Leukemia Research 23 (1999) 357–364

Biologic characteristics of patients with hypocellularmyelodysplastic syndromes

Rajat Goyal a, Huma Qawi a, Irfan Ali a, Saleem Dar a, Suneel Mundle a,Vilasini Shetty a, Yifwayimare Mativi a, Krishnan Allampallam a, Laurie Lisak a,

Jerome Loew b, Parameswaren Venugopal a, Sefer Gezer a, Erwin Robin c,Shelby Rifkin d, Azra Raza a,*

a Rush Cancer Institute, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL, USAb Department of Pathology, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL, USA

c Ingall’s Memorial Hospital, Har6ey, IL, USAd Northwest Community Hospital, Arlington Heights, IL, USA

Received 17 August 1998; accepted 14 November 1998

Abstract

Rates of proliferation and apoptosis as well as expression of tumor necrosis factor alpha (TNF-a), transforming growth factorbeta (TGF-b) and the number of macrophages were measured in bone marrow (BM) biopsies of 33 patients who presented withhypocellular (cellularityB30%) myelodysplastic syndromes (MDS). Results showed that 2/3 of the patients had high apoptosis,high cytokine levels and large number of macrophages in their biopsies while 1/3 did not. Apoptosis and TNF-a levels weredirectly related (r=0.583, P=0.003, n=24) as was apoptosis and the degree of anemia (P=0.033, n=18). A subgroup ofpatients with abnormalities of chromosomes 5 or 7 had higher platelets (P=0.026) and higher apoptosis (P=0.038) whencompared with the rest of the group. Eight patients had no evidence of apoptosis and almost no detectable TNF-a in theirbiopsies. We conclude that within the hypocellular variant of MDS, there may be two distinct sub-groups of patients, one whopresent with high cytokine-mediated intramedullary apoptosis and the other who may be better characterized as having a stem-cellfailure defect since they showed no evidence of apoptosis. © 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Hypocellular myelodysplastic syndromes; Apoptosis; Proliferation; TNF-a; TGF-b; Chromosomes 5 and/or 7

1. Introduction

The myelodysplastic syndromes (MDS) are a groupof hematopoietic disorders that primarily originate in apluripotential bone marrow (BM) stem cell and tend topredominate in the elderly [1]. The vast majority ofpatients present with a refractory anemia that may ormay not be associated with additional cytopenias andsupportive care continues to be the mainstay of therapy

for these variable cytopenias [2]. Several recent ad-vances have been achieved in understanding the biolog-ical basis for the paucity of circulating blood cells, themost significant being the recognition of excessive in-tramedullary apoptotic death of hematopoietic cellsbelonging to all three hematopoietic lineages [3–6].This peculiar mode of suicidal cellular destruction ap-pears to be cytokine mediated in a substantive majorityof patients with tumor necrosis factor alpha (TNF-a)being the prominent pro-apoptotic cytokine involved[7–9]. In summary then, the biological hall-marks ofMDS are rapid intramedullary proliferation of hemato-poietic cells followed by an equally rapid and prema-ture apoptotic death of the vast majority of these cellsin the presence of increased levels of pro-inflammatorycytokines such as TNF-a. It is important to note that

Abbre6iations: MDS, myelodysplastic syndromes; ISEL, in situ endlabeling of fragmented DNA; IudR, lododeoxyuridine; BrdU, bro-modeoxyuridine; TNF-a, tumor necrosis factor alpha; TGF-b, trans-forming growth factor beta; LI, labeling index; BM, bone marrow.

* Corresponding author. Rush Cancer Institute, 2242 West Har-rison Street, Tech 2000 Building-Suite 108, Chicago, IL 60612-3515,USA. Tel.: +1-312-4558474; fax: +1-312-4558479.

0145-2126/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.

PII: S 0 1 4 5 -2126 (98 )00187 -8

R. Goyal et al. / Leukemia Research 23 (1999) 357–364358

these biological characteristics may not be present uni-versally in all MDS patients mainly because there arefive different syndromes grouped under one umbrella.Each syndrome presents with its own unique naturalhistory and prognosis and each has its own spectrumof clinical heterogeneity. Thus it is important to iden-tify various subgroups within MDS and determinewhether these patients have archetypal biologic char-acteristics which would distinguish them from otherMDS patients, thereby allowing hematologists to de-sign therapeutic strategies tailored to suit their individ-ual needs. The present study is one such attempt.

