L. Gribaldo,*,1 I. Malerba,* A. Collotta,* S. Casati,* and A. Pessina

Transcription

1 TOXICOLOGICAL SCIENCES 58, (2000) Copyright 2000 by the Society of Toxicology Inhibition of CFU-E/BFU-E by 3 -Azido-3 -deoxythymidine, Chlorpropamide, and Protoporphirin IX Zinc (II): A Comparison between Direct Exposure of Progenitor Cells and Long-Term Exposure of Bone Marrow Cultures L. Gribaldo,*,1 I. Malerba,* A. Collotta,* S. Casati,* and A. Pessina *European Centre for the Validation of Alternative Methods (ECVAM), Institute for Health and Consumer Protection, European Commission Joint Research Centre, Ispra (VA) Italy; and Institute of Microbiology, University of Milan, Milan, Italy Received April 28, 2000; accepted July 25, 2000 Erythropoiesis occurs in two stages: proliferation amplifies cell number, and differentiation stimulates the acquisition of the functional properties of red blood cells. The erythroid colony-forming unit (CFU-E) amplifies the differentiation process in response to erythropoietic stress in vitro, whereas the burst-forming unit (BFU-E), which is not particularly sensitive to erythropoietin stimulation, gives rise to the CFU-E and, when stimulated, produces morphologically-identifiable erythroid colonies. The aim of this work was to evaluate the toxic effects of the antiviral agent, 3 -azido-3 -deoxythymidine (AZT), the antidiabetic drug, chlorpropamide (CLP), and the heme-analogous compound, protophorphirin IX zinc (II) (ZnPP), on the proliferation of erythroblastic progenitors by using human umbilical-cord blood cells and murine progenitors from long-term bone marrow cultures. All these agents may interfere with the hemopoietic process, causing myelotoxicity as an adverse effect via different mechanisms. Our results showed selective toxicity of the three drugs on the erythroid progenitors (IC 50 : AZT M, ZnPP M, CLP mm), with respect to the myeloid progenitors (IC 50 : AZT 0.8 M, ZnPP M and CLP > 2800 M). The IC 50 values were well correlated with peak plasma levels reached in vivo by the drugs. There was a marked similarity between the drug sensitivities of the human and murine progenitors but differences in toxicity exerted by the drugs on the basis of the time of exposure. Drug treatment of long-term cultures, followed by the clonogenic assay of progenitors collected from them in the absence of the drugs, generally resulted in a lower hematotoxicity. Key Words: CFU-E; BFU-E; CFU-GM; human umbilical cord blood; long-term murine bone marrow cultures. Hematopoiesis, a complex process involving various cell populations and growth factors, ensures the constant and controlled replacement of blood cells. This process of rapid cell renewal and differentiation makes the hematopoietic system a sensitive target for xenobiotics (Metcalf, 1984). Erythropoiesis 1 To whom correspondence should be addressed. Fax: laura.gribaldo@jrc.it. is responsible for red blood cell production via proliferation and differentiation of specific erythroblastic progenitors, the BFU-E/CFU-E (burst and colony-forming unit-erythroid) (Axelrod et al., 1973). The erythroid colony-forming unit (CFU-E), which is erythropoietin-sensitive, amplifies the differentiation process in response to erythropoietic stress (Noble and Sina, 1993). The in vitro culture of erythroid cell lines has revealed that the burstforming unit (BFU-E) is not particularly sensitive to erythropoietin stimulation, but gives rise to the CFU-E and, when stimulated, produces morphologically identifiable erythroid colonies (Stephenson et al., 1971). Many drugs and chemicals (e.g., antivirals, antineoplastics, pesticides, benzene metabolites) can cause hematological disease and interfere with one or more hemopoietic lineages, leading to neutropenia or agranulocytosis, thrombocytopenia, pure red cell aplasia, and in