Beta-defensins: linking innate and adaptive immunity through dendritic and T cell CCR6.

Defensins contribute to host defense by disrupting the cytoplasmic membrane of microorganisms. This report shows that human beta-defensins are also chemotactic for immature dendritic cells and memory T cells. Human beta-defensin was selectively chemotactic for cells stably transfected to express human CCR6, a chemokine receptor preferentially expressed by immature dendritic cells and memory T cells. The beta-defensin-induced chemotaxis was sensitive to pertussis toxin and inhibited by antibodies to CCR6. The binding of iodinated LARC, the chemokine ligand for CCR6, to CCR6-transfected cells was competitively displaced by beta-defensin. Thus, beta-defensins may promote adaptive immune responses by recruiting dendritic and T cells to the site of microbial invasion through interaction with CCR6.

Interleukin-2-induces development of denditric cells from cord blood CD34+ cells.

Dendritic cells (DC) have been shown to develop along a myeloid or lymphoid lineage of differentiation propagated from bone marrow or early thymic precursor cells with hematopoietic cytokines. In our study, we have induced growth and differentiation of DC from cord blood CD34+ cells initiated in interleukin-2 (IL-2) alone or in IL-2 + stem cell factor (SCF) + tumor necrosis factor alpha (TNF-alpha)-supplemented medium and cultured with IL-2 or IL-2 + SCF for 28-35 days. Dendritic morphology and antigenic phenotype of DC grown with IL-2 were characteristic for DC cultured in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF). Growth and differentiation of DC was followed by an increase in expression of MHC II and co-stimulating molecules CD80 and CD86. We have also shown the expression of the IL-2 receptor (IL-2R) gamma-chain in CD34+ cells after 2-3 days of culture with IL-2 alone. The co-expression of the IL-2R alpha, beta, and gamma subunits in both DC cultured with IL-2- or GM-CSF-containing cocktail of cytokines was also shown. The time curve for induction of IL-2R demonstrated low levels of subunit expression at the beginning of culture. The number of CD1a cells co-expressing CD25, CD122, and CDgamma increased to about 24-68 and to 78-95% after 21 and 28-35 days, respectively. Development of natural killer cells was shown along with DC. The proportion of CD56+ cells and cytotoxicity increased in a time-dependent manner.

Evidence for a possible role of Asialo-GM1-positive cells in the graft-versus-leukemia repression of a murine type-C retroviral leukemia.

Studies were performed to examine whether, in addition to T cells, there might be other immune cells also capable of exerting a graft-versus-leukemia (GVL) response following allogeneic marrow transplant. Using an MHC-matched mouse model, consisting of normal B10.S donors and SJL/J Rauscher-retroviral-leukemic recipients, the donor cells were selectively depleted of their Asialo-GM1+ component prior to being infused into the leukemic recipients. The incidence of relapse was then compared against that for matched leukemic control recipients of undepleted cells from the same donors. FCM analysis of the depletion protocol indicated that exposure to anti-Asialo-GM1 antibody eliminated more than half of the donor NK1.1+ cells, but caused no significant losses among the Thy-1+, CD3+, or CD8+ cells. Nevertheless, fatal relapse among the leukemic recipients of the depleted cells was nearly double that found among the leukemic control recipients of undepleted cells, 47.5 vs 25.4% (P = 0.01). In a parallel study, using normal SJL/J recipients, this same depletion protocol was found to have no significant effect on the incidence of graft-versus-host disease (GVHD). These results therefore suggest that Asialo-GM1+ NK cells may be capable of contributing to the suppression of relapse in this type of leukemic recipient of allogeneic marrow, and that this suppression may occur independently of GVHD.

Comparison of the effects of CD3 and CD5 donor T cell depletion on graft-versus-leukemia in a murine model for MHC-matched unrelated-donor transplantation.

