), but th ese have not been mad efinal, and they d o not represent a change in totalacceptable dose from th e earlier guidelines. Theydo, how ever, place less of the emph asis on mu ta-genesis, and relatively more on somatic effectsthan do the 1960 guidelines. The ICRP includesin its risk estimates only gen etic effects occurringin the first two generations after exposure. Thatprobab ly accounts for roughly half of the totalgenet ic effect, whatever its size.

The current occupa tional exposure limit is 5 remtotal body d ose per year, with not more than 3rem total body d ose from occupational exposurein any one qu arter of the year, and a more d etailedbreakdown for different groups of organs. The1977 ICRP recommendations abandon the speci-

fications by organ, and use a weighted whole-body dose.

Today, almost all rad iation exposures of U.S.workers are well below the regulated limits,though there are exceptions. The quantitativelimits set by the ICRP and NCRP and adopted

by Federal groups, are accompanied by the“ALARA principle’ ’—that radiation exposuresshou ld be “as low as reasonably achievable. ”

Exposure to the p ublic are to be limited to be-low 25 millirem (mrem) to the whole body, 75

mrem to th e thyroid, and 25 mrem to any otherorgan. These levels mainly affect the regulationof radionuclides in air an d the d isposal of radio-active w aste. The n um bers come from ICRP andNCRP recommendations, and are based on con-sideration of both genetic and somatic effects.

Agents Other Than Radiation

Congress formally recognized the n eed to p ro-tect against chemical mutagens in the Toxic Sub-stances Control Act of 1976 (TSCA), and againin 1980 in the Com prehensive Environmental Re-sponse, Compensation, and Liability Act (CERCLA,or “Superfund ”). These two laws are adm inisteredby EPA, as are other statutes that includ e broadmandates to protect the public from environ-mental hazards. Other laws designed to protectcitizens from external agents u nder wh ich chem-

ical mu tagens could be regu lated include th e Fed-eral Food, Drug, and Cosmetics Act, administeredby FDA; the Occup ational Safety and H ealth Act,adm inistered by OSHA; the Consum er ProductSafety Act, administered by the Consum er Prod-uct Safety Commission; and the Atomic EnergyAct, through which the Nuclear Regulatory Com-mission is empow ered to p rotect certain w orkersfrom radiation hazards.

Although it is almost certain that chemicals thatmight cause heritable mutations have been regu-lated, no regulations have been w ritten or stand-ards set for these agents because of that prop erty.In a few cases, the mutagenic potential of chemi-cals has been considered by regulatory agencies,but carcinogenic properties have driven stand ard-setting.



Ch. 9—Mutagens: Regulatory Considerations . 115

OSHA has included thorough reviews of mu-tagenicity data in notices of regulatory actions fortwo of the best-publicized chemical hazards of the1980s: ethylene oxide (EtO) and ethylene dibro-

mide (EDB). Tests for heritable mutagenicity inDrosophila and experimental mammals have

yielded positive results in at least some systemsfor both of these chemicals. While results of mutagenicity tests are included in the Federal Reg-ister notices for th ese chemicals, quantitative ex-trapolations for both the final EtO standard (152),and the proposed rulemaking for EDB (151), arebased on pr otecting against carcinogenicity.

REGULATORY ISSUES

Currently there does not ap pear to be a scien-

tific basis for the specifics of regulatory actionagainst mutagens. The following questions faceregulators and the scientific community involvedin mutation research:

1.

2.

3.

4.

What is an app ropriate regulatory d efinitionof a probable human germ-cell mutagen andhow does that definition relate to what isknown about mutagens from epidemiologicstudies and experimental stud ies?Is it possible to derive quantitative estimatesof the risk of heritable mu tations in hu man sfrom experimental evidence in animals orfrom somatic-cell mutation tests in humanbeings? If it is not, w hat kind s of informa-tion are necessary before such extrapolationis possible?How likely is it that a substance will requirea more stringent standard as a mutagen thanit will as a carcinogen or for other toxiceffects?How will information from the new technol-ogies for detecting heritable mutations thatare d escribed in this assessment change ourperception of the kinds of “adverse effects”against which regulation shou ld be d irected?

The discussion in th e remaind er of this chap -ter ad dresses these questions.

EPA has recognized germ-cell mutagenicity asa class of adverse effects, particularly in its respon-sibilities under the Federal Insecticide, Fungicide,and Rodenticide Act, un der w hich it, in ad ditionto OSHA, has acted to regulate exposures to EtOand EDB. Unique among the regulatory agencies,

EPA’s Reprod uctive Effects Assessment Grou p (inthe Office of Research and Development) has pre-pared guidelines for m utagenicity testing, whichare described later in this chapter.

A Regulatory Definition of

a Germ-Cell Mutagen

Strategies for regu lating mu tagens to p rotectpu blic health cannot tod ay rely on da ta from cur-rent stud ies of heritable mutations in hum an be-ings. Just as is the case in reg ulating carcinogens,a regulatory d efinition mu st serve as a substitute,particularly for making jud gments about th e po-tential risks of new su bstances.

Certain lessons can be learned from experiencein regulating carcinogens. (For a discussion of theissues surrounding carcinogen regulation, see

145. ) The m ost convincing ev idence for carcino-genicity, from well-conducted epidemiologic stud-ies, is that hum an beings have d eveloped cancerafter exposure to a given agent. If an increase ingenetic d i sea se could be convincingly show n tobe related to a sp ecific agent, there wou ld certainlybe no p roblem in acting against that agent. Thespirit of the regulatory law s, how ever, embod ythe concept of taking protective action before peo-ple are har med . In th e absence of direct evidencefrom human beings, regulators must rely on in-direct evidence from a variety of experimental testsystems wh ich w ill never b e a bs olutely p redict ive

of effects in h um an beings. A regu latory defini-tion of a m utagen w ill be pragm atic and rely on




information that it is p ossible to collect, and ona nu mber of untested assumptions.

