regulation of erythropoiesis. erythropoietin on...

Journal of Clinical InvestigationVol. 45, No. 8, 1966

Regulation of Erythropoiesis. XVIII. The Effect of Vin-cristine and Erythropoietin on Bone Marrow *

BERNARDS. MORSEt ANDFREDERICKSTOHLMANJR.4(From St. Elizabeth's Hospital and Tufts Medical School, Boston, Mass.)

The Vinca alkaloids, vinblastine and vincristine,have been reported to have relatively little effect insome species on the peripheral platelet level ascompared with their effect on myeloid and eryth-roid elements. Boggs and his associates (1) ob-served marked suppression of erythropoiesis andmyelopoiesis in dogs receiving vinblastine, but thelevel of platelets and megakaryocytes, although notquantified, did not appear to be importantlychanged. One explanation of this dichotomy mightbe that the megakaryocyte is a nondividing celland as such is not susceptible to the influence ofantimitotic agents. If this is the case, then it wouldappear that the megakaryocytic precursor cells arealso little affected. On the basis of morphologicand kinetic observations, the megakaryocytic com-partment in the rat has been divided into threestages. The earlier, or stage 1, megakaryocyteshave a turnover time of 8 to 14 hours (2). Tomaintain the stage 1 megakaryocyte in the vincris-tine-treated animal would require a continual in-put from a relatively undamaged precursor com-partment. If the stem cell is pluripotential, itwould follow that the erythroid and myeloid dam-age is mostly the result of an effect on differenti-ated cells. Alternatively, there may be a differ-ent immediate precursor for the differentiated he-matopoietic elements, as we have suggested (3).In an attempt to distinguish between the alterna-tives, we determined the effect of erythropoietin inthe hypertransfused vincristine-treated rat. Hy-pertransfusion was used to suppress the differentia-tion of erythroid elements so that the effect of vin-cristine on precursor cells could be evaluated. The

* Submitted for publication July 14, 1965; acceptedApril 20, 1966.

Supported in part by grants HTS-5600 and HE-07542from the National Heart Institute.

tU. S. Public Health Service trainee.t Address requests for reprints to Dr. Frederick

Stohlman, Jr., St. Elizabeth's Hospital, 736 CambridgeSt., Boston, Mass. 02135.

results to be reported herein appear to favor dif-ferent immediate precursors and a more primitivepluripotential precursor cell.

Methods

Adult female Sprague-Dawley rats weighing 150 to170 g were given vincristine in a single intravenous doseof 0.3 mg per kg. This dose is just below the lethalrange in this colony; 0.4 mg per kg resulted in a 90%mortality within 10 days. Rats, when transfused, weregiven washed, packed homologous red cells by tail veininjection 5 days before the start of the experiment. Theyreceived red cells in an amount equivalent to 3.5%o ofthe body weight. One hundred units of human urinaryerythropoietin, prepared by alcohol fractionation (4), wasgiven by subcutaneous injection. Blood samples werecollected by cardiac puncture from anesthetized animalsinto Versenate. Peripheral blood values were measuredby standard technics. Leukocyte differential counts werebased on the evaluation of 200 cells in peripheral smears.Bone marrow smears were made from the split femur bya brush technic. The percentage of erythroid precursorswas estimated from differential counts on 500 to 1,000nucleated cells. The mitotic index was measured in thesesame preparations. Megakaryocytic differentials werebased on evaluation of 100 megakaryocytes. Blood vol-ume was measured from the dilution of intravenously ad-ministered 'Cr-labeled washed homologous red cells.Blood samples for red cell wFe incorporation were col-lected 20 hours after intravenous administration of plasma-bound 'Fe. They were lysed with saponin and countedin a well type scintillation counter. The total iron in-corporation was based on direct measurement of theblood volume.

Results

After the administration of vincristine to normalanimals, the marrow smears revealed an accumula-tion of metaphase forms that approximated line-arity for the first 4 hours. Extrapolation of thiscurve to 100% provided an estimate of - 11hours for the generation time of rat erythroid pre-cursors. By 8 hours, 67% of erythroid precursorspotentially capable of dividing and 23%o of granu-locytic precursors potentially capable of dividingwere in metaphase. No postmetaphase forms were

1241

BERNARDS. MORSEAND FREDERICK STOHLMAN,JR.

FIG. 1. PHOTOMICROGRAPHSOF BONEMARROWOF VINCRISTINE-TREATED AND CONTROLANIMALS. Upper left: Marrow from vincristine-treated animals showing abnormalmetaphase, karyorrhexis, and phagocytosis. Center: Phagocytosed normoblast inmetaphase. The clear zone was characteristic of many such cells. Lower left: Mar-row from animal 48 hours after vincristine showing megakaryocytes in all stages andmyeloid and lymphoid elements but virtually no erythroid cells. Upper right: Eryth-roid response in marrow obtained 5 days after administration of erythropoietin to vin-cristine-treated hypertransfused animals. Lower right: Erythroid response of marrowof hypertransfused control animal 48 hours after receiving erythropoietin.

observed in erythroid cells, whereas such formswere continually noted in myeloid cells. Mitoticfigures were not observed among the remainingerythroid precursors at 24 hours. Granulocyticprecursors, in contrast, had a mitotic index of 9%at 24 hours. This decreased to 3% after 48 hours.

