Search

The limited life span of most blood cells requires the continuous production of cells, which in adults exceeds 10(12) cells/day. This impressive production of cells (approximately 4 x 10(16) cells over a lifetime) is achieved by the proliferation and differentiation of committed progenitor cells, which themselves are derived from a population of pluripotent stem cells with self-renewal potential. Paradoxically, the large majority of stem cells in adult bone marrow are quiescent cells. One possibility is that stem cells, like other somatic cells, have only a limited replicative potential (< 100 divisions). This hypothesis is supported by two key observations and the consideration that, in theory, 55 divisions can yield 4 x 10(16) cells. First, it was shown that 'candidate' stem cells purified from fetal and adult tissue showed dramatic functional differences in turn-over time and the ability to produce cells with stem cell properties, Second, these functional differences were found to correlate with a measurable loss of telomere repeats despite the presence of low but readily detectable levels of telomerase in all purified cell fractions. In order to address questions about the role of telomeres in normal and malignant haemopoiesis, we developed a quantitative fluorescence in situ hybridization technique. Here we review the characteristics of this novel tool to assess the number of telomere repeats at the end of individual chromosomes and provide an overview of recent observations.

Efforts to expand the number of stem cells ex vivo are based on the assumption that the self-renewal properties of stem cells are subject to extrinsic control. Although loss of stem cell properties ex vivo has been well-documented, conclusive evidence for a gain of stem cell function in vitro has not been forthcoming. In this short discussion paper, some issues related to the self-renewal of stem cells in relation to future experimental strategies are presented.

In an attempt to improve our gene transfer efficiency into hematopoietic stem cells and to evaluate the capacity of immunoselected CD34+Thy-1+(CDw90) cells to reconstitute hematopoiesis following myeloablation, bone marrow (BM) transplantation was performed using autologous, immunoselected CD34+Thy-1+ cells in rhesus macaques. BM samples were positively selected for cells that express CD34, further subdivided using high gradient immunomagnetic selection for cells that express Thy-1, and transduced using a 7-day supernatant transduction protocol with a replication-defective retroviral vector that contained the human glucocerebrosidase (GC) gene. Circulating leukocytes were evaluated using a semiquantitative polymerase chain reaction (PCR) assay for the human GC gene, with the longest surviving animal evaluated at day 558. Provirus was detected at all time points in both CD20+ B cells and CD2+ dim T cells, but long-term gene transfer was not observed in the granulocyte population. The CD2+ dim population was phenotypically identified as being CD8+ natural killer cells. By day 302 and day 330, both the CD2+ bright and dim cell populations and sorted CD4+ and CD8+ cells had detectable provirus. Vector-derived GC mRNA was detected by reverse transcriptase (RT)-PCR analysis as far out as day 588. Thus, CD34+Thy-1+ cells isolated using high gradient magnetic separation techniques can engraft, be transduced with a replication-defective retroviral vector, and contribute to CD20+ B lymphocytes, CD8+ T lymphocytes, and CD4+ T lymphocytes; making them a suitable cell population to target for gene therapies involving lymphocytes.