EPA is the first U.S. regulatory a gency to ha veprop osed “Gu idelines for Mu tagenicity Risk As-sessmen t” (154). The guid elines require evid enceof: 1) mutagenic activity and 2) chemical inter-actions in the mam malian gonad . The d ecision toregulate is to be based on a “w eight-of-evidence”determination. EPA has proposed no formalmethod for quantitative extrapolation by whichacceptable exposu re levels could be set.

According to the EPA guidelines, evidence formu tagenic activity may come from tests that de-tect point mu tations and structural or num ericalchromosome aberrations. Structural aberrationsinclude deficiencies, duplications, inversions, andtranslocations. In the absence of evidence of heritable mutations in human beings, evidence

from a variety of experimental test systems maybe invoked . For mu tagens that cause point mu-tations, whole animals tests (e.g., the mousespecific-locus test) provide the highest degree of evidence, but these tests are relatively more ex-pensive than short-term tests, and there is alimited capacity for laboratories to perform them.Other tests for point mu tations includ e those inbacteria, eukaryotic micro-organisms, higherplants, insects, and mam malian som atic cells.

Structural chromosome aberrations can be de-tected either in somatic or germ cells in different

assays. The organisms used include higher p lants,insects, fish, birds, and several species of mam-mals. Mutagens that cause num erical changes inchromosomes may be missed by the tests thatdirectly measu re DN A da ma ge. Tests specificallydirected at detecting changes in chromosome nu m-ber are not as w ell developed as are th ose for pointmutations or structural changes in chromosomes.Tests are in various stages of development infungi, Drosophila, mam malian cells in cultu re,and intact mammals, including mammalian germ-cells tests.

Results from tests that measure en dp oints otherthan m utagen icity directly may also be used in judging the potential mutagenicity of a substance.DNA damage, unscheduled DNA synthesis inmam malian som atic and germ cells, mitotic re-combination and gene conversion in yeast, and

sister chromatid exchange in mammalian somaticand germ cells are cited by EPA as tests that pro-vide evidence known to be correlated with m u-tagenicity, though they measure other geneticevents.

Evidence from various kinds of mutagen icitytests is w eighted w ith regard to th e relationshipof the test to hu man germ-cell mutation. Greaterweight will be given to results from tests in: 1)germ cells over soma tic cells, 2) mam malian cellsover su bma mm alian cells, and 3) euk aryotic cellsover prokaryotic cells.

EPA lists two classes of evidence for chemicalinteractions in the mammalian gonad: sufficientand suggestive. Sufficient evidence is from studiesin whole mammals that demonstrate, for exam-ple, unscheduled DNA synthesis, sister chromatidexchange, or chrom osomal aberrations in germcells. Adverse effects on the gonads or on repro-du ctive outcomes after exposure, w hich a re con-sistent with th e substance reaching the gon ads bu twhich do not indicate direct interaction withDNA, are considered as providing suggestiveevidence.

The final step in EPA’s mu tagenicity risk assess-ment is the weight-of-evidence determination,which classifies the evidence for potential germ-cell mu tagen icity as “su fficient,” “su ggestive,” or“limited.” In this step, results of tests plus any in-

formation about effects in human beings is evalu-ated. Sufficient evidence consists of a positivemam malian germ -cell test. In ad dition, positiveresponses in at least two d ifferent test systems,at least one of which is in mammalian cells, andevidence of germ-cell interaction, together con-stitute su fficient eviden ce. Eviden ce of lesser qu an-tity and / or quality of both mu tagenic responseand germ-cell activity constitute suggestive evi-dence. Limited evidence consists of positive re-sults in either mutagenicity assays or tests forchemical interactions in the gonad, but not both.

EPArS guidelines became final in September

1985. Currently and for the foreseeable future, the

greatest value of EPA’s guidelines is the recogni-tion of germ-cell mutagenicity as a legitimate end-point to consider in assessing the p otential adverseeffects of substances in the environm ent.



Ch. 9—Mutagens: Regulatory Considerations 117

Quantitative Extrapolation

If the levels of risk from suspected h um an germ -cell mutag ens is to be estimated in the absence of direct evidence of harm in human beings, datafrom experimental systems must be used in a

“quantitative extrapolation. ” The experimentalsystems are basically those mentioned in the EPAguidelines discussed in the p revious section, andthose discussed elsewhere in this report. The threecategories of tests are: 1) whole animal heritablemutation studies; 2) animal somatic-cell mutationstud ies, either in vivo or in vitro; and 3) hum ansoma tic-cell mu tation stu dies, either in vivo or invitro. Unfortunately, the kind of information (i.e.,measures of hum an m utations) that w ould link results from these three categories of tests to hu-ma n heritable mu tations is scanty. It is encour ag-ing, however, that using tests available now, such

information can be generated a t least for somesubstances. For EtO and EDB, for instan ce, mu -tagenicity data are available in both somatic andgerm-cell systems in animals, and some somaticcell (cytogenetic) data are available from humanbeings exposed at know n occup ational levels.

Obtaining more information to fill in the ar-rows of the extrapolation “parallelograms” pre-sented in chapter 7 of this report should be a highpriority for regulators. In fact, EPA’s Reproduc-tive Effects Assessment Group has collaboratedwith other groups in the Federal Government to

fund such studies (168). Without the kind of in-formation that would come from coordinatedstud ies in several test systems, there is little chan ceof writing a successful regulation that limits ex-posu re to a sp ecific level (short of a com plete banfor an agent acknowledged to be unacceptablyrisky at any level). Given the experience with car-cinogens, a regulation that states an exposure levelwithout adequate experimental evidence and the-ory behind it will not survive a court challenge,wh ich, in the United States today, app ears to bethe final test of a regulation.

