Effect of recombinant human alpha interferon (INF) and antilymphocyte globulin (ALG) on normal and aplastic anaemia haematopoietic progenitor cell growth.

The effect of recombinant alpha interferon (INF) and of antilymphocyte globulin (ALG) to the colony stimulating factor (CSF) production was examined with in vitro culture of the bone marrow of healthy and of aplastic anaemia (AA) persons. In healthy persons the supernatant of lymphocytes preincubated with PHA and ALG was found to show a stimulating effect to clonogenic properties of marrow progenitors, the mentioned effect being not in proportion to the concentration value. Similar properties are shown by interferon in these persons. In patients with aplastic anaemia, a considerable stimulating ALG effect to the granulocytic formation of colonies and a lesser stimulating effect of interferon were shown.     

The suppressive influences of human tumor necrosis factors on bone marrow hematopoietic progenitor cells from normal donors and patients with leukemia: synergism of tumor necrosis factor and interferon-gamma

The influences of human tumor necrosis factor (TNF) (LuKII), recombinant human TNF-alpha, natural human interferon-gamma (HuIFN-gamma), recombinant HuIFN-gamma, and natural HuIFN-alpha were evaluated alone or in combination for their effects in vitro on colony formation by human bone marrow granulocyte-macrophage (CFU-GM), erythroid (BFU-E), and multipotential (CFU-GEMM) progenitor cells incubated at 5% CO2 in lowered (5%) O2 tension. TNF (LuKII) and recombinant TNF-alpha caused a similar dose-dependent inhibition of colony formation from CFU-GM, BFU-E, and CFU-GEMM. Day 7 CFU-GM colonies were more sensitive than both day 14 CFU-GM colonies and day 7 CFU-GM clusters to inhibition by TNF. BFU-E colonies and CFU-GEMM colonies were least sensitive to inhibition with TNF. The suppressive effects of TNF (LuKII) and recombinant TNF-alpha were inactivated respectively with hetero-anti-human TNF (LuKII) and monoclonal anti-recombinant human TNF-alpha. The hetero-anti-TNF (LuKII) did not inactivate the suppressive effects of TNF-alpha and the monoclonal anti-recombinant TNF-alpha did not inactivate TNF (LuKII). The suppressive effects of TNF did not appear to be mediated via endogenous T lymphocytes and/or monocytes in the bone marrow preparation, and a pulse exposure of marrow cells with TNF for 60 min resulted in maximal or near maximal inhibition when compared with cells left with TNF for the full culture incubation period. A degree of species specificity was noted in that human TNF were more active against human marrow CFU-GM colonies than against mouse marrow CFU-GM colonies. Samples of bone marrow from patients with non-remission myeloid leukemia were set up in the CFU-GM assay and formed the characteristic abnormal growth pattern of large numbers of small sized clusters. These cluster-forming cells were more sensitive to inhibition by TNF than were the CFU-GM colonies and clusters grown from the bone marrow of normal donors. The sensitivity to TNF of colony formation by CFU-GM of patients with acute myelogenous leukemia in partial or complete remission was comparable with that of normal donors. When combinations of TNF and HuIFN were evaluated together, it was noted that TNF (LuKII) or recombinant TNF synergized with natural or recombinant HuIFN-gamma, but not with HuIFN-alpha, to suppress colony formation of CFU-GM, BFU-E, and CFU-GEMM from bone marrow of normal donors at concentrations that had no suppressive effects when molecules were used alone.(ABSTRACT TRUNCATED AT 400 WORDS)     

The suppressive effects of recombinant human tumor necrosis factor-alpha on normal and malignant myelopoiesis: synergism with interferon-gamma

The modulation of growth of normal and leukemic myeloid progenitor cells in soft agar cultures by recombinant human tumor necrosis factor-alpha (TNF alpha) and recombinant human interferon-gamma (IFN gamma) was investigated. TNF alpha inhibited colony formation of all colony types representing different maturational stages of normal progenitor cells committed to the myeloid lineage with different orders of sensitivity. Blast-type colonies derived from patients with acute myelogenous leukemia were more sensitive to TNF alpha inhibition than progenitor cells purified from normal bone marrow or bone marrow from patients with stable-phase chronic myelogenous leukemia. The response of most colony types to IFN gamma was poor. However, when IFN gamma was administered together with TNF alpha, synergistically enhanced antiproliferative effects were detected in all colony types tested. The antiproliferative action of IFN gamma on myelopoiesis was enhanced in culture by the presence of autologous monocytes, presumedly by inducing endogenous production of TNF alpha. However, TNF alpha seemed to act directly on the progenitor cells themselves to suppress their clonal growth, rather than involving accessory marrow elements such as monocytes and/or T lymphocytes.     