severe cases, aplastic anemia (Abraham et al., 1989; Lutton et al., 1997; Tepperman et al., 1974). Standard animal toxicology testing uses inhalation or ingestion of preparations to achieve exposure, but the myelotoxicity of xenobiotics measured in these tests does not necessarily predict hematopoietic system injury (Gribaldo et al., 1996). In in vivo tests, it is also difficult to determine which hematopoietic lineages are affected by the toxic compound. At present, in vitro colony-forming assays are used as screens for new drug leads of low hematotoxic potential (Du et al., 1991) and to aid the risk assessment of pesticides and industrial chemicals, but these assays can also be used to select the most appropriate animal species for use in the preclinical evaluation of a compound (Deldar et al., 1995). The predictivity of in vitro data has been shown in validation studies with antineoplastic agents (Parchment et al., 1993), whereas with other classes of compounds, the relationships between plasma drug concentrations and hematotoxicity are not usually so well defined. Moreover, the severity of hematotoxicity depends upon the level of toxicant exposure, which is defined by the levels of hematotoxic metabolites, drug binding by plasma 96

2 INHIBITION OF CFU-E/BFU-E BY AZT 97 proteins, and duration of exposure. For these reasons, it should be determined whether in vitro assay systems that include the microenvironment are better than conventional clonogenic assays for predicting hematotoxicity in vivo. The aim of this work was to evaluate the toxic effects on the proliferation of the erythroblastic lineage of 3 representative drugs among antivirals (3 -azido-3 -deoxythymidine), antidiabetics (chlorpropamide), and heme-analogous compounds (protophorphirin IX zinc [II]). Human umbilical cord-blood cells and murine long-term bone marrow cultures were used as a source of erythroid progenitors. Murine long-term bone marrow cultures and cord blood cells were also used to investigate the toxic effect of the drugs on myeloid progenitors (colonyforming unit-granulocyte/macrophages [CFU-GM]). Two kinds of tests were employed: continuous exposure of human cord-blood cells and murine bone-marrow cells during the assay procedure, and pretreatment of long-term murine bonemarrow cultures (for 24 or 96 h), with subsequent testing of the clonogenic capacity of progenitors in the absence of the drug. Source of Progenitor Cells MATERIALS AND METHODS Human umbilical cord blood cells and murine progenitors collected from murine long-term bone marrow cultures were used as sources of primary cells for the CFU-assays. Long-term murine bone marrow cultures were established by culturing bone-marrow cells (BMC) obtained from femurs and tibias of BDF/1 female and male mice (Charles River, Como, Italy) under Dexter-type conditions (Dexter and Moore, 1977). Adherent layers were established by flushing the contents of the mouse bones and plating cells directly into a 25-cm 2 flask containing long-term medium (MyeloCult-5300 purchased from Stemcell Technologies, Vancouver, BC, Canada). The cultures were maintained at 33 C in 5% CO 2 for 2 4 weeks and then re-fed weekly with MyeloCult-5300, supplemented with 10 6 M hydrocortisone hemisuccinate. Cord blood cells were obtained, frozen, from Poietic Technologies, Inc, (Gaithersburg, MD) and thawed before use. Briefly, 1 ml of cells was rapidly thawed in a water bath at 37 C and diluted in 1 ml of 2.5% human albumin (Fluka, Milan, Italy) and 5% Dextran 40 (Pharmacia Biotech, Nerviano, Italy) in Iscove s Modified Dulbecco s Medium (IMDM) (Gibco, Paisley, UK), 0.22 m filtered solution, and 8 ml of 10% FBS (Gibco, Paisley, UK), 3 units DNase/ml (Boehringer Mannheim, Germany) in IMDM (Gibco) 0.22 m filtered solution. After 10 min, the solution was centrifuged at 