British Journal of Huematofogy. 1988, 68, 159-164

Enrichment of haemopoietic progenitor cells from the marrow of patients with myelodysplasia

PAUL BAINES, G I L L I A N S . MASTERS, CHRISTOPHER LUSH A N D ALLAN JACOBS Department of Haernatology, University Hospital of Wales, Cardiff

Received 19 May 1987: accepted for publication 10 August 1987

Summary. Myeloid and erythroid progenitors from myelo- dysplastic marrows have been separated from accessory cell populations likely to influence their in-vitro growth. Myeloid colony-forming cells were enriched 2 3-fold and erythroid progenitors 5-7-fold but, in both cases, retained their abnormal growth characteristics. After enrichment and removal of lymphocytes and monocytes, erythroid burst formation became markedly more dependent on the addition of 5637 bladder carcinoma conditioned medium as an exogenous source of haemopoietic growth factors. This

An abnormality arising within a clone of pluripotential stem cells is responsible for the development of the preleukaemic, myelodysplastic state (Prchal et aI, 1978; Abkowitz et al, 1984; Raskind et al, 1984). Progenitors from the preleukae- rnic clone exhibit ineffective haemopoiesis in v i v a reduced colony-formation in vitro (May et al, 1985) and ultimately replace normal polyclonal progenitors in the marrow, with resultant haemopoietic failure. The most likely cause of this partial differentiation block is malfunction of the progenitor cells’ own biochemistry but other mechanisms may be involved. In particular, accessory and suppressor populations in the marrow influence haemopoiesis (Linch et al, 1985; Mangan et al, 1985, 1986) and, since increasing numbers of these will derive from the abnormal clone, their role needs to be defined.

We have recently described a procedure which enriches progenitors from normal marrow while eliminating potential accessory/suppressor populations (Baines et al, 198 7). We have now applied this procedure to myelodysplastic marrow with two objectives: firstly, to define the role of suppressor/ accessory cells in this disorder and, secondly, to obtain suspensions enriched for abnormal progenitor cells which are otherwise too scarce to study directly.

Correspondence: Dr Paul Baines, Department of Haematology, University Hospital of Wales. Heath Park, Cardiff CF4 4XW.

suggests that lymphocytes and monocytes usually support erythropoiesis in cultures of myelodysplastic marrow and demonstrates that erythroid burst-forming progenitor cells from myelodysplastic patients can respond to these popula- tions in vitro. Haemopoietic failure in myelodysplasia appears to result from a defect within the preleukaemic clonogenic cell, rather than from an aberration in exogenous factors. The procedure outlined here can be used as a first step towards the isolation of these abnormal progenitors.

METHODS

Marrow donors All five patients studied fulfilled the diagnostic criteria detined by Bennett et a1 (1982) Marrow aspirates were obtained during routine diagnostic sampling from the iliac crest or sternum. All had hypercellular marrows.

Patient 1. An 84-year-old male presented in 1986 with anaemia (Hb 5.5 g/dl. MCV 95 fl). White cell and platelet counts were normal. The initial marrow aspirate was consistent with the diagnosis of refractory anaemia (RA) with 1% myeloblasts, 2% erythroblasts and the karyotype was normal. He was transfusion-dependent from diagnosis but received only supportive therapy. At the time of this study (2 1 January 1986) a repeat marrow aspirate showed progression to refractory anaemia with excess blasts (RAEB) with 13% blasts.

Patient 2. A 64-year-old male presented in 1983 with anaemia (Hb 10.4 g/dl) and neutropenia (WCC 1.4 x 109/1: neutrophils 6%). A marrow aspirate indicated KA with 2% myeloblasts and 20% erythroblasts. No specific therapy was given and marrow morphology at the time of this study (1 8 February 1986) showed 3% myeloblasts and 23% erythro- blasts.