Cimetidine induced pancytopenia. Effect on human CFU-MIX colony formation

We describe a 26 year-old male with a pancytopenia possibly due to cimetidine. Using progenitor cell culture techniques we investigated the mechanism of this bone marrow toxicity. Our results show a cimetidine dose-dependent inhibition of normal human CFU-GM colony formation as described by Fitchen and Koeffler in 1980. No differences in growth inhibition were found between the patients' recovery marrow and the controls. Toxicity on normal human CFU-MIX colony formation was, however, far more pronounced. At concentrations as low as 5 micrograms/ml the numbers of CFU-MIX colonies were decreased by almost 20% and more than 30% in cultures of two normal bone marrow samples. A significant decrease in CFU-MIX colony size was measured even at therapeutic levels (0.5 micrograms/ml). No obvious decrease in CFU-GM colony size was noticed at low concentrations. Experiments with T-cell- and monocyte-depleted bone marrow samples gave similar results: a pronounced inhibition of the CFU-MIX colony formation at low concentrations of cimetidine whereas the CFU-GM formation was less affected. It is therefore very unlikely that Accessory cells play part in the cimetidine induced CFU-MIX inhibition. Our results suggest the existence of H2 histamine receptors on human CFU-MIX (= multipotent progenitor cell). Blocking these receptors prevents the multipotent progenitor cell from going into the DNA-synthesis phase of the cell cycle.     

Functional differences between cells from MLR colonies grown in soft agar cultures

This report describes a method of growing soft agar colonies of human T lymphocytes activated in the MLR. Two types of colonies were demonstrated: lower colonies grew within the agar layer, and upper colonies grew on the surface of the agar layer. Three days of priming the lymphocytes in the MLR and the use of supernatants of day-3 MLR cultures to provide T cell colony growth factor were necessary for optimal colony formation. Lymphocytes obtained from colonies were grown in long-term (2 to 4 weeks) cultures to generate sufficient numbers of cells to be tested in different functional assays. Cells from both types of colonies exhibited PLT activity. Upper colony cells showed considerably higher CML activity than lower colony cells (mean percent cytotoxicity 37 +/- 5 vs 6 +/- 3). Cells from both types of colonies contained radiosensitive suppressor cell activity that inhibited the primary MLR. The suppressor cell effect of lower colony cells was specific for the original stimulator, but upper colony cells displayed nonspecific suppressive effects. For both types of colony cells, it appeared that suppressive effects were unrelated to the CML activity of these cells. These data suggest that the soft agar colony assay offers a promising approach to separate subpopulations of lymphocytes activated in the MLR.     

Establishment of an interleukin 2-dependent T cell line derived from a patient with severe aplastic anemia, which inhibits in vitro hematopoiesis

Circulating mononuclear cells from a patient developing severe aplastic anemia during the course of non-A, non-B hepatitis were found to be virtually entirely composed of in vivo activated suppressor T cells (Ia+T8+). These cells were used to establish a new permanent cell line, termed SMAA, by using phytohemagglutinin, Ebstein-Barr virus-transformed irradiated B cells, allogeneic irradiated peripheral blood mononuclear cells, and recombinant interleukin 2 to investigate the relationship of aplastic anemia-derived circulating T cells to bone marrow failure. SMAA cells, now in continuous culture for more than 9 mo, were shown to inhibit proliferation of purified myeloid progenitors and their differentiation into early and late appearing neutrophil and eosinophil colonies by 90%, whereas monocyte colonies were much less affected. Similarly, growth of erythroid colonies and bursts was almost completely inhibited, as was anti-mu-induced B cell proliferation and lectin-induced T cell proliferation. This inhibition of hematopoiesis was mediated by the release of a soluble factor that was sensitive to acid (pH 2), heat (56 degrees C), and trypsin. Monoclonal and polyclonal antibodies to interferon-gamma could abrogate the inhibitory effects of SMAA supernatant, but more than 10(4) neutralizing U/ml had to be added. The effects of SMAA could be duplicated by adding 10(4) U/ml of purified recombinant interferon-gamma to colony and proliferation assays. The concentration of interferon-gamma in SMAA supernatant was estimated to be greater than 3 X 10(3) National Institutes of Health reference U/ml by immunoradiometric assay. These results demonstrate that some patients with aplastic anemia have circulating T cells that are capable of prolonged in vitro secretion of interferon-gamma causing severe inhibition of in vitro hematopoiesis, and these cells can be expanded into permanent lines for studies on their regulatory properties.     