800 g for 10 min at C. The pellet of 10 6 viable cells/ml was then diluted in 30% FBS-IMDM and used for the clonogenic test, after counting in a hemocytometer. Chemicals Sigma (St. Louis, MO) supplied 3 -Azido-3 -deoxythymidine (AZT) and Chlorpropamide (CLP). Protoporphyrin IX zinc (II) (ZnPP) was purchased from Aldrich Chemicals (Milwaukee, WI). Stock solutions of test compounds were prepared in double-distilled water (AZT), ethanol (CLP), or DMSO (ZnPP). The final concentration of DMSO never exceeded 0.1%. Human Assays BFU-E/CFU-E assay. Cord blood cells were seeded in MethoCult-H4230 (StemCell Technologies), containing FBS (30%), BSA (bovine serum albumine 1%), methylcellulose (0.9%), 2-mercaptoethanol (10 4 M), and glutamine (2 mm). Erythroid stimulation was provided adding 2.5 units/ml human recombinant erythropoietin (Boerhinger Mannheim) in a tube containing 4 ml of methylcellulose, before the addition of 100 l of 44X drug solutions (in 20% FBS-IMDM) and 300 l of cells ( cells/ml). One hundred l of complete medium was added to the control tubes, whereas the same volume of the vehicle used to prepare the drug dilution was added to the solvent tube at the maximum concentration reached in the final dilutions. Finally, 1 ml of methylcellulose-cell suspension was seeded in 35-mm dishes and the cultures were incubated at 37 C in 5% CO 2 for 7 or 15 days. The final concentrations of drugs were: AZT 0.001, 0.05, 0.1, 0.2, and 0.4 M; CLP 187.5, 375, 750, 1500, and 3000 M; ZnPP 5, 10, 20, 40, and 80 M. CFU-GM assay. Cord blood cells were seeded in MethoCult-H4534 (StemCell Technologies). The procedure followed was the same as that described above for BFU-E/CFU-E. The final concentrations of drugs were: AZT 0.001, 0.05, 0.1, 0.2, and 0.4 M; CLP 187.5, 375, 750, 1500, and 3000 M; ZnPP 5, 10, 20, 40, and 80 M. The cultures were incubated at 37 C with 5% CO 2 and saturated humidity. The colonies were counted after 14 days of incubation. Murine assays. Murine progenitors, collected from long-term bone marrow cultures, were washed once, diluted in 30% FBS-IMDM to a density of cells/ml, and then seeded in MethoCult-M3334 (StemCell Technologies) for the BFU-E/CFU-E assay or in MethoCult-M3534 (StemCell Technologies) for the GM-CFU assay. These media are specific for murine cells and contain methylcellulose (1%), FBS (15%), BSA (1%), bovine pancreatic insulin (10 g/ml), human transferrin iron-saturated (200 g/ml), 2-mercaptoethanol (10 4 M), and glutamine (2 mm). Stimulation of the erythroid lineage was obtained by the addition of erythropoietin (3 units/ml), whereas MethoCult-M3534 contains IL-3 (10 ng/ml), IL-6 (10 ng/ml) and SCF (50 ng/ml) to stimulate GM-CFU growth. The procedure was similar to that followed for human assays, with some modifications. The final concentrations of drugs, for direct exposure, were: AZT 0.001, 0.05, 0.1, 0.2, and 0.4, M; ZnPP 5, 10, 20, 40, and 80 M, CLP 187.5, 375, 750, 1500, and 3000 M. The CFU-E/BFU-E cultures were incubated at 37 C in 5% CO 2 for 3 or 10 days, and the scoring of GM-CFU was performed after 7 days of incubation. Drug Pretreatment of Long-Term Murine Bone Marrow Cultures Fifteen days after their establishment, long-term bone marrow cultures were treated with AZT, ZnPP or CLP by adding the drug in the culture medium at a final concentrations of 0.02, 0.22, 2.20 M (AZT), 2.5, 25, 250 M (ZnPP) or 75, 750, 7500 M (CLP). After 24 and 96 h of exposure, cells from supernatants were collected and tested in the methylcellulose clonogenic assay, as described above, without the addition of drugs. Data Analysis Cell proliferation was expressed as a percentage of growth, 100% corresponding to the number of colonies in control dishes (mean ). Dose-response curves were produced by computer using a standard software program (SigmaPlot version 