J Clin Pathol 1991;44:144-146

Immunohistochemical analysis of amylaseisoenzymes in thyroid cancer

M Higashiyama, S Doi, N Tomita, T Monden, M Murotani, Y Kawasaki, T Kobayashi,T Shimano, M Ogawa, S Takai, T Mori

Second Department ofSurgery, OsakaUniversity School ofMedicine, Osaka,JapanM HigashiyamaS DoiN TomitaT MondenM MurotaniY KawasakiT KobayashiT ShimanoM OgawaS TakaiT MoriCorrespondence to:M Higashiyama, Departmentof Surgery, The Center forAdult Diseases, Osaka.Address: Nakamichi 1-3-3,Higashinari-ku, Osaka,Japan, 537Accepted for publication13 September 1990

AbstractThe expression of amylase in vhistological types of thyroid canctstudied by an immunohistochctechnique, using a polyclonalamylase antiserum and two monoantibodies specific for salivarypancreatic-type amylases, respecAmylase was expressed in 21 of 24thyroid cancers by polyclonal anti!analysis. Analysis by monoclonalbodies, however, showed that oi(54%) cases and three (13%) case.tained salivary-type and pancreatiiamylases, respectively. Moreovermunoreactivity for pancreatiamylase was detected only in medcarcinoma; other histological type!positive for salivary-type amylase.results show that thyroid cance:quently expresses amylase, and sithat the differences between anisoenzymes in thyroid cancer marelate with those found between ccorigin of tumour.

Amylase, which hydrolyses 1,4 glubonds of starch, is expressed by severaof malignant cells and tissues, incgastric,' gynaecological,25 lung,68 andmary cancer,9 and multiple myeloma.'0shown recently that thyroid cance

*.4.-RLW

....S

Figure 1 Normal human pancreas and salivary gland showing immunoreactivitmonoclonal antibodies 88E8 and 3B7. (A) Pancreatic tissues showing no reactiv88E8. (B) Pancreatic tissues showing positive reactivity with 3B7. (C) Salivartissues showing no reactivity with 3B7. (D) Salivary tissues showing positive recwith 88E8.

ariouser wasemicalanti-

oclonal{ andtively.(88%)serumanti-

nly 13s con-c-typer, im-c-typelullarys wereTheser fre-

expresses amylase." 12 Although most of thesetumours express the same amylase isoenzymeas that produced in the salivary gland(salivary-type), some tumours express isoen-zymes of pancreatic origin (pancreatic-type)or unknown origin.The normal thyroid gland and various

histological types of thyroid adenoma havealso been shown by biochemical analyses toexpress salivary-type amylase and severalunknown types, but pancreatic-type amylasehas not been detected in these tissues.' '2 It isnot yet clear, however, what type of amylaseisoenzyme is expressed by thyroid cancer. Wetherefore analysed amylase isoenzymes invarious thyroid cancers, using an immuno-histochemical method with three differentanti-amylase antibodies.

uggest Methodsnylase Tissue sections were prepared from formaliny cor- fixed, paraffin wax blocks of 24 specimens ofellular thyroid cancer 14 well differentiated

papillary or follicular carcinomas, two poorlydifferentiated carcinomas, three anaplasticcarcinomas and five medullary carcinomas

cosidic according to Rosai's classification and that ofl types the WHO.'3 14

Aluding For the immunohistochemical analysis ofmam- amylase isoenzymes, we used three types ofIt was anti-amylase antibodies. A rabbit polyclonalr also antiserum, which reacts to both salivary- and

pancreatic-type amylase, was produced asdescribed previously.'" Monoclonal antibody,88E8, specific for salivary-type amylase,was supplied by Boehringer MannheimYamanouchi Co Ltd, Tokyo, Japan.'6 Mono-clonal antibody, 3B7, specific for pancreatic-type amylase was kindly provided by KohjinCo Ltd, Ohita, Japan. '7 We confirmed thespecificity of each monoclonal antibody for