Patient 3. A 56-year-old male presented in 1982 with macrocytic anaemia (Hb 11.6 g/dl and MCV 112 fl). A marrow aspirate showed refractory anaemia with ring sideroblasts (RARS); 3% myeloblasts and 28% erythroblasts (45% of which were ring sideroblasts). He received 13-cis

159

160 Paul Baines et a1 retinoic acid, 20 mg/d, from 1984 and became transfusion- dependent in 1985. At the time of this study (30 April 1986) his marrow contained 2% myeloblasts and 2 5% erythroblasts of which 49% were ring sideroblasts. Karyotype was normal.

Patient 4 . An 86-year-old male presented with macrocytic anaemia in 1983 (Hb 7.9 g/dl and MCV 104 fl) and was found to be RARS. His marrow aspirate contained 45% erythroblasts of which 28% were ring sideroblasts. In 1984 he developed a transitional cell carcinoma of bladder (Grade 1) which was treated cystoscopically and a Dukes A carci- noma of rectum was diagnosed in 1986 and treated by abdominoperineal resection. He has received supportive therapy only for his myelodysplastic condition. Serial mar- row aspirations have shown no evidence of tumour infiltra- tion and, at the time of this study (2 October 1986), the cellular composition of the marrow was similar to that at presentation. White cell and platelet counts remained normal throughout.

Patient 5. A 78-year-old female presented in 1983 with anaemia (Hb 8.2 g/dl) and hepatosplenomegaly. A marrow aspirate was diagnostic of RARS. Therapy with 13-cis retinoic acid was started in 1984 and no disease progression or increase in her minimal transfusion requirements has been seen. At the time of this study (2 1 October 1986) her marrow contained 42% erythroblasts, of which 22% were ring sideroblasts, and 3% myeloblasts.

Normal marrows were obtained from the femurs of haematologically normal patients undergoing hip surgery in accordance with protocols approved by the Ethical Com- mittee of the South Glamorgan Health Authority.

Marrow fractionation Marrow samples were collected in 10 ml Hepes-buffered, modified Eagle’s medium (MEM Wellcome, U.K.) containing 15 U/ml preservative-free heparin at 4°C. The cells were immediately centrifuged at 500 g for 10 min, the supernatant discarded and the pellet resuspended in 5 ml MEM. Bone marrow particles were broken down by aspiration through a 2 5-gauge needle. Nucleated cells were separated by centrifu- gation on Percoll (1.075 g ~ m - ~ , Pharmacia) at 400 g for 2 5 min and the interface-layer nucleated cells recovered and washed twice with MEM at pH 7-4. Interface cells were then incubated at 4°C for 30 min with the monoclonal antibody Campath-I or, in later experiments, with a mixture of Campath-1 and 80H.3 monoclonal antibodies (MoAbs). Campath-1 is a rat IgM MoAb with specificity for lympho- cytes and monocytes (Hale et al, 1983) and was kindly donated by Dr G. Hale, Department of Pathology, University of Cambridge, Cambridge, U.K. 80H.3 was a gift from Dr P. Mannoni of the Regional Cancer Centre, Marseille, France, and is a mouse IgG MoAb which reacts with maturing myeloid cells and monocytes (Mannoni et al, 1982). Cells were washed three times with MEM and 2% FCS and incubated for 20 min at 4°C with a mixture of sheep red cells which had been coated with rabbit anti-rat Igs (2147 Dakopatts) or rabbit anti-mouse Igs (2259 Dakopatts) by the method of Ling et a1 (19 77) using chromic chloride. Rosetted and non-rosetted cells were separated over Ficoll-Hypaque (Pharmacia) at 400 g for 30 min. Interface cells (lymphocyte-

and monocyte-depleted) were washed once and aliquots set aside for morphology (Jenner-Giemsa staining) and culture. Erythrocytes in the cell pellet were lysed with water.