Human B lymphocyte colony responses. I. General characteristics and modulation by monocytes

The present study investigated the ability of human B cell-enriched subpopulations to focally proliferate and form colonies in semisolid cultures after stimulation with staph protein A (SpA). After 6 days of incubation, cultures of B-enriched populations exhibited distinct colonies, the number being dependent on the concentration of SpA and the cell density. Optimal colony responses were 1.6 x 10(3) per 1 x 10(6) B lymphocytes, and greater than 83% of the colony-forming cells expressed surface immunoglobulin (sIg). The depletion of adherent monocytes from the B cell-enriched preparations decreased the colony responses approximately 3-fold compared with the nondepleted B cell populations. Adding optimal numbers of adherent monocytes to the monocyte-depleted B cells restored the colony responses; however, less augmentation was observed in single-layer co-cultures containing greater than optimal numbers of monocytes. Identical experiments in double-layer semisolid cultures revealed that relatively greater numbers of monocytes were required to enhance B cell colony responses. Thus, progressively higher ratios of monocytes to B cells resulted in increasing numbers of colonies and failed to demonstrate the diminished colony responses observed in the single-layer system. These studies demonstrate that human B cells form distinct colonies when activated by SpA and that normal adherent monocytes modulate the magnitude of colony responses. Although monocytes predominately enhance B cell clonal differentiation, the evidence presented also suggests that, to a lesser extent, soluble inhibitory materials are elaborated.     

Human T lymphocyte colonies in agar: a comparison with other T cell assays in healthy subjects and cancer patients.

Colonies of human lymphocytes with T cell characteristics will grow in agar from repeated mitotic divisions with phytohaemagglutinin (PHA) stimulation. The colonies comprise spheres of tightly-packed cells with up to 500-1,000 blast-like cells in each colony. 65% of cells from pooled colonies bound AET-treated sheep red cells. 1,100-2,500 colonies/10(6) peripheral blood lymphocytes developed when cell donors were healthy but lower numbers (350-1,000 colonies/10(6) lymphocytes) were detected in blood from cancer patients. Comparison with other non-specific assays of cell-mediated immunity showed that while 66% of cancer patients were anergic (to five recall antigens) and 78% exhibited depressed mitotic activity in standard cultures with low dose PHA, 100% of these patients revealed T cell colony formation below normal. It is suggested that further studies of T lymphocyte colony-forming cells in healthy people and in a number of disease states may significantly advance our understanding of mechanisms of cell-mediated immunity.     

Depletion of T-lymphocyte subsets using monoclonal antibody and complement: effect on T-colony formation

Using monoclonal antibody and fresh rabbit serum, mononuclear cells from 11 healthy individuals were depleted of either helper or suppressor T-lymphocytes by complement-mediated lysis. The effect on T-lymphocyte colony formation was studied to determine the nature of the T-colony-forming cell. While depletion of either subset of T-lymphocytes significantly reduced T-colony formation compared to controls (P less than 0.01), colony formation did not differ significantly between helper or suppressor T-cell-depleted cells. Thus T-colony formation appears to be a property of both helper and suppressor T-lymphocytes, and both subsets are necessary for optimal colony formation.     