While the general rule in MDS is that the bonemarrow is hypercellular despite the presence of vari-able cytopenias, 10–30% cases present with a hypocel-lular BM [1,2]. In the absence of a recognizablekaryotypic abnormality, such cases are often difficultto distinguish from classic aplastic anemias, althoughpreponderance of hypocellularity in the therapy-relatedmyelodysplasias and its subsequent association withpathognomic cytogenetic abnormalities are often help-ful in negotiating such semantic matters in a smallsubsets of patients. In the present study, we haveattempted to identify biologic characteristics whichwould be uniquely associated with hypocellular MDScases. The most striking features of this study wererelated to findings of high intramedullary apoptosis ofhematopoietic cells in association with high pro-apop-totic cytokines and presence of macrophages in ap-proximately 2/3 of the cases studied, thereby suggest-ing a possible explanation for the hypocellularity atleast in a sub-set of these patients. Hemoglobin wasinversely related to the level of apoptosis in this groupwith patients showing high apoptosis being more ane-mic. Finally, the subset of patients with hypocellularMDS who presented with abnormalities of chromo-somes 5 and/or 7 tended to have a higher plateletcount than the rest of the group as well as a higherlevel of apoptosis.

2. Materials and methods

Thirty-three MDS patients are the subject of thisreport. These MDS patients were studied prior tostarting a treatment protocol, but had been off alltherapies including growth factors and vitamins for atleast 4 weeks. Every individual received a 1 h infusionof IUdR (or in case of an iodine allergy, BrdU) at 100mg/M2 intravenously prior to undergoing a bone mar-row (BM) examination. Informed consent for the infu-sion protocol (MDS 90-02) was obtained from everyindividual. The infusion protocol was reviewed andapproved by the Institutional Review Board (IRB) ofthe Rush-Presbyterian-St. Luke’s Medical Center, theNational Cancer Institute (NCI) and the Food and

Drug Administration (FDA). The IUdR and BrdU forthese studies were supplied by the NCI. Immediatelyupon completion of the infusion, peripheral blood,BM aspirate and BM biopsies were obtained from thepatient and transported on ice to Dr Raza’s labora-tory. The following studies were performed on thetissues.

2.1. Detection of cytokines in the microen6ironment

Levels of two cytokines TNF-a, and transforminggrowth factor beta (TGF-b) were determined semi-quantitatively in the bone marrow biopsies immuno-histochemically. All tissues were fixed in Bouin’ssolution and embedded in glycol methacrylate. Two tothree micron thick sections were obtained and placedon alcian blue coated coverslips. The sections werethen individually labelled for each cytokine using therespective monoclonal antibodies as follows. After thetissues were dehydrated in distilled water for 10 minthey were incubated with freshly diluted 3% H2O2 for30 min and then with pronase 1 mg/ml (Calbiochem,LaJolla, CA) for 45 min. Specimens were rinsed care-fully with 0.15 M phosphate buffered saline (PBS)[0.15 M sodium chloride in 0.1 M phosphate buffer,pH 7.5] after each incubation. Following the last 0.15M PBS rinse they were placed in 0.5 M PBS [0.5 Msodium chloride in 0.1 M phosphate buffer, pH 7.5]for 15 min. The sections were treated with 0.5 M PBScontaining 1.5% horse serum for 60 min to blocknon-specificity. Subsequently the sections were incu-bated with the respective monoclonal anti TGF-b2b3

(1:50) antibody (Oncogene Science, Manhasset, NY)or anti TNF-a (1:180) (Promega, Madison, WI) anti-body diluted in 0.5M PBS containing 1.5% horseserum for 60 min. This was followed by incubating thesections with biotinylated anti-mouse IgG [diluted(1:200) in 0.5M PBS with 1.5% horse serum] for 30min and with the avidin-biotin complex or ABCreagent. The horse serum, biotinylated anti-IgG andABC complex were reagents in the Vectastain EliteABC kit (Vector, Burlingham, CA). After each of theabove incubations, specimens were rinsed in 0.5 MPBS. The color reaction was then developed using0.025%, 3,3% diamino benzidine tetrachloride (DAB) di-luted in 100 ml of 0.5 M Tris buffer, pH 7.5, with 0.01ml 30% H2O2 for 10 min and rinsed with distilledwater. After processing, the coverslips were mountedon glass slides and examined by light microscopy.