The morphologic configuration of erythroid andmyeloid mitoses was characterized by a disordered

arrangement of chromosomes on the metaphaseplate. Chromosomes were frequently blunted, andsome were seen loosely within the cell or joined tothe main nuclear mass by a narrow bridge. Star,exploded, and imploded metaphases were seen.Karyorrhexis and karyolysis were evident both inarrested and mature cells. Phagocytosis was strik-ing between 2 and 18 hours after vincristine ad-

1242

EFFECT OF VINCRISTINE AND ERYTHROPOIETINON BONEMARROW

TABLE I

Megakaryocytic differentials in vincristine-treated normal animals

Megakaryocytet

Stage Stage StageTreatment Time* Animal I II III

hours % % %Control 1 16 29 55

2 20 33 473 22 26 52

Vincristine 17 1 13 36 512 13 21 663 18 27 S5

* Time refers to hours after administration of 0.3 mg per kg of vin-cristine.

t Stages of maturation of megakaryocytes are as described by Ebbeand Stohlman (2). The values are based on the differential count of100 megakaryocytes.

ministration. The phagocytes were large mono-nuclear cells some of which contained as many as10 ingested cells. Erythroid, myeloid, and lymph-ocytic elements were undergoing phagocytosis;some of the phagocytosed cells were in metaphase(Figure 1).

There was a striking depopulation of differenti-ated marrow erythroid elements between the sec-ond and fourth days. Infrequent late normoblastswere the only erythroid elements seen. In con-trast, all stages of megakaryocytic maturation werepresent. Absolute numbers of megakaryocytescannot be accurately quantified, but scanning ofthe preparation indicated the number of mega-karyocytes to be relatively unaffected (Figure 1).Moreover, the effect of vincristine on the indi-vidual megakaryocytic compartments appeared tobe limited. There was relatively little change inthe megakaryocytic differential (Tables I and II).Myelopoiesis was severely depressed but did not

TABLE II

Megakaryocytic differentials in vincristine-treatedhypertransfused animals

Megakaryocytet

Stage Stage StageTreatment Time* I II III

days % % %Control 18.5 25 56.5Vincristine 1 11 23 66

2 11 19.5 69.53 17 22.5 60.55 16 29 558 11.5 21 67.5

* Days refer to the day after 0.3 mg per kg of vincristine.t Stages of maturation are as described by Ebbe and Stohlman (2).

The values are based on a differential of 100 megakaryocytes. Eachvalue refers to the average of two animals.

TABLE III

Peripheral blood values in vincristine-treated hypertransfusedanimals

Daysafter Neutrophils Platelets Reticulocytes

vincristine X 10' X 10-

0 16.643.4* (15)t 8.9±0.35 (20) 0.1 40.02 (51)3 7.1±2.4 (11) 7.140.5 (16) 0.0140.01 (11)4 7.942.2 (12) 6.3±0.38 (18) 0.0540.03 ( 7)5 4.8±40.9 (15) 7.8 ±0.4 (24) 0 (18)6 7.641.2 (18) 7.9±0.37 (24) 0.02±0.01 (16)7 19.0±1.8 (14) 8.8±0.5 (24) 0.0440.02 ( 7)8 27.5±1.4 (14) 10.7±0.48 (24) 0.01±0.01 ( 9)9 31.445.2 (10) 12.14±0.67 (14) 0.0440.02 ( 5)

11 23.4±4.8 ( 8) 11.4±1.0 (11)

* Standard error of the mean.t Numbers in parentheses refer to the number of animals.

reach the same magnitude of depopulation as theerythroid compartment. Minimal myeloid valueswere seen between the third and fourth days.Bone marrow erythroid recovery began on thefifth day and appeared complete by the seventhday, although even then occasional abnormal mi-totic figures were observed.

The changes in peripheral blood parameters af-ter vincristine in hypertransfused animals aregiven in Table III. In hypertransfused animalsonly occasional reticulocytes were seen in theperipheral blood before treatment with vincristine.After vincristine, only a rare reticulocyte was seen.In normal animals the number of reticulocytes fellwithin the first 48 hours, and between the thirdand fifth day only a rare reticulocyte was seen.Reticulocytes reappeared in the peripheral blood

EP

30 * EP0) 2C3o48 EP

o . VCEP

C0

0

0cc_

ccW 10.

z

IMEITL AFTE 0ICITN VE ,O RTR-

a.