Granulocyte-macrophage colony-forming cells (GM-CFCs) These were assayed in Iscoves Modified Dulbecco’s Medium (IMDM Gibco) containing 0.3% Noble Agar (Difco) and supplemented with 1% deionized bovine serum albumin (Sigma), 20% fetal calf serum (Gibco), 5 x M freshly prepared beta-mercaptoethanol (BDH), 5% of IMDM, condi- tioned by 563 7 bladder carcinoma cells in log-phase growth for 7 d in the presence of 5% fetal calf serum, and 1 unit/ml erythropoietin (Epo: TFL-1, Terry Fox Laboratories, Van- couver, Canada). 5637 cells were obtained from the Human Tumor Cell Laboratory, Memorial Sloan Kettering Cancer Center, New York. Unfractionated cells and non-rosetted cells were seeded at 105/ml and 104/ml respectively. Colonies (> 50 cells) and clusters (20-50) were routinely scored under a dissecting microscope after 8 and 14 d of incubation at 3 7OC at 5% C 0 2 in air in a fully-humidified incubator.

Erythroid colony-forming cells (E-CFCs) We used a modification of the technique described by Iscove et a1 (1974). Depending on the fractionation stage, cells were cultured at between 0.25 and 2 x 105/ml in alpha-medium (Flow) containing 0.8% methylcellulose (Dow, A4M Pre- mium), 2.1 g/l bicarbonate (Wellcome), M beta- mercaptoethanol (BDH), lo-’ M sodium selenite (BDH), 10 mg/l vitamin E (Eastman). 100% saturated human transfer- rin (Behringwerke, 10-20 PM), antibiotic/antimycotic solu- tion ( x 100, Gibco), and 2 U/ml human urinary, TFL-1 erythropoietin (Terry Fox Laboratories, Vancouver, Canada). Unless stated otherwise, 563 7 bladder carcinoma condi- tioned medium was added at 2.5% of total culture volume. Pooled, heat-inactivated normal, human AB serum (3 3% v/ v) was added to obtain the final culture mix. Mononuclear cells and non-rosetted cells were seeded at 2 x lo5 and 2.5 x lo4 per ml of culture respectively. Triplicate cultures were incubated at 3 7°C in air with 5% CO2 in a fully humi- dified incubator. Colony forming units-erythroid (CFU-E; 8-50 haemoglobinized cells) were scored after 7 d and burst forming units-erythroid (BFU-E > 50 cells) after 14 d using an inverted microscope.

Calculation of progenitor yields Yields of each progenitor type were calculated as a percent- age of the total number of progenitors in the unfractionated sample loaded onto Percoll. For example:

total GM-CFC in CampathSOH.3 -ve fraction total GM-CFC in UF sample x 100.

RESULTS

Populations removed by fractionation Late erythroblasts and mature myeloid cells (bands and polymorphs) were considerably reduced in the low-density fractions recovered from the interface after Percoll density centrifugation of RA/RAEB (Table I) or RARS (Table 11)

Enriched Myelodysplastic Progenitors 16 1

Table I. Nucleated cell yields and differential counts on RA/RAEB marrows fractionated by immune-rosetting with Campath-1 alone (patients 1 and 2)

Yield (%) Ery (%I W m o (%) Meta/Poly (%I Mo (%I Lym (%I UIB (%)

Patient: 1 2 1 2 1 2 1 2 1 2 1 2 1 2 ~ _ _ -.

(RAEB) (RA) ~ ~~ ~

Unfractionated* 100 100 0 5 74 49 20 21 2 5 4 2 0 0 0 Percoll-interface 1 1 41 3 16 41 48 7 9 16 14 35 13 0 0 Campath -ve 0.5 8.9 1 2 93 90 1 3 3 3 2 2 0 0

Abbreviations: Ery: erythroid: Mylmo: myelocytes, promonocytes and earlier myeloid and monocytic forms: Meta/

* Normalized to 100%. Cytospin preparations of cells were prepared after each fractionation step, stained with Jenner- Poly: metamyelocytes/polymorphs: Mo: monocytes: Lym: lymphocytes: UIB: unidentified blasts.

Giemsa and scored.