2.2. Detection of macrophages

The presence of macrophages was detected immuno-histochemically by the method described above. Themonoclonal antibody used was EBM-11 (Dakopatts,Denmark).

R. Goyal et al. / Leukemia Research 23 (1999) 357–364 359

2.3. Detection of S-phase cells

The in-situ detection of the two thymidine ana-logues IUdR and BrdU administered via intravenousinfusions was carried out by using the protocols andimmunohistochemical methods described before [5,6].After processing and mounting the coverslips withfluoromount, at least 2000 positively labeled S-phasemyeloid cells were counted to determine the labelingindex (LI). Erythroid and megakaryocytic cells wereexcluded.

2.4. Measurement of apoptosis using in-situ endlabeling (ISEL) of fragmented DNA

ISEL was carried out on all the bone marrow biop-sies of the 33 patients as described in earlier studies[5,6]. Briefly, the sections following pre-treatment withsodium chloride sodium citrate (SSC) solution at80°C and with 1% Pronase (1 mg/ml in 0.15 M PBS;Calbiochem, LaJolla, CA) were incubated with a mix-ture of dATP, dCTP, dGTP (0.01M, Promega,Madison, WI), bio-dUTP (0.001M, Sigma) and DNAPolymerase I (20 U/ml, Promega) at 18°C. Incorpora-tion of bio-dUTP was finally visualized using avidin-biotin-peroxidase conjugate (Vectastain Elite ABCKit, Vector, Burlingham, CA) and diamino benzidinetetrachloride. Thus cells labeled positively for ISELshowed brown staining in their nuclei under the lightmicroscope.

2.5. Interpretation of slides

All the slides were observed on a televised screenby several investigators. A subjective quantitativescale was formulated to determine the degree of posi-tivity of the different cytokines (TGF-b, TNF-a),ISEL staining and the cellular component(macrophages) as follows: negative, low, intermediateand high. The intensity of staining/cell was notrecorded. The percentage of cells were not calculated.Data on reproducibility of the results have been pro-vided in the past [5,8].

1–3+LowIntermediate 4–5+High 6–8+

2.6. Statistical analysis

The nonparametric Spearman test was used to de-termine correlations between groups.

3. Results

A total of 33 MDS patients are the subject of thisreport. According to the French–American–British(FAB) classification [10], 22 patients had refractoryanemia (RA), one had RA with ringed sideroblasts(RARS), seven had RA with excess blasts (RAEB), twohad RAEB in transformation (RAEB–t) and one hadchronic myelomonocytic leukemia (CMMoL). Therewere 20 males and 13 females with a median age of 60years for the entire group. All patients had primaryde-novo MDS.

3.1. Biologic characteristics

The cellularity was assessed from BM biopsies andranged between 5–30% (median=20%). Table 1 showsthe details of all the clinical/biological parameters mea-sured. The median white blood cell (WBC) count atpresentation was 3.0×109/l and the median hemo-globin was 9.4 Gm/dl (Table 1). The median labelingindex (LI) for the entire goup was 22% (n=16,range=11–37%), median ISEL was 2 (n=26, range=0–8), median TNF-a level assessed in the BM biopsywas 3 (n=29, range=0–8), median TGF-b was 4(n=28, range=0–8) and the number of macrophagesas measured by EBM-11 antibody were a median of 2(n=30, range=0–8).

3.2. Apoptosis and cytokine studies

Among the 33 hypocellular MDS cases, ISEL studieswere available in 26 cases and 8/26 cases did not haveany evidence of ISEL positivity in the entire biopsy.Ten of the 18 patients who had evidence of apoptosisshowed greater than 4+ positivity. A highly significantrelationship was found between ISEL positivity andTNF-a levels (r=0.583, P=0.003, n=24). Futher-more, in patients who were ISEL-positive versus thosethat were ISEL-negative, there was a significant differ-ence in the level of TNF-a (P=0.005). The medianTNF-a level was 0 for the ISEL-negative patients (n=7) and the median TNF-a was four for the ISELpositive cases (n=17). Finally, level of apoptosis wasalso inversely related to the level of hemoglobin in thesepatients (r= −0.505, P=0.033, n=18).

