Murine stromal cell line HESS-5 maintains reconstituting ability of Ex vivo-generated hematopoietic stem cells from human bone marrow and cytokine-mobilized peripheral blood

Human bone marrow (BM) or mobilized peripheral blood (mPB) CD34(+) cells have been shown to loose their stem cell quality during culture period more easily than those from cord blood (CB). We previously reported that human umbilical CB stem cells could effectively be expanded in the presence of human recombinant cytokines and a newly established murine bone marrow stromal cell line HESS-5. In this study we assessed the efficacy of this xenogeneic coculture system using human BM and mPB CD34(+) cells as materials. We measured the generation of CD34(+)CD38(-) cells and colony-forming units, and assessed severe-combined immunodeficient mouse-repopulating cell (SRC) activity using cells five days after serum-free cytokine-containing culture in the presence or the absence of a direct contact with HESS-5 cells. As compared with the stroma-free culture, the xenogeneic coculture was significantly superior on expansion of CD34(+)CD38(-) cells and colony-forming cells and on maintenance of SRC activity. The PKH26 study demonstrated that cell division was promoted faster in cells cocultured with HESS-5 cells than in cells cultured without HESS-5 cells. These results indicate that HESS-5 supports rapid generation of primitive progenitor cells (PPC) and maintains reconstituting ability of newly generated stem cells during ex vivo culture irrespective of the source of samples. This xenogeneic coculture system will be useful for ex vivo manipulation such as gene transduction to promote cell division and the generation of PPC and to prevent loss of stem cell quality.     

Mesenchymal stem cells secreting angiopoietin-like-5 support efficient expansion of human hematopoietic stem cells without compromising their repopulating potential

Clinical and preclinical applications of human hematopoietic stem cells (HSCs) are often limited by scarcity of cells. Expanding human HSCs to increase their numbers while maintaining their stem cell properties has therefore become an important area of research. Here, we report a robust HSC coculture system wherein cord blood CD34(+) CD133(+) cells were cocultured with mesenchymal stem cells engineered to express angiopoietin-like-5 in a defined medium. After 11 days of culture, SCID repopulating cells were expanded ~60-fold by limiting dilution assay in NOD-scid Il2rg(-/-) (NSG) mice. The cultured CD34(+) CD133(+) cells had similar engraftment potential to uncultured CD34(+) CD133(+) cells in competitive repopulation assays and were capable of efficient secondary reconstitution. Further, the expanded cells supported a robust multilineage reconstitution of human blood cells in NSG recipient mice, including a more efficient T-cell reconstitution. These results demonstrate that the expanded CD34(+) CD133(+) cells maintain both short-term and long-term HSC activities. To our knowledge, this ~60-fold expansion of SCID repopulating cells is the best expansion of human HSCs reported to date. Further development of this coculture method for expanding human HSCs for clinical and preclinical applications is therefore warranted.     

Phenotypic and functional analysis of human fetal liver hematopoietic stem cells in culture

Steady-state hematopoiesis and hematopoietic transplantation rely on the unique potential of stem cells to undergo both self-renewal and multilineage differentiation. Fetal liver (FL) represents a promising alternative source of hematopoietic stem cells (HSCs), but limited by the total cell number obtained in a typical harvest. We reported that human FL nonobese diabetic/severe combined immunodeficient (NOD/SCID) repopulating cells (SRCs) could be expanded under simple stroma-free culture conditions. Here, we sought to further characterize FL HSC/SRCs phenotypically and functionally before and following culture. Unexpanded or cultured FL cell suspensions were separated into various subpopulations. These were tested for long-term culture potential and for in vivo repopulating function following transplantation into NOD/SCID mice. We found that upon culture of human FL cells, a tight association between classical stem cell phenotypes, such as CD34(+) /CD38(-) and/or side population, and NOD/SCID repopulating function was lost, as observed with other sources. Although SRC activity before and following culture consistently correlated with the presence of a CD34(+) cell population, we provide evidence that, contrary to umbilical cord blood and adult sources, stem cells present in both CD34(+) and CD34(-) FL populations can sustain long-term hematopoietic cultures. Furthermore, upon additional culture, CD34-depleted cell suspensions, devoid of SRCs, regenerated a population of CD34(+) cells possessing SRC function. Our studies suggest that compared to neonatal and adult sources, the phenotypical characteristics of putative human FL HSCs may be less strictly defined, and reinforce the accumulated evidence that human FL represents a unique, valuable alternative and highly proliferative source of HSCs for clinical applications.     

