Disparate regulation of human fetal erythropoiesis by the microenvironments of the liver and bone marrow.

The liver and the bone marrow (BM) are the major organs that support hematopoiesis in the human fetus. Although both tissues contain the spectrum of hematopoietic cells, erythropoiesis dominates the liver. Previous studies suggested that a unique responsiveness of fetal burst-forming units erythroid (BFU-E) to erythropoietin (EPO) obviates the need for cytokines with burst-promoting activity (BPA) in fetal erythropoiesis. This potential regulatory mechanism whereby fetal erythropoiesis is enhanced was further investigated. Fluorescence-activated cell sorting was used to isolate liver and BM progenitors based on their levels of CD34 and CD38 expression. The most mature population of CD34+ lineage (Lin-) cells was also the most prevalent of the three subpopulations and contained BFU-E responsive to EPO alone under serum-deprived conditions. Kit ligand (KL) also strongly synergized with EPO in stimulating the growth of these BFU-E. An intermediate subset of CD34++CD38+Lin- cells contained erythroid progenitors responsive to EPO alone, but also displayed synergism between EPO and KL, granulocyte-macrophage colony-stimulating factor (GM-CSF), or interleukin (IL)-3, demonstrating that erythroid progenitors that respond to cytokines with BPA do exist in fetal tissues as in the adult BM. Candidate stem cells (CD34++CD38-Lin- cells) did not respond to EPO. Synergisms among KL, GM-CSF, and IL-3, and to a lesser extent granulocyte colony-stimulating factor (G-CSF) and FLK-2/FLT-3 ligand (FL), supported the growth of primitive multipotent progenitors that became responsive to EPO. These data define the limits of EPO activity in fetal erythropoiesis to cells that express CD38 and demonstrate the potential for various cytokine interactions to be involved in regulating fetal erythropoiesis. Furthermore, a comparison of the responses of liver and BM erythroid progenitors revealed similarity in their responses to cytokines but a difference in the frequency of BFU-E among the three subpopulations examined. A higher frequency of BFU-E among the intermediate and late progenitor subsets in the liver indicates that regulatory factors acting on stem cells and their immediate progeny are partially responsible for the high content of erythropoiesis in the liver. These data implicate a critical role for the microenvironments of the liver and BM in regulating the disparate levels of erythropoiesis in these tissues.

SCID-hu mice for the study of human cancer metastasis.

Cancer metastasis involves dynamic and multistep in vivo processes. While generation of metastatic clones requires genetic alterations in cancer cells, subsequent selection of the clones is heavily influenced by interactions with the surrounding tissue microenvironment. To reproduce the complex cellular interactions that occur in human patients is, however, difficult, and has not been achieved using currently available in vitro systems or conventional animal models. The SCID-hu mouse is generated by surgical implantation of human fetal tissues into mutant mice of the severe combined immunodeficient (SCID) phenotype. The unique feature of this model is that the implanted human tissues maintain their normal architecture and function. Therefore implanted human tissues will provide relevant microenvironments for the growth and metastasis of human cancer cells. The SCID-hu mouse model, which was specifically designed for the study of human cancer biology, enables experimental investigation of cellular events involved in cancer metastasis on the basis of interactions between human cancer cells and the human tissue microenvironment. It has been demonstrated that various types of human cancer cell lines generate tumors in implanted human bone marrow and lung, organs frequently involved in metastasis in patients, upon intravenous inoculation. Tumorigenic activity in SCID-hu mice faithfully reflects the clinical features of the original cancer. Tumor formation and selection of high tumorigenic variants occur in a species-specific manner. Furthermore, it was shown that metastatic tumor formation is regulated by both cancer cells and conditions in the host organs. Conditioning of animals by either whole-body irradiation or interleukin 1alpha treatment prior to cancer cell inoculation induced metastatic tumor formation by certain small cell lung cancer (SCLC) cell lines specifically in human bone marrow. A novel gene has been identified by comparing gene expression profiles between high and low tumorigenic SCLC cells in human lung. This gene is preferentially expressed in low metastatic lines, and transfection of the gene into highly metastatic cells results in suppression of metastasis. Recent studies have shown that the gene product is involved in the apoptosis induction pathway. Collectively, our results indicate that the SCID-hu mouse will serve as a unique platform technology with which to investigate cellular events involved in human cancer metastasis, as well as to identify genes playing important roles in the growth and metastasis of human cancer, in the context of interactions between human cancer cells and human tissue environments.