TNFa levels were available in a total of 29/33 pa-tients with a median value of 3+ (Table 1). Nine ofthese 29 patients showed no detectable TNF-a at theprotein level in the BM biopsies. A significant differ-ence in the biological characteristics of patients withany detectable TNFa levels was noted in comparisonwith patients who were entirely negative. The mostsignificant difference was related to the level of apopto-sis as already mentioned, TNF levels being high whenISEL was high (n=18, P=0.013) and low when ISEL

R.

Goyal

etal./

Leukem

iaR

esearch23

(1999)357

–364

360

Table 1Clinical/biological characteristics of patients with hypocellular myelodysplastic syndromes

Sex Age FAB Cellularity BM-Bx-blasts* BM-Asp-blasts WBC HGB Platelet LI ISEL TNF-a TGF-b EBM-11

M1. 55 RA 20 3 0 2.1 13.1 143 Ne 4 8 2 6M 55 RA 20 5 0 2.42. 9.6 34 11 0 0 3 0M3. 50 RA 20 5 0 3.9 7.7 31 Ne 8 8 4 0M 62 RA Hypo 5 5 5.54. 9.0 131 Ne Na Ne Ne NeM 70 RA 10-20 5 5 4.55. 8.5 18 Ne Na 3 Na 4M 59 RA Hypo 5 5 1.66. 9.2 225 26 8 4 1 1M 69 RA 20 5 5 3.0 10.7 3977. Ne Na 5 7 6M 75 RA 20 5 5 1.98. 9.5 20 Ne 2 2 4 0

9. M 58 RA Hypo 5 3 4.0 10.9 54 27 1 2 5 310. M 59 RA 10-20 5 1 4.4 11.2 237 Na 4 1 1 1

F 77 RA 5 5 5 3.911. 9.1 131 Ne 0 0 Ne 012. F 53 RA 30 1 2 3.7 8.9 856 Na 4 8 8 2

F 49 RA 10 5 0 5.413. 9.2 292 21 2 3 4 4F 76 RA Hypo 5 0 1.114. 10.1 14 Ne 0 0 0 NeF 77 RAEB-t Hypo 25 5 2.415. 9.9 224 18 8 4 2 0M 46 RAEB 10 16 31 1.7 13.3 225 28 016 1 1 4F 79 RAEB 20 8 1 5.017. 10.7 309 37 0 0 5 5

18. F 47 CMMoL 10 10 22 0.9 7.4 10 24 8 8 4 619. M 58 RA 10 5 0 4.2 10.4 10 11 0 0 3 0

M 66 RA 5 2 2 1.020. 7.9 33 19 Na 0 0 021. M 81 RA Hypo 5 0 3.3 9.0 197 14 8 8 2 7

M 83 RA 10 5 5 5.922. 9.7 72 23 1 6 3 0F 55 RA 20 5 0 3.123. 12.6 225 24 1 3 4 3F 53 RA 30 5 2 2.424. 7.9 10 Ne 0 Na Ne 0M 67 RAEB 10-20 35 6 1.6 9.4 3225. Na 2 Ne 0 4M 50 RAEB 10-20 25 5 2.926. 12.8 193 Ne 1 5 8 0

27. M 66 RAEB 20 9 5 8.6 8.1 628 29 0 7 0 428. M 57 RAEB 20 5 10 1.7 7.8 12 12 2 3 7 0

F 16 RAEB Hypo 5 5 0.829. 9.0 33 11 5 2 3 3M30. 60 RAEB-t 20 33 30 1.1 10.0 33 Na 5 0 0 0F 74 RA 20 0 0 4.431. 9.1 77 Na Na Na Na NaF 72 RA 5 1 1 2.4 9.2 3932. Na Na 0 Ne 0F 77 RARS 20 5 5 4.2 10.5 28033. Na Na 0 1 3

* BM-Bx-blasts, % blasts in bone marrow biopsy; Ne, not evaluable; BM-Asp-blasts, % blasts in bone marrow aspirate; LI, labeling index (% S-phase cells); WBC, cells×109/l; TNF-a: tumornecrosis factor alpha; Hgb, hemoglobin in gm/dl; TGF-b, transforming growth factor beta; Platelets, platelets×109/l; EBM-11, surface antigen for macrophage; Na, not available for evaluation.