Studies were designed to prospectively evaluate the effects of selective depletion for donor T cells strongly expressing the CD3 and CD5 pan-T antigens on the incidence of leukemia relapse following bone marrow transplantation. This evaluation was performed under controlled conditions in a mouse model for MHC-matched unrelated-donor transplantation, employing Rauscher leukemic SJL/J mice as the recipients and leukemia-resistant B10.S mice as the donors. Selective donor cell depletion for CD3 and CD5 was accomplished ex vivo prior to transplantation by incubation with the appropriate monoclonal antibody plus complement. When untreated, Rauscher leukemia resulted in a 97% fatality incidence. This was reduced to 30% by the transplant of non-depleted B10.S cells, with another 37% recipients dying from GVHD and graft failure. CD3 depletion reduced the GVHd deaths to 6% but increased relapse to 62%. Conversely, CD5 depletion had no effect on relapse or on GVHD but did significantly increase graft failure, thus negatively affecting survival. Evaluation of the results, done in conjunction with flow cytometry analysis of the effects of CD3 versus CD5 depletion on the donor cells, suggests that the T cells involved in suppressing leukemic relapse in these studies, and hence contributing to the GVL response, most probably had a phenotype of CD3+, CD5-.

Effect of donor and recipient gender disparities on fatal graft-vs.-host disease in a mouse model for major histocompatibility complex-matched unrelated-donor bone marrow transplantation.

Sex chromosome-linked minor histocompatibility determinants have been shown to affect the incidence and severity of graft-vs.-host disease (GVHD) in both humans and animals. On the basis of earlier studies done in mice and humans, it has often been assumed that this effect is due to a simple response of female donor cells recognizing recipient male HY antigens as foreign and reacting against them. However, the data of various clinical groups have not always supported this assumption. Moreover, since most of the earlier mouse studies focused only on the single transplant direction of female into male and/or were done under totally syngeneic conditions, the possibility of a GVHD response based on donor recognition of the recipient female HX antigen as foreign was never fully addressed. We have therefore reexamined the question in a more clinically relevant allogeneic transplantation setting, using a major histocompatibility complex (MHC)-matched, unrelated-donor mouse model. Five different donor/recipient sets were paired in all four possible gender combinations. The results indicated that, in addition to GVHD reaction against male HY, reaction against female HX was also possible. The results also showed that when the total level of GVHD due to autosomal chromosome minor histocompatibility disparities is extensive, it may masks the influence of gender-related factors on GVHD. Finally, the data also suggest the possibility that the sex chromosome-linked minor histocompatibility determinants may be polymorphic and thus capable of multiple allele expression.

Effect of the YCD3-1 anti-mouse monoclonal on murine GvHD.

The potential for applying the YCD3-1 rat-anti-mouse IgG2b anti-CD3-epsilon monoclonal antibody (MAB) to the study of graft-versus-host disease (GvHD) in mouse models has been examined. This MAB, unlike the previously developed hamster-anti-mouse-CD3 MABs, had been reported to exhibit strong cytolytic properties when applied in vitro in the presence of complement. Therefore, it was of interest to determine whether it could be effectively used for T-cell depletion in mice to suppress GvHD in the same manner as the anti-human-CD3 MABs have been applied in clinical transplantation. Evaluation of the effectiveness of this antibody was carried out both under fully allogeneic MHC-mismatched and under unrelated-donor MHC-matched marrow transplant conditions. For both types of transplantation, depletion of the donor cells with YCD3-1 plus complement prior to their injection into lethally irradiated recipients significantly suppressed GvHD, resulting in survivals of 75-79%, as compared to 8-13% in the controls that received undepleted cells from the same donors (p < 0.0001). These results suggest that the YCD3-1 MAB may have a potential for use as a negative selection agent in the further definition of the roles of the various T-cell subtypes, as well as the possible roles of natural-killer cells, in future studies into the mechanisms of GvHD, and of the graft-versus-leukemia effect.

Thyroid hormones in human follicular fluid and thyroid hormone receptors in human granulosa cells.

OBJECTIVE: To demonstrate the presence of thyroid hormone in human follicular fluid (FF) and the binding of antithyroid hormone antibodies in human granulosa cells (GCs). DESIGN: Follicular fluids and GCs collected from women undergoing oocyte retrieval after superovulation. SETTING: In Vitro Fertilization-America/Allegheny General Hospital and Reproductive Sciences Research Laboratories, the Department of Obstetrics and Gynecology, The Medical College of Pennsylvania/Allegheny Campus. MAIN OUTCOME MEASURES: Follicular fluid levels of triiodothyronine (T3) determined by a microparticle enzyme immunoassay and FF levels of thyroxine (T4) determined by a fluorescence polarization immunoassay. Three anti-thyroid receptor antibodies were used to determine the presence of thyroid receptor. The binding of these antibodies in GCs was assessed by fluorescent microscopy and flow cytometry. RESULTS: Both T3 and T4 were present in the FF of eight patients studied. A large majority of the samples of individual fluids fell within the normal range for serum. There was a positive correlation between serum T4 values and FF T4 values. The three antithyroid receptor antibodies showed positive nuclear staining of GCs by fluorescent microscopy. The antibody to all thyroid hormone receptors yielded 35% positive cells by flow cytometry, and the site specific antibody for either the alpha-1 or beta-1 receptors yielded 78% and 44% positive cells, respectively. CONCLUSION: These data demonstrated, for the first time, the presence of T3 and T4 in human FF and the presence of T3 binding sites in human GCs and suggest a role for thyroid hormone in the regulation of human GCs.