Administration of Flk2/Flt3 ligand induces expansion of human high-proliferative potential colony-forming cells in the SCID-hu mouse.

The effects of Flk2/Flt3 ligand (FL) administration on human hematopoiesis were investigated using SCID-hu mice transplanted with human fetal bone fragments. Treatment with recombinant human FL induced significant increases in the frequencies of the high-proliferative potential colony-forming cells and low-proliferative potential colony-forming cells in steady-state human bone marrow. FL also promoted the expansion of high-proliferative potential colony-forming cells and low-proliferative potential colony-forming cells in the human bone marrow during the recovery phase after irradiation, which was evident in increases in the frequencies as well as in the absolute numbers of colony-forming cells. Furthermore, higher percentages of CD33+ CD15- cells were found in the marrows treated with FL as compared to that of controls, indicating that FL hastened the recovery of at least some aspect of myelopoiesis after irradiation. These results indicate that FL induces the expansion of primitive hematopoietic progenitor cells in vivo and, therefore, may be useful in treating patients to promote an early hematopoietic recovery after cytoablative therapies.

Ex vivo manipulations alter the reconstitution potential of mobilized human CD34+ peripheral blood progenitors.

The phenotype and functions of CD34+ cells isolated from peripheral blood (PB) of steady-state healthy volunteers (ssPB-CD34), and of patients or healthy volunteers after mobilization (mPB-CD34) were investigated. ssPB-CD34+ cells contain a lymphoid cell population that co-express T or B cell markers, while mPB-CD34+ cells lack this population. After 5-day culture, significantly higher levels of expansion in cell, CD34+ cell, and HPP-CFC numbers were induced in ssPB-CD34+ cells, as compared to mPB-CD34+ cells. Hematopoietic reconstitution potential of these ex vivo manipulated CD34+ PBPC was evaluated in SCID-hu mice. It was found that ssPB-CD34+ cells retained the potential to reconstitute human bone marrow (BM), as well as thymus implanted in SCID animals. In contrast, only very low levels of reconstitution were detected in human hematopoietic tissues injected with cultured mPB-CD34+ cells. Reconstitution was restricted to myeloid cells, and no B cell reconstitution in bone marrow, or T cell reconstitution in thymus was achieved by these cells. The loss of B cell reconstitution potential of mPB-CD34+ cells was shown to be induced in a time-dependent manner during culture. These results indicate that mPB-CD34+ cells have different phenotypic and functional properties from ssPB-CD34+ cells. This may affect the efficacy of cell and gene therapy with mobilized PBPC.

Successful reconstitution of human hematopoiesis in the SCID-hu mouse by genetically modified, highly enriched progenitors isolated from fetal liver.

Highly purified CD34++CD38-Lin- hematopoietic progenitors isolated from human fetal liver were infected with the murine retroviral vector, MFG nls-LacZ, which encodes a modified version of the Escherichia coli beta-galactosidase gene. Progenitors that were cocultured with the packaging cell line could reconstitute human bone marrow or thymus implanted in SCID-hu mice. Expression of the beta-galactosidase gene was observed in primitive and committed clonogenic progenitors, mature myeloid, B-lineage cells, and T-lineage cells for up to 4 months after injection into SCID-hu mice. Furthermore, hematopoietic reconstitution by genetically modified progenitor cells could be achieved by the injection of the cells generated from as few as 500 CD34++CD38-Lin- cells, suggesting efficient retroviral gene transfer into fetal liver progenitors.