R. Goyal et al. / Leukemia Research 23 (1999) 357–364 361

Table 2Biologic correlates of TNF-a in patients with hypocellular myelodysplastic syndromes

TNF-negative TNF-positive P

n nMedianMedian

6 3 18 0.013ISEL* 00.061947TGF-b* 1

3 20EBM-11* 0 8 0.06

* ISEL, in situ end labeling of fragmented DNA; TNF-a, tumor necrosis factor alpha; TGF-b, transforming growth factor beta; EBM-11,surface antigen for macrophage 11.

was negative (P=0.013, n=6). TGF-b was similarlyhigh (median=4, n=19) in the TNF-positive groupand low (median=1, n=7) in the TNF-negativegroup (P=0.06). Macrophages were quite prominentin the high TNF-positive/TGF-b-positive group (me-dian=3, n=20) and low in the TNF-negative/TGF-bnegative group (median=0, n=8). These relation-ships were not statistically significant (P=0.06). Thesummary of the data is shown in Table 2.

3.3. Cytogenetic studies

Karyotypes were determined in all 33 patients.Thirteen patients had normal chromosomes and 20had an abnormal karyotype. Among these 20 pa-tients, 13 had abnormalities of chromosomes 5 (n=9)and 7 (n=4), one had del (20)(q11 q13) and threehad trisomy 8. The remaining three patients had vari-able anomalies listed in Table 3 (del 16,+15, del 9,del 15,+21 and marker chromosome). There were nostatistically significant differences between individualswho had normal cytogenetics versus those who hadany abnormal metaphases. However, when the subsetof hypocellular MDS cases who presented with ab-normalities of chromosomes 5 or 7 were compared toall other patients (normal as well as any other kary-otypic abnormality), two statistically significant differ-ences were noted.Firstly, the platelet count was higher for patients with5 or 7 abnormalities (median 247×109/l, n=13 vs.101×109/l, n=20, P=0.026) as shown in Table 4.Secondly, this subgroup of patients also had a statis-tically higher level of apoptosis in their bone marrowwhen compared to normal plus any chromosomal ab-normality except that of 5 or 7 group (medianISEL=4.0, n=10 vs. median 1.0, n=16, respec-tively, P=0.038). These data are shown in Table 4.

4. Discussion

Hypocellular myelodysplastic syndromes have beenrecognized as a distinct variant of MDS however, no

differences in the natural history, prognosis or re-sponse to treatment have been appreciated betweenhypocellular versus normo/hypercellular MDS [11–15]. The main source of confusion continues to be thedifficulty of discriminating hypocellular MDS fromaplastic anemia patients, in the absence of chromoso-mal abnormalities, however recent suggestions to treatthese hypocellular MDS patients with immunosup-pressive aplastic anemia-like therapies have resulted inmodest successes thereby obviating the need to cate-gorically make this distinction in every case [16]. Thepresentation of an identical clinical syndrome in thesetting of such starkly contrasting bone marrow cellu-larity as seen in hypo versus hypercellular MDS how-ever, poses an interesting biological challenge. Thepresent study was therefore undertaken to determinewhether the recently recognized biological characteris-tics commonly associated with hypercellular MDS arealso a consistent feature in patients who present withthe hypocellular variant. What we found in this studyof 33 hypocellular MDS cases was in fact quite simi-lar to what has already been described for the morefrequently encountered hypercellular variety. In sum-mary, 2/3 of these patients demonstrated the presenceof intramedullary apoptosis accompanied by higherlevels of TNF-a, TGF-b and macrophages when com-pared with 1/3 of the group that did not have apop-tosis. Profundity of anemia was related to highapoptosis and patients with abnormalities of chromo-somes 5 or 7 appeared to have more apoptotic cellsin their marrows compared to patients who had anormal karyotype and cytogenetic abnormalities af-fecting chromosomes other than 5 and/or 7.