Normal hematopoietic progenitor cells and malignant lymphohematopoietic cells show different susceptibility to direct cell-mediated MHC-non-restricted lysis by T cell receptor-/CD3-, T cell receptor gamma delta+/CD3+ and T cell receptor-alpha beta+/CD3+ lymphocytes

To evaluate the capability of NK cells and cytotoxic T lymphocytes to interact with normal hematopoietic progenitor cells (HPC), as compared to neoplastic lymphohematopoietic cells, we investigated inhibition of colony growth of these cell populations in semi-solid culture systems, after incubation with cloned cytotoxic effector cells. Three different types of cloned effector cells were investigated: TCR-/CD3- NK cells, TCR-gamma delta+/CD3+ cells, and TCR-alpha beta+/CD3+ cytotoxic T lymphocytes. Effector cells showed differential levels of tumor cell colony inhibition, but no MHC-non-restricted lysis of normal HPC was observed. Pre-stimulation of normal HPC by culturing on established stromal layers had no effect. Cell-mediated lysis of HPC only occurred by Ag-specific MHC-restricted lysis by CTL, or by antibody-dependent cellular cytotoxicity. In cell mixing experiments, irradiated tumor cells, but not normal bone marrow cells inhibited tumor cell lysis. Furthermore, cloned effector lymphocytes were able to specifically eliminate malignant cells from tumor contaminated bone marrow without damaging normal HPC. When fresh leukemic cells were used as targets, growth of acute myeloblastic leukemia colonies was inhibited after incubation with several cytotoxic effector clones, whereas chronic myeloid leukemia precursor cells showed limited sensitivity to MHC-non-restricted cytolysis. These results indicate that MHC-non-restricted cytolysis by NK cells is selectively directed against neoplastic cells and not against normal HPC.     

Restoration of nonclonal hematopoiesis in chronic myelogenous leukemia with interferon alpha

Studies have shown that recombinant human alpha interferon (rIFN alpha) inhibits the growth of colonies of multipotential stem cells from human bone marrow. This report demonstrates that rIFN alpha inhibits the growth of such colonies from the bone marrow of patients with chronic myelogenous leukemia (CML) to a greater extent than from bone marrow of healthy individuals. It also shows that T lymphocyte colonies subcloned with interleukin 2 (IL-2) from CML mixed colonies were inhibited more by rIFN alpha than were similar colonies subcultured from normal mixed colonies. The report demonstrates that the Ph' chromosome is present in such T cell colonies subcultured from CML mixed colonies. When mixed colonies were grown from CML bone marrow in the presence of rIFN alpha, Ph' negative colonies were observed, whereas no such Ph' negative mixed colonies grew from a similar number of bone marrow cells incubated without rIFN alpha. These observations confirm that T lymphocytes derived from bone marrow stem cells are from the CML clone, and that the inhibition of growth of Ph' positive colonies, by rIFN alpha permits the growth of residual normal stem cells. The disappearance of the Ph-chromosome in subclones of T lymphocytes supports the notion of nonclonal hematopoiesis in patients with CML.     

Effect of serum on cells cultured from human mammary tumors.

Clusters of cells derived from biopsy specimens of human mammary ductal carcinomas form two morphologically distinct epithelial colonies in culture, designated as E and E'. The proportion of E' cell clusters that attached and formed colonies ranged from 0.3 to 13.0% with different tumors. Attachment was independent of tumor grade. Microscopic observations revealed that the survival of E' cell colonies was limited to approximately 10 days with rapid cell degeneration commencing about 7 days. A comparison of sera showed that colony formation by cells from malignant tumors during the 1st week of culture was maximum in the presence of fetal bovine serum. Human serum alone was 70 to 100% less effective in promoting E' colonies. The most significant finding was that human serum from normal donors inhibited E' colony development in the presence of FBS. Although human serum was less effective than FBS in promoting colony formation by clusters of E cells, an inhibition was not observed. Inhibitory activity could not be attributed to either antagonistic hormones or the source of human serum. These results demonstrate that normal human serum contains a factor(s) that exhibits an inhibitory activity specific for human epithelial cells (E') derived from malignant tumors.     

Effects of recombinant transforming growth factor-beta 1 on hematologic recovery after treatment of mice with 5-fluorouracil