Even with good experimental d ata, some of thesame problems that p lague extrapolating from ani-mals to hu mans to determine acceptable exposurelevels for carcinogens are certain to hinder quan-titative extrapolation for estimating levels of mu tagenic risk at specific levels of exposu re. In

carcinogen extrapolation, there still are un resolvedcontroversies about the appropriate conversionfactors between species and about th e shape of dose-response curves. The latter is important be-cause most animal bioassays test extremely highdoses in relation to the animals’ body weights,

while humans are generally exposed at lowerlevels over longer periods of time. Though the de-tails of extrapolation for mutagenicity differ fromthose for carcinogenicity, the pr oblems w ill un-doubtedly be similar. Right now, there is notenough information about the relationships be-tween results in various test systems to addressintelligently the practical problems of actually per-forming extrapolations.

Given the app ropriate information, it may be-come possible to carry out quan titative extrapo-lations for m utagen icity. When the tim e comes,

the regulatory agencies will need to requ ire thatthe app ropriate tests are done, either by m anu-facturers or by the Federal Government. The test-ing requirement may take various forms, whichmay vary by statute.

Mutagenicity and Carcinogenicity

As men tioned previously, w hile the regulatoryapp aratus exists for acting against hu man germ-cell mutagens, in fact no regulations based ongerm-cell mutagenesis exist except for radiationexposure limits. There are several reasons for this.

First, there are no proven human germ-cell muta-gens, and only a limited num ber of presum ptivehuman mutagens known from animal tests. Sec-ond, it has often been thought that regulationsbased on demonstrated carcinogenicity wouldautomatically protect against mutagenicity aswell. In fact, however, this may not always betrue. Voytek (167) repor ted a preliminary assess-ment indicating that the risk of heritable geneticdisease in the first generation after exposu re toEDB was greater than the lifetime risk of cancerin the exposed individu als, based on a n extrapo-lation from animal data.

It is wid ely held that a som atic mu tation is anecessary step in the development of cancer.Many substances that are m utagenic in bacteria,in cells in cultu re, and in Dr osoph ila also are car-cinogen ic in labor atory r ats, mice, or both. Short-




term tests based on m utagenicity in these lowerorganisms are therefore used as screens for car-cinogenicity. The most widely used screening testfor carcinogen icity is the Ames test, wh ich m eas-ures mutagenicity in strains of the bacteriumSalmonella. Under some statutes (e.g., TSCA),

negative results in short-term tests, meaning thatthe substance is not mutagenic in those systems,can obviate the need for a long-term bioassay, asavings of up to $1 million to a manufacturer(145). Positive results in short-term tests, mean-ing that the substance is mutagenic in these sys-tems, have not been accepted as a basis for regu-lation un der an y statute, but they have p robablyhalted the d evelopm ent of new chemicals. Man-ufacturers know that p ositive mu tagenicity testsmay trigger the requirement for a long-term bio-assay, which in a high prop ortion of cases willturn out positive. The product, whatever it is,

might never get to m arket. Rather than risk a fi-nancial loss, many manufacturers simply will notproceed w ith that produ ct.

New Methods for Measuring MutationRates and Their Potential Effects onRegulation

The new techniques for detecting heritable mu-tations and the somatic-cell techniques that even-tually may be used to predict germ-cell mutagen-esis will lead to the consideration of effects thatare increasingly removed from measurable or evenhyp othetical adverse health effects in hu man s.Mutagenicity endpoints may be detected thatcould be m ore sensitive than th ose currently u sedto predict carcinogenicity. In the regulatory con-text, jud gm ents w ill hav e to be mad e, most likelyin the absence of certain knowledge of effects,

about the appropriate actions to be triggered bydemonstrations of various kinds of changes inDNA, detectable by various analytic methods,that can be convincingly linked to specific ex-posures. From a public health standpoint, it ismost appropriate to act under the assumption that

mu tations of any kind ar e deleterious, and thatenvironmental agents at levels that cause anyreliably detected chan ges in the DN A shou ld besubject to available regulatory controls. This doesnot get aroun d th e problem of defining “safe” or“acceptab le” exposure levels, however. Animal ex-periments w ill be needed to explore the qu antita-tive relation betw een subtle changes d etectableanyw here in DNA and th e levels of adv erse ef-fects that might be observed in the animal.

At p resent, testing for safety is a significant partof the research and developm ent investment innew prod ucts, wh ether they are chemicals, dru gs,or food add itives; determ ining the risks of sub-stances already in the environment is a significanttask for the Federal Government. If they becomeavailable, new mutagenicity testing technologiesusing experimental animals could either imposesignificant new testing requirements in add itionto those already in place, or could replace someexpensive and not entirely reliable tests for car-cinogens. Regardless of the method s used to de-tect mutations, the relation between specific typesof changes in DNA and health effects w ill haveto be studied experimentally to shed light on the

mea ning of a “positive result. ” Until this know l-edge is available, the ability to detect an effectwithout knowing th e likelihood of any health con-sequences for human beings will remain a thornypublic policy question which scientists, regulators,and politicians must address.



Appendixes



Appendix A

Federal Spending for Mutation Research

A quick review of the origins of biotechnology il-lustrates th e d ifficulty of identifying th e specificareasof research that would further our know ledge of hu-

man m utations, and, in turn, the problems of estimat-ing the current amou nt of Federal expend itures for suchresearch.

In 1968 and 1969, Ham ilton O. Smith, a microbiol-ogist at Johns Hopkins University, had no thought of studying hu man heritable mutations. No venture cap-ital firm w as looking for biotechnology compan ies toinvest in, and had it been, it would have been dis-appointed. No such company existed,

Smith w as studying h ow som e bacteria can take upmolecules of DNA from their growth medium andrecombine it into their genetic material. He kn ew th atthe bacteria w ould not take up DNA from other spe-cies of bacteria or from viruses, and h e add ed som e

viral DNA to the experiment expecting that it wouldremain inert and intact while the bacterial DNA u n-derw ent recombination. Instead, the viral DNA wasrapidly degraded. Looking at that result, Smith hy-pothesized that a bacterial enzyme(s) could discrimi-nate betw een foreign DN A (from the virus), which itdegrad ed and self DNA (from th e same kind of bacte-ria), wh ich it left alone. He p urified the enzy me an dshowed that it recognized particular sequences of nu cleotides on the viral DNA and cut it at those sites(128).

