Evaluation of Marrow Granulocytic Reservesin Normal and Disease States

By CHARLES G. CRADDOCK, JR., SE\IouR PERRY, Lufz E. VENTZKE ANDJOHN S. LAWRENCE

\Vith the technical assist(znce of ?JIari/ H. Baker and Gloria Pa iii

T H I S RE1ORT 1)rese1ts a summarization of recent findings with respectto the kinetics of granulocyte physiology, and an attempt to interpret

changes in granulocyte levels in light of these concepts.The principles upon which the granulocyte data reported are based are as

follows:

1. In the normal steady state granulocytopoiesis is confined to the marrow. There isgeneral agreement that cell division in the granulocytic series does not occur after themyelocyte stage.

2. In the normal liuiiiaii in the steady state the granulocyte remains in the marrow fora I)eriOd of several days, probal)hy five to six days, after the last division of granulocyteprecursors. After maturation occurs, the majority of granulocytes are released into thel)lOOd in an orderly fashion. This maSS of granulocytic cells ( which we have called themarrow granulocyte reserve, or NIGR, and which is incapable of proliferation ) comprises1l)Out five-sixths of the marrow granulocytic cell population. It is evident that this intra-nedullary mass of maturing cells is iuaiiy tines larger than the circulating mass ofgranulocytcs. This was shovn in leukopheresis experiments in normal and heavilyirradiated dogs.3

Yoffey estimates th ratio of uiarrow to blood granulocytes at about one hiuiulred toOne in the guinea pig. Recently, Donohue et al.5 have estimated a ratio of 25 to 30 to 1in the human, em)loying an ingenious method for quantitating the marrow crythroidelements with Fe. Osgood arrive(l at a comparable relationship on the basis of hiscalculations.

3. The rise in circulating granulocyte l4ulnber in rcspone to such stimuli a leuko-pheresis or the intravenous injection of bacterial endotoxin is due to the accelerated entryof granulocytes into the blood from the marrow granulocyte reserve.:i,T This was establishedby labeling granulocyte desoxyribose nucleic acid (l)NA) and taking advantage of thetime interval when the marrow granulocyte population was highly labeled relative to thecirculating cells. Induction of a demand for increased granulocytes in the periphery at atime when the MGR is completely labeled will cause an acceleration in the rise incirculating granulocyte l)NAPa2 specific activity. Accordingly, the majority of cells con-tributing to the peripheral demand came from the highly labeled marrow population andnot from the unlabeled cells in various peripheral organs.

4. In the normal steady state, granulocytes which enter the peripheral blood from the

From the Department of Medicine, School of Medicine, Unicersity of California atLos Angck?s, and the Hematology Research Laboratory, Wadsworth Hospital, VeteransAdministration Medical Center, Los Angeles, Calif.

Work performed tinder contract #H-1069, U.S. Public Health Service, and partly sssp-ported by a grant from Parke, Davis & Go. and the Gladys Bow yer Foundation.

Presented in part at the Annual Meeting of the National Blood Club, Atlantic City,May 4, 1958.

Submitted July 28, 1958; accepted for publication July 11, 1959.

840

EVALUATION OF MARROW GRANULOCYT1C RESERVES 841

marrow may circulate for a variable time. The average time granulocytes spend in thecirculation is about 10 hours, as will be (hiscusse(l subsequently. The peripheral circuha-tion of granulocytes may l)e divide(l into two main compartments: ( a ) those freely cir-ctilating in the larger vessels, and ( I) ) those sequestered or trapped in the capillarynetworks of various organs. The lung seems to play a role of major importance in thistrapping of granulocytes.5 There is an interchange of cells between these two main

peripheral compartnents. The work of Ambrus an(1 Ambrustt has demonstrated the re-lationship of circulating granulocyte number to removal or release of cells by the lung.On the basis of earlier experiments with leukopheresis in dogs, it was estimated that thenumber of granulocytes harbored in the capillary beds throughout the body of a normaldog was slightly less than the number freely circulating.1 Failure to take cognizance ofthe interchange between the freely circulating compartment and cells in capillary bedshi:is led to the belief that granulocytes leave and re-enter the blood stream manytinies61#{176}11 (luring their total life span.

6. Granulocytes which have moved across the endothelial barrier between blood andtissues (10 not re-enter the circulation, at least not in numbers that approach thoseleaving the blood stream. No evidence has been presented showing a free interchangeof granulocytes between blood and tissues. The evidence against such an interchange isconsiderable: ( a ) Leukopheresis experiments in irradiated dogs. These failed to show re-entry of granulocytes which had migrated into areas of infection.1 ( b ) Experiments suchas that shown in figure 1 . Here the preponderance of I)NA-P2 labeled cells had left theblood stream prior to the time of stimulation, and these cells failed to re-enter the bloodto cause a rise in granulocyte DNA_pie. ( c ) Experiments such as those in figure 2. Herethe granulocytes were pulled into a peritoneal exudate and not removed from the body.The DNAP:2 specific activity of the granulocytes in the peripheral blood does notcorrespond with that in the peritoneal exudate, indicating a lack of rapid interchangebetween the granulocytes in these two compartments.