Flow cytometry method for the analysis of membrane-associated human chorionic gonadotropin, its subunits, and fragments on human cancer cells.

A quantitative flow cytometry method for the analysis of membrane-associated human chorionic gonadotropin (hCG), its subunits, and fragments on human cancer cells was developed using a double-antibody reaction; a flow cytometry with a 2-W argon laser, standard settings, and filters for fluorescein isothiocyanate use; commercially available software; and the ectopic hCG producer CCL 2 HeLa cells from the American Type Culture Collection (ATCC) as a cell control to standardize the reagents and for overall quality control. Twenty-two monoclonal antibodies (MoAb) and immunoglobulin G fractions from three rabbit polyclonal antisera were tested for effects of antibody concentration (titration), reproducibility at different levels of epitope expression, and variability of epitope expression to select appropriate primary antibodies. Based on the results of the various tests, three polyclonal immunoglobulin G antibodies and a panel of nine MoAb directed to epitopes located in five different regions on the hCG molecule were selected as first antibodies. Their specificity was determined by using two unrelated MoAb of the same isotype at the same concentration to replace the primary MoAb and by a competition experiment. The unrelated MoAb also were used for the selection of the appropriate control fluorescence profile needed for the software. The unique characteristics of this method were: the use of living cells, standardized reagents, internal and external quality control, and the highest sensitivity, which could detect as few as 10(3) molecules of fluorochrome per cell. Serial analyses of the ATCC CCL 2 HeLa cells and two of its variants and of the eutopic hCG producer JEG-3 choriocarcinoma cells revealed the expression of membrane-associated epitopes of intact hCG, its subunits, and fragments by a high percentage of the cells, indicating that the expression of these sialoglycoproteins by these two different types of cancer cells is a common phenotypic characteristic.

Expression of membrane-associated human chorionic gonadotropin, its subunits, and fragments by cultured human cancer cells.

The expression of human chorionic gonadotropin (hCG), its subunits, and fragments on the cell membrane of cultured human cancer cells was investigated using a flow cytometric method. This method uses living cells; a double-antibody reaction; a flow cytometer with an argon laser, standard settings, and filters for fluorescein isothiocyanate; commercially available software; the American Type Culture Collection (ATCC) CCL 2 HeLa cell line as cell control and overall quality control; polyclonal rabbit antisera raised against the hCG dimer, its alpha subunit (hCG alpha), and its beta subunit (hCG beta); and a panel of monoclonal antibodies (MoAb) recognizing different epitopes on the intact hCG molecule, its subunits, and fragments. The purified immunoglobulin G fractions from the polyclonal antisera were used to estimate the total expression of the membrane-associated glycoproteins; the MoAb were used to detect the expression of epitopes of the hCG dimer, its subunits, and fragments. The results of the analyses done on cells from 74 established cancer cell lines of different types and origins (including 52 carcinomas, 10 sarcomas, 4 leukemias, 6 lymphomas, and 2 retinoblastomas) showed variable degrees of reactivity in a great percentage of cells in all cell lines studied with MoAb directed against different conformational epitopes of intact hCG (hCG-holo), hCG beta, hCG beta-free, the carboxy terminal peptide (CTP) of hCG beta, and an epitope of hCG alpha. The expression of the membrane-associated epitopes of hCG and its subunits was found to be a phenotypic marker characteristic of all evaluated cultured human cancer cell lines, irrespective of their type or origin. There were, however, quantitative and qualitative differences in the expression of the different epitopes. Thus, hCG beta, free and as part of hCG-holo, recognized by the MoAb against hCG beta-CTP, was expressed by a high percentage of cells of most cell lines. There was great variability in the expression of hCG-holo, recognized by MoAb B109. For this reason some groups of cancers expressed larger amounts of incompetent hCG alpha and/or hCG beta than others. Cell lines derived from adenocarcinomas of the lung were the only exception to this general finding; the expression of small amounts of hCG-holo was caused by a low degree of hCG alpha synthesis.