The paradox of variable cytopenias despite hypercel-lular BMs in MDS in general has been explained on thebasis of excessive intramedullarly cytokine-induceddeath of hematopoietic cells [5–7,17]. Obviously, thiscannot apply universally to all MDS patients since atleast 1/3–1/2 of MDS cases do not show evidence ofexcessive apoptosis [17]. In those patients who have ahypercellular BM but no evidence of apoptosis, it isconceivable that the rapidly proliferating hematopoieticcells are being retained in the marrow for abnormally

R. Goyal et al. / Leukemia Research 23 (1999) 357–364362

Table 3Cytogenetic characteristics of patients with hypocellular myelodysplastic syndromes

Cytogenetics ISEL*GMA TNF-alpha*

472/95 46, XY 0 046, XY408/96 0 0

419/97 46, XY na 08191/93 846, XY

46, XY236/97 1 6176/96 46, XX 1 3

46, XX337/96 0 na46, XY506/97 2 ne

390/97 46, XY 1 546, XY 7023/9546, XY238/96 2 3

502/95 46, XX 5 205163/95 46, XY

46, XY, del(5)(q13q33)[19]/46, XY[1]191/97 4 1194/97 46, XX, del(5)(q22)[4]/46, XX [16] 0 0

46, XX[3]/47-48,X,-X,del(5)(q?12q21) 8 4234/93198/98 0na46, XX[19]/NCA:46, XX,del (5) (?q15?q35) [1]

4 846,XX, del(5)(q15q33)[18]/47, XX,+8[2]55/9846,XX,del(5)(q31q33)[26]/46,XX[2]/ NCA 47,XX,+8 2 3161/97

00329/94 46,XX[2]/46,XX,del(5)(q22q35)[6]/47,XX,+8[2]na46, XX, del(5)(q15q33)[18]/47, XX,+8[2] na208/98546,XY,del(5)(q23q33),del(9)(q21q32)[10]/ 46,idem,?del(20)(q13.1)[4]/ 46,XY[5] NCA:46,XY,del(9) (q21q32) na504/97

46,XY,der(7)t(1;7)(q10;p10)[18]/46,XY[2]516/97 4 846,XY,[9]/46,XY,der,(7)t(1;7)(q21;q11)[7]340/94 8 846,XY[6] /46,XY,del(7)(q22q32)[19]5/93 8 4

1113/94 8845xy,-7[7]/46xy[9]0 045,X,-Y[5]/47,XY,+8[10]/47, idem,del(16)(q22)[2]/46,XY[3]211/96

60,XXY,+1,add(2)(q33?),+8,+9x2,+10,+11,(p11.2) +13,+14,+15,+20,add(21)(p11.2),+22x2,+ 0 1107/98mar[21]/ 46,XY,[3]46, XX,del(13) (q12q22)[8]/47, XX,+8[2]/46, XX[10] na 0199/9846, XY, del(20)(11.2q13.1)[20]477/97 na 346,X,-Y,+15 na98/95 ne

278/97 246, XY, del(15)(?q21q24)[2]/46, XY[18] 2384/96 247,XY,+21[3]/46,XY[17] 1

* na, not available for evaluation; ne, not evaluable; ISEL, in situ end labeling of fragmented DNA; TNF-alpha, tumor necrosis factor alpha.

long periods of time. Overexpression of adhesionmolecules could result in such a probability. The situa-tion of hypocellularity in MDS on the other hand,could be explained by excessive apoptosis which is notmatched by an equal rapidity of cell-birth therebytipping the balance in favor of cellular destruction.Hypocellularity in the absence of apoptosis is a moreintriguing challenge and this small group of patients

may be more like the aplastic anemia patients in thatthey suffer more from a stem-cell failure disorder ratherthan apoptotic death of their progeny following rapidproliferation. The description of MDS progenitor cellas having a proliferation defect is compatible with suchhypocellular cases and no evidence of apoptosis. In thepresent study, of the eight patients who showed noapoptosis by ISEL, the LI was low (11%, 11%) for twoand high (28, 37 and 29%) for three patients while itcould not be reliably measured in three patients due totechnical problems posed by the extremely hypocellularmarrows. In other words, there are no unique prolifera-tive properties associated with this group of MDSpatients who have a hypocelluar BM with no detectableapoptosis. It must be noted here that only one tech-nique of in situ end labeling was utilized in the presentstudy to document apoptosis. It is equally probablethat these patients also have excessive apoptosis whichis not detectable by the tests used. If only large molecu-lar weight DNA segments were produced for example,