Findings such as this do not preclude the re-entry of some granuhocytes froiii certainextravascular areas. They do indicate, however, that the primary direction of granulocytemovement is from the marrow into the circulating blood, into capillary beds, and theninto the tissues. This would not seem incompatible with the known functions of thesecells. The work of Hollingsworth et th. I C indicates the importance of granulocytes seques-ttred in capillary beds in regard to bacterial phagocytosis and that significant numbersmay be in capillary l)e(ls (Ie5l)it(. circulating agranulocytosis.

The fact that the granulocytes circulating in the blood do not freely exchange withthose outside means that the circulating cells do iiot represent the average age of allgranuhocytes in the periphery. No data exist as to the size of the extravascular pool ofgranulocytes or as to the time which the granulocyte survives in the tissues. ( The termtissues includes all areas, secretions, body cavities, etcetera where granulocytes maytoter.) It is reasonable to believe this time period will vary enormously depending uponthe environmental conditions at the site at which the granulocyte moves across thecapillary wall. It is illogical to assume that the granulocytes in bronchial secretions are ofthe same average age as those just entering the blood from the marrow. Therefore, thedisappearance rate of DNA-labeled cells is not a reflection of the time granulocytes stir-vive in the blood and tissues.

Kline and Clifton,13 and subsequently several other investigators,1014 have interpretedthe disapperance rate of labeled granulocytes from the blood as indicating a life spanof the granulocyte of 13 to 20 days. Similar figures have been obtained15 with Cr and5:15 labeling.16 Ottesen1#{176} concluded on the basis of his experiments that the average ageof the circulating granulocyte WaS 9 days, implying a much longer total life span. Allof these experimenters assumed that the circulating granulocytes were in equilibriumwith those in the tissues, an(l all ignored the very large intrameduhlary mass of maturinggranulocytes which was also lal)ehed by these technics.

An additional observation which indicates the error in interpreting granulocyte DNAlabeling in terms of survival time of granulocytes is that the disappearance rate of DNA-

842 CRADDOCK, JR., PERRY, VEN12KE AND LAWRENCE

udate.

.0-so

1

DOG 4-90DATE 2-12-56 to 2-22-56WT. 53 lbs

101500)

DOG # 5- III

MARROW BLOOD PERITONEALCAVITY

A 500 ml. SterilePMN DNA P Nutrient BrothBlood 1.5 c/rn/vWBC 12 350 1mm3

(0.186 xIO#{176}Circ.)

Fig. 1.-In dogs rendered leukopenic by ir-radiation, the adminis.tration of typhoid aftermost of the P32-labeledleukocytes have left theperipheral blood doesnot result in an in-crease in leukocyteDNA-P32 activity.

AT 6 HOURS

A

PMN DNA P (O219x10#{176}Total)Blood 12.9 c/rn/v Exudate 3.0 c/m/xWBC= 7,400/mm3

AT 7 HOURS

BC 31,000 1mm5______________ (O.124x10 Total)1BIood 4 c/rn/a Exudate 4.4 c/mlr

WBC 7,900/mm3

Fig. 2.-Creation of peritone-al exudate with removal ofgranulocytes into the exudate.This was done at a time whenthe DNA.labeled granulocyteswere first entering the bloodfrom the marrow. Note that theblood granulocyte specific ac-tivity rises much faster thanthat in the peritoneal cavity.Also note the fall in bloodleukocyte number despite themovement of cells into the ex

EVALUATION OF MARROW GRANULOCYTIC RESERVES 843

labeled cells is not greatly accelerated in conditions known to be associated with increasedgranulocyte destruction. In patients with pyogenic infections,17 in animals subjected toleukopheresis or peritonitis,18 the fall-off of DNA-labeled granulocytes is not markedlyaltered. The rise and fall of blood granulocyte DNA-P32 is primarily a reflection of theturnover of the MGR. The labeling of the MGR DNA with p32 follows a definite patternand depends upon the period of time which labeled precursors are available and uponthe dilution of the DNA label by cell division. The granulocytes which first enter theMGR after labeling are the most highly labeled, and as they move through this matur-ing pool they are succeeded by cells whose DNA is less highly labeled as a result of divisionof myelocytes which are in turn, less highly labeled. This process may he depicted dia-grammatically as shown in figure 3.