2.0). The concentration inhibiting the growth of 50% of CFU-GM, CFU-E and BFU-E (IC 50 ) was calculated by applying the Reed and Muench formula (Reed and Muench, 1938). Results are reported as mean SEM values of at least 3 experiments performed in triplicate. Statistical differences between means were evaluated by using Student s t-test. Differences were considered significant when p 0.05 and p RESULTS Continuous Exposure of Progenitor Cells The toxic effect of drugs on CFU-E/BFU-E colony formation from human cord blood was evaluated after direct exposure to the toxicants. Table 1 shows the IC 50 values calculated

3 98 GRIBALDO ET AL. TABLE 1 Comparison between IC 50 ( M) on CFU-E, BFU-E, Total Erythroid Colonies, and CFU-GM after Direct Drug Exposure of Human Umbilical Cord-Blood Cells CFU-E BFU-E TOT-E * ** * CFU-GM values in this case were 9.93 M for CFU-E, 12.3 M for BFU-E, and M for TOT-E. (Table 4). As is shown in Tables 3 and 4, the effects of AZT and ZnPP were greater on murine erythroblastic progenitors than on CFU-GM precursors (IC 50 /24 h for AZT: 1.03 M and 4.51 M; for ZnPP: 54.5 and M; IC 50 /96 h for AZT: 2.15 and 8.55 M; for ZnPP: and 37.8 M for TOT-E and CFU-GM, respectively), whereas no differences were seen for CLP toxicity after 24 h of pre-treatment (IC 50 : 2.38 mm for Note. IC 50 values (mean SEM) were calculated by the Reed and Muench formula on at least 3 experiments carried out in triplicate. *Significantly lower (p 0.01) than IC 50 on CFU-GM. on at least 3 experiments for each drug. The statistical evaluations did not demonstrate any significant selectivity in drug toxicity for CFU-E or BFU-E, although CLP appeared to be slightly more toxic to CFU-E than to BFU-E progenitors. Figure 1 represents the dose-response curves for AZT (A), ZnPP (B), and CLP (C) on murine progenitors from long-term bone marrow cultures and human cord-blood cells after continuous exposure to the toxicant in the methylcellulose assay. The slopes of the curves were similar for ZnPP and CLP, whereas human cells are less sensitive to AZT than the murine cells. There were no significant differences between the IC 50 values, except for AZT, although the anti-diabetic (CLP) seemed to be more toxic to the BFU-E murine progenitors. The toxic effects of ZnPP, CLP, and AZT on human CFU-GM were evaluated and the IC 50 values were 0.8 M for AZT, M for ZnPP, and 2800 M for CLP (Table 1). For murine CFU-GM progenitors, the IC 50 values were 2.5 M for AZT, 80.0 M for ZnPP, and 2.9 mm for CLP (Table 2). Drug Pretreatment of Long-term Murine Bone Marrow Cultures The IC 50 values obtained for TOT-E (total erythroid colonies) and CFU-GM colony formation after drug pre-treatment of long-term cultures for 24 and 96 h are summarized in Tables 3 and 4. Each drug exerted a lower toxicity when long-term cultures were treated for 24 h than after continuous exposure of murine erythroid progenitors to the xenobiotics (Fig. 2). After prolongation of the treatment of the long-term cultures to 96 h, the IC 50 values with AZT were 3.25 M for CFU-E, 1.7 M for BFU-E and 2.15 M for TOT-E, whereas CLP became more toxic after 96 h of pre-treatment, but nevertheless was less toxic than after continuous exposure (IC 50 values: 1.91 mm for CFU-E, 1.39 mm for BFU-E, and 1.73 mm for TOT-E (Table 4). On the contrary, the toxic effect exerted by ZnPP after the 96-h pre-treatment of murine long-term cultures was greater than that observed after direct exposure (Fig. 2). The IC 50 FIG. 1. Concentration-dependent (micromolar) inhibition of colony formation by erythroid progenitors (CFU-E/BFU-E) resulting from in vitro exposure of human umbilical-cord blood cells and murine bone-marrow cells to AZT (A), ZnPP (B), and CLP (C). Error bars indicate SEM. **Significantly less sensitive (p 0.01) than murine bone marrow.

