British Jnurnnl of Haematnlngy, 1993, 85, 20-24

Haemopoietic CD34 + progenitor cells are not infected by HIV-1 in vivo but show impaired clonogenesis

ANDREA DE LUCA, LUCIANA TEOFILI,* ANDREA ANTINORI, MICHELA STEFANIA IOVINO,* PAOLA MENCARINI, ELENA VISCONTI, ENRICA TAMBURRINI, GIUSEPPE LEONE* A N D LUIGI ORTONA Istituti di Clinica delle Malattie Infettive e di *Semeiotica Medica, Universita Cattolica del Sacro Cuore. Rorna, Italy

RPceived 22 March 1993; acceptedfor publication 26 March 1993

Summary. We evaluated the role of CD34+ bone marrow progenitor cells in vivo, in the pathogenesis of AIDS-related haematological abnormalities. The clonogenic activity of CI>34+ cells from seven patients with HIV-1 infection, without bone marrow involving opportunistic infections or neoplasms, was assessed in semisolid cultures. The number of CFU-GM was significantiy reduced as compared to the controls ( P = 0.01 7). independently from myelotoxic ther- apy, while the number of BFU-E was not. The presence of retroviral sequences in CFU-GM colonies from four patients and in the total population of CD34+ cells from six patients

with advanced stage HIV infection was investigated using the polymerase chain reaction. The presence of HIV- 1 sequences was also searched for in a purified suspension of CD34 + cells after 3 weeks liquid culture. All these cells were always HIV-1 negative, while viral sequences were always detected in bone marrow mononuclear cells from these and other patients. The number of HIV-1 DNA copies decreased with increasing enrichment. At most 1: 10000 CD34+ cells are infected in vivo. Other mechanisms than direct viral infection of progeni- tor cells must account for the defective haemopoiesis in HIV-I infected patients.

Haematological abnormalities, such as cytopenia of the different lineages and myelodysplasia, are a common feature in patients with HIV-1 infection (Scadden et al, 1989). The pathophysiology of these impairments is incompletely under- stood: several mechanisms have been suggested. Deficient colony generation, by partially purified bone marrow (BM) progenitor cells from AIDS patients, corrected in part after T cell depletion. has been described (Stella et al. 1987). Direct infection of BM precursors has been demonstrated by in situ hybridization (Busch et ul. 1986), and cells derived from purified progenitors have been successfully infected in vitro (Folks et al, 1988). Nevertheless, recent in vivo studies on purified CD34 antigen bearing cell populations, which include committed haemopoietic progenitors and pluripotent stem cells, failed to demonstrate proviral DNA (von Laer et al. 1990; Davis et al, 1991). A number of mechanisms for abnormal haemopoiesis. indirectly induced by HIV-1, have been suggested. These include soluble factors released by BM cells (Leidermann et al, 198 7), anti-HIV-1 antibodies (Dona- hue et af. 1987), cytokines with inhibitory effects (Molina ut al. 1989) or the abnormal expression of haemopoietic growth factors (Scadden et al, 1989).

Correspondence: Dr Andrea De Luca, Istituto di Clinica delle Malattie Infettive, Universita Cattolica del Sacro Cuore, L.go A. Gemelli 8- 00168 Roma, Italy.

We here report the results of a study on the clonogenic activity of CD34 + cells purified from bone marrow and on the presence of HIV-1 DNA sequences in BM derived cells after different steps of purification. including CD34 + cells, from patients with late stage HIV infection.

MATERIALS AND METHODS

Patients. BM aspirates were obtained from the posterior iliac crest of 11 HIV-1 infected patients (HIV+) (age 32 .8f1 .9 years: 10M. 1F) and 1 5 healthy donors (age 29 f 3.4 years: 10M, SF) after informed consent. Peripheral blood parameters of HIV + patients were: haemoglobin 9 .34f0 .45 g/dl: white blood cells 4 .47f1 .23 x 109/1: absolute neutrophil counts 3.17 1 0 . 9 4 x 10y/l; platelets 167&97 x 109/1: CD4 lymphocytes 97.1&50.2 x 106/1. Cells from seven HIV + patients were used in the clonogenic assays (for peripheral blood picture see Table I). BM cells from nine HIV + patients (six from group IV C1, three from group IV C2 of the CDC classification, CD4+ lymphocytes 6- 103 x 10b/l) were used in the PCR experiments. Three of the 11 patients were given zidovudine 500 mg/d; no other potentially myelotoxic drug was administered. Bone marrow involving opportunistic infections or neoplasms were excluded by cytology and microbiological cultures.