The size of the MGR is so much greater than that of the circulating mass of granulocytes( fig. 4. ) that acute fluctuations in the latter cause very little alteration in the turnover of theMGR. The turnover time of the MGR in the normal human is five to six days. Estimatingthe size of the MGR at about 140 x 1010 granulocytes in the 70 Kg. man, as others havedone,56 this would mean that sufficient numbers of cells are entering the peripheral bloodfrom the marrow ( 1 X 1O#{176}per hour) to replace those circulating every 6 to 12 hours. Sincethe evidence previously cited does not indicate recirculation of granulocytes, it appcars thatthe average time a granulocyte spends in the blood stream is of this order of magnitude.It should be re-emphasized that this is an average figure and that some granulocytes maybe held in capillary beds much longer and that others may go into the tissues soon afterentering the blood. This randomness of granulocyte removal from the blood agreeswith the data of Athens et al.19 using DFP32 labeling.

This concept of granulocytopoiesis envisions the blood granulocytes as thetop layer of a larger intramedullary mass of granulocytes which are pro-liferating and growing in an orderly and predictable manner, in the normalsteady state. The granulocytes which leave the marrow and enter the bloodand then the tissues are expendable cells, in the sense that their functionis to be completed shortly after this time. An acute demand for granulocytesis met by the release of cells from the maturing marrow granulocyte pool(MGR) at a rate exceeding the basal rate. Since the size of the MGR is somuch larger than the mass of granulocytes in the peripheral circulation, thelatter can be doubled or tripled in size by a slight acceleration of the releaseof cells from the MGR.

The nature of the stimulus to the marrow and the mechanisms involvedin the release of granulocytes are beyond the scope of this paper. It wouldseem, as stated in an earlier publication,7 that the fundamental stimulus in-volves the removal of cells. Leukopheresis, or the removal of white cells with-out the introduction of known toxic factors, seems to be a sufficient stimulus.The possible humoral mediation remains to be demonstrated convincingly.

With these concepts in mind, an attempt has been made to obtain a clearerview of blood granulocyte alterations. The acute leukocytosis observed afterintravenous injection of bacterial endotoxin will not occur unless the marrowcontains granulocytes (3). The material employed here, Pyrexal, is a lipopoly-saccharide from Salmonella abortus equi.#{176}

This material in doses of 0.1 to 0.2 gamma intravenously causes very littlein the way of systemic symptoms, except, rarely, slight transient chilliness.

We are indebted to Wander & Co. for generously supplying this agent, and especiallyto Dr. Fred Schultz for his cooperation.

00000 0)0

0 00000

00000 cxixo

0 cX000

P32

..

I

I V.

I

#{149}

#{149}

I

0)000(Q(JQ0000#{149}ocy3o0000y9

0900

- CONC

DNA P

PER CELLI OF

Fig. 3A (top),B (middle) andC hottom). -Schematic rep-resentation ofthe dilution ofthe DNA labelby myeloid pro-hiferation. Bythe time themarrow reserve

of nonproliferat-ing granulocytesis labeled, the

beled cells enterthe blood first.most highly la-

PROLIF. NON PROLIF

MYELOID MARROW

CONC

Ill OFDNA PPERCELL #{149}I.U.I

CONCOF

_ __.u_luII _____ ----IDNA PPER CELL - - -

S44 CRADDOCK, JR.. PERRY, VENKE AND LAWRENCE

The hematologic effects of Pvrexal are nonspecific and resemble those observedafter intravenous injection of a variety of bacterial toxins and foreign proteins.The material has been employed in Europe by Heilmeyer2#{176} and Eichenbergerand associates21 for the assessment of marrow function. The present studies

DAY 5 DAY 6.7

p

OxXG DAY 8.9oGO3GG___

-sc-pQGQQOGQGX%X3cxce _cx1XJ#{248}

-0

EVALUATION OF MARROW GRANULOCYTIC RESERVES $45

BLOOD PMN Fig. 4. - Schematic6#{174}_ DNA P32 representation of the- )#{128}X XC) type of DNA specific OG!X activity curve which #{248}GQ would be expected in:K DG#{128} the blood granulocytes G#{128} as a result of the DNA(3 _____________ QQx _____________ .

*COCC -C J label by proliferation- -5- -* -/ -> and release of cells

MARROW BLOOD TISSUES from a labeled marrowRESERVE mass of nonproliferat-

ing cells.

_ LDAY

extend the use of this agent to a variety of hematologic diseases and describecertain refinements in interpretation.

The following main points concerning the response to Pvrexal, many ofwhich have been observed by 1 are noteworthy.