4 INHIBITION OF CFU-E/BFU-E BY AZT 99 TABLE 2 Comparison between IC 50 ( M) on CFU-E, BFU-E, Total Erythroid Colonies and CFU-GM after Direct Drug Exposure of Murine Bone Marrow Progenitors CFU-E BFU-E TOT-E * * * CFU-GM 2.5** Note. IC 50 s values ([x bar] SEM) have been calculated by the Reed and Muench formula on at least 3 experiments carried out in triplicate. *Significantly lower (p 0.01) than IC 50 on CFU-GM. **Data not published. TABLE 4 Comparison between IC 50 ( M) on CFU-E, BFU-E, Total Erythroid Colonies, and CFU-GM after 96-h Pretreatment of Murine Long-Term Bone Marrow Cultures CFU-E BFU-E TOT-E * ** ** CFU-GM Note. IC 50 values (mean SEM) have been calculated by the Reed and Muench formula on at least 3 experiments carried out in triplicate. */**Significantly lower (p 0.01/p 0.05) than IC 50 on CFU-GM. TOT-E and 2.47 mm for CFU-GM), or when the pre-treatment was prolonged to 96 h (IC 50 : 1.73 and 2.44 mm for TOT-E and for CFU-GM, respectively). DISCUSSION Bone marrow is the principal source of blood cell formation in normal adults, and is responsible for the continuous production of red cells, white cells, and platelets (Lord and Testa, 1988). Cells representing earlier stages of hematopoietic cell differentiation are distinguished by functional endpoints that measure the number and types of morphologically recognizable progeny they are able to generate. Erythropoiesis is the part of hematopoiesis responsible for red blood cell production by cell proliferation and the differentiation of specific erythroblastic progenitors, BFU-E/CFU-E. (Metcalf, 1984). The toxic effects of drugs on human cord-blood progenitors and murine precursors were evaluated by direct exposure to chemicals showing a similar toxicity to murine and human CFU-E/BFU-E with ZnPP and CLP, whereas AZT was less toxic to human than to murine erythroid cells. The selective toxicity of drugs with respect to erythroid or myeloid progenitors has also been evaluated. All 3 compounds tested were TABLE 3 Comparison between IC 50 s ( M) on CFU-E, BFU-E, Total Erythroid Colonies, and CFU-GM after a 24-h Pretreatment of Murine Long-Term Bone Marrow Cultures CFU-E BFU-E TOT-E * ** CFU-GM Note. IC 50 values (mean SEM) have been calculated by the Reed and Muench formula on at least 3 experiments carried out in triplicate. */**Significantly lower (p 0.01/p 0.05) than IC 50 on CFU-GM. significantly more toxic to erythroid progenitors than to GM- CFU progenitors in both species, whereas there was no selectivity with respect to drug toxicity to CFU-E or BFU-E with murine or human cells. Anemia and neutropenia are the most common effects observed in vivo during treatment with antivirals, which seem to have a direct suppressive effect on heme synthesis (Langtry and Campoli-Richards, 1989). Also, high doses of heme-analogous compounds, which are used for treatment of heavy metal-induced hematotoxicity, may cause anemia because of their modulatory effects on the enzymes involved in the heme-synthesis pathway (Lutton et al., 1991). A pure red cell aplasia, where only erythroid precursors are absent from the bone marrow, has been reported with the use of antidiabetics (Recker and Hynes, 1969). Our in vitro data confirmed the in vivo clinical adverse effects observed with the administration of these drugs, mainly with AZT. AZT was found to inhibit de novo infection of T cells by HIV-1 at a concentration of 0.5 M (Mytsuya et al., 1985). AZT is well absorbed from the gut, with an average oral bioavailability of approximately 60%. After an oral dose of 200 mg, peak plasma levels of 3 to 4 M can be obtained in 30 to 90 min (Langtry and Campoli-Richards, 1989). Our results showed an IC 50 in vitro of 0.35 M on human erythroid progenitors and of 0.13 M on murine erythroid progenitors, confirming that severe hematotoxicity is the primary adverse effect after the administration of AZT. It has been reported in the literature that zinc porphyrin is toxic to both myeloid and erythroid cell growth, even at low concentrations with an IC 50 in vitro of 10 M on human erythroid colony formation by bone marrow precursors (Lutton et al., 1997). Our results showed an IC 50 of M on human erythroid precursors and 22 M on murine cells, confirming the myelotoxicity of this compound on both species. In an elderly man receiving high doses of CLP as an antidiabetic, he was reported with neutropenia and marrow pure white and red aplasia. In the patients investigated, the serum CLP level in the acute phase of disease was 100 g/ml (about 360 M) (average concentration) (Levitt, 1987). The IC 50 value, which we calculated with our results, was at