Table 11. Nucleated cell yields and differential counts on RARS myelodysplastic marrow fractionated by immune-rosetting with both Campath and 80H.3 (patients 3-5)

Yield nucl. cells Ery Mylmo Meta/Poly Mo Lym UIB (%) (%I (%I (%) (%I (%) (%I

~- ~

Unfractionated 100 21f 9 1 3 f 4 6Of13 4 + 2 2 f 2 0 . 3 f 0 . 3 Percoll-interface 37+7 42f14 1 6 f 7 2 6 f 8 1 3 f 1 3 + 2 0 Campath/80H.3 -ve 3 f 0 . 2 4 2 f 1 0 4 1 f 9 7 f 5 3 f 2 4 f 2 4 1 2

Legend as for Table I , n = 3 , blasts in Campath/SOH.3 -ve fraction=9.4+3.4.

marrows and the proportion of monocytes and lymphocytes increased (except in RARS marrows where mature lympho- cytes were scarce). Mature monocytes and lymphocytes were eliminated during the immune-rosetting procedure with either Campath-1 alone or with combined Campath-1/8OH.3 treatment: only immature forms of these cells remained. Hypogranular myelocytes predominated in Campath - ve fractions from RA/RAEB marrows and megaloblastoid ery- throblasts co-predominated with hypogranular myelocytes in the Campath/80H.3 -ve cells from RARS marrows. Blasts were scarce in RA cell suspensions following immune depletion with Campath-1 alone, but formed 9% of all cells present in RARS suspensions following immune rosetting with both Campath and 80H.3.

Progenitor growth in fractionated marrow Myeloid progenitor growth is tabulated only for those patients (3-5) whose marrows were immune-rosetted follow- ing combined Campath-1 /80H.3 treatment.

Granulocyte-macrophage cluster-forming cells (GM-Clus FCs) and colony-forming cells (GM-CFCs). scored after 14 d of culture, were enriched 16- and 23-fold, respectively in the non-rosetted fraction following MoAb treatment (Fig I ). Data for myeloid cluster and colony-forming cells, scored after 8 d of culture, are not tabulated but these were enriched 25- and 29-fold respectively. After 14 d, cluster numbers predomi-

nated over colonies, both in cultures of unfractionated and fractionated marrow. This contrasts with the ratio between the means of clusters and colonies in cultures of unfractio- nated, normal marrow given in Fig 1 (n= 5). Yields of progenitors were 48% and 66% for day 14 myeloid cluster and colony-forming cells and 72% and 78% for day 8 cluster- forming cells. Fractionation did not affect the pattern of response by myeloid clone-forming cells to several concentra- tions of 5637CM.

Since it has proved difficult to score E-CFCs accurately in cultures of unfractionated marrow, enrichments and yields of these progenitors have been calculated using clone numbers in the Percoll-interface suspensions as starting values. Conse- quently, the final values in Fig 2 do not include any enrichment obtained, or reduction in yield incurred, during Percoll fractionation. Our normal ranges for CFU-E and BFU- E per lo5 mononuclear cells (MNC), prepared from the Percoll-interface, are given in Fig 2 and this shows that erythroid progenitor numbers are severely reduced in the five patients studied. The proportion of both CFU-E and BFU-E increased after immune-rosetting but this increase was no greater than that expected from the physical removal of so many nucleated cells and in in no case did progenitor yields exceed 100%. In patient 2, E-CFC yields did approach lOOo/, but this patient formed very few colonies and subsequent calculations are unreliable. Clone size and degree of haemo-

162 Paul Baines et a1

"OO.

200.

100-

0 .

10,000

u)

0 t 1000,

z u)

\

E - .- C Q ul

P 2 100.

10

-1 00 3 1 -100

I

6,-; jj8... 0

a UF cells 0 Enriched cells

96 progenitor yield in enriched cells + Normal UF mean 2 SE 16x

T 1 h -

GM-CLUS FC

23x

F 1': -

GM-CFC

Fig I . Day-14 myeloid cluster and colony growth after fractionation of myelodysplastic marrow from patients 3-5. GM-Clus FC = granu- locyte-macrophage cluster-forming cells: GM-CFC = rnyeloid colony- forming cells: UF = unfractionated marrow cells: Enriched cells = cells enriched for progenitors by immuno-rosetting; % yield progeni- tors=% of all progenitors initially loaded on to Percoll recovered in enriched suspension: 0 =mean (i 1 SE) of progenitor cell numbers/ lo5 unfractionated marrow cells from five normal donors.

globinization were frequently reduced in cultures of myelo- dysplastic marrow compared to normal marrow cultures. These parameters were not visibly altered following fractio- nation.