Colony-forming cells expressing high levels of CD34 are the main targets for granulocyte colony-stimulating factor and macrophage colony-stimulating factor in the human fetal liver.

The effects of the granulocyte (G) and macrophage (M) colony-stimulating factors (CSFs) on the growth of purified subpopulations of human fetal liver progenitors were investigated. In contradiction to the characterization of these cytokines as CSFs acting late in the course of hematopoiesis, both G-CSF and M-CSF were most potent in promoting the growth of fetal liver colony-forming cells (CFCs) that express high levels of CD34 and CD38 (CD34++CD38+) and are depleted of cells expressing a panel of lineage markers (Lin-). Cultures of these cells in serum-deprived conditions generated a mean of 11.2 and 39.1 low-proliferative potential (LPP)-CFCs per 1.0 x 10(3) CD34++CD38+Lin- cells grown in G-CSF and M-CSF, respectively. Cultures of more mature progenitors, isolated based on a lower level of CD34 expression (CD34+ Lin-), generated few LPP-CFCs and 6.3 and 4.7 clusters per 1.0 x 10(3) CD34+Lin- cells in response to G-CSFs and M-CSF, respectively. G-CSF was also found to synergistically enhance colony growth by either kit-ligand (KL) or fit-3/flk-2 ligand (FL) in cultures of CD34++CD38+Lin- cells as well as the more primitive compartment of CD34++CD38-Lin- cells. Synergism between G-CSF and KL or FL was also observed in liquid cultures of CD34++CD38-Lin- cells. The effects of G-CSF on CD342++CD38-Lin- cells were further demonstrated by the ability of G-CSF to support the short-term survival of these cells in clonal cultures. In contrast, M-CSF did not affect the growth or survival of CD34++CD38-Lin- cells, a finding that was also supported by the observation that the receptor for M-CSF (CD115 or fms) was only expressed on CD34++CD38+Lin- cells. G-CSF receptor expression and flt-3/flk-2 expression were detected by flow cytometry on both the CD38- and CD38+ subpopulations of CD34++Lin- cells, but these receptors were not detected on CD34+ cells. Receptors for KL (CD117) and interleukin-3 (CD123), for which the ligands are active on a broad range of fetal liver progenitors, were detected on cells expressing both high and low levels of CD34. These data help to define the potential roles of cytokines in human fetal hematopoiesis.

Phenotypic and functional evidence for the expression of CD4 by hematopoietic stem cells isolated from human fetal liver.

Expression of the CD4 antigen was observed on human fetal liver, fetal bone marrow (BM), and umbilical cord blood progenitors expressing high levels of CD34. Using clonal and liquid-culture assays, CD4+ CD34++ Lin- (lineage = CD3, CD8, CD10, CD14, CD15, CD16, CD19, CD20, and glycophorin A) fetal liver progenitors were found to have a greater proliferative potential than CD4- CD34++ Lin- progenitors, whereas the CD4- fraction was more enriched for erythroid progenitors. Both the CD4+ and the CD4- progenitor subpopulations also gave rise to multilineage engraftment upon transplantation into human fetal bone fragments, supportive of B-lymphoid and myeloid growth, or into human fetal thymic fragments, supportive of T-cell growth, implanted in scid/scid (SCID) mice. However, in SCID-hu mice transplanted with graded doses of donor cells ranging from 2.0 x 10(2) to 2.0 x 10(4) cells, BM reconstitution by the CD4+ fraction of CD34++ Lin- cells was more frequent than by the CD4- fraction when low numbers of cells were injected. These functional data strongly suggest that stem cells reside among CD4+ CD34++ Lin- fetal liver cells. This hypothesis was further supported by the observations that CD4+ CD34++ Lin- fetal liver cells were enriched for CDw90+ (Thy-1), CD117+ (kit), CD123+, HLA-DR+, CD7-, CD38-, CD45RA-, CD71-, CD115- (fms), and rhodamine 123(dull) cells, a phenotypic profile believed to represent fetal stem cells. Furthermore, all CD4+ CD34++ Lin- fetal liver cells also expressed CD13 and CD33.