Cells. BM cells were purified by a modification of the

20

CD34 + Progenitors and HIV-7 2 1 Table I. Peripheral blood picture of the seven HIV+ patients tested for clonogenic activity of CD34 + bone marrow cells.

Age (yr) CD4 + x 10'/l Hb (g/dl) WBC x 1 Oy/1 ANC x 1 OY/1 Plt x 10y/l

1 27 81 10.4 3.77 2.688 183 2 42 21 8.4 1.65 1.011 41 3 27 13 10.5 1.61 1.320 382 4 25 68 9.5 2.10 1.440 91 5 31 98 10.6 1.84 0.957 117 6 32 6 8 .8 2.70 2.052 120 7 34 19 11.3 2.78 1,028 139

method of Gabbianelli et a1 ( 1 990). Briefly, bone marrow mononuclear cells (BMMC), obtained by centrifugation through a Ficoll-Hypaque density gradient (Pharmacia. Uppsala, Sweden), underwent three plastic adherence cycles and were further enriched in progenitors on a triple gradient Percoll ( d = 1.054, 1.066, 1.077 g/ml) (Pharmacia). Cells with the lowest density (< 1.054 g/ml) (very low density MC, VLDMC) were collected and incubated with an anti-My 10 (CD34) IgGl mouse monoclonal antibody (HPCA-1, Becton Dickinson, San Jose, Calif.). CD34-positive and CD34-nega- tive fractions were separated using sheep anti-mouse IgG 1 coated immunomagnetic beads and a magnetic particle concentrator (both from Dynal, Oslo, Norway). The popula- tion of CD34+ cells was >90% pure as assessed by flow cytometry (FACScan, Becton Dickinson) after detachment from the beads: it did not contain detectable CD4+ or CD14 + cells. BMMC, VLDMC, CD34 - and CD34 + fractions were counted and independently processed.

Colony assays and liquid cultures. CD34+ cells were sus- pended in IMDM (Flow, Irvine, Scotland) with 40% FCS (Flow), 0.9% methylcellulose (Sigma). 5 x lo-' M 2-mercap- toethanol (Sigma), 0.7 ng/ml pure human transferrin satu- rated with FeC13 (Sigma) and following recombinant human growth factors: 3 U/ml erythropoietin (Boehringer. Mann- heim, Germany). 10 ng/ml GM-CSF, 100 U/mlI1-3 (all from Genzyme, Boston, Mass.): in some experiments 100 ng/ml basic fibroblast growth factor (bFGF) (Genzyme), a cytokine with activity on undifferentiated progenitors (Gabbianelli ef al, 1990), was added. The cells in the semisolid medium were plated in duplicate on 35 mm dishes (Falcon, Becton Dickinson, Lincoln Park, N.J.) and incubated at 3 7°C in a 5% COz humidified atmosphere.

After 14 d. CFU-GM and BFU-E colonies with >50 cells were counted under an inverted microscope according to their morphologic appearance and were scored as per cent of viable plated CD34 + cells.

In four patients the CFU-GM were individually picked from the dishes under microscopic visualization using microcapil- lary tips, washed, counted and used for PCR experiments.

CD34+ cells from three patients were seeded at a concen- tration of 20000/ml in IMDM+lO% FCS+10 ng/ml GM-CSF or 10 ng/ml G-CSF (Genzyme) and 10 ng/ml recombinant human stem cell factor (SCF) (Genzyme). After an incubation of 21 d, cells were rescued, counted and processed for DNA amplification.

DNA amplifications. DNA was prepared from cells according to conventional lysis and digestion protocols with 60 pl/ml Proteinase K (Sigma) and nonionic detergents (Kawasaki, 1990). Two sets of primers were used for the HIV-1 DNA polymerase chain reaction: SK38/39 for the gag and SK68/ 69 for the env region. 36 amplification cycles with 30 s at 9 5"C, 30 s at 5 5°C and 1 min at 72OC were performed, with 10 min final extension at 72OC. in a thermal cycler (Perkin Elmer, NorwAlk, Conn.). Amplified products were blotted on a nylon filter (Gene Screen Plus, NEN Du Pont, Boston, Mass.) and hybridized with 5'end 32P labelled SK19 and SK70 probes. After high stringency washes the hybridization products were visualized by overnight exposure to a XAR-5 film (Eastman Kodak, Rochester, N.Y.) at -8OOC with intensifying screens (Abbott et al, 1988). With our assay, we could detect a minimum of 10 DNA copies of HIVZ6 (Perkin Elmer) in a background of 10' PCR negative mononuclear cells. The quality of the DNA in the samples was assessed using HLADQA1 primers (GH26 and GH27) and probe (RH54). Samples without or with a poor HLA-DNA signal were not considered. Rules to avoid sample or reagent contamination were strictly applied (Kawasaki, 1990). Several negative controls containing the different buffers and reagents, except the template, were included in each amplifi- cation run.