Highly efficient ex vivo expansion of human hematopoietic stem cells using Delta1-Fc chimeric protein

Ex vivo expansion of hematopoietic stem cells (HSCs) has been explored in the fields of stem cell biology, gene therapy, and clinical transplantation. Here, we demonstrate efficient ex vivo expansion of HSCs measured by long-term severe combined immunodeficient (SCID) repopulating cells (SRCs) from human cord blood CD133-sorted cells using a soluble form of Delta1. After a 3-week culture on immobilized Delta1 supplemented with stem cell factor, thrombopoietin, Flt-3 ligand, interleukin (IL)-3, and IL-6/soluble IL-6 receptor chimeric protein (FP6) in a serum- and stromal cell-free condition, we achieved approximately sixfold expansion of SRCs when evaluated by limiting dilution/transplantation assays. The maintenance of full multipotency and self-renewal capacity during culture was confirmed by transplantation to nonobese diabetic/SCID/gammac(null) mice, which showed myeloid, B, T, and natural killer cells as well as CD133(+)CD34(+) cells, and hematopoietic reconstitution in the secondary recipients. Interestingly, the CD133-sorted cells contained approximately 4.5 times more SRCs than the CD34-sorted cells. The present study provides a promising method to expand HSCs and encourages future trials on clinical transplantation.     

Characterization of serum-free ex vivo-expanded hematopoietic stem cells derived from human umbilical cord blood CD133(+) cells

The development of ex vivo expansion of hematopoietic stem cells (HSCs) is a promising approach to restore the required bone marrow function of patients with hematological disorders. Previously, we have reported the development of an optimized serum-free and cytokines-limited defined medium using statistic methodology for umbilical cord blood-derived HSC expansion. The aim of this study was to analyze further the characteristics and functions of cells in vitro and in vivo when cultured in this defined medium. After a 7-day batch culture, the average absolute fold expansions for CD133(+) cells, CD34(+)CD133(+) cells, CD34(+)CD38() cells, CD133(+)CD38(-) cells, CD34(+)CXCR4(+) cells, CD133(+)CXCR4(+) cells, and long-term culture-initiating cells were 21-, 20-, 723-, 618-, 160-, 384-, and 8-fold, respectively. The high enrichment of CD38(-) cells and CXCR4(+) cells of the CD34(+) subpopulation provided a very early uncommitted HSC proliferation and homing ability. Furthermore, the expanded cells showed a high level of telomerase activity to maintain their telomere length and repopulated the lethally irradiated NOD/SCID mice in vivo. These results indicated that the cytokines limited expanded cells from CD133(+) cells could substantially support simultaneous expansion of various stem/progenitor cells and engraft with the expanded cells from a low number of HSCs initially.     

Generation of natural killer cells from serum-free, expanded human umbilical cord blood CD34+ cells

Natural killer (NK) cells are important effectors of the innate immune system, which exhibits cytolytic activity against infectious agents and tumor cells. NK cells are derived from CD34(+) hematopoietic stem cells (HSCs). Human umbilical cord blood (UCB) has been recognized as a rich source of HSCs. Previously, we have reported an optimized serum-free medium for ex vivo expansion of CD34(+) cells from UCB. In this study, the serum-free, expanded CD34(+) cells were tested to differentiate into NK cells and their induction kinetics. After 5 weeks of induction, the induced NK cells were characterized by analysis of surface antigens, IFN-gamma secretion, and cytotoxicity against K562 cells. The results indicated that NK cells derived from the serum-free, expanded CD34(+) cells exhibited both characteristics and functions of NK cells. Furthermore, the serum-free, expanded CD34(+) cells showed a significantly higher NK cell differentiation potential than freshly isolated CD34(+) cells. NK cells induced from serum-free, expanded CD34(+) cells showed a higher concentration of IFN-gamma secretion and ability of cytotoxicity than those from freshly isolated CD34(+) cells. Therefore, ex vivo-expanded CD34(+) cells in optimized serum-free medium could differentiate into NK cells and provided a promising cell source for immunotherapeutic approaches.     