Requirement of BFU-Efor exogenous CM after marrow fractionation In epo-supplemented cultures of marrow suspensions de- pleted of accessory cells by Campath-1 or Campath- 1 + 80H.3, the addition of conditioned medium significantly increased erythroid burst numbers per lo5 nucleated cells from a mean of 1.4 to 18 (medians = 0 and 20 respectively: P= 0.016; Mann-Whitney non-parametric test). This con- trasts with epo-supplemented cultures of marrow cells taken before rosetting (Fig 3 ) where BFU-E numbers were not increased when conditioned medium was added as an exogenous source of growth factors. CFU-E colony numbers were not significantly increased by the addition of condi- tioned medium, either before or after fractionation. GM-CFC growth required conditioned medium, even in unfractio- nated marrow cultures (data not shown).

DISCUSSION

We have eliminated fully-differentiated myeloid and lym- phoid populations from myelodysplastic marrow as effec- tively as we have done from normal marrow (Baines et al, 1987).

CFU-E EFU-E

Patient1

Patient 2

Early myeloid cells predominated in our final, non-roset- ted, suspension and, although these were hypogranular, this does not seem to reflect a myelodysplastic abnormality since the myelocytes which predominated in similarly-prepared suspensions from normal marrow were poorly-granulated too (Baines et al, 1987). Possibly we select a subpopulation of hypogranular, early myeloid cells using this procedure but it seems more likely that these cells degranulate during the separation procedure. The high proportion of erythroid cells recovered following combined Campath-1/80H.3 MoAb treatment and immune rosetting of RARS marrow reflects the dyserythroid nature of these cells since many of these are low-density and are retained at the Percoll-interface: these cells will not be removed by combined MoAb immune- rosetting.

Blast forms were undetectable in the Campath -ve cells prepared from RA/RAEB marrows but formed 9% of all cells present in the Campath/80H.3 -ve cells recovered from RARS marrow. This figure, though, is still below that obtained for Campath/80H.3 - ve suspensions from normal

Enriched Myelodysplastic Progenitors 16 3

monocytes and a considerable proportion of lymphocytes could be clonal in origin, although evidence for clonally- derived lymphocytes, using genetic marker studies, is scarce and inconsistent (Kere e t al, 198 7; Prchal et al, 19 78). Despite subtle numerical and functional changes in T lymphocyte subpopulations (Bynoe et al. 1983; Knox et ul, 1983: Kerndrup eta!, 1984), our data support those of Greenberg & Mara (1 9 79) and Merchav ef al(1987) which suggested that accessory cells in myelodysplasia continue to produce hae- mopoietic growth factors. Francis et a1 (1983) have also reported that endogenous marrow CSA production was detectable in the majority of MDS patients, although the source of this activity was not defined and the CSA level varied considerably from patient to patient.

Overall, our data suggest that progenitors from myelodys- plastics receive, and can respond to, support from some elements of the marrow environment. Although other stromal elements may inhibit progenitor growth in myelo- dysplastic marrow, it now seems more probable that the haemopoietic failure results from an intracellular defect in the aberrant clone. In patients with hypercellular marrows, this defect is not manifest until differentiation is well advanced since the presence of large numbers of clusters suggests that limited myeloid proliferation and differentiation does occur. Further, in our sideroblastic patients, large numbers of haemoglobinized cells were present in the marrow suggesting that erythroid differentiation had pro- ceeded thus far. RARS patients, in particular, display con- siderable ineffective erythropoiesis (Barosi eta!, 1978) which suggests that erythroid growth fails in vitro, not because erythroid progenitors are absent, but because they fail to proliferate and mature to a stage where haemoglobinized clusters or colonies can be observed.