In the low dosage used, there is, in the normal human, a slight fall in totalcount, which is inconstant, followed by a rise, which is invariable ( fig. 5 ) . Theearly fall, occurring immediately after injection, is due usually to lvrnphopenia( fig. 6 ) , although there is often a transient fall in granuloctes at the sametime. The leukocytosis, which reaches a maximum in 6 to 10 hours, involvesgranulocytes only. The lymphocytes usually remain low and there is an ab-solute eosinopenia. It is necessary to express the response to Pyrexal in termsof granulocyte numbers rather than total white cell count. Instances were en-countered in which changes in total \VBC count might be considered in-significant and yet the granulocyte increase was adequate, the total WBCbeing unchanged because of lymphopenia. This is apparent in figure 5., inwhich the maximum total WTBC in one instance was only 125 per cent and inmost instances between 140 per cent and 160 per cent. Calculating the per-centage increase in terms of granulocvtes yielded a maximum of 180 per centor more in all normals, the average increase being about 200 per cent.

The normal values for absolute number of circulating segmented neutrophilsare 3000 to 5800 per cu.mm. with an average of 4000 per cu.mm.22 With theuse of these values, the normal response to Pvrexal involves an increment ofat least 2400 segmented granulocytes per cu.mrn. The average normal responsewas an increase of 4000 granulocytes per cu.mrn. Another characterislic ofthe normal response is the absence of a left shift in the granulocytes, noincrease in the percentage of immature forms being observed with this type ofstimulation at the dosage employed. As in all values pertaining to circulatingleukocvte numbers, there is considerable latitude of normality. At present, we

Segmented Neutrophils

Lymphocytes

846 CRADDOCK, JR., PERRY, VENT2KE AND LAWRENCE

220 s-a WBCx---..x GRANULOCYTES200 0-.-o Lymphocytes

ISO

60

40 r::T:

0 I 2 3 4 5 6 7 232425 o I 2 3 4 5 6 7 8i4

TIME IN HOURS TIME IN HOURS

Fig. 5 (at left).-The change in total WBC count after Pyrexal (0.1 to 0.2 gammaintravenously) in eight normal persons. The maximum increase in segmentedneutrophils is 180 per cent or more in all instances.

Fig. 6 (at right).-The percentage change in circulating white cells in a typicalnormal subject after 0.1 gamma of Pyrexal. The rise in granulocytes involves anincrement of 6376 cells per cu.mm., or an absolute granulocyte count of 10,932per cu.mm. at the peak response.

belieVe the response to Pyrexal is not significant uiiless it is far below the re-

51)0h1 oI)served in normals to date.The data to be presented represent examples of how the Pyrexal test has

been employed. The results are of interest in terms of correlation with theover-all clinical picture and with the concepts discussed previously. They arenot to he interpreted as showing a predictable pattern in any given diseasesituation since this will require more data than are preseiited here.

EXAMPLES OF ThE PYREXAL TEST IN HEMATOLOGIC DISORDERS

A plastic AnemiaFigures 7a and h show the results in two cases of aplastic anemia. In the

first case (7a) the patient required infrequent transfusions and exhibited amild leukopenia and thrombocytopenia. The peripheral blood picture showeda modlerate left shift in the neutrophilic series and the marrow was markedlyhypocellular in all elements. The acute response of granulocytes to Pyrexal wasmarkedly depressed. The later slow rise by 24 hours has been noted in manyinstances of marrow depression and will he discussed subsequently.

The second case (7h) is an example of a patient with a comparable leuko-penia associated with a drug-induced pancvtopenia. At the time of thePvrexal test the patient was recovering, though still leukopenic, and the mar-row appeared normal. The response was qualitatively and quantitatively nor-mal in every respect. In contrast with the previous case, at no time was anyleft shift observed in the granulocytic series, a finding compatible with theinierpretation of adequate marrow reserves of granulocytes.

Lymphomas

The granulocyte response to Pyrexal has been entirely normal in 10 casesof Hodgkins disease with normocellular marrow and without marked splen-

293.-. WBC

X x GRANULOCYTES

o---oLYMPHOCYTES

Segmented Neutrophils

Bonded NeutrophIs

Metomyelocy ten

Lymphocytes

EVALUATION OF MARROW GRANULOCYTIC RESERVES 847

. -. wBc 0n-- 0 GRONULGCYTES 240

n- . -n LYMPHOCYIES 220 0-.---GRANULOCYTES160 ! n---oLYMPHOCYTES

. 200 14C 1 - Segmented N#{149}utophIs

ir .- ISO #{149}-2C .. ., ...-:1 Ft- l\ . Lymphocytes

10< ./ 160 !\\-. I . , . - S,g,eted Neut,optIs 140 :

Sc . _-,y_ ! Bonded Neutoph.Is I6C M,to,ny,iocytes 20 / S.