4 isoenzymes by immunostaining sections*t,l. prepared from normal human pancreas and

salivary gland (fig 1).Sections were stained with a biotin-

streptavidin immunoperoxidase technique(BioGenex Laboratories Biotin-Streptavidin

p Immunostaining Kit; Dublin, California).Briefly, after endogenous peroxidase wasblocked by immersion in 0-3% hydrogenperoxide, sections were exposed to 3% normalgoat serum, followed by incubation with

ty with primary antibody (rabbit polyclonal x 500;lity with 88E8 x 10; 3B7 x 10) for one hour at rooma temperature. Sections were then incubated

with a biotinylated goat anti-rabbit IgG or a

144

AWA

i, :.

:..

..

-, A..t..

.1, .:

.1

I"Z4, U)

on February 1, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.2.144 on 1 February 1991. D

ownloaded from

Immunohistochemical analysis of amylase isoenzymes in thyroid cancer

Immunoreactivity of anti-amylase antibodies

Immunoreactivity

88E8 3B7Thwroid cancer histological type Polyclonal (salivary type) (pancreatic type)

Well differentiated papillary or follicular carcinoma 14/14 12/14 0/14Poorly differentiated carcinoma 2/2 1/2 0/2Anaplastic carcinoma 1/3 0/3 0/3Medullary carcinoma 4/5 0/5 3/5

biotinylated goat anti-mouse IgG, followed bytreatment with a streptavidin-horseradishperoxidase conjugate. Peroxidase activity wasmade visible using the diaminobenzidine/H202 reaction. Sections were counterstainedwith haematoxylin.

ResultsResults are shown in the table. Twenty one(88%) cases of thyroid cancer expressedvarying degrees of immunoreactivity for

4

< -v .,. 4; 4. ...* , S.W/.t.,.gf-7.

~a showing immunoreactivity for amylase. Positiveiwith polyclonal antiserum (A) and 88E8 (B), but nowith 3B7 (C).

~~~~~~~~~~~~J 4

74

~~~~_" ...R."A W -;-F

skiss*

s£'s.~~~~

(

Figure 3 Medullary carcinoma showing immunoreactivity for amylase. Positiveimmunostainings were obtained with polyclonal antiserum (A) and 3B7 (C), but no

immunostaining was observed with 88E8 (B).

polyclonal anti-amylase antiserum. Immuno-reactive intensity for amylase was strong insome cases of well differentiated carcinomas,while expression was not detectable in twoanaplastic carcinomas and one medullary car-cinoma. According to immunohistochemicalanalyses using monoclonal antibodies, 13(54%) cases were positive for salivary-typeand three (13%) for pancreatic-type amylase.Only medullary carcinomas showed immuno-reactivity for pancreatic-type amylase whilethe other histological types were positive forsalivary-type amylase (figs 2 and 3).

DiscussionIt has been shown that papillary carcinoma ofthe thyroid gland expressed various degrees ofamylase activity." 12 In this study we showedthat other histological types also frequentlyproduced amylase. Two anaplastic and onemedullary carcinoma, however, contained noamylase, whereas almost all well differentiatedpapillary or follicular carcinomas expressedamylase, some showing particularly strongimmunoreactive intensity. It is suspected thatamylase expression closely correlates with thedegree of tumour differentiation.