Finally, determination of the defect involved may necessit- ate direct investigation of progenitor cell metabolism which will be difficult since progenitor numbers are so low in marrow. The enrichment procedure described here may be a useful first step in the purification of progenitors from myelodysplastic patients in preparation for these kinds of studies.

8 0)

0 .3

8 * a 8 * * o

cm

- I €

CFU-E

0 -CM tCM

i

a

0x8 n

O O O J X )

I E

BFU-E

- Fig 3. Requirement of CFU-E and BFU-E for exogenous conditioned medium after fractionation of myelodysplastic marrow. Legend as for Fig 2.

marrow (22%) (Baines et al. 1987) and this is probably a consequence of the accumulation of cells of intermediate maturation which will dilute more primitive forms in these hypercellular marrows. In addition, since these cells tend to be retained at the Percoll interface, this will reduce the proportion of blasts in the low-density fraction still further.

Since progenitors were enriched in these immune-depleted suspensions, we were able to seed fewer cells per culture, so reducing the lymphocyte/monocyte burden. Despite this, defective myeloid and reduced erythroid pi-ogenitor growth was still apparent. Myeloid cluster-forming cells still predomi- nated over colony-forming cells and we did not recover more CFU-E or BFU-E than were expected, following fractionation, which would have indicated suppression. On the contrary, after removal of monocytes and lymphocytes, the require- ment of BFU-E for exogenous burst promoting activity (BPA) increased. There is no evidence that the antibodies used here might remove BFU-E with only a low requirement for BPA, thus enriching preferentially for BFU-E with a high BPA requirement. It seems more probable that lymphocytes and monocytes support erythropoiesis in cultures of myelodys- plastic marrow in the same way as they do in cultures of normal marrow (Baines et af, 1987; Linch eL al, 1985). although whether they do this as efficiently is difficult to ascertain accurately since erythroid growth is so poor in cultures of myelodysplastic marrow. If accessory cell function was only partially impaired, this could cause a gradual fall in clonogenic cells in these patients.

It may take 6-7 years for the malignant clone in chronic myeloid leukaemia to dominate haemopoiesis (Kamada & Uchino, 19 78; Andreef, 1986) and the myelodysplastic clone is unlikely to manifest itself earlier. Over a decade, most

ACKNOWLEDGMENT

This work was supported by grants from the Leukaemia Research Fund.

REFERENCES

Abkowitz, J.L., Fialkow, P.J., Niebruge, D.J., Raskind. W.H. & Adamson, J.W. (1984) Pancytopenia as a clonal disorder of a multipotent hemopoietic stem cell. journal of Clinical Investigation, 73, 258-261.

Andreef, M. ( 1 986) Cell kinetics of leukemia. Seminars in Hematology.

Baines, P., Masters, G.S.. Booth. M. &Jacobs, A. (1987) Enrichment of progenitor cells from human marrow. Experimental Hematology. 15, 809-813.

Barosi. G., Cazzola, M., Morandi, S . , Stefanelli. M. & Perugini, S. (1978) Estimation of ferrokinetic parameters by a mathematical

23, 300-314.

164 Paul Baines et a1

model in patients with primary acquired sideroblastic anaemia. British Journal of Haematology, 39, 4 0 9 4 2 3 .

Bennett, J.M., Catovsky, D.. Daniel, M.T., Flandrin, G.. Galton, D.A.G.. Gralnick, H.R. & Sultan, C. (1982) Proposals for the classification of the myelodysplastic syndromes. British Journal of Haematology, 51, 189-1 99.

Bynoe, A.G., Scott, C.S.. Ford, P. &Roberts, B.E. (1983) DecreasedT- helper cells in the myelodysplastic syndromes. British Journal of Haematology, 54, 97-102.

Francis, G.E., Miller, E.J., Wonke. B., Wing, B.A., Berney. J.J. & Hoffbrand, A.V. (1983) IJse of bone marrow culture in prediction of acute leukaemic transformation in preleukaemia. Lancet, i,

Greenberg, P.L. & Mara. B. (1979) The preleukemic syndrome: correlation of in-vitro parameters of granulopoiesis with clinical features. American Journal of Medicine, 66, 951-958.