0 2 3 4 5 6 7 8 9DAYS

FIG. 2. THE PERCENTAGEOF ERYTHROIDPRECURSORSIN.-THE MARROWOF HYPERTRANSFUSEDANIMALS GIVEN ERYTH-ROPOIETIN (EP) , VINCRISTINE (VG), ERYTHROPOIETINIMMEDIATELY AFTER VINCRISTINE, (VCEP), OR ERYTHRO--PoIETI 48 HOURSAFTER VINCRISTINE (VC48EP).

1243


EP4

601

!i

2

0r.

40

20

* o EPo VC48

/ VCEPVC

1IZt

0 2 3 4 5 6 7 8

DAYS

8 EP

FIG. 3. THE PERCENTAGEOF INTRAVENOUSLYINJECTEDIRON INCORPORATEDWHEN'9FE WASGIVEN INTRAVENOUSLY20 HOURSBEFORETHETIME OF SACRIFICE. All animals were

hypertransfused 5 days before injection of vincristine(VC). Erythropoietin (EP) was given either immedi-ately (VCEP) or 48 hours after vincristine (VC48EP).

on the sixth day; on the seventh and eighth daysthe reticulocyte value increased above the normalrange. The nadir of the platelet count was ap-

proximately 70%o of the control value (Table III).This decrease, although significant, was strik-ingly out of proportion to the absolute reticulocy-topenia and profound granulocytopenia. Granulo-cytic recovery began on the sixth and progressedto values considerably above normal by the eighthto tenth days.

As previously mentioned, the turnover rate ofimmature megakaryocytes is 8 to 14 hours (2).In view of this and the minimal effect of vincristineon megakaryocytes, it appeared that precursor

w

tI-

w

I-zw

0.

EP* EP

VC48 EPX VCEPa VC

DAYS

FIG. 4. THE PERCENTAGEOF RETICULOCYTES IN THE

PERIPHERAL BLOOD OF HYPERTRANSFUSEDANIMALS GIVEN

ERYTHROPOIETIN (EP), VINCRISTINE (VC), ERYTHRO-POIETIN IMMEDIATELY AFTER VINCRISTINE (VCEP), ORERYTHROPOIETIN48 HOURSAFTERVINCRISTINE (VC48EP).

cells were not substantially affected. To test thishypothesis, we assessed the stem cell response toerythropoietin at intervals after treatment of hy-pertransfused rats with vincristine. Bone marrow

erythroid precursor, 20-hour ""Fe uptake, andreticulocyte values from these studies are given inFigures 2 to 4 and Tables IV to VI.

Erythropoietin when given to control animalsevoked a significant increase in bone marrow eryth-roid precursors within 24 hours; erythroid hyper-plasia reached a maximum on the second or thirdday. The maximal 59Fe incorporation of 58%o was

attained when 59Fe was given on the second day,and peak reticulocytosis was observed on thefourth day. In contrast, erythropoietin adminis-tered immediately after vincristine failed to elicit

TABLE IV

Reticulocyte response in vincristine- and erythropoietin-treated hypertransfused animals*

Dayst VC EP VCEP VC48EP Control

0 0.1 i 0.02t (51)§1 0.3 0.1(4) 0 (3)2 1.1:1=0.3(6) 0 (6)3 0.01 i 0.01 (11) 2.6 1= 0.2 (12) 0 ( 5) 0.02 i 0.01 (10)4 0.05 ±0.03( 7) 4.6±0.4( 6) 0 (-9) 0.24i0.06( 7)5 0 (18) 2.6=1=0.3 (3) 0 (15) 1.1 -40.02( 7)6 0.02 z 0.01 (16) 0.24 1 0.06 (15) 0.7 ( 3)7 0.04 :0.02 ( 7) 0.2 ( 2) 0.27 ±0.05 (14) 0.05 ( 3)8 0.01 i 0.01 ( 9) 0.19 1 0.07 ( 9)9 0.04 + 0.02 (5) 0.10 (3)

* VC, vincristine; EP, erythropoietin; VCEP, erythropoietin and vincristine given at time 0; VC48EP, erythropoietingiven 48 hours after vincristine.

t Days refer to days after treatment; in case of VC48EP, the days after EP.t 1 SEM.§ Numbers in parentheses refer to the number of animals. All animals were hypertransfused, including controls.