5 100 GRIBALDO ET AL. FIG. 2. Comparison between IC 50 values ( M) obtained with direct exposure of total erythroid colony-forming unit and CFU-GM to the drugs or 24- or 96-h pre-treatments of murine long-term bone marrow progenitors. IC 50 values ( SEM) have been calculated by the Reed and Muench formula on at least 3 experiments carried out in triplicate. */**Significantly different (p 0.05/p 0.01) from 96-h pre-treatment IC 50 value. / Significantly different (p 0.05/p 0.01) from direct exposure IC 50 value. /, significantly different (p 0.05/p 0.01) from direct exposure IC 50 value. 960 M on CFU-E development. This value correlates, as order of magnitude, with plasma level, shown before, confirming a possible toxicity of this drug at the highest doses of administration. In this investigation, we evaluated the role of primary cultures of murine bone marrow stroma in modulating the in vitro toxic effects of these drugs to murine CFU-E/BFU-E and CFU-GM. AZT appeared to be the most toxic compound to the CFU-GM murine progenitors, as has been reported in the literature (Deldar and Stevens, 1993), and it was significantly more toxic to erythroid progenitors, even after 24 and 96 h of exposure, as was ZnPP. Conversely, CLP did not exert a selective toxic effect on either erythroid or myeloid progenitors after 24 or 96 h of pre-incubation. The metabolic capacity of bone marrow stroma has been shown in long-term exposure experiments with AZT and CLP, where the IC 50 increased after 24 and 96 h of treatment, both for erythroid and CFU-GM progenitors. Further investigation is needed in the case of ZnPP, since the decrease in toxicity after 24 h of exposure was not maintained at the exposure time of 96 h. This result could be due to the toxicity of ZnPP to stromal cells, as observed in the morphological evaluation of the long-term cultures (data not shown), or to a biotransformation of ZnPP in a more toxic metabolite. These results reflect the basic role of the total hemopoietic environment. Without stromal cells, hematopoietic precursors are more sensitive than whole bone marrow culture to xenobiotics (Pisciotta, 1978). The importance of stromal cells seems to depend both on the production of growth factors and on metabolic detoxification (Gribaldo et al., 1999; Pessina et al., 1999). In fact, colony-stimulating factors can inhibit the apoptosis of progenitors (Crompton, 1991), and it is important to consider whether their high levels in vitro may be partially protecting the progenitors from being killed. Also of significance is the fact that stromal cells have been shown to have cytochrome P450-mediated metabolic activities (Myers and Flesher, 1990); thus, future work needs to focus on the identification of a possible contribution of metabolites to the hematotoxicity observed. On the other hand, duration of exposure has to be taken into account: prolonged exposure with unstable compounds may be misleading, because the breakdown products might be myelotoxic, or less toxic than the parent compound. Our results suggested that 24-h exposure does not represent the real in vivo situation, confirming that continuous exposure would be the optimal experimental design when we want to evaluate and predict drug hematotoxicity using in vitro tests. REFERENCES Abraham, N. G., Bucher, D., Niranjan, U., Brown, A. C., Lutton, J. D., Distenfeld, A., Ahmed, T., and Levere, R. D. (1989). Microenvironmental toxicity of azidothymidine: Partial sparing with hemin. Blood, 74, Axelrod, A. A., McLeod, O. L., Shreeve, M. M., and Heath, D. S. (1973).