:L 6 24 1Ec(. TIME IN 101)05 40

20 I 2 4 5

T( IN HOURS

Fig. 7a (top left).-The acute leukocyte re-sponse involves an increment of only 150 gran-ulocytes per cu.mm., a markedly deficient re-spouse in a patient with known marrow de-

0) pression. Fig. 7b (top, right).-Leukocyte responsew to Pyrexal recovering from hone marrow aplasia. The increment in granulocytes was 3700 per cu.mm. Fig. 8 (bottom, left).-Lack of acute re-0 sponse to Pyrexal in a patient with Hodgkins

disease receiving x-ray and mustard therapy.The elevation of count at 24 hours is fre-quently seen in patients with marrow depres-sion and is not a normal response.

omegaly. Patients who have marrow depression as a result of therapy exhibita deficient response. Figure 8 is an example of this in a patient who had re-ceived a course of nitrogen mustard two weeks previously. The delayed in-crease by 24 hours is again noted.

Figure 9 shows the response in a patient with diffuse lymphosarcoma withmarrow depression, invasion of the marrow by lymphosarcoma, splenomegaly,mild anemia and thrombocytopenia. Although there was not a marked totalleukopenia, there was a relative lymphocyte increase and a definite left shiftin the granuloctic series. The granulocytes failed to increase after Pyrexal.Subsequently the patient received a total dose of 7.5 mg. of triethylene mela-mine in divided doses over a three week period. The tumor was markedlysensitive to the agent but at the same time the patient became severely agran-ulocytic and almost expired. He later entered a state of complete remission.

This case demonstrates that extreme depletion of reserves of granulocytescan occur without profound peripheral leukopenia. The morphology of themarrow, the lack of response to Pyrexal and the left shift in peripheralgranulocytes all indicated this. The remaining granulopoietic tissue could main-tain the granulocytes in the peripheral blood but there was no marrow reserve

wz-JL,JU)4

O t 2 3 6TIME IN HOURS

L4Fig. 9.-Leuko-

cyte response to Py-rexal. See text fordetails.

848 :BADDOCK, JR., PERRY, \EN-rLKE AND LAWRENCE

.-. WBC

x-----x GRANULOCYTES

o---o LYMPHOCYTES

- Segmented Neutrophils

= Banded Neutrophils

Metomyelocytes

Myelocytes= Lymphocytes

of granulocytes. Suppression of granulocyte production for even a brief timeresulted in severe agranulocytosis.

Leukemia

Figures 10, 11, 12 and 13 show responses to Pyrexal in selected cases ofacute and chronic leukemia. They demonstrate the necessity of calculatingthe absolute increase in granulocytes in determining the adequacy of marrowgranulocytic reserves. Thus, in figure 11, although the percentage rise ingranulocytes is not marked, the total increment in granulocytes exceeds thenormal. In figure 13 the graniilocte response appears fairly good on a percent-age basis, but involves only a very small increase in total number of granulo-cvtes. Figure 12 is of interest in that this patient with typical chronic lym-)hocytic leukemia shows a normal response to Pyrexal despite marked lym-phocytic infiltration of the marrow. The preservation of marrow granulocyticreserves and the ability to develop a granulocytic leukocytosis on demand maybe factors in the generally benign course of patients with chronic lymphaticleukemia.

Aberrant Responses

\\Then splenomegaly with hyperspienism exists, the interpretation of theresponse to Pyrexal becomes more difficult. The spleen may be removing leuko-cvtes from the circulation more rapidly than normal and may possibly exert adepressant effect on the marrow. An example of the effect of an enlargedspleen is shown in the study of a case of congestive splenomegaI associatedwith posthepatitic cirrhosis and portal hypertension, in which hepatic functiontests were normal. Ferrokinetic and Cral studies in this patient showed thespleen to be removing red cells from the circulation at an accelerated rate prior

to surgery. The patient underwent portocaval anastomosis because of esopha-geal varices, the spleen being left intact. The spleen rapidly decreased in size,and repeat isotope studies showed no further removal of red cells by thespleen. Figure 14 shows the response to Pyrexal before and after portocavalanastomosis. The time interval between the two studies was 10 days. Thequalitative similarity but quantitative difference of the granulocytic response

=TIME IN HOURS

Fig. 10. - Leuko-cyte response toPyrexal. The gran-ulocyte incrementwas 877 cells percu.mm., a very de-ficient response.

Fig. 11. - Leuko-cyte response toPyrexal. The incre-ment in segmentedgranulocytes was13800 per cu.mm.,indicating adequatemarrow reserves ofthese cells.