Table 4Relationship between cytogenetics and apoptosis

Platelet (mean9ISEL* (median)SEM)

247.1959.8Abnormalities of Chromo- 4 (n=10)(n=130)some 5/7

Normal/others 1 (n=16) 100.7932.5(n=20)

P value 0.0260.038

* ISEL, in situ end labeling for fragmented DNA.

R. Goyal et al. / Leukemia Research 23 (1999) 357–364 363

then ISEL as well as gel electrophoretic techniqueswould fail to detect these and a pulsed field elec-trophoretic assay would be required. In fact, we havepreviously demonstrated in a group of eight patientswho showed no evidence of apoptosis either by ISELor gel electrophoresis, six out of eight showed thepresence of high molecular weight DNA fragmentswhen studied by pulsed field electrophoresis [18]. Insummary therefore, all we can say is that eight pa-tients with hypocellular MDS did not show ISEL pos-itivity as a sign of apoptotic death of hematopoieticcells in their bone marrows. On the other hand, themajority of hypocellular MDS cases have high cell-birth, high cell-death and high pro-apoptotic cytoki-nes very similar to the majority of hypercellular MDScases. Thus, the fact that cases of hypocellular MDShave a similar clinical course and responsiveness totherapy as hypercellular MDS is matched by the bio-logic characteristics measured here.

In the present study, two other associations are ofinterest. First, patients who had evidence of apoptosisin their marrows showed a statistically significant in-verse relationship with hemoglobin levels. We havenot shown such a correlation in the hypercellular vari-ety of MDS patients. Second, patients with abnormal-ities of chromosomes 5 and/or 7 appeared to have ahigher incidence of apoptosis in their bone marrowsthan those patients who showed either a normal kary-otype or some other cytogenetic abnormality. Onceagain, we have not found this association betweenabnormalities of 5 and/or 7 chromosomes and apop-tosis in the more commonly encountered hypercellularvariety of MDS. Both of the above observations arestatistically significant only in the hypocellular MDSindicating that hypocellular MDS cases are more ho-mogeneous in their characteristics than are hypercellu-lar patients. This should come as no surprise. Withineach category of MDS, there exists remarkable het-erogeneity, an obvious example being the cellularity ofthe marrow. By separating the hypocellular MDS pa-tients with specific karyotypic abnormalities, the pop-ulation being analyzed is clearly far more ‘pure’ orless heterogeneous and therefore the relationships be-tween biological parameters are far more significant.In the case of hypercellular MDS, even if we separatepatients by karyotype, it is possible that in some indi-viduals it is the erythroid precursors that are con-tributing to the cellularity in a major way as opposedto others in whom myeloid or megakaryocytic precur-sors may be more prominent. Thus, we would have afar more heterogeneous marrow in a hypercellular set-ting as compared to a hypocellular situation. The re-sulting weak or no relationships between biologicalcharacteristics are only to be expected in such cases.

In conclusion, a group of 33 hypocellular MDScases were studied for a variety of biological and clin-

ical characteristics. The resulting insights should helpus understand the genesis of cytopenias in MDS bet-ter. The ultimate hope of course is that such an im-proved understanding will be translated into bettertherapies for our patients.

Acknowledgements

The authors wish to thank the nurses of the Hema-tology Division, Rush Cancer Institute for their caringand thoughtful attitude towards protocol patients andSandra Howery and Lakshmi Venugopal for their ex-cellent administrative/secretarial assistance.

References

[1] Vallespi T, Imbert M, Mecucci C, Preudhomme C, Fenaux P.Diagnosis, classification and cytogenetics of myelodysplasticsyndromes. Hematologica 1998;83(3):258.

[2] Karp JE. Molecular pathogenesis and targets for therapy inmyelodysplastic syndrome (MDS) and MDS-related leukemias.Curr Opinion Oncol 1998;10:3.