These experiments, wh ich identified and character-ized the first known restriction enzyme, opened thedoor to the d iscovery of hund reds of others. The use

of restriction enzymes, allowing precise cutting of DN A and precise joining of DNA pieces from differ-ent p arts of an organ ism’s genome as well as joiningof pieces from different organisms, led to fund amen-tal advances in the basic approach of research inmolecular biology and genetics (146,147). Accord-ingly, these enzymes are basic ingredients of the newtechnologies for studying human heritable mutationsdescribed in this report,

However, someone in 1969 who had tried to iden-tify research projects that w ould contribu te to our u n-derstanding of human m utation rates would almostcertainly not have includ ed Smith’s. His w as basic re-search directed at understanding the mechanism of

recombination in an organism far removed from h u-man s. This OTA report p rovides add itional examplesof the difficulty of estimating how mu ch money is be-ing spent on research that m ay be important to un-derstanding human mutations.

Current Federal Expenditures for

Mutation Research

OTA queried Federal research and regu latory agen-cies about their supp ort of research directed at und er-standing human mutations. Questions were askedabout the amount of money spent specifically on hu-man m utation research and the amount spent on “re-lated” areas of biological research. The first categorywa s tightly defined—the research had to focus on d e-velopment or app lications of method s for detectingand/ or counting hum an somatic or heritable mutationsin vivo. The category of “related” research w as mu chbroader, including studies examining genetic contri-butions to such common diseases as diabetes and ar-thritis, studies of mutagenesis and repair mechanisms,developm ent of instru men ts for measuring cellular ef-

fects of various kind s, and tests for m utagenic activ-ity in lower organisms as well as in cultured hu mancells.

The expenditures shown in table A-1 must be treatedwith some reservation about their accuracy, the totalof $14.3 million estimated as the am oun t spent on hu -man mutation research being more precise than the ap-proximately $207 million estimated for related geneticresearch. The $14.3 million may be an overestimate—

Table A-1 .—Federal Expenditures for ResearchRelated to Human Mutations

Human Related

mutation genetic

Federal aqency research research

Department of Energy $ 6,220,000a$ 30,123,000

Department of Health and HumanServices:

Centers for Disease Control 277,000 1,387.000National Institutes of Health, 7,244,000b 156,959,000CFood and Drug Administration 1,200,000

Environmental Protection Agency 606,000 1,189,000

National Science Foundation 0 16,336,000

T o t a l 14,347,000 207,194,000aThl~ includes $1 ,e,50 000 (or half of the total $3 700,000 atcurrent exchange rates) spent annu-

ally for the U S contribution to the Radlatlon Effects Research Foundation the group that coor.

dmates ongoing genetic studies of the survwors of the Nagasaki and Hlroshlma bombings andthetr offspr!ngbThls total was obtained by inspecting grant Mles and reading grant abstracts to ldefltlfy thOSe

that focused on human mutahon researchcThls figure IS fhe budget for the (%netlcs Program m the National lnStltW? Of General Medcal

Sctences It does not include all genetics research at NIH some of which goes on m other Institutes

SOURCE Off Ice of Technology Assessment

721




some projects studying clinical effects of mutationsmay provide no information about d etection—and the$207 million for related genetic research is probablyan und erestimate, since mu ch genetic research is sup -ported by p arts of the National Institutes of Health(NIH) not queried. For comparison, the total NIHbudget for the fiscal year 1985 was about $4 billion.

Most of the NIH funding for mutation research is di-rected at identifying mutations in at-risk families(where sp ecific tests for specific mu tations are ap pro-priate), wh ereas DOE sup ports a search for more gen-eral methods to d etect mu tations.

Successes of the research efforts in hu man mu tationresearch and related genetic research w ill probablygenerate a requirement for additional funding in thenear futu re. It w ill be a costly ventu re to take any n ewtechnique (such as those d escribed inch. 4) and app lyit to stud ies of popu lations su fficiently large to pro-vide a good possibility of yielding inform ation about

hu man m utations. For instance, the ongoing monitor-ing of the Nagasaki and H iroshima p opulations usingclinical examinations and protein analyses (see ch. 3)costs about $3.7 million annu ally to the U.S. and Jap-anese governments together (48).



Appendix B

Acknowledgments and

Health Program Advisory Committee

HEALTH PROGRAM ADVISORY COMMITTEESidney S. Lee, Chair

President, Milbank Memorial Fund

H. David BantaProject DirectorSTG Project on Future Health TechnologyThe Netherlands

Rashi FeinProfessorDepartment of Social Medicine and Health PolicyHarvar d Medical School

Harvey FinebergDeanSchool of Pu blic HealthHarvard University

Patricia KingProfessorGeorgetown Law Center

Joyce C. Lashof DeanSchool of Pu blic HealthUniversity of California, Berkeley

Alexander Leaf Professor of MedicineHarvard Medical SchoolMassachusetts General Hospital

Frederick M ostellerProfessor and ChairDepartmen t of Health Policy and Managem entSchool of Pu blic HealthHarvard University

Norton NelsonProfessorDepartment of Environmental Med icineNew York Un iversity Medical School

Robert Oseasohn

Associate DeanSchool of Pu blic HealthUniversity of Texas

Nora PioreSenior Fellow and Advisor to the PresidentUnited Hosp ital Fund of New York

Dorothy RiceRegents LecturerDepartm ent of Social and Behav ioral SciencesSchool of NursingUniversity of California, San Francisco

Richard RiegelmanAssociate ProfessorGeorge Washington UniversitySchool of Medicine

Walter RobbVice Presiden t & General Man agerMedical Systems OperationsGeneral ElectricMilwaukee, WI

Frederick C. RobbinsUniversity ProfessorDepartment of Epidemiology and BiostatisticsSchool of Medicine

Case Western Reserve UniversityCleveland, OH

Frank E. Samuel, Jr.PresidentHealth Industry Manufacturers’ Association

Rosemary StevensProfessorDepartment of Historyand Sociology of ScienceUniversity of Pennsylvania

123




ACKNOWLEDGMENTS

OTA would like to thank the m embers of the advisory pan el and the contractors who provid ed informationfor this assessment. In addition, OTA acknowledges the following individuals for their assistance in reviewingdra fts or supp lying information.