Antigen-specific cytotoxic T cells mediate human fetal pancreas allograft rejection in SCID-hu mice.

Human allograft rejection was studied in SCID mice transplanted with human fetal liver and thymus tissue (SCID-hu mice). These SCID-hu mice have functional, mature T cells with a polyclonal TCR repertoire. Within 12 to 36 wk after construction, SCID-hu mice were transplanted with an HLA-mismatched human fetal pancreas. In contrast to control SCID mice transplanted with pancreas alone, cellular infiltration, induction of HLA-DR on pancreatic epithelial cells, and tissue destruction of the allogenic pancreata were observed in SCID-hu mice. In addition, human insulin was not detected in the serum of SCID-hu mice in which pancreas rejection occurred. The infiltrating cells were mainly human CD3+ T lymphocytes of thymic origin, expressing the CD45RO isoform. T cell lines and CD4+ T cell clones obtained from the rejected tissues proliferated vigorously when stimulated with EBV-transformed B cell lines of pancreas donor origin. Furthermore, the majority of these CD4+ T cell clones displayed strong allospecific cytotoxicity. In addition, CD8+ T cell clones cytotoxic for EBV-transformed B cell lines of pancreas donors were isolated. Blocking experiments with anti-HLA mAbs and panel studies with HLA-matched cell lines showed that these CD4+ and CD8+ T cell clones were specific for the HLA class II and class I molecules, respectively, expressed by the pancreas donor. These data indicate that human T lymphocytes developing in SCID-hu mice are able to mount in vivo responses against allogenic organs, resulting in tissue infiltration and rejection. In addition, they show that both CD4(+)- and CD8(+)-allospecific CTL can be isolated from rejected allogenic pancreata.

Human T-and B-cell functions in SCID-hu mice.

SCID mice transplanted with human fetal liver and thymus (SCID-hu Thy/Liv) provide a unique in-vivo model to study human T-cell development and clonal selection mechanisms. This SCID-hu mouse model can be adapted to study the role of thymic epithelial cells, or bone marrow-derived cells in transplantation tolerance. In addition, these mice have circulating human T cells, which mediate human allograft rejection in vivo. SCID-hu mice constructed with fetal bones and thymus (SCID-hu BM/Thy) have both circulating human T and B cells, and can be used to study human B-cell development and B-cell functions. In addition, human T-B-cell interactions resulting in human lg production and the modulating effects of cytokines and cytokine receptor antagonists on this process, can be monitored. Collectively, this information indicates that the SCID-hu mouse is a powerful and versatile model to study human immune responses in vivo.

Regulatory roles of the ligand for Flk2/Flt3 tyrosine kinase receptor on human hematopoiesis.

The biological activities of the ligand for the Flk2/Flt3 receptor tyrosine kinase (FL) on human hematopoietic cells are reviewed. In in vitro studies, FL shows relatively few effects by itself on the proliferation and differentiation of hematopoietic cells, but exhibits a potent costimulatory activity in enhancing the proliferation of progenitor cells of multiple lineages. FL promotes the growth of clonogenic myeloid progenitor cells in the presence of other cytokines known to be active on myeloid progenitors, including GM-CSF, interleukin 3 (IL-3), kit ligand (KL), M-CSF and G-CSF. In addition, FL synergizes with IL-7 in inducing the proliferation of pro-B cells, whereas FL has little effect on the growth of clonogenic erythroid progenitors. Furthermore, FL induces the in vitro expansion of the high proliferative potential colony-forming cells (HPP-CFC) and stimulates the proliferation of long-term culture-initiating cells (LTC-IC), suggesting an activity on the proliferation of putative stem cells. Thus, FL plays important roles in regulating the proliferation of hematopoietic progenitor cells and, therefore, may have therapeutic applications.