Transforming growth factor beta 1 (TGF-beta 1) has been shown to inhibit bone marrow colony formation after in vitro treatment as well as after in vivo administration to normal mice. These data suggest that TGF-beta might either protect, or further depress, progenitor cell levels in mice exposed to a cell cycle-active drug such as 5-fluorouracil (5FU). rTGF-beta 1 was administered repeatedly by either the i.v. or i.p. routes to mice during the hyperproliferative state of the bone marrow that occurs 7 to 9 days after the i.v. administration of 150 mg/kg 5FU. The formation of both multilineage and the more differentiated (CFU-c) colonies was inhibited by 20 to 40%/culture, and 66 to 93%/mouse. When multiple doses of rTGF-beta 1 were administered systemically immediately before the injection of 5FU, the resulting rebound in the number of CFU-c and multilineage colonies containing granulocyte, erythroid, megakaryocyte, and macrophage lineage colonies per culture was markedly inhibited by 30 to 77%, whereas the total number of CFU per mouse was inhibited up to 93%. This effect was maximal when rTGF-beta 1 was administered at daily doses of greater than or equal to 5 micrograms/mouse for at least 3 days. This inhibition of the recovery of the bone marrow from 5FU treatment induced by rTGF-beta 1 was a delayed transient response because by day 16 the progenitor cell numbers and bone marrow cellularity were identical to the 5FU-treated marrow controls.     

The influence of T lymphocyte subsets and humoral factors on colony formation by human bone marrow and blood megakaryocyte progenitor cells in vitro

Cellular and humoral influences of T lymphocytes on human megakaryocyte colony formation in vitro were assessed by using a microagar system. Megakaryocyte colony formation from nonadherent low density T lymphocyte-depleted (NALDT-) bone marrow cells was increased significantly after the addition of aplastic anemia serum (AAS) or purified megakaryocyte colony-stimulating factor (Meg-CSF). The addition of conditioned medium obtained from phytohemagglutinin-stimulated T lymphocytes replaced, at least partially, the requirement for AAS or purified Meg-CSF for the growth of megakaryocyte colonies. The cellular influence of T lymphocytes and T lymphocyte subsets on megakaryocyte colony formation was assessed by removing either T cells from nonadherent peripheral blood mononuclear cells with monoclonal OKT4, OKT8, or OKT3 antibodies plus complement, or by adding back populations of bone marrow or blood T4+ or T8+ lymphocytes, isolated by means of fluorescence-activated cell sorting, respectively, to NALDT--bone marrow or -blood cells. When sorted T cell subpopulations were added to a fixed number of NALDT--bone marrow or -peripheral blood cells in the presence of AAS or Meg-CSF, T4+ cells enhanced megakaryocyte colony formation and T8+ cells decreased it. These studies demonstrate that although the stimulation of megakaryocytic progenitor cells by Meg-CSF may not require the presence of monocytes or T lymphocytes, T4+ lymphocytes enhance and T8+ lymphocytes down-regulate megakaryocyte colony formation induced by Meg-CSF. These observations suggest that the immune system is capable of modulating the proliferative response of human megakaryocytic progenitor cells to Meg-CSF.     

Colony formation by subpopulations of human T lymphocytes. VI. Further studies on colony phenotype, function, and cloning efficiency

Phytohemagglutinin (PHA)-induced colony formation in semisolid agar medium by human peripheral blood T lymphocytes showed an increasing cloning efficiency with decreasing numbers of cultured cells. Ninety percent of CD4+ cells (inducer/helper phenotype) and 20% of CD8+ cells (cytotoxic/suppressor phenotype) formed colonies when cultured at 10-200 cells/ml culture in the presence of sheep red blood cells (SRBC) and a source of interleukin-2 (IL-2). Probably all T-colony-forming cells, but none of the subsequent colony cells, expressed the Leu-8 antigen. The cloning efficiencies of FACS-sorted cells expressing the natural killer antigenic phenotypes Leu-7+ and CD16+ were found to be less than 1%. The costimulatory effect of red blood cells for colony formation was specific for SRBC and not observed in the presence of red cells obtained from seven other species including man. All T-lymphocyte colonies obtained from unseparated peripheral blood mononuclear cells expressed the CD25 antigen (IL-2 receptor) and colonies were always composed of either CD4+ or CD8+ cells. None of the colony cells expressed the Leu-8 or the CD16 antigens. By their specific morphology in agar culture the majority of colonies composed of CD4+ cells were easily recognized, but but approximately one-third of the CD4+ colonies could not be distinguished from colonies composed of CD8+ cells. On expansion of individual colonies in liquid subculture in the presence of interleukin-2, approximately 15% of the colonies developed natural killer (NK)-like cytotoxic activity, being capable of direct killing of K562 tumor cells. It is concluded that the present method for growing human T colonies exhibits the same cloning efficiency as the most efficient liquid culture systems. Individual T colonies are composed exclusively of T inducer/helper or T cytotoxic/suppressor cells, they are never of mixed phenotype, and they do not contain cells of natural killer phenotype. Regulatory mechanisms influencing colony formation are operating between and within the various subsets of T lymphocytes.     