Hale, G.. Bright, S . , Chumbley. G.. Hoang. T., Metcalf, D.. Munro, A.J. & Waldmann. H. (1983) Removal of T cells from bone marrow for transplantation: a monoclonal antilymphocyte antibody that fixes human complement. Blood. 6 2 , 873-882.

Iscove, N.N.. Sieber, F. & Winterhalter, K.H. (1974) Fxythroidcolony formation in cultures of mouse and human bone marrow: analysis of the requirement for erythropoietin by gel filtration and affinity chromatography on agarose concanavalin. American Journal of Cellular Physiology, 83 , 309-322.

Kamada. N. & Uchino. H. (1978) Chronologic sequence of clinical and laboratory findings characteristic of chronic myelocytic leukemia. Blood. 51, 843-850.

Kere, J., Ruutu, T. & de la Chapelle, A. (1987) Monosomy 7 in granulocytes and monocytes in myelodysplastic syndrome. New England Journal of Medicine, 316, 499-503.

Kerndrup. G.. Meyer. K., Ellegaard. J. & Hokland, P. (1984) Natural killer (NK)-cell activity and antibody-dependent cellular cytotoxi- city (ADCC) in primary preleukaemic syndrome. Leukaemia Research, 8, 239-247.

Knox. S.J.. Greenberg. B.R.. Anderson, R.W. & Rosenblatt. L.S. (1983) Studies of T-lymphocytes in preleukemic disorders and acute non-lymphocytic leukemia: in-vitro radiosensitivity. mito-

1409-1412.

genic responsiveness, colony formation and enumeration of lymphocytic subpopulations. Blood, 6 1 , 4 4 9 4 5 5 .

Linch. D.C., Lipton, J.M. &Nathan, D.G. (1 985) Identification ofthree accessory cell populations in human bone marrow with erythroid burst promoting properties. Journal of Clinical Investigation, 75,

Ling. N.R., Bishop, S.A. & Jefferis, R. (1 977) Use of antibody-coated red cells for the sensitive detection of antigen and in rosette tests for cells bearing surface immunoglobulins. Journal of Immunological Methods, 15, 279-289.

Mangan, K.F.. Zidar. B., Shadduck, R.K., Zeigler, Z. & Winkelstein. A. (1 985) Interferon-induced aplasia: evidence for T cell mediated suppression of hematopoiesis and recovery after treatment with horse antihuman thymocyte globulin. American Journal of Hemato- logy, 19, 401-413.

Mangan. K.F.. Volkin, R. & Winkelstein, A. (1986) Autoreactive suppressor cells in the pure red cell aplasia associated with thymoma and panhypogammaglobulinemia. American Journal of Hematology, 23 , 167-173.

Mannoni, P.. Janowska-Wieczorek, A., Turner, A.R.. McGann. L. & Turc, J.M. (1982) Monoclonal antibodies against human granulo- cytes and myeloid differentiation antigens. Human Immunology, 5,

May, S.J., Smith, S.A., Jacobs, A., Williams, A. & Bailey-Wood, R. (198 5 ) The myelodysplastic syndrome: analyses of laboratory characteristics in relation to the FAB classification. British Journal ofHaematology, 59, 311-319.

Merchav. S . , Nagler, A., Sahar. E. &Tatarsky, I. (1987) Productionof human pluripotent progenitor cell colony stimulating activity (CFU-GEMM CSA) in patients with myelodysplastic syndromes. LPukaemia Research, 11, 273-279.

Prchal. J.T., Throckmorton, D.W., Carroll, A.J.. Fuson. E.W.. Gams, R.A. & Prchal. J.S. (1978) A common progenitor for human myeloid and lymphoid cells. Nature, 274, 590-591.

Raskind. W.H., Tirumali, N., Jacobsen, R., Singer, J. & Fialkow. P.J. (1984) Evidence for a multistep pathogenesis of a myelodysplastic syndrome. Blood, 63, 1318-1323.

12 78-1284.

309-320.