[3] Clark DM, Lampert IA. Apoptosis is a common histopatholog-ical finding in myelodysplasia: the correlate of ineffective hema-topoiesis. Leuk Lymph 1990;2:415.

[4] Yoshida Y. Apoptosis may be the mechanism responsible forthe premature intramedullary cell death in myelodysplastic syn-dromes. Leuk 1993;7(1):144.

[5] Raza A, Gezer S, Mundle S, Gao X-Z, Alvi S, Borok R, et al.Apoptosis in bone marrow biopsy samples involving stromaland hematopoietic cells in 50 patients with myelodysplasticsyndromes. Blood 1995;86(1):268.

[6] Raza A, Gregory SA, Preisler HD. The Myelodysplastic syn-dromes in 1996: complex stem cell disorders confounded bydual actions of cytokines. Leuk Res 1996;20(11/12):881.

[7] Koike M, Ishiyama T, Tomoyaso S, Tsurooka N. Spontaneouscytokine over-production by peripheral blood mononuclear cellsfrom patients with myelodysplastic syndromes and aplastic ane-mia. Leuk Res 1995;19(9):639.

[8] Shetty V, Mundle S, Alvi S, Showel M, Broady-Robinson L,Dar S, et al. Measurement of apoptosis, proliferation and threecytokines in 46 patients with myelodysplastic syndromes. LeukRes 1996;20(11/12):891.

[9] Verhoef GEG, DeSchouwer P, Ceuppens JL, Van Damme J,Goossens W, Boogaerts MA. Measurement of serum cytokinelevels in patients with myelodysplastic syndromes. Leukemia1992;6(12):1268.

[10] Bennett JM, Catovsky D, Daniel MT, Flandrin G, GaltonDAG, Gralnick HR, Sultan C. Proposals for the classificationof the myelodysplastic syndromes. Br J Haematol 1982;51:189.

[11] Nagai K, Kohno T, Chen YX, Tsushima H, Mori H, Naka-mura H, Jinnai I, Matsuo T, Kuriyama K, Tomonaga M,Bennett JM. Diagnostic criteria for hypocellular acuteleukemia: a clinical entity distinct from overt acute leukemiaand myelodysplastic syndrome. Leuk Res 1996;20(7):563.

[12] Saad ST, Vassallo J, Arruda VA, Lorand-Metze I. The role ofbone marrow study in diagnosis and prognosis of myelodys-plastic syndrome. Pathologica 1994;86(1):47.

[13] Toyama K, Ohyashiki K, Yoshida Y, Abe T, Asano S, HiraiH, Hirashima K, Hotta T, Kuramoto A, Kuriya S. Clinical andcytogenetic findings of myelodysplastic syndromes showing

R. Goyal et al. / Leukemia Research 23 (1999) 357–364364

hypocellular bone marrow or minimal dysplasia, in compari-son with typical myelodysplastic syndromes. Int J Hematol1993;58(1–2):53.

[14] Kampmeier P, Anastasi J, Vardiman JW. Issues in thepathology of the myelodysplastic syndromes. Hematol OncolClin North Am 1992;6(3):501.

[15] Kitagawa M, Kamiyama R, Takemura T, Kasuga T. Bonemarrow analysis of the myelodysplastic syndromes: histologi-cal and immunohistochemical features related to the evolutionof overt leukemia. Virchows Arch B Cell Pathol Incl MolPathol 1989;57(1):47.

[16] Elghetany MR, Hunduall SD, Gardner FH. Peripheral blood

picture in primary hypocellular refractory anemia and idio-pathic acquired aplastic anemia: an additional tool for differ-ential diagnosis. Haematologica 1997;82(1):21.

[17] Raza A, Mundle S, Shetty V, Alvi S, Chopra H, Span L, etal. A paradigm shift in myelodysplastic syndromes. Leukemia1996;10:1648.

[18] Alvi S, Borok R, Mundle S, Gao X-Z, Shetty V, Hines C,Robin E, Rifkin S, Klein M, Alston D, Hernandez B, HsuW-T, Gezer S, Gregory S, Raza A. Detection of high molec-ular weight DNA as evidence of apoptosis in myelodysplasia.Proc of the American Society of Hematology 37th AnnualMeeting. Blood 1995, 86 (10), 334a, Abst c1324.

.