6 INHIBITION OF CFU-E/BFU-E BY AZT 101 Properties of cells that produce erythrocitic colonies in vitro.inproceedings of the Second International Workshop on Haemopoiesis in Culture. (W. A.Robinson, Ed.), p Grune and Stratton, New York. Crompton, T. (1991). IL-3-dependent cells die by apoptosis on removal of their growth factor. Growth Factors 4, Deldar, A., Stevens, C. E., and Zack, P. M. (1995). Differential hematotoxicity by a nucleoside in progenitor cell assays from the rat, dog, non-human primates, and human. Vet. Pathol. 32, Deldar, A., and Stevens, C. E. (1993). Development and application of in vitro models of hematopoiesis to drug development. Toxicol. Pathol. 21, Dexter, T. M., and Moore, M. A. (1977). In vitro duplication and cure of haemopoietic defects in genetically anaemic mice. Nature 269, Du, D. L., Volpe, D. A., Grieshaber, C. K., and Murphy, M. J., Jr. (1991). Comparative toxicity of fostriecin, hepsulfam, and pyrazine diazohydroxide to human and murine hematopoietic progenitor cells in vitro. Investig. New Drugs 9, Gribaldo, L., Buere,n, A., Deldar, P., Hokland, P., Meredith, C., Moneta, D., Mosesso, P., Parchment, R., Parent-Massin, D., Pessina, A., San Roman, J., and Schoeters, G. (1996). The use of in vitro systems for evaluating haematotoxicity. ECVAM Workshop Report 14, Reprinted with minor amendments from ATLA, 24, Gribaldo, L., Catalani, P., and Marafante, E. (1999). Metabolism of doxorubicin in long-term bone marrow cultures and SR-4987 stromal established cell line. Drug Metabol. Drug Interact. 15, Langtry, H. D., and Campoli-Richards, D. M. (1989). Zidovudine: A review of its pharmacodynamic and pharmacokinetic properties, and therapeutic efficacy. Drugs 37, Levitt, L. J. (1987). Chlorpropamide-induced pure white cell aplasia. Blood 69, Lord, B. I., and Testa, N. G., Eds. (1988). Hemopoiesis: Long-term Effects of Chemotherapy and Radiation. Marcel Dekker, New York. Lutton, J. D., Abraham, N. G., Drummond, G. S., Levere, R. D., and Kappas, A. (1997). Zinc porphyrins: Potent inhibitors on hematopoieses in animal and human bone marrow. Proc. Natl. Acad. Sci. U.S.A. 94, Lutton, J. D., Chertkov, J. L., Levere, R. D., and Abraham, N. G. (1991). Comparative effect of heme analogues on hematopoiesis in lymphoproliferative disorders. Leukemia Lymphoma 5, Metcalf, D. (1984). The Haemopoietic Colony-Stimulating Factors. Elsevier, Amsterdam. Mitsuya, H., Weinhold, K. J., Furman, P. A., St. Clair, M. H., Lehrman, S. N., Gallo, R. C., Bolognesi, D., Barry, D. W., and Broder, S. (1985). 3 -Azido- 3 -deoxythymidine (BW A5090): An antiviral agent that inhibits the infectivity and cytophatic effect of human T-lymphotropic virus type III/lymphadenopathy-associated virus in vitro. Proc. Natl.Acad. Sci. U.S.A. 82, Myers, S. R., and Flesher, J. W. (1990). Metabolism of the carcinogen 3-methylcolantrene in human bone marrow preparations. Drug Met. Disp. 18, Noble, C., and Sina, J. F. (1993). Usefulness of the in vitro bone marrow colony-forming assay in cellular toxicology. In Vitro Toxicol. 6, Parchment, R. E., Huang, M., and Erickson-Miller, C. L. (1993). Roles for in vitro myelotoxicity tests in preclinical drug development and clinical trial planning. Toxicol. Pathol. 21, Pessina, A., Piccirillo, M., Mineo, E., Catalani, P. Gribaldo, L., Marafante, E., Neri, M. G., and Raimondi, A. (1999). Role of SR-4987 stromal cells in the modulation of doxorubicin toxicity in vitro granulocyte-macrophage progenitors (CFU-GM). Life Sci. 65, Pisciotta, V. (1978). Drug-induced agranulocytosis. Drugs 15, Recker, R. R., and Hynes, H. E. (1969). Pure red blood cell aplasia associated with chlorpropamide therapy: Patient summary and review of the literature. Arch. Intern. Med. 123, Reed, L. J., and Muench, H. A. (1938). A simple method of estimating fifty percent endpoints. Am. J. Hyg. 27, Stephenson, J. R., Axelrod, A. A., McLeod, D. L., and Shreeve, M. M. (1971) Induction of colonies of hemoglobin-synthesizing cells by erythropoietin in vitro. Proc. Natl. Acad. Sci. U.S.A. 68, Tepperman, A. D., Curtis, J. E., and McCulloch, E. A. (1974). Erythropoietic colonies in cultures of human marrow. Blood, 44,