7\L;7\-..H

0 I 2 3 4 5 6 74

Iiz-Jus54

St

60

140w

120-Jw 100U)4m 80

60

40

20

3OC

28C

26C

24

22

20

Z 181

1614 4

0- 121

0

80

60

40

20

set

x

//

.-. WBC/ x x GRANULOCYTES

n---o LYMIHOCyTES

- Segmented Neutropho

Banded NeutrophIs= Lymphocytes

OL345678L4TIME IN HOURS

EVALUATION OF MARROw GRANULOCYTIC RESERVES 849

TIME IN HOURS

#{149}. WBC

X_.XGRANULOCYTES

o---o LYMPHOCYTES

Segmented Neutrophils

Banded NeutrophilsMetomyelocytesMyelocytes

PromyelocyteeBlosts

Lymphocytes

.-. WBC

x x GRANULOCYTES

WBC 119000/mmDIFFERENT IAL:

Basophls 4Blasts 4Promyetocytes I

Myelocytes 20

Metomyelocytes 20Bonded Neutrophlls 41

Segmented Neutrophlls $Lymphocytes 2

Fig. 12.-Leukocyte response toPyrexal. Note that the granulocyticincrease was normal quantitativelyas well as qualitatively, the peak attwo hours involving an increment of8293 segmented granulocytes per cu.mm.

#{149}_. WBC

x-----x GRANULOCYTES

a---. LYMPHOCYTES

Segmented Neutrophtls

TIME IN HOURS

After Porto Covol Shunt

-- o___ . --o__ - - - -

x

-c

240 I 2 34 5 67

850 CRADDOCK, JR., PERRY, VENTZKE AND LAWRENCE

Fig. 13.-Leukocyte response toPyrexal. Note that the maximum risein granulocytes involves an incrementof only 735 cells per cu.inm.

z-alaJU)4

10 Before Porto Covol Shunt

9

8 #{149}-.WBC-G RAN ULO C Y TE S

7 0_._0LYMPHOCYTES

b600

TIME IN HOURS

Fig. 14.-Leukocyte response to Pyrexal before and after portocaval anastomosisin a patient with pancytopenia associated with congestive splenomegaly due toposthepatitic cirrhosis. Pancytopenia disappeared following surgery.

to Pyrexal is clearly shown. One interpretation of this is that the spleen nolonger exerted its parasitic effect on red cells, and presumably white cellsand platelets, as shown by the labeled red cell data. These findings could alsobe compatible with a humoral influence of the spleen on the marrow, and donot serve to elucidate this much debated point. The results are presented toindicate the difficulties in interpretation of hematologic data of this type inthe presence of hypersplenism. Since splenomegaly is a prominent featureof many conditions in which evaluation of marrow function is important,this represents a handicap in the application of this type of study.

INSTANCES OF DIVERGENCE OF MARROW MORPHOLOGY AND FUNCTION

The reserves of marrow granulocytes cannot always be gauged from scrutinyof marrow morphology alone. The marrow space is only partly filled by cellprecursors, and extension of the hematopoietic space can readily occur. Suchmust have been the case in a dog in which marked erythroid hyperplasia and

Fig. 15. - Leukocyte re-sponse in a dog with marrowerythroid hyperplasia pro-duced by repeated phlebot-omy. The response is normalin every respect. See text fordetails.

U

assz

-JassU)4

LI.0

3OC

250

200

I 5C

IO(

50

.-. NORMAL DOG 35-52WBC 8000/mm

,----m ANEMIC DOG 15-121Initial WBC 9800/mm

0

,4k/,1

//1

0 2 3 4 5 6 7 8IO 142224

TIME IN HOURS

t

S

EVALUATION OF MARROW GRANULOCYTIC RESERVES 851

Fig. 16.-Leukocyte response to repeated injections of typhoid vaccine in thesame animal shown in figure 17. The response is normal for each occasion.

preponderance was induced by repeated phlebotomy (600 to 1175 ml. perweek) for four months. This dogs nutritional condition and general healthwere excellent throughout this time. He was subjected to procedures producingleukocytosis on several occasions and responded normally (figs. 15 and 16).Platelet recovery was likewise normal (fig. 17). The intense erythroid hyper-plasia, which might suggest encroachment upon granulocyte production onhistologic examination, did not take place at the expense of marrow reservesof granulocytic tissue. This example is presented to emphasize that morphologic

I 20

60

40

TINE IN HOURS

852 (;RADDOCK, JR.. PERRY, VENLKE AND LAWRENCE

U)I-LI-JusI-4-5

a.usz-JusU)4

0

20

.. AVERAGE NORMAL(9)Initial Platelet 280,000/mm

ANEMlC DOG 15-121InItial Platelet 499800/mm

Fig. 17.-Response of blood platelets to depletion by leukopheresis in a dogrendered anemic (see figures 15 and 16) as compared with normal.

examination does not always permit assessment of myelol)oiesis. This dis-crepancy was also apparent in the case of chronic lymphocytic leukemia(fig. 12).