The FLK2/FLT3 ligand synergizes with interleukin-7 in promoting stromal-cell-independent expansion and differentiation of human fetal pro-B cells in vitro.

The effects of a novel cytokine FLK2/FLT3 ligand (FL) on human fetal bone marrow-derived CD34+CD19+ pro-B cells were analyzed in a stromal-cell-independent, serum-deprived culture system. FL, like interleukin-3 (IL-3), synergized with IL-7 in promoting pro-B cell growth, and differentiation of these cells into CD34-CD19+clgM+slgM- pre-B cells, whereas a small proportion of these cells even differentiate into more mature slgM+ B cells. In contrast, KIT ligand (KL) and granulocyte-macrophage colony-stimulating factor (GM-CSF) were ineffective in promoting IL-7-dependent pro-B cell growth and differentiation. Maximal levels of pro-B cell expansion, generally resulting in 15- to 30-fold increases in cellularity, were obtained in cultures supplemented with optimal doses of FL + IL-7 + IL-3. The addition of mouse bone marrow stromal cells further enhanced the proliferation and differentiation of pro-B cells obtained in the presence of these three cytokines. Under these conditions, cultures could be maintained for more than 4 weeks, and in general 40- to 50-fold increases in cell numbers were observed by 3 weeks of culture. The percentages of clgM+ and slgM+ B cells increased 1.5- to 3-fold and 2-fold, respectively, suggesting that stromal cells may provide additional costimulatory signals for human B-cell growth and differentiation that are different from IL-7, IL-3, and FL. Collectively, our results indicate that FL, in contrast to KL, strongly promotes long-term expansion and differentiation of human pro-B cells in the presence of IL-7 or in combination of IL-7 and IL-3, which is a novel property of this hematopoietic growth factor.

IL-4 induces human B cell maturation and IgE synthesis in SCID-hu mice. Inhibition of ongoing IgE production by in vivo treatment with an IL-4/IL-13 receptor antagonist.

The effect of cytokine treatment on the in vivo maturation and Ig isotype switching of human B cells was studied in a modified SCID-hu mouse model. SCID mice, subcutaneously cotransplanted with small fragments of fetal human thymus and bone (SCID-hu BM/T mice) generated all human leukocyte lineages including T and B lymphocytes, macrophages, and granulocytes. All SCID-hu BM/T mice spontaneously produced human IgM and IgG, whereas IgE and IgA were detected in 37 and 80% of the mice, respectively, indicating that productive human T-B cell interactions resulting in Ig isotype switching occur in these mice. Administration of IL-4 to SCID-hu BM/T mice enhanced human B cell maturation, as judged by the increase in the percentages of CD45+, CD19+ bone marrow B cells expressing CD20, CD23, CD40, sIgM, and sIgD. Furthermore, these cells were also functionally more mature because they spontaneously produced human IgG/IgG4 in vitro and could be induced to secrete human IgE by addition of anti-CD40 mAb alone. In contrast, B cells isolated from PBS-treated mice only produced significant Ig levels after stimulation with anti-CD40 mAb in the presence of exogenous IL-4. IL-4 administration also induced human IgE synthesis in 44% of the mice, which had no serum IgE before treatment. More importantly, ongoing human IgE synthesis in SCID-hu BM/T mice was suppressed by > 90% following administration of an IL-4 mutant protein, which acts as an IL-4 and IL-13 receptor antagonist. These results suggest that IL-4/IL-13 receptor antagonists have potential clinical utility in treating human atopic diseases associated with enhanced IgE production.

Species-specific metastasis of human tumor cells in the severe combined immunodeficiency mouse engrafted with human tissue.