Discrepancies between the degree ofamylaseexpression by each antibody, such as 14/14(100%) by polyclonal antibody compared with12/14 (86%) by 88E8 monoclonal antibody inwell differentiated carcinoma, are explained bythe following: monoclonal antibody specific forsalivary-type amylase may possess lowersensitivity than polyclonal antiserum whichdetects both types. In fact, all cases judgedpositive for salivary-type amylase by poly-clonal antibody but negative by monoclonalantibody were faintly stained. The same isequally true of discrepancies in medullarycarcinoma: the sensitivity of monoclonalantibody specific for pancreatic-type amylasemay also be lower than that of polyclonalantibody.A few malignant cells in some cases of

papillary or follicular carcinoma also showedvery weak staining with the antibody 3B7 (datanot shown). This may be associated with thespecificity or the sensitivity of the antibody.The numbers of these cells were also too smalland the intensity too faint to draw definitiveconclusions about their true pancreatic-typeamylase positivity. The clinical importance ofthese observations in the cases of papillary orfollicular carcinoma is unclear.Although not all cases of staining in papillary

or follicular carcinoma were clear cut, theproperties of isoenzymes in thyroid cancer are

-

Figure 2 Papillary carcinomimmunostainings were obtaineeimmunostaining was observed z

145

on February 1, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.2.144 on 1 February 1991. D

ownloaded from

Higashiyama, Doi, Tomita, Monden, Murotani, Kawasaki, Kobayashi, Shimano, Ogawa, Takai, Mori

of special interest. Only medullary carcinomashowed pancreatic-type immunoreactivity;other types of thyroid cancer showed salivary-type amylase. Medullary carcinoma is nowconsidered to originate from C-cells, whileother types are of follicular cell origin.18 Accor-dingly, the type ofisoenzyme seems to correlatewith the cells of origin.

Recently, Tomita et al, using molecularbiological analysis, detected a novel type ofhuman amylase in a lung carcinoid, which wasencoded by the amy3 gene.'9 Interestingly,both medullary carcinoma of the thyroid glandand lung carcinoid tumour are regardedas neuroendocrine tumours histologically.Medullary carcinoma may therefore expressthis new amylase rather than pancreatic-typeamylase. Unfortunately, immunohisto-chemical analysis of the expression of amylaseisoenzymes in lung carcinoid, using thesespecific monoclonal antibodies, has not beendone. The monoclonal antibody 3B7 againstpancreatic-type amylase cross-reacted withthis novel amylase isoenzyme while the mono-clonal antibody 88E8 against salivary-typeamylase did not.'920 An antibody to differen-tiate the novel amylase isoenzyme encoded bythe amy3 gene from pancreatic-type amylasehas not been made as yet because the two are sosimilar. '920 On the other hand, the two differenttypes of amylase-encoding genes, AMY2A andAMY2B, were expressed in normal human

212pancreas. 22 While the former encodes"original" pancreatic-type amylase, the latteris identical with the amy3 gene.23 Accordingly,although histological localisation of this noveltype of amylase isoenzyme has not been pos-sible, it may be distributed not only in neuro-endocrine tumours but also in normal tissues,including neuroendocrine cells or epithelialcells. In fact, Samuelson et al providedevidence that normal liver tissues also containmRNA of the amy3 gene.24 The distribution ofthe amy3 gene product needs to be elucidated.

In conclusion, we have shown that mostthyroid cancers produced amylase. Moreover,medullary carcinomas were immunoreactivefor pancreatic-type amylase whereas otherhistological types were immunoreactive forsalivary-type amylase. To elucidate the type ofamylase isoenzyme in thyroid cancer molecularbiological analyses are needed in addition tobiochemical or immunohistochemical studies,or both. More discriminatory and sensitive

antibodies to amylase isoenzymes, includingthe amy3 gene product, must also be found.We thank Drs K Kataoka and K Ito for supplying the antibody3B7, and for their advice.

1 Nomura H, Tokumitsu S, Takeuchi T. Ultrastructure,cytochemical, and biochemical characterisation of alpha-amylase produced by human gastric cancer cells in vitro.JNCI 1980;64:1015-24.

2 Bruns DE, Mills SE, Savory J. Amylase in fallopian tubeand serous ovarian neoplasms. Immunohistochemicallocalisation. Arch Pathol Lab Med 1982;106:17-20.

3 Ueda G, Yamasaki M, InoueM, et al. Immunohistochemicalstudy of amylase in common epithelial tumors of theovary. Int J Gynecol Pathol 1985;4:240-4.