Co-culture of umbilical cord blood CD34+ cells with human mesenchymal stem cells

Insufficient numbers of hematopoietic stem cells (HSCs) and hematopoietic progenitor cells (HPCs) sometimes limit allogenic transplantation of umbilical cord blood (UCB). Ex vivo expansion may overcome this limitation. Mesenchymal stem cells (MSCs), as non-hematopoietic, well-characterized skeletal and connective-tissue progenitor cells within the bone marrow stroma, have been investigated as support cells for the culture of HSCs/HPCs. MSCs are attractive for the rich environmental signals that they provide and for immunological compatibility in transplantation. Thus far, HSC/MSC co-cultures have mainly been performed in 2-dimensional (2D) configuration. We postulate that a 3-dimensional (3D) culture environment that resembles the natural in vivo hematopoietic compartment might be more conducive for regulating HSC expansion. In this study, we compared the co-culture of HSCs and MSCs in 2D and 3D configurations. The results demonstrated the benefit of MSC inclusion in HSC expansion ex vivo. Direct contact between MSCs and HSCs in 3D cultures led to statistically significantly higher expansion of cord blood CD34+ cells than in 2D cultures (891- versus 545-fold increase in total cells, 96- versus 48-fold increase of CD34+ cells, and 230- versus 150-fold increase in colony-forming cell assay [CFC]). Engraftment assays in non-obese diabetic/severe combined immunodeficiency mice also indicated a high success rate of hematopoiesis reconstruction with these expanded cells.     

Pre-differentiated human committed T-lymphoid progenitors promote peripheral T-cell re-constitution after stem cell transplantation in immunodeficient mice

T-cell re-constitution after allogeneic stem cell transplantation (alloSCT) is often dampened by the slow differentiation of human peripheral blood CD34(+) (huCD34(+) ) hematopoietic stem cells (HSCs) into mature T cells. This process may be accelerated by the co-transfer of in vitro-pre-differentiated committed T/NK-lymphoid progenitors (CTLPs). Here, we analysed the developmental potential of huCD34(+) HSCs compared with CTLPs from a third-party donor in a murine NOD-scid IL2Rγ(null) model of humanised chimeric haematopoiesis. CTLPs (CD34(+) lin(-) CD45RA(+) CD7(+) ) could be generated in vitro within 10 days upon co-culture of huCD34(+) or cord blood CD34(+) (CB-CD34) HSCs on murine OP9/N-DLL-1 stroma cells but not in a novel 3-D cell-culture matrix with DLL-1(low) human stroma cells. In both in vitro systems, huCD34(+) and CB-CD34(+) HSCs did not give rise to mature T cells. Upon transfer into 6-wk-old immune-deficient mice, CTLPs alone did not engraft. However, transplantation of CTLPs together with huCD34(+) HSCs resulted in rapid T-cell engraftment in spleen, bone marrow and thymus at day 28. Strikingly, at this early time point mature T cells originated exclusively from CTLPs, whereas descendants of huCD34(+) HSCs still expressed a T-cell-precursor phenotype (CD7(+) CD5(+) CD1a(+/-) ). This strategy to enhance early T-cell re-constitution with ex vivo-pre-differentiated T-lymphoid progenitors could bridge the gap until full T-cell recovery in severely immunocompromised patients after allogeneic stem cell transplantation.     

Application of porous glycosaminoglycan-based scaffolds for expansion of human cord blood stem cells in perfusion culture

In vitro expansion of hematopoietic stem cells (HSCs) has been employed to obtain sufficient numbers of stem cells for successful engraftment after HSC transplantation. A three-dimensional perfusion bioreactor system with a heparin-chitosan scaffold was designed and evaluated for its capability to support maintenance and expansion of HSCs. Porous chitosan scaffolds were fabricated by a freeze-drying technique and N-desulfated heparin was covalently immobilized within the scaffolds using carbodiimide chemistry. CD34+ HSCs isolated from umbilical cord blood by immunomagnetic separation were cultured within the porous scaffold in a perfusion bioreactor system. Control cultures were maintained on dishes coated with similar heparin-chitosan films. Oxygen uptake was measured during the culture period. After 7 days of culture, scaffolds were harvested for analysis. Cellular phenotype and HSC characteristics were evaluated via flow cytometry and colony forming unit assays. The results indicate good cell retention and proliferation within the perfused scaffolds. Oxygen consumption in the perfusion bioreactor system increased continuously during the culture, indicating steady cell growth. Cells from the perfused scaffold cultures showed higher percentages of primitive progenitors and exhibited superior colony forming unit performance as compared to cells from static cultures. In addition, perfusion culture at low oxygen (5%) enhanced the expansion of CD34+ cells and colony-forming activity compared to high oxygen (19%) cultures. The results suggest that perfusion culture of cord blood CD34+ cells under bone marrow-like conditions enhances HSC expansion compared to static cultures.     