We have attempted to model human metastatic disease by implanting human target organs into the immunodeficient C.B-17 scid/scid (severe combined immunodeficiency; SCID) mouse, creating SCID-hu mice. Preferential metastasis to implants of human fetal lung and human fetal bone marrow occurred after i.v. injection of human small cell lung cancer (SCLC) cells into SCID-hu mice; the homologous mouse organs were spared. Clinically more aggressive variant SCLC cells metastasized more efficiently to human fetal lung implants than did cells from classic SCLC. Metastasis of variant SCLC to human fetal bone marrow was enhanced in SCID-hu mice exposed to gamma-irradiation or to interleukin 1 alpha. These data indicate that the SCID-hu mice may provide a model in which to study species- and tissue-specific steps of the human metastatic process.

FLK-2/FLT-3 ligand regulates the growth of early myeloid progenitors isolated from human fetal liver.

The effects of the recently identified FLK-2/FLT-3 ligand (FL) on the growth of purified human fetal liver progenitors were investigated under serum-deprived culture conditions. FL alone was found to stimulate modest proliferation in short-term cultures of CD34++ CD38+ lineage (Lin)- light-density fetal liver (LDFL) cells and the more primitive CD34++ CD38- Lin- LDFL cells. However, the low levels of growth induced by FL were insufficient for colony formation in clonal cultures. Synergism between FL and either granulocyte-macrophage colony-stimulating factor (GM-CSF), interleukin-3 (IL-3) or KIT ligand (KL) was observed in promoting the growth of high-proliferative potential (HPP) colony-forming cells (CF) and/or low-proliferative potential (LPP)-CFC in cultures of CD34++ CD38+ Lin- and CD34++ CD38- Lin- LDFL-cells. FL, alone or in combination with other cytokines, was not found to affect the growth of CD34+ Lin- LDFL cells, the most mature subpopulation of fetal liver progenitors investigated. The growth of the most primitive subset of progenitors studied, CD34++ CD38- Lin- LDFL cells, required the interactions of at least two cytokines, because only very low levels of growth were observed in response to either FL, GM-CSF, IL-3 or KL alone. However, the results of delayed cytokine-addition experiments suggested that individually these cytokines did promote the survival of this early population of progenitors. Although two-factor combinations of FL, KL, and GM-CSF were observed to promote the growth of early progenitors in a synergistic manner, neither of these factors was found to make fetal liver progenitors more responsive to suboptimal concentrations of a second cytokine. Only myeloid cells were recovered from liquid cultures of CD34++ CD38- Lin- LDFL cells grown in the presence of combinations of FL, KL, and GM-CSF. These results indicate that FL is part of a network of growth factors that regulate the growth and survival of early hematopoietic progenitors.

Progress in the ex vivo expansion of hematopoietic progenitors.

In this review we describe how studies on the cytokine-stimulated growth of murine bone marrow (BM) progenitors have lead to the observations that large increases in progenitor numbers can be achieved in short-term cytokine-stimulated liquid cultures. Transplantation of these ex vivo expanded murine BM cells was shown to decrease the number of BM cells required to confer radioprotection and to increase the recovery rate of both myeloid and erythroid peripheral blood cells. The ex vivo expansion of murine BM cells does not however, markedly diminish stem cells capable of long-term hematopoietic reconstitution. Investigations on the expansion of human BM, peripheral blood, umbilical cord blood and fetal hematopoietic progenitors have demonstrated that clinically useful increases in progenitor numbers from these tissues are possible. Thus, ex vivo progenitor expansion may soon be of use in transplantation protocols to accelerate hematopoietic reconstitution and in gene therapy protocols if hematopoietic stem cells can be maintained during ex vivo culture.

Engraftment of human hematopoietic precursor cells with secondary transfer potential in SCID-hu mice.