F. Alan And ersen

U.S. Food a nd Drug Adm inistration

Benjamin J. Barn har tU.S. Department of Energy

Fred Bergman nNational Institute of General Medical Sciences

David BotsteinMassachusetts Institute of Technology

Noel Bouck North western University School of Medicine

Michele CalosStanford University School of Medicine

L.L. Cavalli-SforzaStanford University School of Medicine

George ChurchUniversity of California at San Francisco

Mary Ann DanelloU.S. Food a nd Drug Adm inistration

Amiram DanielHealth Industry Manufacturers Association

Freder ick J. DeSerresNational Institute of Environmental Health Sciences

Robert DixonU.S. Environmental Protection Agency

William Fitzsimm onsNational Institute of General Medical Sciences

W. Gary FlammU.S. Food an d Drug Ad ministration

Joseph Fraum eniNational Cancer Institute

Helen GeeNational Institutes of Health

Margaret GordonNational Institutes of H ealth

Ronald H artNational Center for Toxicological Research

John HeddleLudw ig Institute for Cancer Research (Canad a)

Richard Hill

U.S. Environmental Protection Agency

Abraham HsieOak Ridge Na tional Laboratory

Seymour JablonNational Research Coun cil

Irene JonesLawrence Livermore National Laboratory

Moin KhouryJohns Hopkins University

Norman MansfieldNational Institutes of H ealth

Karen MessingUniversity of Quebec at Montreal

Alexander MorleyFlinders Medical Center (Australia)

John J. Mulv ihillNational Cancer Institute

DeLill NasserNational Science Foundation

Neal S. NelsonU.S. Environmental Protection Agency

Per OftedalUniversity of Oslo

Doug Ohlendorf E.I. du Pent de Nemours & Co.

Dilys ParryNational Cancer Institute

Claud e PickelsimerCenters for Disease Control

William L. RussellOak Ridge National Laboratory

F. Raymond SalemmeE.I. du Pent de Nemours & Co.

Michael M. SimpsonCongressional Research Service

David A. SmithU.S. Department of Energy



App. B—Acknowledgments and Health Program Advisory Committee • 125

Gail Stetten Larry R. Valcovic

Johns H opkins H ospital U.S. Environmental

J. Timothy Stout Peter E. Voytek

Baylor College of Medicine U.S. Environmental

Gary StraussU.S. Environmental Protection Agency

Protection

Protection

Agency

Agency



References



— ———

136 “ Technologies for Detecting Heritable Mutations in Human Beings

Group , U.S. Environmental Pr otection Agen cy,personal communication, Aug. 7, 1985.

169. Warburton, D., Stein, Z., and Kline, J., et al.,“Chromosome Abnormalities in SpontaneousAbortion: Data From the N ew York City Stud y,”E’rnbryortic and Fetal Death, I .H. Porter and E.B.H ook (eds.) (New York: Acad emic Press, 1980),

p p . 261-287.170. Waters, M., Allen, J., Doerr, C., et al., “Appli-cation of the Parallelogram Ap proach To Pre-

dict Genotoxic Effects in Hu man s, ” typescript,1985.

171. Weathera l l , D. J., The New Genetics and ClirIi-

cal Practice (London: Nuffield Provincial Hos-pitals Trust, 1982).

172. Yang, T. P., Patel, P. I., Chinau l t , A. C., et al.,“Molecular Evidence for New Mu tation at the

hprt Locus in Lesch-Nyhan Patients, ” Nature310:412-414, 1984.



Index



Index

Achondroplasia (dwarfism), 7, 30, 31Ad hoc studies, epidemiologic, 14, 103, 104, 107Adenine (A), 23, 24, 27, 64, 66Aden osine d eaminase deficiency, 105Agarose gel electrophoresis, 59, 60, 61, 63, 68

Agency for Toxic Substances and Disease Registry(ATSDR), 18-19

Agent Oran ge, 18Alanine, 27ALARA prin ciple, 114Alkylating agen ts, 93-94Alleles, 27, 44, 77, 82Alpha-globin chain, 47Alpha-thalassemia, 29American Society of Human Genetics, 19Ames test, 15, 118Amino acids, 9, 25-27, 29, 44, 46, 47, 49, 57, 58,

78; see also specific amino acidsAmniocentesis, 41

Animals, mutation studies in, 3-5, 6-7, 19, 48, 78,81-87, 91-100, 117-118

Aniridia, 7Antibodies, in somatic mutation studies, 77-78Apert’s syndrome, 31Arginine, 27Asparag ine, 27Aspartic acid, 27Atomic Bomb Casualty Commission, 48Atomic bomb survivors, effects of radiation on,

9-10, 32, 48-50, 77, 86, 106, 108Atomic Energy Act, 114Atomic Energy Commission (AEC), 82, 114Autorad iographic assay, in somatic mutation

studies, 74-75Au torad iograp hy , 56, 59, 61, 64, 65, 68, 74-75Autosom es, 27, 38, 42, 44, 46, 49

Bacteria, 81, 91, 92, 94, 116, 117, 118Benzo[a]pyrene, 97Beta-globin chain, 47, 77Biochemical analyses, 37, 44-48, 49-50, 51, 52Biologic extrapolation, see Extrapolation,

qualitativeBiological effects of ionizing r ad iation, 113Biological samp les, bank ing of, 17-18Blood cells, somatic mutation studies on, 73-78Blood proteins

detection of variants in, 44-48, 49-50electrophoretic analysis of, 8-9, 10, 37, 44-46,