Adherent cells generated during long-term culture of human umbilical cord blood CD34+ cells have characteristics of endothelial cells and beneficial effect on cord blood ex vivo expansion

Hematopoiesis depends on the association of hematopoietic stem cells with stromal cells that constitute the hematopoietic microenvironment. The in vitro development of the endothelial cell from umbilical cord blood (UCB) is not well established and has met very limited success. In this study, UCB CD34(+) cells were cultured for 5 weeks in a stroma-free liquid culture system using thrombopoietin, flt3 ligand, and granulocyte-colony stimulating factor. By week 4-5, we found that firmly adherent fibroblast-like cells were established. These cells showed characteristics of endothelial cells expressing von Willebrand factor, human vascular cell adhesion molecule-1, human intracellular adhesion molecule-1, human CD31, E-selectin, and human macrophage. Furthermore, when comparing an ex vivo system without an established endothelial monolayer to an ex vivo system with an established endothelial monolayer, better expansion of total nucleated cells, CD34(+) cells, and colony-forming units (CFUs)-granulocyte-macrophage and CFUs-granulocyte-erythroid-megakaryocyte-macrophage were found during culture. This phenomenon was in part due to the fact that a significant reduction of apoptotic fractions was found in the CD34(+) cells, which were cultured on the adherent monolayer for up to 5 weeks. To gather quantitative data on the number of endothelial cells derived from a given number of CD34 cells, we performed limiting dilution assay by using Poisson distribution: the number of tested cells (linear scale) producing a 37% negative culture (logarithmic scale) is the number of cells containing one endothelial cell. By this method, one endothelial cell may be found from 314 CD34(+) cells after 5 weeks of culture. These results suggest that the UCB CD34(+) cell fraction contains endothelial cell precursors, establishing the hematopoietic microenvironment and providing the beneficial effects through downregulating apoptosis on UCB expansion protocols. These observations may provide insight for future cellular therapy or graft engineering.     

Repopulating activity of ex vivo-expanded murine hematopoietic stem cells resides in the CD48-c-Kit+Sca-1+lineage marker- cell population

A better understanding of the biology of cultured hematopoietic stem cells (HSCs) is required to achieve ex vivo expansion of HSCs. In this study, clonal analysis of the surface phenotype and repopulating activity of ex vivo-expanded murine HSCs was performed. After 7 days of culture with stem cell factor, thrombopoietin, fibroblast growth factor-1, and insulin-like growth factor-2, single CD34-/lowc-Kit+Sca-1+lineage marker- (CD34-KSL) cells gave rise to various numbers of cells. The proportion of KSL cells decreased with increasing number of expanded cells. Transplantation studies revealed that the progeny containing a higher percentage of KSL cells tended to have enhanced repopulating potential. We also found that CD48 was heterogeneously expressed in the KSL cell population after culture. Repopulating activity resided only in the CD48-KSL cell population, which had a relatively long intermitotic interval. Microarray analysis showed surprisingly few differences in gene expression between cultured CD48-KSL cells (cycling HSCs) and CD48+KSL cells (cycling non-HSCs) compared with freshly isolated CD34-KSL cells (quiescent HSCs), suggesting that the maintenance of stem cell activity is controlled by a relatively small number of genes. These findings should lead to a better understanding of ex vivo-expanded HSCs.     