1244neX Ibe

9- ;"I

EFFECT OF VINCRISTINE ANDERYTHROPOIETINON BONEMARROW

TABLE V

Erythroid precursors in bone marrow in vincristine- and erythropoietin-treated hypertransfused animals*

Dayst VC EP VCEP VC48EP Control

0 3.4 ± 0.3$ (51)§1 5.6 ± 0.9 (4) 1.2 ± 0.1 ( 3)2 25.9±1.1( 6) 1.6+0.5( 6)3 0.1 0.1 (11) 27.6 ± 1.1 (12) 0.8 ± 0.2 ( 5) 14.3 + 1.5 (10)4 1.0±0.2 ( 7) 20.8±2.1 ( 6) 2.5 ±0.1 (9) 32.5 i 1.4( 7)5 1.5 i 0.3 (18) 15.5 ± 1.8 ( 3) 8.9 ± 0.7 (15) 3.9 ± 0.7 ( 7)6 2.3 i 0.5 (16) 8.6 ± 1.0 (15) 2.0 ± 0.7 ( 3)7 2.1 ±0.3( 7) 6.9 ( 2) 2.440.4 (14) 0.8±0.2( 3)8 1.5 +0.5 (9) 1.5 ±0.2 (9)9 1.1±t0.3( 5) 1.0±0.1( 3)

*VC, vincristine; EP, erythropoietin; VCEP, vincristine given at the same time as erythropoietin; VC48EP,erythropoietin given 48 hours after vincristine.

t Days refer to days after treatment; in case of VC48EP, the days after EP.t ± 1 SEM.§ Numbers in parentheses refer to the number of animals.

a detectable bone marrow erythroid response untilthe fifth day, at which time erythroid precursors

increased to 9%. Significant elevations of 59Feincorporation and reticulocytes were not observeduntil the sixth and seventh days. The values atthis time were 8.4%o for iron incorporation and0.24%o for reticulocytes. Whenerythropoietin was

given 48 hours after vincristine, the erythroid re-

sponse occurred earlier, but was not manifest un-

til 24 to 48 hours after that seen in erythropoietin-treated controls. Vincristine treatment alone re-

sulted in values consistently lower than transfusedcontrols for the 9-day course of the experiment.

Not only was the onset of the erythropoietic re-

sponse delayed in vincristine-treated animals, butit was also diminished in magnitude. Further,

there appeared to be a dichotomy between the de-gree of erythroid activity in the bone marrow andthe peripheral blood response as judged by reticu-locytes or iron incorporation or both. These ob-servations suggested a substantial degree of in-effective erythropoiesis. The presence of bizarremultinucleated erythroblasts, observed in bonemarrow smears prepared at the peak of the delayedresponse, is in concert with such a notion. Thesecells were not seen in erythropoietin-treatedcontrols.

Discussion

The data reported herein are consonant withprevious reports that a major effect of vincristineis to arrest mitosis in metaphase. The estimate of

TABLE VI

Iron incorporation* in hypertransfused and vincristine- and erythropoietin-treated ratst

Day* VC EP VCEP VC48EP Control

0 3.5 ± 0.2§ (51)1 4.8±0.9( 4)2 36.9 46.4( 6) 0.8±0.1( 6)3 1.9 ± 0.3 (11) 57.7 ± 1.7 (12) 1.6 ± 0.2 ( 5) 2.7 ± 0.3 (10)4 1.4 ± 0.1 ( 7) 51.9 ± 1.9 (6) 1.2 ± 0.1 (9) 6.1 ± 1.2 (7)5 1.6 ± 0.1 (18) 32.8 ±i: 5.1 ( 3) 1.3 ± 0.1 (15) 30.5 + 4.2 (7)6 1.6 ± 0.1 (16) 8.4 ± 1.6 (15) 12.8 ±- 1.0 ( 3)7 2.7±0.2( 7) 7.9 (2) 6.0±0.5 (14) 1.7+0.3( 3)8 2.4 ± 0.3 (9) 3.3 ± 0.4 (9)9 2.840.9( 5) 1.3±0.2( 3)

* Measured 20 hours after injection of plasma-bound 59Fe.t VC, vincristine; EP, erythropoietin; VCEP, vincristine and erythropoietin at time 0; VC48EP, erythropoietin 48

hours after vincristine.t Day refers to day of sacrifice; in the case of VC48EP, the days after EP.§ ±- 1 SEM.

11 Numbers in parentheses refer to the number of animals.

1245


11 hours for the generation time of rat erythroidprecursors is in concert with values of 9 to 15hours derived from either radioiron (5) or triti-ated thymidine (6) labeling in the dog. Ana-telo-phase stages were not observed in the nucleatederythroid cell during the first 24 hours after ad-ministration of vincristine. In view of this, andthe virtually complete depopulation of the nucle-ated erythroid compartment, it appears that escapefrom the mitotic block produced by vincristine isa rare event in the erythroid series.

The accumulation of arrested myeloid mitoseswas much slower. Approximately 20%o' of themyeloid cells were in mitosis after 8 hours; alonger generation time for myeloid than erythroidcells has also been estimated from thymidine data(7). Ana-telophase forms could be observedamong the myeloid mitoses throughout the first12 hours. Although the frequency of this mitoticstage was diminished, its continued presence sug-gests that some of the arrested myeloid cells arecapable of escaping the metaphase block. A con-siderable degree of karyorrhexis and phagocytosisof arrested nucleated precursors was also notedand gives further evidence of their decreased via-bility. The increased phagocytic activity of thebone marrow is probably secondary to the so-called"odor of death," but a primary vincristine effecton the macrophage has not been eliminated.