The role of CD4 and CD8 T cells in the graft-versus-leukemia response in Rauscher murine leukemia.

Using a mouse model for MHC-matched unrelated donor transplantation, the relative influences of the CD4 and CD8 T cell subtypes on graft-versus-leukemia (GVL) were examined in a murine erythroleukemia induced in SJL/J mice by the injection of Rauscher virus. Following leukemia induction, the mice were given 9.5 Gy of total body irradiation (TBI) and injected with mixed marrow and spleen cells from normal MHC-matched--but minor histocompatibility mismatched--B10.S donors. Prior to their injection these donor cells were selectively depleted ex vivo for either CD4, CD8 or Thy-1 by exposure to the appropriate monoclonal antibody (MoAb) plus complement. Following transplant the recipients were observed for 20 weeks, along with parallel control groups, for survival, leukemia relapse, graft failure and graft-versus-host disease; 98% of the controls receiving no transplantation therapy died of leukemia. Among the controls that received TBI plus undepleted B10.S cells 30.9% died of leukemia relapse, but another 34.2% survived free of any clinical evidence of their leukemia. Donor cell depletion for Thy-1 increased the relapse to 68.8%, while survival fell to 10.4%. CD8 depletion resulted in a relapse of 55.6%, with a survival of 19.4%. By contrast, CD4 depletion had no effect on relapse, but did significantly increase the incidence of graft failure. At the end of the 20 weeks additional tests were run to determine whether those transplant survivors that had remained leukemia-free were also free of any residual Rauscher virus. Those tests showed that they were not.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparative hematopoietic toxicity of doxorubicin and 4'-epirubicin.

4'-Epirubicin is an anthracycline analog of doxorubicin which has been shown to be similar to doxorubicin in its anti-tumor activity but significantly lower in its cardiotoxicity. Therefore, it has been proposed as a potential clinical substitute for doxorubicin. Using the hematopoietic colony-forming unit, spleen (CFU-S) assay technique, direct comparison was made of the hematopoietic toxicity of the two drugs in vivo in a mouse model, and 4'-epirubicin was found to be significantly (P less than 0.01) less toxic than doxorubicin. On a milligram per kilogram basis, the dose of 4'-epirubicin required to achieve a given level of hematopoietic progenitor cell kill was approximately 50% larger than that required for doxorubicin. Early CFU-S recovery following 4'-epirubicin exposure was also stronger than that achieved following doxorubicin, as was short-term peripheral white blood cell recovery. These findings confirm previous clinical suggestions that the acute toxicity of 4'-epirubicin toward hematopoietic progenitor cells might be less than that of doxorubicin. At the same time, however, when given in doses near their lethal limit, both drugs were shown to induce a chronic hematopoietic suppression. This was evident in the depressed long-term CFU-S levels following high doses of either drug, as well as in chronically depressed white blood cell levels following high-dose 4'-epirubicin.

Comparative effects of various T cell subtypes on GVHD in a murine model for MHC-matched unrelated donor transplant.

Studies were undertaken to achieve a further understanding of T cell subtype involvement in minor-histocompatibility graft-versus-host disease (MiHL-GVHD) as it may occur in MHC-matched unrelated donor (MUD) transplantation. For this, the H-2 identical, but minor-histocompatibility disparate, B10.S----SJL/J mouse model was employed, using a 50/50 mixture of B10.S spleen and marrow cells to induce GVHD in the SJL/J recipients. Utilizing dual labeling flow cytometry analysis, the relative distributions of the various T cell subtypes within the spleen and marrow of the B10.S donor strain were determined. The effects of selectively depleting for pan-T (Thy-1+), CD-4 (L3T4+), CD-8 (Lyt-2+), or CD-5 (Lyt-1+) cells were then evaluated and the results were compared with the incidence and severity of GVHD in the recipients. The data reinforced the results of previous studies indicating that a significant element of MiHL-GVHD is dependent on CD-8 cytolytic T cells which may operate independently of any helper cell input. However, they also indicated that the presence of CD-4 helper cells can accelerate the response. Furthermore, they suggested that the induction of MiHL-GVHD in MUD transplants may not be limited to the activity of fully matured T cells, but that other immature T cell subtypes, lacking in both the CD-4 and CD-8 markers, may likewise be involved.