COMMENT

The (leVelOI)flleIlt of an acute rise in blood granulocyte level in responseto a stimulus such as that caused by Pyrexal intravenously will occur only iftliere exist reserves of granulocytes which can be quickly released into theblood. Normally the bone marrow contains the hulk of granulocytes whichcan be quickly mobilized, and whether other tissues may assume this capacityof holding granulocytes in reserve remains to be determined. This would notseem to he the case in agnogenic myeloid metaplasia of the spleen, in whichthe normal delay before the appearance of DNA-labeled cells after givingP is not observed (fig. 18).

Since leukopenia can exist without depression of marrow reserves of gEan-ulocytes, the response to Pyrexal would seem to offer valuable additional in-formation in the interpretation of leukopenia. A normal response is strongevidence of adequate reserves of granulocytes. A negative or deficient re-sponse, however, must be correlated with the clinical findings as well as withblood and marrow granulocyte morphology. Failure to develop leukocytosismay be due simply to the rapid removal of cells from the blood, as in leuko-pheresis, or as in the experiment in figure 2, in which huge numbers of cellswere entering the peritoneal cavity despite a fall in circulating concentration.An analogous situation might prevail in patients with infection or hyper-splenism. In the absence of such complications, a deficient response to Pyrexal

6 78 91011 I23I4

EVALUATION OF MARROW (;RANULOCYTIC RESERVES 853

0

z

a

E

0

8

6

4

2

#{176}-#{176} Specific Activity DNA-P

X Average Specific Activity4 Normal Individuals

HOURS 0 I 2 34 5 6 7

DAYS 4 5

Fig. 18. - The appearance ofDNA.Plabeled granulocytes inpatient with myeloid metaplasiaof the spleen and myelofibrosis ascompared with normal. Note thevery early appearance of labeledgranulocytes.

together with a left shift in the 1)eriPhertl l)lOOdl granulocytes strongly sug-gests deficient intramedullary reserves of granulocytes.

SUMMARY

Recent COIlceI)tS of the relationship of the blood granulocyte mass to themairow reserve of granulocytes have been reviewed. Evidence has been

I)resentedl to show that the marrow is the chief area of granulocyte reservesor stores. The development of acute leukocytosis in response to a stimulussuch as the intravenous injection of l)acterial endotoxin depends U1)Ofl releaseof cells from the intramedullary pool of granulocytes.

The turnover of the marrow granulocyte reserve ( MGR ) is an orderly-irocess in the steady state, and deterniines the form of the curve of DNA-labeled granulocvtes in the 1)eriI)lertl 1)100(1. From estimates of the turnovertime of the MGR it appears that the granulocyte spends an average time ofabout 10 hours in the peripheral blood. Cranulocytes dO not appear to recircu-late once they have left the eripheral blood and have entered the tissues.However, granulocytes may be sequestered within capillary beds for variable

and may re-enter the circulating blood from such areas. Such cells arenot to be considered as having re-entered the blOOdi from the tissues.

The intravenous injection of a purified bacterial lipopolysaccharide as astimulus to acute leukocytosis is described. The possbile usefulness of this

proceditire in assessing the MGR is discussed.

SuIIARIo IN IXTERLINGUA

Es presentate un revista (IC recente conceptos concernente le relation interle massa del granulocytos in Ic sanguine e le reserva de granulocytos in lemedulla ossee. Es presentate observationes ciue iridica (1W le medulla es lemajor area pro Ic immagasinage de reseervas d granulocytos. Le disvelop-pamento die leucocytosis acute in responsa a un stimulo como per exemploIc injection intravenose de endotoxina bacterial depende del liberation decellulas ab le fundos intramedullari (IC granulocytos.

Le movimento del reservas medullari de granulocytos (RMG) es tinprocesso regulate quando le organismo se trova in un stato de stabilitate e

854 CRADDOcK, JR., PERRY, VEN1ZKE AND LAWRENCE

determina le forma del curva del presentia in le sanguine peripheric degranulocytos marcate con acido disoxyribonucleic. Super Ic base de estima-tiones del tempore de transition del RMG il pare que le granulocytos passaal media circa 10 horas in le sanguine peripheric. Apparentemente le granu-locytos non re-entra in circulation post que illos ha quitate Ic sanguine periph-eric pro entrar in le tissus. Tamen, granulocytos pote esser sequestrate duranteun variabile periodo de tempore in le vasculatura capillari, e ab tal areas illospote re-entrar in le circulation. Tal cellulas non es considerate como habenteretornate ab Ic tissus.

Le injection intravenose de tin purificate lipopolysaccharido bacterial comostimulo de leucocytosis acute es describite. Le utilitate possibile de istemanovra in Ic evalutation del RMG es discutite.

REFERENCES

1. Craddock, C. G., Jr., Adams, \V. S.,Perry, S., Skoog, V. A., and Law-rence, J. S.: Studies of leukopoiesis:The technique of leukopheresis andthe response of myeloid tissue innormal and irradiated dogs. J.Lab.&Clin.Med. 45:881, 1955.