Statistical analysis. After an initial analysis of data, the non- parametric Mann-Whitney rank sum test was used to compare results in the clonogenic assays, because of their non normal distribution. The same data were reanalysed after 'normalization' (Edwards, 198 5) using a two-tailed Student t-test. Spearmann's correlation rank coefficient was used to correlate results in the clonogenic assays to CD4 + lymphocytes and peripheral blood cell counts.

RESULTS

Yield of purijkations Purifications of CD34+ cells from 11 HIV+ patients with advanced stage HIV disease yielded a 0.13 &0.05% CD34+ cell population from total BMMC this was not significantly different than the yield from the 15 HIV - subjects, which was 0 .34f0 .09% CD34+ cells (P=0.106). These data were confirmed by FACS analysis which showed a similar content of CD34+ cells in BMMC from two patients and controls (2-5%).

22 Andrea De Luca et a1

Fig 1. Percentage of viable CD34+ cells from seven HIV+ patients and seven controls generating colonies in semisolid medium. Results are expressed as meanfSD. CFU-GM HIV+: 3.71i~5.22; CFU-GM controls: 1 1.20 f 8.18 P= 0.01 7. BFU-E HIV + : 1.2 1 f 1.58: BFU-E controls: 2.2913.27 P>O.O5 (Mann-Whitney U test).

Table 11. Percentage of plated CD34-t cells generating CFU-GM colonies with different growth factors combinations. Means of duplicate cultures from three HIVf patients are shown. Growth factors concentrations as in Methods. Differences between cultures without or with bFGF calculated by the Mann-Whitney U test. Epo =Erythropoietin: bFGF = basic fibroblast growth factor: ns = not significant.

With bFGF Mean f SE (mean f SE) P value

EPO 0.52f0.19 4 .84 f3 .94 ns EPO + GM-CSF 3.14f1.77 4 .86 f2 .28 ns EPO + IL- 3 l . l l f 0 . 5 8 5 .04 f2 .74 ns Epo + GM-CSF + IL-3 1.2 1 + 0.3 7 4.39 f 2.5 5 ns All the cultures 1.41rt0.37 4 .79 f1 .26 0.016

Clonogenesis The CD34 + progenitors from seven HIV + patients genera- ted significantly fewer CFU-GM than those from seven controls (Mann-Whitney U P=0.017; Student t P=0.03) (see Fig l), whereas the percentage of BFU-E producing CD34+ cells was not significantly reduced. There was no difference in the percentage of CFU-GM between patients treated with AZT and those who were not (3-8%, SD 5 .1 v 3.6%, SD 6.1). The highest percentage of GM colonies was obtained in cultures containing bFGF. Comparing the sum of all colonies, when testing 24 cultures with and 24 without bFGF. in this way there was a statistically significant increase in cultures containing this growth factor (Mann-Whitney U P=0.017: Student t P=0*024) (see Table 11). The number of cultures was too small to allow evaluation of the efficacy of single growth factors. There was no correlation between

Fig 2. Hybridization analysis of amplified HIV DNA sequences from bone marrow cells of an HIV positive subject. 300 000 cells were used for each amplification with gag primers and probe.

results obtained in the clonogenic assays and the peripheral blood picture or the absolute number of CD4 + lymphocytes in the patients.