Evidence for two distinct mechanisms in acute fatal graft-versus-host reaction across minor histocompatibility barriers.

The roles of Lyt-1+ and Lyt-2+ T cells in the mechanisms of minor histocompatibility graft-versus-host reaction (MiHL-GvHR), as well as the influence of the source tissues from which those T cells were drawn, have been examined. Using SJL/J recipients H-2 matched to B10.S donors, the responses obtained transplanting donor spleen cells alone, spleen cells mixed with marrow, or lymph nodes mixed with marrow, and treated with anti-Thy-1, anti-Lyt-1, and/or anti-Lyt-2 monoclonal antibodies (MABs) were compared. The results indicated that both Lyt-1+ and Lyt-2+ cells may contribute to MiHL-GvHR and that, at least in part, they may play separate roles. It was also found that when the T cells were derived from the spleen, as opposed to the lymph nodes, there were substantial differences between the observed GvHR survival patterns and in the relative influences of Lyt-1+ versus Lyt-2+ cells on the resultant survival. With the spleen transplant, the Lyt-1+ cells exerted a dominant influence, but with the lymph node transplant, the influence of Lyt-2+ cells was dominant. There was also evidence to suggest the possibility of a Lyt-1 helper-cell contribution to the MiHL-GvHR exhibited by this transplant combination. Finally, it was found that the relative influences of Lyt-1+ and Lyt-2+ cells on MiHL-GvHR were expressed at two distinctly places in the survival curves, the former being seen in the early phase of acute GvHR and the latter at a later phase of the acute response.

Flow cytofluorometric analysis of choriogonadotropin-like material on the surface of human and mouse malignant cells.

Quantitation by flow cytofluorometry of the distribution of human choriogonadotropin (hCG)-like material on the surface of various human and mouse tumor cells grown in tissue culture and as solid tumors has been done using fluorescein-tagged rabbit antisera (IgG fraction) to intact hCG and, in one experiment, by use of two monoclonal antibodies specific for hCG. Fibroblasts were used as a negative (nontumorigenic) cell control, and a rabbit antiserum to human hemoglobin was used as reagent control. All malignant cells tested stained more intensely with the anti-hCG serum than with the antihuman hemoglobin serum. Positive reaction with the monoclonal antibodies specific for hCG provided strong evidence that the material stained was identical to hCG. Heterogeneity of the expression of the hCG-like material was notable both within each cell line and between different cell lines. This heterogeneity was not associated with cell-cycle phase. 3T3 fibroblast-like cells in vitro were originally negative for hCG but acquired reactivity with anti-hCG serum after ten passages.

Effect of propane sultone pretreatment on Friend virus leukemogenesis in mice.

Propane sultone (PS) injected i.p. 24 or more hours before Friend leukemia virus increased the incidence of lymphoma in SJL/J mice and at a higher dose increased the incidence of erythroleukemia in B10SJF1 mice. PS at the same time also decreased hematopoietic stem cell clonogenicity.

Effect of cyclophosphamide on Friend virus leukemogenesis in virus-sensitive and virus-resistant mice.

The influence of cyclophosphamide (CY) on Friend virus leukemogenesis was studied in SJL/J, C57BL/10J, and C57BL/10J X SJL/J F1 (hereafter called B10SJF1) mice. All three differ in their susceptibility to the viral oncogenic effect. Immunosuppressive doses of CY, which by themselves produced no cancer, were followed 2 days later by injection of Friend leukemia virus. The virus doses were the same as used previously. Although in other experiments preinjection of various chemical carcinogens augmented leukemogenesis by Friend leukemia virus in SJL/J mice, in the present study, pretreatment by CY had no such effect. In contrast, CY increased Friend erythroleukemia incidence from 15 to 100% in B10SJF1 mice and from 0 to 85% in C57BL/10J mice. The disease in C57BL/10J mice had a 190-day incubation period, which is approximately 5 times that in the SJL/J and B10SJF1 mice. During this latent period, the C57BL/10J mice harbored infectious Friend leukemia virus in their plasma.

Comparative toxicity of mitoxantrone and doxorubicin on hematopoietic stem cells.