50, 51, 83

molecu lar ana lysis of, 46-48see also, Enzymes; Hemoglobin proteins

Cancer, 3, 7, 9, 15, 31-32, 38, 50, 73, 77, 104-105,108-109, 117

Cancer registries, 105, 109Case-control studies, in epidemiologic testing, 107Cells, see Blood cells; Chorionic villi cells; Germ

cells; Red blood cells; Somatic cells; Stemcells; TG

rcells; White blood cells

Centers for Disease Control, 18Chemicals

denaturing, 12, 59, 61effects on DNA, 23, 59, 81effects on germinal mutations, 7, 9effects on heritable mutations, 6-7effects on somatic mutations, 7, 74-78in extrapolation models, 93-97, 100as mutagenic agents, 6-7, 9, 32, 48, 91-97, 106“stand ard s” for, 17

Chemotherapy, 108Children, genetic diseases in, 7-8, 30-32, 37, 38-40,

47-48, 49/ 50, 109, 110Chorionic villi cells, 41Chromosomes

abnorm al at birth, 41-43banding methods, 8, 28, 41, 42, 43, 50biochem ical analyses of, 49-5ocytogenetic analysis of, 37, 40-43, 49, 52, 55, 98DNA and, 23-25, 27-28, 41, 43health effects of abnormal, 40-43, 49mutations in, 5, 6, 8, 9, 13, 29-30, 31, 32,

40-43, 49, 51, 68, 82, 92, 95numerical abnormalities, 8, 29-30, 31, 41, 49,

116pulsed field gel electrophoresis and, 68

staining methods, 8, 28, 41, 42, 43, 50structural abnormalities, 8, 29, 31, 41, 49studies in mice, 82surveillance of abnormal, 105translation of, 30, 82

Chronic hemolytic anemia, 46, 47Clonal assay, in somatic mutation stud ies, 75Cloning, 58-59, 60, 64-65, 75Codons, 6, 25-27, 29Cohort studies, in epidemiologic testing, 107, 109Comprehensive Environmental Response,

Compensation, and Liability Act of 1980(CERCLA), 3, 15, 114

Congenital hypothyroidism, 105

Consumer Product Safety Act, 114Consum er Prod uct Safety Comm ission, 114Cyanosis, 47Cyclophosphamide, 97Cysteine, 27

139




Cytogenetic analysis, of chromosomeabn orm alities, 37, 40-43, 49, 52, 55, 98

Cytosine (C), 23, 24, 27, 64, 66Cytotoxic drugs, 108

Deoxyribonucleic acid (DNA)chromosomes and, 23-25, 27-28, 41, 43

current m ethods for stud ying, 6-10denaturing of, 11-12, 59-64double helix, 24, 25, 27double-stranded, 12-13, 23, 59-68EPA guidelines and, 116extrapolation stud ies and, 91-1OOFederal expenditures on research, 121-122function and organization of, 23-28genetic nature of, 3, 4, 5-6genomic, 6, 11, 55, 57-68heteroduplexes, 12, 55, 61, 62, 63, 64-65, 67homoduplexes, 61, 63, 67hybridization, 66-67nontranscribed, 45

nu cleotide chains in, 23-25organizational hierarchy of, 6replication of, 23-25, 74sequencing, 23-25, 55, 57-58, 60, 61-64, 66-67,

104single-stran ded, 12-13, 59-68structure of, 23-25, 81use of new technologies in examining, 10-14, 17,

55-69, 74-76Departm ent of Defense, 114Department of Energy (DOE), 16, 17, 19, 55, 114,

122Department of Transportation, 114Diabetes, 3

Dioxin, 106Disease registries, 104-106, 107Diseases

dominant, 44environm ental factors in, 31genetic, 3, 5, 7-8, 9, 30-33, 37, 38-40, 42, 44-45,

47-48, 49, 50, 93, 94, 103-110, 108-109, 110,115

“multifactorial, ” 31recessive, 44see also Sentinel phenotypes and specific diseases

DNA-RNA heteroduplex analysis, 12, 64-65Dominant lethal test, 81-82Dominant m utation tests, 83Dose-response relationships, in quantitative

extrapolation studies, 97-99Down syndrome, 3, 8, 30, 31, 42Drosop hila, 48, 81, 82, 96, 116, 117Duchenne muscular dystrophy, 3, 30

Electrophoresis, 8-9, 10, 11-12, 13, 37, 44-46, 49,50, 51, 55, 68, 83, 84

see also Gel electrophoresis and specific types of

gel electrophoresis

Embryos, in rodent tests, 81-82Endonucleases, restriction, use in detecting

mu tations, 11

Environmental Protection Agency (EPA), 15, 16,19, 114, 115, 116, 117

Enzymesbiochem ical analyses of, 49-5oHPRT, 74kinetic studies of, deficiency variants, 44, 45-46quantitative studies of deficiency variants, 45-46,

83, 84restriction, 11, 13, 58-64, 66, 67, 68, 75RNaseA, 12role in protein synthesis, 25-27somatic mutation stud ies on, 73

Epidemiologic studies, 9-1o, 14, 18, 49, 103-11oEpstein-Barr viru s, 104

Escherichia coli, 58, 60Ethylene d ibrom ide (EDB), 115, 117Ethylene oxide (EtO), 115, 117Ethylnitrosourea (ENU), in qu antitative

extrapolation studies, 97, 98-99Eukaryotes, 92, 94, 95, 116Exposure r egistries, 106, 107Extrapolation

aims of, 91-92from animal to human d ata, 86-87

defined, 14-15limits of, 99-100qualitative, 15, 91, 94-97, 99-100quantitative, 15, 91, 97-100, 115, 117theoretical models for, 15, 93-94

Federal Food, Drug, and Cosmetics Act, 114Federal Insecticide, Fun gicide, a nd Rodenticide

Act, 115Federa l Rad iation Council (FRC), 113, 114Fetuses, chromosome abnormalities in, 40, 42, 82Fluorochrome, 77Food and Drug Adm inistration (FDA), 19, 114Fruit flies, see Drosop hilaFungi, 92

Galactosemia, 105Gamma radiation, 97Gel electrophoresis, 11-12, 13, 55, 56, 58, 59, 65

see also Agarose gel electrophoresis; One-dimensional denaturing gradient gelelectroph oresis; Pulsed field gelelectrophoresis; Two-dimensional denaturing