To determine whether the sensitivity of the human long-term culture-initiating cell (LTC-IC) assay could be increased, we have evaluated a spectrum of different fibroblast cell lines for their abilities to influence the number of cells detectable as LTC-IC, to influence LTC-IC maintenance, and/or to influence LTC-IC differentiation into colony-forming cells (CFC) in cocultures containing various sources of LTC-IC. In a series of initial experiments with highly purified subpopulations of CD34+ cells from normal human marrow, no significant difference could be found between any of 3 different murine stromal fibroblast cells in terms of their support of either LTC-IC detection (CFC production) or maintenance (over a 6-week period), and all were equivalent to primary human marrow feeders (HMF). On the other hand, murine M2-10B4 fibroblasts engineered to produce high levels of both human granulocyte colony-stimulating factor (G-CSF) and interleukin-3 (IL-3; 190 and 4 ng/mL, respectively), either alone or mixed 1:1 with SI/SI fibroblasts engineered to produce high levels of soluble Steel factor (SF), with or without production of the transmembrane form of SF (60 and 4 ng/ mL, respectively), stimulated the production of up to 20-fold more CFC in LTC of cells from normal human marrow, G-CSF-mobilized blood or cord blood when compared with parallel cocultures containing HMF. Limiting dilution analysis of the CFC output from all three sources of LTC-IC showed that most of this increase was due to an ability of the engineered feeders to increase the plating efficiency of the LTC-IC assay (approximately 14-fold for marrow LTC-IC and approximately 4-fold for cord blood or mobilized blood LTC-IC). Analysis of the phenotype of these additionally recruited LTC-IC from marrow showed they had the same primitive CD34+CD45RA-CD71- phenotype as conventionally defined LTC-IC. The limiting dilution studies also showed that the average number of CFC produced per LTC-IC was additionally and independently increased to yield values of 18 CFC per LTC-IC in marrow, 28 for LTC-IC in cord blood, and 25 for LTC-IC in G-CSF-mobilized blood. Replating of cells from primary LTC with different feeders into secondary LTC-IC assays containing the best combination of engineered feeders showed that LTC-IC maintenance could be significantly enhanced (up to 7-fold as compared with primary cocultures containing HMF). However, this enhancement was still not sufficient to amplify the number of LTC-IC present after 6 weeks above the input value. Thus, engineering murine fibroblasts to produce sufficient SF, G-CSF, and IL-3 can markedly enhance the detection as well as the maintenance in vitro of a very primitive population of human progenitor cells present in normal adult marrow, mobilized blood, and cord blood by providing the most sensitive assay conditions thus far described. The present findings also provide new evidence of biologic heterogeneity between different cell populations that can be operationally identified as LTC-IC, thus re-emphasizing the importance of limiting dilution analyses to distinguish between quantitative and qualitative effects on these cells.

Vertebrate chromosomes terminate in variable numbers of T2AG3 nucleotide repeats. In order to study telomere repeats at individual chromosomes, we developed novel, quantitative fluorescence in situ hybridization procedures using labeled (C3TA2)3 peptide nucleic acid and digital imaging microscopy. Telomere fluorescence intensity values from metaphase chromosomes of cultured human hematopoietic cells decreased with the replication history of the cells, varied up to six-fold within a metaphase, and were similar between sister chromatid telomeres. Surprisingly, telomere fluorescence intensity values within normal adult bone marrow metaphases did not show a normal distribution, suggesting that a minimum number of repeats at each telomere is required and/or maintained during normal hematopoiesis.

Varying, limiting numbers of unseparated or purified cells (Ly-5.1), either from 14.5-day-old fetal liver (FL) or from adult bone marrow (BM) were coinjected with 10(5) unseparated BM cells (Ly-5.2) into lethally irradiated adult C57B1/6 recipients (Ly-5.2). The kinetics of donor cell repopulation of the lymphoid and myeloid compartments by Ly-5.1+ donor hematopoietic stem cells (ie, competitive repopulation units [CRU]) were monitored at various time points after the transplantation by Ly-5 analysis of the peripheral white blood cells (WBC). Recipients that had received on average less than 2 adult BM or FL CRU did not show a significant difference in the level of donor-reconstitution when analyzed 4 weeks after the transplantation, However, at 8 and 16 weeks, the FL recipients showed a significantly higher percentage of donor-derived nucleated peripheral blood cells than did the recipients of adult BM cells. Analysis of individual mice showed that approximately 80% of the recipients of FL CRU showed an increase in mature WBC output between 4 and 8 weeks after transplantation, whereas this occurred in less than 40% in the recipients of adult BM cells. In addition to this effect on mature cell output, the cellularity of the reconstituted BM was significantly higher in recipients of FL CRU than in recipients of adult BM CRU, even at 7 to 9 months after transplantation, which is consistent with an increased clonal expansion of FL CRU. When marrow cells from primary recipients of FL CRU were injected into secondary recipients, a significantly higher percentage of these mice showed donor-reconstitution of their lymphoid and myeloid compartments (P < .01) and to a greater extent (P < .008) as compared with mice that had received marrow cells from primary recipients of similar numbers of adult BM CRU. Taken together, these results show that individual FL CRU exhibit a greater proliferative activity in vivo than similar cells from adult BM that is accompanied by a greater production of daughter CRU.