Mitoxantrone (DHAD) is a recently developed cancer chemotherapeutic drug proposed as a possible substitute for the older established chemotherapeutic, doxorubicin (adriamycin, or ADR). We have directly compared the toxicity of DHAD and ADR against normal hematopoietic progenitors in a mouse model. Using doses that produced equal depressions in spleen weight, we examined the recovery patterns for pluripotent hematopoietic stem cells (CFU-S), myeloid cell progenitors (CFU-GM), and reticulocyte (Retic) production. The spleen weight depression assay indicated that, on a mg/kg basis, DHAD was more toxic to the organ than ADR, with 17.5 mg of ADR required to produce the same level of effect as 10 mg of DHAD. Recovery of splenic mass after exposure to these doses was also poorer in the DHAD-treated mice. CFU-S studies showed that the initial direct killing effects of pluripotent stem cells by the two drugs were equivalent, but that CFU-S recovery was better after ADR exposure than after DHAD exposure. By 12 days after exposure to ADR, the number of CFU-S per spleen had not only regained normal levels, but exceeded the normal by a factor of 2. In contrast, in the DHAD-treated mice the number was only half normal at this same time. The results suggest that there is a delay in recovery of the pluripotent stem cell compartment after DHAD exposure that may be due to some type of unrepaired damage to the support tissue in the spleen on which the CFU-S are dependent. Analysis of the effects of DHAD and ADR on CFU-GM indicated an initial toxic effect that was roughly equivalent for the two drugs at the doses used. However, the timing of the points of maximum suppression were different--earlier after exposure to DHAD than after ADR. Subsequently, the recovery patterns were quite similar for both drugs, and at 14 days the CFU-GM numbers were virtually normal. Reticulocyte assay indicated that both ADR and DHAD severely depressed red blood cell production. Recovery was rapid and complete by nine days, however, with significant overshoots, especially in the case of ADR exposure. Serial white blood cell (WBC) counts were also carried out. Reduction in total WBC number was evident between two and 11 days after exposure and more severe with DHAD than with ADR. However, neither the extent of suppression nor its duration accurately reflected the events occurring in the CFU-S or CFU-GM progenitor compartments.(ABSTRACT TRUNCATED AT 400 WORDS)

Effect of split dose TBI on hematopoietic stem-cell survival in combined radiation-drug exposures.

Studies were carried out to determine if a priming dose of total body irradiation (TBI) given before the first drug exposure in chemo-radiation protocols similar to those used in marrow transplantation would reduce the survival of hematopoietic stem cells. The cytotoxic drugs employed were cyclophosphamide (CY) and piperazinedione (PIP), both of which are currently used in the clinic for ablation of the host marrow prior to transplantation therapy for leukemia. The effects were evaluated in a normal and a leukemic mouse model using the endogenous colony-former technique. Splitting the TBI to give part of the total dose before the first dose of drug was found to enhance stem cell kill in some instances, but not in others. The optimum proportion of TBI given as the first dose did not appear to exceed 100 rads. When a higher proportion of the total TBI was given as the initial dose there was an indication of a protective effect on the stem cells with the PIP-TBI protocols, but similar protection was not observed with the CY-TBI protocols. When CY and PIP were combined together in the same protocol it was found that a simple inversion of the order of these two drugs could result in a six-fold difference in the extent of stem cell ablation achieved, indicating that with multiple drug protocols the drug sequencing itself could be equally important as the manner in which the radiation is given.

Comparison of CFU-GM response to piperazinedione versus cyclophosphamide in normal mice.

The effects of cyclophosphamide and piperazinedione on marrow granulocyte-macrophage precursor (CFU-GM) cells were compared in a mouse model. Endogenous colony-forming unit ( ECFU -S) assays after drug exposure were utilized in selecting doses of the two drugs with approximately equal hematopoietic stem cell ablative potential. The selected doses were 20 mg/kg of piperazinedione and 325 mg/kg of cyclophosphamide. Injection of these doses resulted in an initial (3-h) marrow CFU-GM depression of 90% in the cyclophosphamide-treated mice and 98% in the piperazinedione-treated mice. Recovery began within 6-24 h and proceeded in a similar fashion for both groups of animals, peaking at levels near normal control values for CFU-GM at three days after exposure. Both groups showed a subsequent decline to a second nadir at nine days followed by a second recovery toward normal levels. The data suggest that the two drugs affect the CFU-GM in a similar fashion.