Human fetal bone fragments implanted subcutaneously in immunodeficient (SCID) mice maintain active human hematopoiesis. In this study, we show that this human hematopoietic microenvironment supports the engraftment and differentiation of HLA-mismatched, CD34+ primitive hematopoietic progenitor cells isolated from fetal and adult human bone marrow (BM). The BM CD34+ cells were depleted of CD2, CD14, CD15, CD16, glycophorin A, and CD19 lineage-committed cells (CD34+Lin-). Donor cell engraftment was manifested by the presence of B (CD19+) and myeloid (CD33+) cells of donor HLA phenotype. Successful engraftment was observed as early as 4 weeks after fetal BM donor cell injection and sustained for at least 12 weeks, with engraftment success rates of 100% (11/11 grafts) and 92% (11/12 grafts) at 8 and 12 weeks, respectively. Mixed BM chimerism of donor and endogenous cells was consistently observed in SCID-hu bones successfully engrafted with HLA-mismatched CD34+Lin- donor cells. Preconditioning of the SCID-hu bone with a single dose of sublethal (350 rad) whole body irradiation (WBI) immediately before cell injection enhanced the repopulation of the bone grafts with donor cells and, in some instances, resulted in complete repopulation. After WBI, as few as 500 fetal bone marrow CD34+Lin- cells injected in the human bone grafts resulted in donor-derived hematopoiesis. Donor progenitor cells recovered from the SCID-hu bone grafts 8 weeks postinjection had the capacity to repopulate secondary groups of HLA-disparate fetal human bones in SCID-hu mice with B and myeloid cells as well as CD34+ cells in some recipients. In addition, these cells repopulated fetal human thymus fragments in SCID mice with donor thymocytes including immature CD4+CD8+ and mature CD4+CD8- as well as CD4-CD8+ subsets. These results indicate that the fetal human bone implants of SCID-hu mice can support the maintenance of a cell population that has both multilineage potential and repopulating potential for periods of time as long as 16 weeks. The SCID-hu bone model consistently supported the engraftment of both fetal and adult CD34+Lin- cells without the administration of exogenous human cytokines to these animals. This model is currently being used to permit the isolation and characterization of candidate human hematopoietic stem cells (HSCs) and provide important information critical for human HSC therapy in humans.

Human fetal bone marrow early progenitors for T, B, and myeloid cells are found exclusively in the population expressing high levels of CD34.

Experimentation on human stem cells is hampered by the relative paucity of this population and by the lack of assays identifying multilineage differentiation, particularly along the lymphoid lineages. In our current study, phenotypic analysis of low-density fetal bone marrow cells showed two distinct populations of CD34+ cells: those expressing a high density of CD34 antigen on their surface (CD34hi) and those expressing an intermediate level of CD34 antigen (CD34lo). Multiple tissues were used to characterize the in vitro and in vivo potential of these subsets and showed that only CD34hi cells support long-term B lymphopoiesis and myelopoiesis in vitro and mediate T, B, and myeloid repopulation of human tissues implanted into SCID mice. CD34lo cells repeatedly failed to provide long-term hematopoietic activity in vivo or in vitro. These results indicate that a simple fractionation based on well-defined CD34 antigen levels can be used to reproducibly isolate cells highly enriched for in vivo long-term repopulating activity and for multipotent progenitors, including T- and B-cell precursors. Additionally, given the limited variability in the results and the high correlation between in vitro and in vivo hematopoietic potential, we propose that the CD34hi population contains virtually all of the stem cell activity in fetal bone marrow and therefore is the population of choice for future studies in hematopoietic stem cell development and gene therapy.

Human hematolymphoid cells in SCID mice.

The severe combined immunodeficient C.B.-17 scid/scid (SCID) mouse has been widely used to study the normal processes of murine lymphoid differentiation. To create an in vivo model of the human hematolymphoid system, this mouse strain has been engrafted with human organ systems (the SCID-hu mouse) or with human peripheral blood mononuclear cells (the hu-PBL-SCID mouse). These mouse models have now been characterized and used to analyze human infectious diseases, hematopoiesis, malignancies and vaccines.