Previous studies have shown that stem cells able to competitively reconstitute the hematopoietic system of lethally irradiated mice (competitive repopulating units [CRU]) can be obtained in highly purified form from adult mouse bone marrow (BM) by the isolation of cells with a Sca-1+Lin-WGA+ phenotype. We now report on the phenotypic characteristics of CRU from day-14.5 murine fetal liver (FL). Our results confirm previous reports of similarities between the two CRU populations but also reveal a few striking differences. Both were found to express the Sca-1 antigen (SCA-1+ and surface molecules that bind wheat germ agglutinin (WGA+), and both show an absence or low expression of a number of markers characteristic of mature hematopoietic cells: B220, Gr-1,ly-1 and Ter119 (together termed Lin*-). Limiting dilution analysis of recipients transplanted with purified Sca-1+Lin*- FL cells with intermediate forward- and side-scatter properties showed that the frequency of CRU in this FL subpopulation was one in 39 cells. This represents an enrichment of approximately 450-fold over the labeled but unseparated FL starting population (one in 17,300 total FL cells). These FL CRU also resembled their counterparts in adult BM in that they expressed high levels of MHC class I and CD43 and intermediate levels of heat-stable antigen (HSA) and c-kit and did not express, or expressed at a low level, Thy-1.2, CD71, and the antigen recognized by the Fall-3 monoclonal antibody (mAb). In contrast, a high percentage of the Sca-1+Lin*- cells isolated from 14.5-day-old FL stained with the AA4.1, anti-Mac-1, and the anti-CD45RB mAbs and retained Rhodamine 123 (Rh123(bright)), whereas the Sca-1+Lin-WGA+ CRU-containing fraction of adult BM cells was found to be AA4.1-, Mac-1-, CD45RB-, and Rh123(dull). These differences in phenotype between CRU in FL and adult BM indicate changes that occur during ontogeny in cells that are similar with respect to their totipotentiality and long-term repopulating potential and complement parallel observations of functional differences between these two populations of CRU.

The loss of telomeric DNA may serve as a mitotic clock which signals cell senescence and exit from cell cycle. Telomerase, and enzyme which synthesizes telomeric repeats de novo, is required to maintain telomere lengths. In humans, significant telomerase activity has been found in cells with essentially unlimited replicative potential such as reproductive cells in ovaries and testes, immortal cell lines and cancer tissues, but not in most normal somatic cells or tissues. We have now examined telomerase expression in subpopulations of hematopoietic cells from adult human bone marrow using a sensitive polymerase chain reaction-based telomeric repeat amplification protocol. Telomerase activity was found at low levels in the highly enriched primitive hematopoietic cells (CD34+CD71loCD45RAlo) and was increased transiently when these cells were cultured in the presence of a mixture of cytokines. In contrast, the early progenitors (CD34+CD71+) expressed telomerase activity at a higher level which was subsequently downregulated in response to cytokines. Telomerase activity remained low in the more mature CD34-cells upon exposure to cytokines. Taken together, our results suggest that telomerase is expressed at a basal level in all hematopoietic cell populations examined, is induced in a primitive subset of hematopoietic progenitor cells and is downregulated upon further proliferation and differentiation of these cells. We have previously observed telomere shortening in cytokine-stimulated primitive hematopoietic cells. The low and transient activation of telomerase activity described here thus appears insufficient to maintain telomere lengths in cultured hematopoietic cells.