PCR From six patients (CD4+ lymphocytes 6-103 x 106/1), we obtained enough purified CD34 + cells for the PCR. In two patients we could test 0.6 x lo5 cells with gag primers: in the third we could test up to 3 x lo5 cells with gag and 1.5 x lo5 with env primers, in the three other patients we tested 1 x lo5 CD34+ cells. All results were negative. Tests for HIV-1 were negative after 3 weeks liquid culture of CD34+ cells in the presence of GM-CSF or G-CSF and SCF (see Table 111). HLADQAl was positive in all these samples. BMMC and the CD34 - fractions from the same patients were positive by gag or both primers ( 0 . 6 - 3 . 0 ~ 1 0 ~ cells, according to the number of CD34 + tested). Intensity of the autoradiographic signal decreased with increasing enrichment for progenitor cells (Fig 2). BMMC and CD34- cells from five further patients were all PCR positive. Pools of 29-105 of indivi- dually picked CD34 + cells derived CFU-GM (minimum 50 cells per colony) were tested (i.e. 1.8-9-4 x lo3 cells per patient). The four samples from the four patients were PCR negative, while BMMC from these patients were positive with gag primers. With our DNA amplification we could detect 10 HIV DNA copies in a background of lo5 cells in all the experiments.

DISCUSSION

We investigated the involvement of CD34 + progenitor cells purified from bone marrow in the pathogenesis of haematolo- gical abnormalities during advanced stage HIV infection.

We first calculated the percentage of CD34 + cells in the aspirated bone marrow mononuclear cells and found a slight reduction, which did not reach statistical significance, in 11

CD34 + Progenitors and HZV-I 2 3 Table 111. PCR results of different bone marrow derived cell populations from nine HIV+ patients.

C D 3 4 + after CFU-GM Patients* BMMC CD34 - cells 0 3 4 + cellst liquid culture$ from CD34 +

__

1 gag + /env + gag + /env - gag - nt gag - 2 gag + gag + gag - nt nt 3 gag + /env + gag + jenv + gag - /env - nt nt 4 gag + gag + gag - gag - nt 5 gag + gag + gag - gag - nt 6 gag + gag + gag - gag - nt 7 gag + gag + nt nt gag - 8 gag + gag + nt nt gag - 9 gag + gag + nt nt gag -

* CD4+ lymphocyte range 6-103 x 106/1. t 60000 to 300000 cells tested. $100000 cells tested. Gag= amplication with primers S K 3 8 / 3 9 and hybridization with probe S K 1 9 from the gag region of HIV- 1. Env=primers S K 6 8 / 6 9 and probe S K 7 0 from the env region of HIV-1. nt=not tested.

HIV-1 infected patients compared to controls. Then we assessed the in vitro clonogenic response of these cells from seven AIDS patients to recombinant growth factors. Our results clearly show that their ability to generate GM colonies is significantly reduced. The role of bone marrow-involving opportunistic infections or of myelotoxic drugs must be ruled out in the patients studied. Previous reports on the clonoge- nesis of BM cells in HIV + patients were based on less purified cell populations, containing T cells or monocytes, which are known to be infected in vivo and to inhibit colony formation (Stella et al, 1987; Leidermann et al, 1987). Though BFU-E: formation was preserved, the impaired CFU-GM formation shows that at least part of the haemopoietic defect is due to a functional deficit of CD34 + progenitors. Recent work based on HIV + thrombocytopenic patients showed a selective defect of CFU-Meg, but not of CFU-GM or BFU-E, colony formation by freshly isolated CD34 + cells (Zauli et al, 1992). The discrepancy with our data can be explained by the fact that we studied patients in a more advanced stage of disease, as shown by their lower CD4+ cell counts, most of them having haematological defects involving different cell lines. Haemopoietic growth factors, particularly those with a selective activity on most undifferentiated cells, such as bFGF (Gabbianelli et al, 1990), induced increased colony formation when added to the cultures, showing that these cells are still responsive to growth factors, though less than cells from HIV - subjects.

The question whether there is a direct infection of progenitors was assessed by performing a DNA PCR on CD34-t cells purified from BM. We could not detect viral sequences in these cells. From the sensitivity of our assay (10 HIV-1 DNA copies diluted in lo5 PBMC), we can calculate that fewer than 1 : 6000 CD34 + cells in the first two patients and at the most 1 : 10 000 in the other four were infected by HIV-1: thus, direct infection of CD34+ cells is even less frequent than estimated by other authors, who could test only few cells from ARC/AIDS patients (Davis et al, 1991) or HIV + thrombocytopenic patients (Zauli et al, 1992), or who did not define the sensitivity of their assay (von Laer et al, 1990). Furthermore, cells derived from CD34+ cells after 2 1

d liquid culture, in the presence of growth factors, did not contain HlV-DNA.