Vincristine caused almost complete depopulationof erythroid elements in the marrow but not ofmegakaryocytic elements. Megakaryocytes, likedifferentiated erythroid precursors, are apparentlyfed from a more primitive pool that has not beenidentified morphologically. Such a notion is sup-ported by kinetic studies of rat bone marrow af-ter flash labeling with tritiated thymidine (2).Thirty minutes after a single dose of intravenousthymidine - 25%o of the earliest recognizable(stage 1) megakaryocytes were labeled. Twenty-four hours later, more than 90%b of this groupwere labeled. Grain count data indicated that celldivision rarely occurred in this compartment.Thus the major portion of the labeled early mega-karyocytes had to be. derived from a labeled, un-recognizable precursor pool in DNAsynthesis atthe time of injection of thymidine. From thisstudy, the turnover rate of the early megakaryo-cyte was estimated to be of the order of 8 to 14hours and the total maturation time 60 hours with

a minimum of 43 and a maximum of 75 hours.If differentiation of precursor cells stopped, de-population of stage 1 megakaryocytes would beanticipated within 16 hours, stage 2 within 36hours, and virtually complete depopulation of theentire megakaryocytic compartment within 60hours. The megakaryocytic differential (TablesI and II) showed a slight decrease in stage 1megakaryocytes relative to control values duringthe first 24 hours. The values for stage 1 mega-karyocytes, however, were within the range ofnormal values previously reported in this colony(2). Thereafter there were only minor changesin the pattern of megakaryocytic differential.These considerations together with the presenceof substantial numbers of megakaryocytes in thebone marrow (Figure 1) lead to the conclusionthat differentiation of substantial numbers ofmegakaryocytes and maturation were not impor-tantly affected by the administration of vincristine.The relatively minor fall in the platelet count ob-served in the peripheral blood would be in con-cert with such a notion. The alternative explana-tion would be that megakaryocytic maturationstopped with a resultant "biological standstill."This is not only unlikely biologically, but appearsuntenable on other grounds. Had this obtained,the abrupt cessation of platelet production wouldhave resulted in an earlier and more pronouncedthrombocytopenia. The platelet life-span in thiscolony approaches an exponential with a tj of

1.3 days (8). Accordingly, if platelet produc-tion stopped for 4 days, as would be implied bysuch a biologic standstill, a 90%o decrease in theperipheral platelet count would be anticipated.This was not observed (Table III). Finally, itwas evident from morphologic evaluation of mega-karyocytes that platelet shedding and hence plate-let production continued during the period of ob-servation (Figure 1). This in itself would implycontinued maturation. From these considerationswe conclude that differentiation and maturationof the megakaryocytes continued during the pe-riod after administration of vincristine.

The depopulation of the erythroid compartmentfor a period of - 4 days indicated that differentia-tion into this compartment was not taking place.Moreover, during this period differentiation intothe erythroid compartment could not be inducedby massive doses of erythropoietin.

1246


The continued differentiation of precursor cellsinto the megakaryocytic but not the erythroidcompartment could be explained if the immediateprecursor cells for both cell lines differed, inwhich case the megakaryocytic precursors wouldbe presumed to be less sensitive to the effects ofvincristine than the erythroid precursors. Al-ternatively, a selective block of erythroid differen-tiation in a common precursor cell could explainthe observations. Although a final decision be-tween these two alternatives cannot be made withthe data at hand, the inability to induce erythroiddifferentiation with massive doses of erythropoietintogether with the morphologic abnormalities seenin erythroid elements, but not in megakaryocytesduring the recovery phase, argues in favor of dif-ferent precursor cells.

For more than 50 years, a vast literature hasaccumulated concerning the questions of whetherthe differentiated hematopoietic cells are partiallyor completely self-sustaining and whether a unipo-tential or pluripotential stem cell exists. It be-came evident from the studies of Alpen, Cranmore,and Johnston (5) on iron labeling of pronormo-blasts in bled dogs, together with studies of theeffects of hypertransfusion, starvation, -and eryth-ropoietin in rodents (9), that the erythroid com-partment normally is not self-sustaining but is con-tinually fed from a more primitive precursor com-partment. The studies of Ebbe and Stohlman(2) and Feinendegen, Odartchenko, Cottier, andBond (10) would indicate that the same holdstrue for the megakaryocytic compartment. Al-though the data are not so clear-cut in the case ofthe myeloid elements, it seems likely that this com-partment also is fed from a precursor pool (7, 11).The argument concerning whether the precursorcell is unipotential or pluripotential appeared tohave been answered by the study of radiationchimeras by Becker, McCulloch, and Till (12) andof the Ph chromosomes in chronic myelogenousleukemia by Whang and her colleagues (13).Both of these studies indicate that at some level apluripotential precursor cell exists, the markerchromosome being common to all elements. Theidentity of the common precursor is moot. Yof-fey (14) and Cudkowicz, Bennett, and Shearer(15) have suggested a small lymphocyte. Thomas,Fliedner, Thomas, and Cronkite (16) suggested asmall mononuclear cell, but were uncertain of its