Progress in understanding the abnormal regulation of hematopoiesis in chronic myelogenous leukemia (CML) would be facilitated if neoplastic cells, at all stages of the disease, could be studied in an animal model. In this report, we show that irradiated severe combined immunodeficient (SCID) mice can be transplanted with both normal (Philadelphia chromosome [Ph]-negative) and neoplastic (Ph+) cells from CML patients with either chronic or blast phase disease. Mice transplanted with peripheral blood (PB) or bone marrow (BM) cells from 9 of 12 chronic phase CML patients were well engrafted with human cells including multilineage colony-forming progenitors and CD34+ cells for at least 90 days posttransplantation. Repeated posttransplant injections of cytokines did not enhance the number of engrafted human cells. Interestingly, approximately 70% of the human progenitors found in the engrafted SCID BM were Ph-, suggesting that the growth of primitive normal cells is favored in this in vivo transplant model. A similar number of normal cells were found in mice transplanted with either PB or BM cells, suggesting that elevated numbers of primitive normal cells are present in CML PB. When cells from patients with CML in either myeloid or lymphoid blast crisis were transplanted into SCID mice, the BM of these mice was more rapidly repopulated and to a higher level than that observed with transplants of chronic phase cells. Moreover, all human colony-forming progenitors present in the BM of mice transplanted with blast crisis cells were Ph+, and the majority of cells showed the same morphological features of the blast crisis cells originally transplanted. These experiments provide a starting point for the creation of an animal model of CML and establish the feasibility of using this model for the future characterization of transplantable CML stem cells during disease progression.

Previous studies have suggested that Steel factor (SF) can influence the behavior of many types of hematopoietic progenitor cells both in vivo and in vitro, although whether these may include the most primitive populations of totipotent repopulating cells remains controversial. To approach this question, we measured the number of Sca1+Lin-WGA+ cells, the number of cells with demonstrable myeloid (long-term culture-initiating cell [LTC-IC]) or both myeloid and lymphoid (LTC-IC(ML)) potential in 4- to 5-week-old long-term cultures containing irradiated primary marrow feeder layers, and the number of multilineage long-term in vivo repopulating cells (competitive repopulating unit [CRU]) present in the marrow of W42/+ or W41/W41 mice compared to +/+ controls. There was no significant effect of either of these W mutations on the number of Sca1+Lin-WGA+ cells and, in W41/W41 mice, neither LTC-IC nor LTC-IC(ML) populations appeared to be affected. On the other hand, although W41/W41 and W42/+ cells could both be detected in the in vivo CRU assay, their numbers were markedly reduced (17- and seven-fold, respectively) in spite of the fact that both of these W mutant genotypes contained near normal numbers of day-9 and -12 colony-forming units-spleen (CFU-S). In vitro quantitation of erythroid (burst-forming units-erythroid [BFU-E]), granulopoietic (CFU-granulocyte/macrophage [CFU-GM]), multilineage (CFU-granulocyte/erythrocyte/monocyte/macrophage [CFU-GEMM]), and pre-B clonogenic progenitors (CFU-pre-B) also revealed no differences in the numbers (or proliferative potential) of any of these cells when W41/W41 or W42/+ and normal mice were compared, although day 3 BFU-E from both types of W mutant mice showed no response to the typical enhancing effect exerted by SF on their +/+ counterparts. Taken together, these findings are consistent with the view that SF activation of c-kit receptor-induced signaling events is not a rate-limiting mechanism controlling red blood cell production during normal development until hematopoietic cells differentiate beyond the day-3 BFU-E stage. Nevertheless, normal hematopoietic stem cells do appear to be responsive to SF, since their W mutant counterparts display a disadvantage in the in vivo setting which is exaggerated under conditions of hematopoietic regeneration. On the other hand, alternative mechanisms also appear to contribute to the regulation of hematopoietic stem cell numbers in vivo and to their detection as LTC-IC in vitro.

The development of in vitro conditions that promote a numerical expansion of hematopoietic stem cells (HSCs) with long-term reconstituting ability has been a long-standing goal in experimental hematology. In previous studies, we showed that input numbers of such cells, i.e., competitive repopulating units (CRU), could be maintained for 2 weeks when purified Sca-1 + Lin-WGA + adult bone marrow (BM) cells were cultured in serum-free and stromal cell-free cultures containing Steel factor (SF), interleukin 6 (IL-6), and erythropoietin (Epo). In separate studies, we showed that limiting numbers of purified fetal liver (FL) cells that are highly enriched for CRU display a higher proliferative and self-renewal potential in vivo compared with similar cells purified from adult BM. These findings prompted us to explore the possibility of achieving a numerical expansion of purified FL cells in culture. Although we observed an extensive increase in the number of nucleated FL cells in all culture conditions tested, none of the cultures, including cultures in serum-containing medium and cocultures on a preestablished feeder layer of BM stromal cells of S17 cells, sustained the expected expansion or even supported the maintenance of input numbers of FL CRU. Single cell cultures showed that the production of nucleated cells by purified Sca-1 ++ Lin.-AA4.1 + FL cells stayed behind that of purified Sca-1 + Lin-WGA + adult BM cells. Taken together, our results show that a variety of culture conditions tested, including conditions that support maintenance and limited expansion of adult BM CRU do not support the production of repopulating stem cells from FL. Because such expansion can be observed in vivo, this system appears useful to search for novel culture conditions and-perhaps yet unidentified-cytokines or other microenvironment-related factors that may be required for FL CRU to prevent loss of repopulation potential in vitro and allow these cells to exhibit their expected self-renewal potential.