Mitochondrial DNA sequence heterogeneity of single CD34+ cells after nonmyeloablative allogeneic stem cell transplantation

We applied a single-cell method to detect mitochondrial DNA (mtDNA) mutations to evaluate the reconstitution of hematopoietic stem cells (HSCs) and committed progenitor cells after nonmyeloablative allogeneic stem cell transplantation in humans. In a total of 1,958 single CD34(+) cells from six human leukocyte antigen-matched sibling donor and recipient pairs, individual CD34(+) clones were recognized based on the observed donor- or recipient-specific mtDNA sequence somatic alteration. There was no overall reduction of mtDNA heterogeneity among CD34(+) cells from the recipient after transplantation. Samples collected from two donors over time showed the persistence of certain CD34(+) clones marked by specific mutations. Our results demonstrate the feasibility of distinguishing donor and recipient individual CD34(+) clones based on mtDNA mutations during engraftment. HSCs were not limited in number, and similar mtDNA heterogeneity levels suggested representation of the total stem cell compartment during rapid hematopoietic reconstitution in the recipient. Disclosure of potential conflicts of interest is found at the end of this article.     

体外分化胚胎干细胞为造血干细胞重建造血功能的研究

目的：探讨体外定向分化胚胎干细胞（ESCs）为造血干细胞（HSCs）对体内造血功能的重建作用。方法：将小鼠E14.1胚胎干细胞采用“三步诱导法”在体外分化发育为HSCs，造血克隆形成(CFU)实验观察其体外造血集落形成情况，免疫磁珠分选纯化HSCs移植给经亚致死剂量γ射线照射的雌性SCID小鼠，观察其植入及小鼠造血功能恢复情况。结果: 经过分阶段诱导，多种造血刺激因子联合应用能有效促进ESCs定向分化发育为HSCs,流式细胞仪检测HSCs特异性表面标志物CD34+/Sca-1+表达最高为（58.64±4.20）%，CFU培养能形成较多的红系、粒系/巨噬细胞系及混合细胞集落, Wright-Giemsa 染色显示为原始的造血细胞。此阶段的HSCs经分选纯化后移植给经γ射线照射后的小鼠，移植组小鼠+10 d造血功能开始恢复，观察40 d后除血小板恢复较慢外，白细胞、红细胞、血红蛋白等指标已接近正常，植入率为71.4%，存活率为43.0%，染色体检测证实已由受体鼠的XX转为供体鼠的XY。结论: 采用分阶段诱导的方法，可在体外定向诱导小鼠ESCs分化发育为HSCs，此来源的HSCs可以有效重建体内造血功能。     

Immunophenotype and functional characteristics of human primitive CD34-negative hematopoietic stem cells: the significance of the intra-bone marrow injection

The biology of hematopoietic stem cell (HSC) is a current topic of interest which has important implications for clinical HSC transplantation as well as for the basic research of HSC. The most primitive HSCs in mammals, including mice and humans, have long been believed to be CD34 antigen (Ag)-positive (CD34(+)) cells. In fact, bone marrow (BM), peripheral blood (PB), and cord blood (CB) stem cell transplantation studies indicate that a CD34(+) subpopulation in the BM, PB, or CB can provide durable long-term donor-derived lymphohematopoietic reconstitution. Therefore, CD34 Ag was used to identify/purify immature HSCs. However, Osawa et al. reported that murine long-term lymphohematopoietic reconstituting HSCs are lineage marker-negative (Lin(-)) c-kit(+)Sca-1(+)CD34-low/negative (CD34(low/-)), which are called CD34(low/-) KSL cells. Recently, human CB-derived CD34(-) HSCs, a counterpart of murine CD34(low/-) KSL cells, were successfully identified using an intra-bone marrow injection (IBMI) method. This review will update the concept of the immunophenotype and the functional characteristics of human primitive CD34(-) HSCs. In addition, the significance of the application of the IBMI technique in clinical HSC transplantation is also discussed. Recent rapid advances in understanding the biological nature of HSCs may make it possible to fully characterize the most primitive class of human HSCs in the near future.     

Phenotypic and functional reversal within the early human hematopoietic compartment