PRUITIVEPRECURSOR

COMPARTMENTt COONYFORMNG

CELL

IMMEDIATEPRECURSOR

COMPARTMENT

DIFFERENTIATEDCOMPARTMENT

FIG. 5. A TENTATIVE MODEL FOR HEMATOPOIESIS.Mega, megakaryocyte; E, erythroid; M, myeloid. Seetext for discussion.

identity. It is of interest to note that the smallperipheral blood lymphocyte capable of growth intissue culture does not have the Ph chromosome.There may be functional differences, however, be-tween the small peripheral lymphocyte and thatof the bone marrow. In fact, evidence has beenadduced by Osmond and Everett (17) and Robin-son, Brecher, Lourie, and Haley (18) to indicatethat growth potentials may differ between twotypes of morphologically similar lymphoid cells.Irrespective of the morphologic identity of thecolony-forming cell, there is a question that this isthe immediate precursor cell for each of the he-matopoietic elements. The observations describedherein, as well as those of megakaryocytic labeling(2) and the transplantation studies of Blackett,Roylance, and Adams (19) in sublethally irradi-ated animals, suggest that a more differentiatedintermediate precursor may exist. Similarly theobservations of Lewis and Trobaugh (20), al-though not interpreted as such, suggest to us adifferent immediate precursor cell for each cellline. Accordingly we should like to propose themodel for precursor cells shown in Figure 5.

In this model, there is a pluripotential precursorcell similar to the colony-forming cell describedby Till and McCulloch (21). This cell in turngives rise to a more differentiated cell that is theimmediate precursor for the individual morpho-logically identifiable cell lines. The previously dis-cussed evidence indicates that the immediate pre-cursor cell for megakaryocytic elements differs

1247


from that of myeloid and erythroid elements. Thedata do not permit a final judgment of whetherthe immediate precursor cell for myeloid anderythroid cells differs, but it seems likely that thisis the case. In favor of this are the apparent dif-ference in generation time for the two elementsreported previously (7), as well as herein, togetherwith the difference in the rate and degree of de-population observed under the experimental con-dition used in the present study.

The relationships of the immediate precursorcells to the primitive pluripotential cell, and themechanism of regulation, must remain at present toa large extent speculative, although some of thenotions to be presented are subject to experimentalverification. Weshould like to suggest that the im-mediate precursor compartments are to a largeextent self-maintaining, but when damaged andperhaps in some instances of severe stimulation,e.g., phenylhydrazine-induced anemia, they are re-populated from the primitive pluripotential com-partment. A feedback mechanism, probably neg-ative, to the primitive compartment is thus im-plicated. An important segment of the immediateprecursor compartment presumably is dormantand is triggered into cell cycle either by differ-entiation, e.g., into erythroid elements, or deathof cells within the compartment. Accordingly,increased demand for one or another formed ele-ment could be met in large measure by a decreasedturnover time of the immediate precursor com-partment. Critical to adequate homeostasis wouldbe a control mechanism to prevent the depopulationof the immediate precursor cell. Differentiation ofimmediate precursor cells does not occur untila critical level of cells is achieved. The signal fordifferentiation into the erythroid compartmentis presumably erythropoietin, and there may beanalogs of this for the other formed elements. Itwould follow, then, that the erythropoietin-sensi-tive cell is the immediate precursor and not theprimitive pluripotential cell.

Gurney, Lajtha, and Oliver (22) suggested theplethoric rodent as an experimental model to studythe extent of "stem cell" damage in radiation in-jury. The virtue of the "stem cell tolerance test"is that the polycythemic rodent has virtually com-plete suppression of erythropoiesis and a pre-dictable response to a given dose of erythropoietin.This preparation, then, theoretically should permit

an estimate of the degree of damage to the immedi-ate precursor compartment. In view of the dif-ference between the megakaryocytic and erythroidresponses to vincristine, we applied this experi-mental model to evaluate the effect of vincristineon the immediate precursor of erythroid cells.When erythropoietin was given to the polycythe-mic animal, within the first several days of theadministration of vincristine there was a subopti-mal response to erythropoietin. One can infer,then, that vincristine damaged the immediateerythroid precursor compartment. If vincristine'seffect was confined mainly to the dividing cell,then a substantial portion of the immediate pre-cursor cells was in cycle even when the demandfor differentiation was reduced by hypertransfu-sion. Extensive damage to the immediate pre-cursor compartment is further supported by themorphologic evaluation of the bone marrow dur-ing the initial period of response to erythropoietin.Abnormal multinucleated forms were seen, as wereabnormal mitoses. Further, there was a dichotomybetween the bone marrow response and peripheralblood, suggesting ineffective erythropoiesis. Thedegree of ineffective erythropoiesis indicated thatthere was substantial residual damage to the im-mediate precursor compartment even after re-covery had progressed to the point of permittingdifferentiation.