Index • 141

gradient gel electrophoresis; Two-dimensionalpolyacrylamide g el electrophoresis

Genesdominant, 27, 44, 51globin, 29, 46, 47-48, 55hprt, 55, 74-76

mutations in, 5, 6-7, 29, 38-40, 42-43, 44, 47-48,50, 51, 55, 57-68, 74-76, 92, 95recessive, 27, 44

Genomes and genomic sequencing, 6, 11, 57-58,66-67, 68

Germ cells, 6, 7, 9, 13, 27, 30, 31, 37, 38, 39, 48,49, 80-84, 91-100, 108, 115-116, 117; see alsoGerminal mu tations

Germinal mutationsin cells, 7

defined, 6electrophoretic studies on, 51new, 37-38rad iation-ind uced , 9-10, 48-49

relationship to somatic mutations, 73see also MutationsGlossary of Acronyms and Terms, vii-xiiGlutamic acid, 26, 27, 29Glutamine, 27Glycine, 27Glycophorin A, 76-77Glycoph orin assay, in som atic mu tation stud ies,

76-77Guanine (G), 23, 24, 27, 64, 66“Guid elines for Mu tagenicity Risk Assessment, ”

116

Health effects

of chromosome abnormalities, 40-43of mutations, 4, 30-32, 108-109see also Children; Diseases; Hu man s; Infants

Heart disease, 3, 25, 32Hem oglobin A (HbA), 46Hemoglobin assays, in somatic mu tation studies,

77-78Hem oglobin C, 77Hemoglobin M (HbM), molecular analysis of, 44,

46-48, 51Hemoglobin S, 77Hemoglobins, unstable, molecular analysis of, 44,

46-48, 51Hemophilia, 3

Heritable mu tationschemicals and, 6-7, 9, 32, 48, 91-97

defined, 6detection and measu rement of, 11-13epidemiologic studies in, 9-10, 14, 18, 49,

103-110

estimating effects in hum ans from animal d ata,86-87

extrapolation of data to predict risk in humans,91-100

factors that influence ind uced rates in m ice,84-86

health effects of, 30-32, 108-109indu ced by chemicals, 6-7kinds and rates of, in human s, 32-33, 37-52

laboratory studies on, 81-87new technologies and, 3, 10-14, 55-69rodent tests for, 81-84spontaneous, studies in hu mans, 7-IO

see also MutationsHeritable translocation tests (HTTs), 82, 94-96, 98Heteroduplexes, 12, 55, 61, 62, 63, 64-65, 67Heterozygotes, 44, 45, 77, 82Histidine, 27, 47Histidinem ia, 105Hodgkin’s disease, 108

Hom ocystinu ria, 105Homoduplexes, 61, 63, 67Homologous sequences, 64Homozygotes, 44, 82hprt, 55, 74-76 Humans

animal stud ies and, 84-87epidemiologic studies in, 9-10, 14, 18, 49,

103-110Federal expenditures on mutation research in,

121-122frequencies of hprt – T-lymphocytes in, 76health effects of mutations in, 30-32, 108-109kinds and effects of mutations in, 4, 5-6, 29-32

kinds a nd rates of heritable mutations in, 32-33,37-52mu tagenic risk pr edictions by extrap olation,

91-100mutation studies in, 3-4, 7-14, 19, 44-52, 55-69,

73-78new technologies for detecting m utations in,

10-14, 55-69, 73-78persistence of new mutations in, 32-33, 37sentinel phenotypes in 7-8, 37, 38-40, 45, 50, 51,

52, 55, 105, 106, 109spontan eous m utation rates in, 32-33, 51see also Children; Diseases; Infants

Hu ntington d isease, 31

Hypoxanthineguanine phosphoribosyl transferase(HPRT), 74

Infants, genetic diseases in, 7-8, 30-32, 37, 38-40,42, 44-45, 49, 50, 109

Insects, 91, 92, 116




International Comm ission for Protection AgainstEnvironmental Mutagens and Carcinogens(ICPEMC), 55

International Comm ission on RadiologicalProtection (ICRP), 113, 114

International dioxin registry, 106Isoleucine, 27

Karyotypes, 28, 41, 42Kinetic studies, of enzyme deficiency variants,

45-46

Laboratory stud ies, of heritable mu tation rates,81-87; see also Animals

Lambda (virus), 59, 60Leucine, 27Leukemia, 9, 32, 109Lysine, 26, 27

Maple syru p u rine disease, 105Marfan’s syndrome, 31, 109Measuremen t techniques, validation of, 104Messenger RNA (mRNA), 25, 26Methemoglobinemia, 47Methemoglobinemic cyanosis, 46Methionine, 27Methyl isocyanate, 106Mice

effects of radiation on, 82-86, 97in extrapolation tests, 93, 96factors affecting mutations in, 84-86heritable mutation tests in, 82-84mutation effects in females, 85-86mutation effects in males, 84-85spermatogonial studies in, 84-85, 97-98unscheduled DN A synthesis in sperm atids of,

97-98Mine Safety and Health Ad ministration (MSHA),

114Monitor ing, in ep idem iologic testing, 14, 103-104,

106, 107Mortality and morbidity, 31, 41, 49, 82Mouse spot test, 96Mutagenicity assays, 91-1oo; see also Mutations

and specific tests for mut agenicit yMutants, in somatic mutation studies, 74-78; see

also MutationsMutations

animal studies, 3-5, 6-7, 19, 48, 78, 81-87,117-118at birth, 41in chromosomes, 5, 6, 7, 8, 9, 13, 29-30, 31, 32,

40-43, 49, 51, 68, 82, 92at conception, 41

current methods for studying, 6-10, 37-52dominant, 32, 38, 44environmental factors in, 3, 4, 37, 73, 96, 106,