In previous studies, we described long-term cultures of subpopulations of CD34+ cord blood cells that were fractionated using antibodies specific for CD45RA and CD71. In the present study, we plated the most primitive CD34+CD45RAlowCD71low cells individually and analyzed the proliferation and expansion kinetics of the various hematopoietic progenitors included in this subpopulation. The proliferation capacity of single progenitors was assessed by the total number of cells produced, and their expansion capacity was assessed by the production of CD34+ and colony-forming cells. Monolineage progenitors (erythroid, granulocyte, and macrophage progenitors) showed the lowest expansion capacity, which correlated with a relatively short life span (< 37 days) of the generated colonies. Among these cells, erythroid progenitors showed the highest proliferative potential. Based on their relative cell content and proliferation/expansion capacities, two types of both granulomacrophagic (GM) and multipotent (Mix) progenitors were identified. GM progenitors producing colonies with > 70% granulocytes had a low expansion capacity and gave rise to colonies that were sustained for < 42 days. A second type of GM progenitors, giving rise to colonies with > 90% macrophages, had a higher expansion capacity and the colonies they produced were sustained for > 72 days. On the other hand, Mix progenitors producing colonies with > 70% erythroid cells had a lower expansion capacity and life span (< 46 days in culture), than Mix progenitors producing colonies with > 90% myeloid cells (life span of > 72 days). These results demonstrate, at the single-cell level, a correlation between the type of progenitor cell and its expansion capacity. In turn, the expansion capacity of the progenitors was correlated with the life span of the colonies produced. Furthermore, our results suggest that in cord blood, at least two types of bipotential myeloid (GM) progenitors and two types of multipotent (Mix) progenitors can be distinguished, that have respectively low and high proliferative and expansion capacity.

Despite the increasing use of bone marrow transplantation (BMT) as a treatment for a wide variety of diseases, the numbers and types of cells required for both rapid and sustained recovery of hematopoiesis are not well defined. To investigate further the potential of transplants consisting of highly enriched populations of long-term repopulating cells, we transplanted a series of Ly-5.2 mice given various doses (750, 800, 850, 900, or 950 cGy) of total-body irradiation (TBI) with 30 or 90 Sca-1+Lin-WGA+ marrow cells isolated from congenic Ly-5.1 donors. As expected, mature progeny derived from these cells, belonging to both myeloid and lymphoid compartments, could be detected with increasing case in recipients given radiation doses from 750 to 900 cGy TBI. Surprisingly, expression of this potential was significantly reduced in mice that had received 950 cGy TBI. This contrasts with the capacity of the same number of purified Sca-1+Lin-WGA+ cells to generate readily detectable progeny in 950 cGy treated mice given a simultaneous transplant of 10(5) normal marrow cells or 2 x 10(5) serially passaged marrow cells. We suggest that this variable behavior of purified stem cells in differently treated recipients may reflect radiation dose-dependent differences in the types or levels of expression of factors that regulate transplanted stem cell proliferation and differentiation in vivo and that above a certain threshold radiation dose, this may result in an irreversible loss of long-term reconstituting potential. Regardless of the nature of the underlying mechanism, this study shows that the extent of donor repopulation after BMT can be a function not only of the number of stem cells transplanted but also of the conditioning of the recipient and whether other cell types are also injected.