Effect of serum on cells cultured from human mammary tumors

Summary Clusters of cells derived from biopsy specimens of human mammary ductal carcinomas form two morphologically distinct epithelial colonies in culture, designated as E and E′. The proportion of E′ cell clusters that attached and formed colonies ranged from 0.3 to 13.0% with different tumors. Attachment was independent of tumor grade. Microscopic observations revealed that the survival of E′ cell colonies was limited to approximately 10 days with rapid cell degeneration commencing about 7 days. A comparison of sera showed that colony formation by cells from malignant tumors during the 1st week of culture was maximum in the presence of fetal bovine serum. Human serum alone was 70 to 100% less effective in promoting E′ colonies. The most significant finding was that human serum from normal donors inhibited E′ colony development in the presence of FBS. Although human serum was less effective than FBS in promoting colony formation by clusters of E cells, an inhibition was not observed. Inhibitory activity could not be attributed to either antagonistic hormones or the source of human serum. THese results demonstrate that normal human serum contains a factor(s) that exhibits an inhibitory activity specific for human epithelial cells (E′) derived from malignant tumors. Supported by NCI Contract CB-33898 and a Fellowship from the Imperial Cancer Research Fund, London, England.     

B lymphocyte regulation of human hematopoiesis

Epstein Barr virus (EBV)-transformed B lymphoblastoid cell lines (BLCL) were derived from seven different individuals. The ability of BLCL supernatants to stimulate hematopoietic colony formation in vitro was tested in a conventional stem cell assay system. Supernatants promoted the growth of single (GM, E, MK) as well as multi-lineage (GEMM) colonies in bone marrow cultures. Our results indicate that EBV-transformed B lymphocytes produce cytokines that affect in vitro stem cell proliferation and differentiation. These studies demonstrate the regulatory potential of activated B lymphocytes in human hematopoiesis.     

B Lymphocyte Colony Formation in vitro: Ultrastructural Development of Individual Colonies

B lymphocyte colony development in agar culture was studied using an electron microscope, and more than 3,000 colony cells were identified and photographed. In early cultures (day 4) lymphoblasts dominated the colonies. From day 5 onwards plasmablasts and small lymphocytes were present in the colonies. From day 6 onward mature plasma cells were observed in increasing numbers. On day 9 of culture the colonies started to degenerate and on day 10 of culture approximately 70% of the colony consisted of pyknotic and degenerating cells. Topographically, the degenerating cells were concentrated in the center of the colony whereas proliferation took place in the periphery. Colony growth occurred in an exponential fashion, the number of viable colony cells being maximal on day 8 of culture (400–600 cells/colony). At this time the frequencies of the four B cell categories were: lymphoblasts 72%, plasmablasts 20%, plasma cells 6%, and small lymphocytes 2%. Recloning experiments showed that dispersed colony cells were capable of forming only small cell clusters. It is concluded that B lymphocyte colony formation reflects a series of B cell developmental stages including the formation of the end cell categories of this lymphocyte lineage.     

Interferon-gamma inhibits proliferation, but not commitment, of murine granulocyte-macrophage progenitors.

We investigated the effects of interferon gamma (IFN-gamma) on the growth of murine hematopoietic progenitors. IFN-gamma inhibited granulocyte colony-stimulating factor (G-CSF)- and interleukin-3 (IL-3)-dependent colony growth by granulocyte-macrophage (GM) progenitors derived from the bone marrow cells of normal mice. However, the number of IL-3-dependent GM colonies formed by the bone marrow cells of 5-fluorouracil (5-FU)-treated mice was not influenced by the addition of IFN-gamma. Replating experiments suggested that IFN-gamma suppressed GM colony growth directly and that it exerted an inhibitory effect on the proliferation, but not on the commitment, of GM progenitors. In contrast, IFN-gamma failed to suppress colony growth by mast cell progenitors. Erythroid and megakaryocytic progenitors exhibited different responses to IFN-gamma depending on mouse strains. These results suggest that potent negative regulators are not always inhibitors of hematopoietic progenitors.