Normally a detectable response to erythropoietinis seen in the bone marrow of hypertransfused ani-mals within 24 hours, and by 48 to 72 hours sub-stantial numbers of reticulocytes have been re-leased into the peripheral blood. In contrast, invincristine-treated animals this response was de-layed. Whenerythropoietin was given at the timeof treatment, a small but significant increase inbone marrow erythroid cells was not seen until5 days after treatment. Since the reticulocytosiswas negligible, it is presumed that most of theerythroid precursors died in situ. When erythro-poietin was given 24 to 48 hours after vincristine,the extent of ineffective erythropoiesis did not ap-pear to be so great, but there was again a delay inresponse. The length of the delay and the mag-nitude of the response were related to the intervalbetween administration of vincristine and erythro-poietin.

The delay in response to erythropoietin couldbe explained by continued availability of extra-

1248


cellular erythropoietin. Alternatively, one mightsuggest that erythropoietin entered the cell duringthe first 12 hours after injection and interactedwith its intracellular target and that the informa-tion imparted thereby was stored until regenerationof the immediate precursor compartment reachedthe point at which differentiation could once againoccur. The half-time for plasma clearance oferythropoietin in normal rats has been shown tobe of the order of 2.5 to 3 hours (4). In rats withsuppressed erythropoiesis, clearance is somewhatslower, the half-time being -6 hours (23).From these observations it would appear unlikelythat erythropoietin would be available from extra-cellular sources for several days, as would be neces-sary if this were the explanation of the delayed re-sponse observed when the hormone was given attime 0. More likely the alternative, intracellularstorage of information or perhaps erythropoietinitself, obtains. The difference in the magnitudeof response between animals given erythropoietinat time 0, as compared with 48 hours, presumablyreflects death from vincristine of erythropoietin-"activated" immediate precursor cells within thecompartment. Dilution of information throughdivision may also play a minor role.

A delayed response to erythropoietin is prob-ably a general phenomenon of the regeneratingmarrow and not peculiar to vincristine-treated ani-mals. There are studies on other regeneratingsystems that may be interpreted in this light.Preliminary studies in our laboratory indicate thatanimals given actinomycin D or cytoxan behavelike vincristine-treated animals in this respect.Although observations have not been made in hy-pertransfused irradiated animals, there is evidencein normal irradiated animals that such a delayedresponse occurs (24). Erythropoietin given im-mediately after sublethal irradiation produced aneffect that was not manifest for several days.Storage of information until regeneration permitsdifferentiation, as described herein, may prove tobe a more attractive explanation than the originalinterpretation that a negative feedback was re-moved by differentiating immediate precursor cellswhich then died after one or two abortive divisions.Studies on the effect of erythropoietin of the bonemarrow of hypertransfused irradiated animals car-ried out for several days beyond that in the "stemcell tolerance test" should provide a conclusive an-

swer as to which alternative obtains in the ir-radiated animal.

Summary

Studies are presented on the hematopoietic re-sponse of rats to vincristine and the effect of eryth-ropoietin on red cell production in these animals.The megakaryocytic compartment appeared to berelatively unaffected during the period of observa-tion, which was interpreted as indicating that therewas continued differentiation of precursors intothis compartment. In contrast, there was a strik-ing depopulation of the erythroid compartment be-tween the second and fourth days during whichdifferentiation of precursors could not be achievedeven with massive doses of erythropoietin. Con-tinued differentiation of cells into the megakaryo-cytic compartment but not the erythroid compart-ment suggests different immediate precursor cells;a model for this is presented and discussed. Inthe hypertransfused vincristine-treated rat therewas a delayed response to erythropoietin. It issuggested that the immediate precursor cells forthe erythroid compartment are capable of storinginformation imparted by erythropoietin until suchtime as regeneration permits differentiation. Mor-phologic evaluation of the bone marow suggeststhat even after recovery has proceeded to a pointat which differentiation once again occurs, there isresidual damage that leads to ineffective erythro-poiesis.

Acknowledgments

We gratefully acknowledge the technical assistance ofDonald Howard, Janet Donovan, and Sheila Sax. Wewish to thank Dr. J. M. McGuire of the Eli Lilly Com-pany, who generously supplied the vincristine, and Dr.William C. Janssen of Lakeside Laboratories, who sup-plied the Imferon.

References1. Boggs, D. R., J. W. Athens, 0. P. Haab, P. A.

Cancilla, S. 0. Raab, G. E. Cartwright, and M. M.Wintrobe. Leukokinetic studies. VII. Morphologyof the bone marrow and blood of dogs givenvinblastine sulfate. Blood 1964, 23, 53.