We also tested CFU-GM colonies derived from CD34 + cells from four patients. If a CFU-GM contains integrated proviral DNA, all the cells of this colony are expected to be infected. To test the largest possible number of CFU-GM, cells from the individually picked colonies of each patient were pooled together, so that, from a single patient, 29-105 CFU-GM were examined in the same amplification. In this way, we tested larger samples of CFU-GM capable of forming colonies in vitro than in other studies, where only single colonies (Molina et al, 1990) or pools of a maximum of five (Davis et al, 199 1) were examined. Nevertheless, we could not detect any HIV-1 DNA in these cells, while BMMC of these patients were all PCR positive. It could be argued that infected progenitors could be incapable of generating colonies and would thus escape detection, but HIV-DNA was absent also in the total purified CD34 + cell population. Very recently, other authors (Kaczmarski et al, 1992) have shown the presence of HIV- DNA both in colonies derived from direct culture (four subjects) and in those from long-term culture (five subjects) of bone marrow mononuclear cells from AIDS patients. These results are not in contrast to the data presented in the present work, because a different cell population (total mononuclear cells instead of purified CD34 + cells) has been studied, and it is possible that progenitor cells have been contaminated in vitro, as stated by the authors, or that HIV-DNA sequences were present in the culture matrix, since neither T-cell depletion nor a negative control consisting in a mock aspiration from the cultures (Davis et al, 1991) has been performed. Furthermore, the same authors found no statisti- cally significant difference between HIV and control marrows with respect to the progenitor colony counts, but only two CFU-GM and five CFU-LTC colony assays were done: other authors, testing more sampies, found a significant T cell dependent reduction in colony growth from AIDS patients (Stella et al, 1987).

Other mechanisms than direct infection of progenitors must account for the impaired colony formation in our study. CD34+ cells, though substantially free from HIV-1 DNA,

24

come from a bone marrow microenvironment, which has been deeply influenced by the presence of the retrovirus. This is confirmed by the strong positivity of other BM cells in this and other studies using PCR (von Laer et al, 1990; Davis et al. 1991). Other cells, such as macrophages (Gartner et al, 1986) or stromal fibroblasts (Scadden et al, 1990). may act as reservoir for the virus in the bone marrow: these infected cells can influence the function of progenitors by an impaired production of regulatory cytokines (Scadden et al, 1989: Molina et al, 1989). HIV-1 replication in the BM micro- environment could thus, through an indirect mechanism, suppress haemopoiesis, possibly inducing apoptosis of CD34+ cells (Williams et al, 1990). The positive effect of bFGF on the differentiation of CD34 + cells, when combined with the other growth factors, suggests that still a fraction of undifferentiated stem cells from HIV+ patients can be rendered responsive to growth factors which display their function on more differentiated cells. Furthermore, CD34 + cells derived from patients’ bone marrow do not contain HIV- 1 and are capable of generating an HIV negative progeny in vitro, in the presence of growth factors. These results show that it is worth investigating the potential therapeutical use of growth factors acting on undifferentiated cells for the treatment of multilineage cytopenias in HIV + patients (Miles et al, 1991).

Andrea De Luca et al

ACKNOWLEDGMENTS

We are grateful to Dr M. Genuardi (Istituto di Genetica, UCSC) for assistance during probes purification and hybridization, to Dr P. Pezzotti (Centro operativo AIDS, Istituto Superiore di Sanita, Rome) for statistical suggestions and to Professor P. Engelfriet (CLB. Amsterdam) for revision of the manuscript. This work was supported by grant 5206 069 from the Progetto AIDS, Istituto Superiore di Sanita, Minister0 della Sanita, 1989-90.

REFERENCES

Abbott, M.A., Poiesz, B.J., Byrne, B.C., Kwok, S., Snisky, J.J. &Ehrlich, G.D. (1988) Enzymatic gene amplification: qualitative and quanti- tative method for detecting proviral DNA amplified in vitro. Journal oflnfectious Diseuses, 158, 1158-1168.

Busch. M., Beckstead, J., Gantz, D. & Vyas. G. (1986) Detection of human immunodeficiency virus infection of myeloid precursors in bone marrow samples from AIDS patients. Blood, 68, (Suppl. 1). 122a.

Davis, B.R., Schwartz, D.H., Marx, J.C., Johnson, C.E., Berry, J.M., Lyding, J., Merigan, T.C. & Zander, A. (1991) Absent or rare human immunodeficiency virus infection of bone marrow stem/ progenitor cells in vivo. Journal of Virologg, 65, 1985-1990.