108-109

Federal expenditures on research, 15, 18, 117,121-122

in genes, 5, 6-7, 29, 38-40, 42-43, 44, 47-48, 50,

51, 55, 57-68, 74-76, 92, 95genetic inheritance and, 23-33germinal, see Germinal mu tationshealth effects of 4, 30-32, 108-109heritable, see Heritable mu tationshuman studies, 3-4, 7-14, 19, 44-52, 55-69,

73-78, 103-110kinds of, and effects in humans, 4, 5-6, 29-32,

55

new, persistence of in humans, 32-33, 37, 41new technologies for d etecting in hu man s, 10-19,

55-69options for research on, 16-19in proteins, 3, 8-9, 25-27, 29, 44-48, 49, 56-57recessive, 32, 33regulatory considerations, 15, 17, 113-118somatic, see Somatic mutationsspontaneous, 5, 7, 32-33, 51, 83test methods for research in, 14-16X-linked (sex-linked), 32, 37, 38

National Cancer Institute, 104National Center for Toxicological Research, 19National Coun cil on Radiation Protection an d

Measurements (NCRP), 113, 114National Environmental Protection Act of 1970,

114

National Institutes of H ealth (NIH), 122National Research Council (NRC), 32-33National Toxicology Program (NTP), 19Neu roblastoma, 109Nuclear Regulatory Commission (NRC), 106, 108,

114Nucleotides and nucleotide sequencing, 5-6, 8, 11,

12, 23-25, 29, 44, 46, 47, 57-58, 59, 61, 66,68, 93

Occupational Safety and Health Act, 114, 115Occup ational Safety and H ealth Adm inistration,

114, 115

Office of Technology Assessment (OTA), 4, 15,121

One-dimensional denaturing gradient gelelectrophoresis, 55, 59-61, 63

Oocytes, in mice, 85-86Osteogenesis imperfect, 30



Index . 143

Parallelogram, as extrapolation model, 93, 97, 117Peripheral blood lymphocytes (PBLs), 97PFGE, see Pulsed field gel electrophoresisPhenotypic data

limitations of, 40, 45mu tation rates estimated from, 39-40sur veillance of, 38-40

see also Sentinel phenotypesPhenylalanine, 26, 27Phenylketonuria (PKU), 45, 105Polyacrylamide, use in protein separation, 5.5,

56-57Polymorphisms, 58-59, 64Polypeptide chain, 26, 37, 46; see also ProteinsProkaryotes, 94, 95Proline, 27

Proteinsbiochemical analysis of, 37, 44-48, 49-50, 51, 52DNA and, 23-27, 45, 46, 50, 57mutations in, 3, 8-9, 25-27, 29, 44-48, 49, 56-57one-dimensional separation of, 8-9, 44-45, 49, 56

synth esis of, 25-27variant, mutation studies on, 8-9, 44-50, 55,

56-57, 73-74two-dimensional separation of, 9, 56-57see also Amino acids; Blood proteins; Enzymes;

Hemoglobins, Nucleotides and nucleotidesequencing

Public Health Service, 18Pulsed field gel electrop horesis (PFGE), 13, 55, 68

Qualitative studies, 15, 91, 94-97, 99-100Quantitative studies, 15, 45-46, 55, 91, 97-100,

115, 117

Radiationeffects on atomic bomb survivors, 9-10, 32,

48-50, 77, 86, 106, 108effects on DNA, 23effects on fruit flies, 82effects on hum an germ cells, 9effects on mice, 82-86, 97effects on p regnan t wom en, 9effects on U.S. citizens, 86-87, 106in extrapolation studies, 97-1oo

in Marshall Island s pop ulations test, 50in somatic mutation stud ies, 74mutagenic effects of, 73U.S. Government efforts to protect humans

from, 113-114Radiation Effects Research Foundation, 16, 48, 104Radioactive thym idine, 74Red blood cells, somatic mutation studies on, 73,

76-78

Regulatory considerationsdefinition of germ-cell mutagen , 115-117mutagenicity and carcinogenicity, 117-118

n ew meth o d s fo r measu r in g mu ta t io n r a tes , 1 1 8

protection against agents o ther than radiat ion ,

11 4

q u a n t i t a t i v e e x t r a p o l a t i o n , 1 1 7

rad ia t io n p ro tec t io n , 1 1 3 -1 1 4role of U.S. Government, 15, 17, 113-118

Research and pu blic policy, use of newtechnologies in, 14-16

Restriction fragment length polymorphisms(RFLPs), 55, 58-59, 60

Retinoblastoma, 7, 31, 109Ribonuclease A (RNaseA), 64, 65Ribonuclease cleavage of m ismatches in

RNA/ DNA h eterodu plexes, 12, 55, 64-65Ribonucleic acid (RNA)

heter odup lexes, 55, 64-65role in protein synthesis, 26, 45

Risk assessm ent, EPA guid elines for, 116

RNA, transcribed, 45RNA/ DNA heteroduplexes, ribonuclease cleavage

of mismatches in, 12, 55, 64-65Rodents, heritable mutation tests in, 81-84

Salmonella, 118Sentinel phenotypes, 7-8, 37, 38-40, 45, 50, 51, 52,

55, 105, 106, 109; see also Phenotypic dataSerine, 27Sickle-cell anemia, 105Sister-chrom atid exchange (SCE), 99, 1166-TG, use in somatic mutation studies, 74-75Somatic cel]s, 7, 13-14, 27, 31-32, 73, 91-100, 115,

116, 117; see also Somatic mutations

Somatic mutationsdefined, 6detection and measu rement of, 4, 13-14effects of, 31-32new m ethods for measuring in hu man s, 73-78see also Mutations

Specific locus tests (SLTs), 50, 82-83, 85, 94-96,97-98, 116

Spermatogonial studies in mice, 84-85, 97-98Spontaneous mutation rates, 5, 7, 32-33, 51, 83Squirrel monkeys, 85Stem cells, in mice, 84, 85Subtractive hybridization, 12, 55, 66-67“Superfund,” 18, 114

Sur veillance, in ep idem iologic testing, 14, 103,104-106, 107

T-lymp hocytes, 74-76Technologies, new

technologies for detecting heritable mutations in human beings

Documents