2. -, Perry, S., and Lawrence, J. S.: Thedynamics of leukopoiesis and leuko-cytosis, as studied by leukopheresisand isotopic techniques. J.Clin.Invest.35:285, 1956.

3. Perry, S., Weinstein, I. M., Craddock,C. G., Jr., and Lawrence, J. S.: Thecombined use of typhoid vaccine and32 labeling to assess myelopoie-sis. Blood 12:549, 1957.

4. Yoffey, J. NI.: The mobilization andturnover times of cell populations inblood and blood forming tissue. J.Histochem.& Cytochem. 4:516, 1956.

5. l)onohue, D. NI., Reiff, R. H., Houson,NI. L., Betson, Y., and Finch, C. A.:Quantitative measurement of theErythrocytic and granulocytic cells ofthe marrow and blood. J.Clin.Invest.37:1571-1577, 1958.

6. Osgood, E. E.: Number and distribu-tion of human liemic cells. Blood9:1141, 1954.

7. Craddock, C. G., Jr., Perry, S., andLawrence, J. S.: Newer concepts inleukocyte physiology and their clin-ical significance. A.M.A.Arch.Int.Med.100:183, 1957.

8. Bienrian, H. R., Kelly, K. H., Cordes, F.L., Byron, R. L., Poihemus, J. A., andRapaport, S.: The release of leuko-

cytes and platelets from the pulmo-nary circulation by epinephrine. Blood7:683, 1952.

9. Ambrus, J. C., and Ambrus, C. NI.: Reg-ulation of the elimination of leuko-cytes. In Brookhaven Symposia in Bi-ology #10, Homeostatic Mechanisms,84-95, 1958.

10. Ottesen, J.: On the age of human whitecells in peripheral blood. Acta physiol.Scandinav. 32:75, 1954.

11. Osgood, E. E.: Control of peripheralconcentration of leukocytes. In Brook-haven Sympos. in Biology #10Homeostatic Mechanisms, pp. 31-51,1958.

12. Hollingsworth, J. \V., Finch, S. C., andBeeson, P. B.: The role of transfusedleukocytes in experimental bacteremiaof irradiated rats. J.Lab.& Clin.Med.48:227, 1956.

13. Kline, D. L., and Clifton, E. E.: Life-span of leukocytes in man. J.Appl.Physiol. 5:79, 1952.

14. Hamilton, L. D.: Nletabolic stability ofPNA and DNA in human leukemicleukocytes; The function of lympho-cytes. In Rebuck, J. W., Ed.: TheLeukemias, Etiology, Pathophysiologyand Treatment. Henry Ford Hos-pital Symposium. New York, Aca-

demic Press, 1957, p. 381.15. McCall, NI. S., Sutherland, I). A., Eisen-

traut, A. NI., and Lanz, H.: The tag-ging of leukemic leukocytes withradioactive chromium and measure-ment of the in vivo cell survival.

J.Lab.& Clin.NIed. 45:5, 1955.

EVALUATION OF MARROW GRANULOCYT1d; RESERVES 855

16. \\eisberger, A. S., and Levine, B.:Incorporation of radioactive L-cystineby normal and leukemic leukocytesin vivo. Blood 10:1082-1094, 1954.

17. Perry, S., Craddock, C. C., Jr., an(lLawrence, J. S.: Rates of appearanceand disappearance of white 1)100(1cells in normal and in various diseasestates. J.Lab.& Clin. NIed. 51:4, 1958.

18. Craddock, C. C., Jr., Perry, S., andLawrence, J. S.: Control of the steadystate proliferation of leukocvtes. InStohlman, F., Ed.: The Kinetics ofCellular Proliferation. New York,Grune & Stratton, 1959, pp. 242-259.

19. Athens, J. \V., Nlauer, A. NI., Ashen-brucker, H., Cartwright, C. E., and

\Vintrobc,studies. I.

leukocytes

phosphate1959.

20. Heilmeycr, I.: A test of leukopoieticfunction of the hone marrow. l)ent-sche med. Wchnsch. 2:683, 1957.

21. Eichenherger, Von E., Schinidhauser-Kopp, NI., Hurin, H., Fricsay, NI.,

an(1 \Vestphal, 0.: Biological effectsof a highly purifid bacterial pyrogenfrom Salmonella ahortus e(Iui, Sib-weiz.med.Wchnschr. 85:1 190, 1955.

22. \Vintrobe. NI. H.: Clinical Hematology,

ed. 4. Philadelphia, Lea and Febiger,

1956.

.sl. NI.: Leukokinetk

A method for labeling(Vith diisopropylfiuoro-

(I)FP:12). Blood 14:303,