The fate of phenotypically defined human hematopoietic stem cells (hHSCs) in culture and the link between their surface marker expression profile and function are still controversial. We studied these aspects of hHSC biology by relating the expression of the early lineage markers (ELM) CD33, CD38, and CD71 on the surface of human umbilical cord blood (UCB) CD34(+) cells to their long-term nonobese diabetic/severe combined immunodeficient (NOD/SCID) mouse repopulation activity (LT-SRA). In uncultured UCB samples, LT-SRA was largely confined to the small CD34(+)ELM(-) cell fraction. CD34(+) cells expressing ELM markers at their surface usually lacked LT-SRA. After culturing UCB CD34(+) cells for 6 days in serum-free medium and on a feeder layer of Rat2 cells, the number of CD34(+)ELM(-) cells stayed roughly the same or showed a slight increase and the LT-SRA was preserved, suggesting a close association between LT-SRA and the CD34(+)ELM(-) phenotype. Indeed, transplantation of CD34(+)ELM(-) cells isolated from cultured UCB CD34(+) cells resulted in long-term hematopoietic reconstitution of conditioned NOD/SCID mice, whereas CD34(+)ELM(+) cells derived from the same cultures were devoid of LT-SRA. Remarkably, roughly 1% of the cells recovered from cultures initiated with isolated CD34(+)ELM(+) cells had lost ELM surface expression. Concurrently, the cultured CD34(+)ELM(+) cells acquired LT-SRA, suggesting that hematopoietic stem cells (HSCs) may arise by the dedifferentiation of early hematopoietic progenitor cells. The latter finding challenges the paradigm of unidirectional hematopoietic differentiation and opens new opportunities for HSC expansion prior to transplantation.     

Effect of addition of FLT-3 ligand and megakaryocyte growth and development factor on hemopoietic cells in serum-free conditions

The aim of this study was to clarify the mechanisms that regulate hematopoietic cell expansion in vitro by identifying defined culture conditions. We report the results of experiments with CD34(+) cells from cord blood (CB, n = 13), bone marrow (BM, n = 4), and mobilized peripheral blood stem cells (PBSC, n = 5) using two combinations of cytokines: (A) granulocyte colony-stimulating factor (G-CSF), interleukin-3 (IL-3), interleukin-6 (IL-6), stem cell factor (SCF), erythropoietin (EPO), insulin-like growth factor-1 (IGF-1), basic fibroblast growth factor (FGF-b) and (B) combination A plus FLT3 ligand (FL) and megakaryocyte growth and development factor (PEG rhMGDF). Cultures of immunoselected CD34(+) cells were performed in serum-free liquid medium without serum substitutes. The area under the curve (AUC) obtained by plotting the logarithm of the total number of viable cells, CD34(+) cells, and CFC per well, toward the week of culture was used as an index of cell expansion. With CB, a significant difference was obtained between the two combinations of cytokines with regard to the total number of viable cells, GM-CFC, and CD34(+) cells. The difference between the two combinations of cytokines obtained with BM was significant with respect to the total number of viable cells and CD34(+) cells but not for the erythroid and myeloid progenitors. When CD34(+) cells from peripheral blood stem cells (PBSC) were cultured in presence of the two combinations of cytokines, the difference in terms of AUC was not statistically significant. Our data indicate additional effects in terms of proliferation and expansion of hematopoietic cells in serum-free conditions when FL and polyethylene glycol (PEG) rhMGDF are included in culture and suggest a differential activity of these cytokines on cells from different hematopoietic sources.     

Cobblestone area-forming cells in human cord blood are heterogeneous and differ from long-term culture-initiating cells

The long-term culture-initiating cell (LTC-IC) assay is a physiological approach to the quantitation of primitive human hematopoietic cells. The readout using identification of cobblestone area-forming cells (CAFC) has gained popularity over the LTC-IC readout where cells are subcultured in a colony-forming cell assay. However, comparing the two assays, cord blood (CB) mononuclear cell (MNC) samples were found to contain a higher frequency of CAFC than LTC-IC (126 +/- 83 versus 40 +/- 31 per 10(5) cells, p = 0.0001). Overall, 60% of week-5 cobblestones produced by CB MNC were not functional LTC-IC and were classified as "false." Separation of CB MNC using immunomagnetic columns showed that false cobblestones were CD34(-)/lineage(+). Purified CD34(+) cells, as expected, gave very similar readouts in the two assays, with 4,084 and 3,468/10(5) cells being CAFC and LTC-IC, respectively. CD34(-)/lineage(-) cells did not form cobblestones or become CD34(+) on stroma or in cytokine culture. Human CB MNC contain a population of mature lineage(+) cells, possibly mature T or B cells, which, although producing cobblestone areas (CA), are not functional LTC-IC. The CAFC readout by this method, therefore, is unreliable for estimation of primitive hematopoietic cells by limiting dilution analysis in whole human CB or MNC and also may not detect CD34(-) CA stem cells.