2. Ebbe, S., and F. Stohlman, Jr. Megakaryocytopoiesisin the rat. Blood 1965, 26, 20.

3. Morse, B., and F. Stohlman, Jr. A delayed stemcell response to erythropoietin (abstract). Blood1964, 24, 846.

1249


4. Stohlman, F., Jr., and D. Howard. Humoral regula-tion of erythropoiesis. IX. The rate of disap-pearance of erythropoietin from the plasma inErythropoiesis, L. 0. Jacobson and M. Doyle,Eds. New York, Grune & Stratton, 1962, p. 120.

5. Alpen, E. L., D. Cranmore, and M. E. Johnston.Early observation on the effects of blood loss inErythropoiesis, L. 0. Jacobson and M. Doyle, Eds.New York, Grune & Stratton, 1962, p. 184.

6. Bond, V. P., N. Odartchenko, H. Cottier, L. E.Feinendegen, and E. P. Cronkite. The kinetics ofthe more mature erythrocytic precursors studiedwith tritiated thymidine in Erythropoiesis, L. 0.Jacobson and M. Doyle, Eds. New York, Grune& Stratton, 1962, p. 173.

7. Cronkite, E. P., and T. M. Fliedner. Granulocyto-poiesis. New Engl. J. Med. 1964, 270, 1347, 1403.

8. Ebbe, S., F. Stohlman, Jr., J. Donovan, and D.Howard. Platelet survival in the rat as measuredwith tritium-labeled diisopropylfluorophosphate. J.Lab. clin. Med. In press.

9. Fried, W., L. F. Plzak, L. 0. Jacobson, and E.Goldwasser. Studies on erythropoiesis. III. Fac-tors controlling erythropoietin production. Proc.Soc. exp. Biol. (N. Y.) 1957, 94, 237.

10. Feinendegen, L. E., N. Odartchenko, H. Cottier, andV. P. Bond. Kinetics of megakaryocyte prolifera-tion. Proc. Soc. exp. Biol. (N. Y.) 1962, 111, 177.

11. Patt, H. M., and M. A. Maloney. An evaluation ofgranulocytopoiesis in Cell Proliferation, L. F.Lamerton and R. J. M. Fry, Eds. Philadelphia,F. A. Davis, 1963, p. 157.

12. Becker, A. J., E. A. McCulloch, and J. E. Till.Cytological demonstration of the clonal nature ofspleen colonies derived from transplanted mousemarrow cells. Nature (Lond.) 1963, 197, 452.

13. Whang, J., E. Frei III, J. H. Tijo, P. P. Carbone,and G. Brecher. The distribution of the Phila-

delphia chromosome in patients with chronicmyelogenous leukemia. Blood 1963, 22, 664.

14. Yoffey, J. M. Cellular equilibria in blood and blood-forming tissues. Brookhaven Symposia in Biology.No. 10. Homeostatic Mechanisms. 1957, p. 1.

15. Cudkowicz, G., M. Bennett, and G. M. Shearer.Pluripotent stem cell function of the mouse marrow"lymphocyte." Science 1964, 144, 866.

16. Thomas, E. D., T. M. Fliedner, D. Thomas, andE. P. Cronkite. The problem of the stem cell:observations in dogs following nitrogen mustard.J. Lab. clin. Med. 1965, 65, 794.

17. Osmond, D. G., and N. B. Everett. Radioautographicstudies of bone marrow lymphocytes in vivo and indiffusion chamber cultures. Blood 1964, 23, 1.

18. Robinson, S., G. Brecher, I. Lourie, and J. Haley.Leukocyte labeling in rats during and after con-tinuous infusion of tritiated thymidine: implicationsfor lymphocyte longevity and DNA reutilization.Blood 1965, 26, 281.

19. Blackett, N. M., P. J. Roylance, and K. Adams.Studies of the capacity of bone-marrow cells to re-store erythropoiesis in heavily irradiated rats. Brit.J. Haemat. 1964, 10, 453.

20. Lewis, J. P., and F. E. Trobaugh, Jr. Hmmatopoieticstem cells. Nature (Lond.) 1964, 204, 589.

21. Till, J. E., and E. A. McCulloch. A direct measure-ment of the radiation sensitivity of normal mousebone marrow cells. Radiat. Res. 1961, 14, 213.

22. Gurney, C. W., L. G. Lajtha, and R. Oliver. Amethod for investigation of stem-cell kinetics. Brit.J. Haemat. 1962, 8, 461.

23. Stohlman, F., Jr. Unpublished observations.24. Stohlman, F., Jr. Humoral regulation of erythro-

poiesis. VI. Mechanism of action of erythropoietinein the irradiated animal. Proc. Soc. exp. Biol.(N. Y.) 1961, 107, 751.

SPECIAL NOTICE TO SUBSCRIBERS

Post Offices will no longer forward the Journal when you move.

Please notify The Journal of Clinical Investigation, BusinessOffice, 10 Stoughton Street, Boston, Mass. 02118, at once whenyou have a change of address, and do not omit the Zip Codenumber.

1250

regulation of erythropoiesis. erythropoietin on...

Documents