Donahue, R.E., Johnson, M.M., Zon, L.I., Clark, S.C. & Groopman, J.E. (1 98 7) Suppression of in vitro haematopoiesis following human immunodeficiency virus infection. Nature, 326, 200-203.

Edwards, A.E. (1985) Experirnental Design in Psychological Research, 5th edn. p. 85. Harper and Row, New York.

Folks, T.M., Keffler. S.W., Orenstein, J.M.. Justement. J.F.. Jaffe, E.F. & Fauci, A.F. (1988) Infection and replication of HIV-1 in purified progenitor cells of normal human bone marrow. Science, 242,

Gabbianelli, M., Sargiacomo, M., Pelosi. E., Testa, U.. Isacchi, G. & Peschle, C. (1990) ‘Pure’ human hematopoietic progenitors: permissive action of basic fibroblast growth factor. Science, 249,

Gartner, S., Markovits, P.. Markovits, D.M.. Kaplan, M.H., Gallo. R.C. & Popovic, M. (1986) The role of mononuclear phagocytes in HTLV-III/LAV infection. Science, 223. 215-219.

Kaczmarski, R.S., Davison. F.. Blair, E., Sutherland. S.. Moxham. J.. McManus, T. & Mufti. G.J. (1992) Detection of HIV in haemato- poietic progenitors. British Journal of Haemutalogy. 82, 764-769.

Kawasaki, E.S. (1990) Sample preparation from blood, cells and other fluids. PCR Protocols (ed. by M. A. Innis et al), pp. 146-152. Academic Press, London.

Leiderman, I.Z., Greenberg, M.L., Adelsberg, B.R. & Siegal. F.P. (1987) A glycoprotein inhibitor of in vitro granulopoiesis asso- ciated with AIDS. Blood, 70, 1267-1272.

Miles, S.A.. Lee, K., Hutlin. L., Zsebo, K.M. & Mitsuyasu. R.T. (1991) Potential use of human stem cell factor as adjunctive therapy for human immunodeficiency virus-related cytopenias. Blood, 78,

Molina, J.M., Scadden, D.T., Bym, R.A., Dinarello, C.A. & Groopman. J.E. (1989) Production of tumor necrosis factor and interleukin-1 by monocytic cells infected by human immunodeficiency virus. Journal of Clinical Investigation. 84, 733-737.

Molina, J.M., Scadden. D.T., Sakagmchi, M., Fuller, B., Woon, A. & Groopman, J.E. (1990) Lack of evidence for infection of or effect on growth of haematopoietic progenitor cells after in vivo or in vitro exposure to human immunodeficiency virus. Blood, 76, 2476- 2482.

Scadden. D.T., Zeira. M., Woon, A., Wang, 2.. Schieve. L.. Ikeuchi, K., Lim, B. & Groopman, J.E. (1990) Human immunodeficiency virus infection of human bone marrow stromal fibroblasts. Blood, 76, 3 1 7-32 2.

Scadden, D.T., Zon, L.I. & Groopman, J.E. (1989) Pathophysiology and management of HIV-associated hematologic disorders. Blood,

Stella, C.C., Gamer, A. & Hoelzer, D. (1987) Defective in vitro growth of hemopoietic progenitor cells in the acquired immunodeficiency syndrome. Journal of Clinical Investigation, 80, 286-293.

vonLaer. D., Hufert. F.T.Y., Fenner, T.E., Schwander, S., Dietrich. M., Schmitz, H. & Kern, P. (1990) CD34+ hematopoietic progenitor cells are not a major reservoir of the human immunodeficiency virus. Blood, 76, 1281-1286.

Williams, G.T., Smith, C.A.. Spooncer, E.. Dexter, M. & Taylor, D.R. (1 990) Hematopoietic colony stimulating factors promote cell survival by suppressing apoptosis. Nature, 343, 76-80.

Zauli, G., Re. M.C., Davis, B., Sen, L., Visani, G., Gugliotta, L., Furlini, G. & La Placa. M. (1992) Impaired in vitro growth of purified (CD34 + ) hematopoietic progenitors in human immunodeficiency virus-1 seropositive thrombocytopenic individuals. Blood, 79,

91 9-922.

1561-1 564.

3200-3208.

74,1455-1463.

2680-2687.