4 Ueda G, Yamasaki M, Inoue M, et al. Immunohistochemicaldemonstration of amylase in endometrial carcinomas.Int J Gynecol Pathol 1986;5:47-51.

5 Lee YS, Raju GC. The expression and localisation ofamylase in normal and malignant glands of theendometrium and endocervix. J Pathol 1988;155:201-5.

6 Yokoyama M, Natsuizaki T, Ishii Y, et al. Amylase-produc-ing lung cancer. Ultrastructural and biochemical studies.Cancer 1977;40:766-72.

7 Tomita N, Matsuura N, Horii A, et al. Expression of-amylase in human cancers. Cancer Res 1988;48:3292-6.

8 Nakamura Y, Tomita N, Nishide T, et al. Production ofsalivary type -amylase in human lung cancer. Gene1989;77:107-12.

9 Weitzel JN, Pooler PA, Mohammed R, et al. A unique case ofbreast carcinoma producing pancreatic-type isoamylase.Gastroenterology 1988;94:519-20.

10 Hata H, Matsuzaki H, Tanaka K, et al. Ectopic production ofsalivary-type amylase by a IgA- -type multiple myeloma.Cancer 1988;62:151 1-5.

11 Otsu N, Oami H, Okawa M, et al. Amylase in the thyroidgland. Histochemistry 1982;74:21-6.

12 Mizukami Y, Matsubara F, Hashimoto T, et al. Amylase inthyroid tissues from normal and various thyroid diseases.Biochemical and immunohistochemical study. ActaPathol Jpn 1987;37:253-60.

13 Rosai J, Saxen EA, Woolner L. Undifferentiated and poorlydifferentiated carcinoma. Semin Diagnostic Pathol1985;2:123-36.

14 Hedinger CE, Williams ED, Sobin LH. Histologicaltyping ofthyroid tumours. 2nd ed WHO Berlin: Springer, 1988.

15 Takatsuka Y, Kitahara T, Matsuura K, et al. Radio-immunoassay for human pancreatic amylase: comparisonof human serum amylase by measurement of enzymaticactivity and by radioimmunoassay. Clin Chim Acta 1979;97:261-8.

16 Gerber M, Naujoks H, Lenz H, et al. A Monoclonalantibody that specifically inhibits human salivarya-amylase. Clin Chem 1987;33:1158-62.

17 Minamiura N, Ito K. Discriminative assay method ofhumansalivary and pancreatic -amylase using their specificmonoclonal antibodies. Journal Japan Society StarchScience 1988;35:159-63.

18 Williams ED. Histogenesis of medullary carcinoma of thethyroid. J Clin Pathol 1966;19:114-8.

19 Tomita N, Horii A, Doi S, et al. A novel type humana-amylase expressed in lung carcinoid tumor. Gene1989;76: 11-8.

20 Shiosaki K, Takata K, Omichi K, et al. Identification of anovel -amylase by expression of a newly cloned humanamy3 cDNA in yeast. Gene 1990;89:253-8.

21 Groot PC, Bleeker MJ, Pronk JC, et al. Human pancreaticamylase is encoded by two different genes. Nucleic AcidsRes 1988;15:4724.

22 Groot PC, Bleeker MJ, Pronk JC, et al. The human -amylasemultigene family consists of haplotypes with variablenumbers of genes. Genomics 1989;5:29-42.

23 Groot PC, Mager WH, Frants RR, et al. The humanamylase-encoding genes amy2 and amy3 are identical toAMY2A and AMY2B. Gene 1989;85:567-8.

24 Samuelson LC, Wiebauer K, CaldwellRM, et al. Expressionof human amylase genes: recent origin of a salivaryamylase promoter from an actin pseudogene. NucleicAcids Res 1988;16:8261-76.

146

on February 1, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.2.144 on 1 February 1991. D

ownloaded from