CD44 is a widely expressed, multifunctional, cell-surface glycoprotein that has been implicated in the regulation of normal hematopoiesis. In addition, expression of particular isoforms of CD44 has been associated with malignant transformation and/or the acquisition of metastatic potential. In this study, we used two recently developed monoclonal anti-CD44 antibodies, one reactive with an epitope shared by many CD44 isoforms and the other with an epitope unique to CD44 isoforms containing amino acids encoded by the alternatively spliced exon v10, to compare the expression of CD44 on primitive hematopoietic cells from the marrow of normal individuals and their neoplastic counterparts present in the peripheral blood of patients with chronic myeloid leukemia (CML). Multiparameter fluorescence-activated cell sorter (FACS) analysis and cell sorting studies showed that CD44 is normally expressed at high to very high levels on both long-term culture-initiating cells (LTC-IC) and granulopoietic colony-forming cells (granulocyte-macrophage colony-forming units [CFU-GM]). In contrast, primitive erythropoietic progenitors (burst-forming units-erythroid [BFU-E]) in normal marrow were more homogeneous in their expression of CD44, and very few (less than 5%) showed the very high levels of CD44 seen on 20% to 25% of LTC-IC and CFU-GM. Antibody staining showed the expression of exon v10-containing CD44 isoforms to be restricted to a small subpopulation (4% to 8%) of morphologically recognizable mature (CD34-) myeloid cells within the light-density fraction of normal marrow cells. Reverse transcription-polymerase chain reaction (RT-PCR) analysis showed the presence of two exon v10-containing mRNA species. In CML, a significantly greater proportion of the circulating neoplastic CFU-GM expressed very high levels of CD44, and these CFU-GM were accompanied by an increased number of light density v10+ cells, including some that coexpressed CD34. Nonmalignant hematopoietic progenitors mobilized by prior chemotherapy and growth factor treatment of patients with Hodgkin's disease or acute myeloid leukemia in remission showed no changes in CD44 expression relative to normal marrow progenitors. These results provide evidence of early differentiation-associated changes in CD44 expression during normal hematopoiesis in vivo that may be deregulated in the neoplastic clone of patients with CML.

In this report, we describe a modification of the assay for long-term culture-initiating cells (LTC-IC) that allows a subset of murine LTC-IC (designated as LTC-ICML) to express both their myeloid (M) and lymphoid (L) differentiative potentials in vitro. The modified assay involves culturing test cells at limiting dilutions on irradiated mouse marrow feeder layers for an initial 4 weeks under conditions that support myelopoiesis and then for an additional week under conditions permissive for B-lymphopoiesis. All of the clonogenic pre-B progenitors (colony-forming unit [CFU] pre-B) detected in such postswitch LTC appear to be the progeny of uncommitted cells present in the original cell suspension because exposure of lymphoid-restricted progenitors to myeloid LTC conditions for > or = 7 days was found to irreversibly terminate CFU-pre-B production and, in cultures initiated with limiting numbers of input cells (no progenitors of any type detected in > 70% of cultures 1 week after the switch), the presence of CFU-pre-B was tightly associated with the presence of myeloid clonogenic cells, regardless of the purity of the input population. Limiting dilution analysis of the proportion of negative cultures measured for different numbers of input cells showed the frequency of LTC-ICML in normal adult mouse marrow to be 1 per 5 x 10(5) cells with an enrichment of approximately 500-fold in the Sca-1+ Lin-WGA+ fraction, as was also found for competitive in vivo repopulating units (CRU) and conventionally defined LTC-IC. LTC-ICML also exhibited the same resistance to treatment in vivo with 5-fluorouracil (5-FU) as CRU and LTC-IC, thereby distinguishing these three populations from the great majority of both in vitro clonogenic cells and day 12 CFU-S. The ability to quantitate cells with dual lymphoid and myeloid differentiation potentials in vitro, without the need for their prior purification, should facilitate studies of totipotent hematopoietic stem cell regulation.

We have previously characterized the proliferative response of primitive CD34+ cells, purified from adult bone marrow, umbilical cord blood, and fetal liver, to a mixture of hematopoietic stimulators (steel factor [SF], interleukin-3 [IL-3], IL-6, and erythropoietin [Epo]) in serum-free liquid cultures. In the present study, we assessed the effects of the hematopoietic inhibitors, macrophage inflammatory protein-1 alpha (MIP-1 alpha), transforming growth factor-beta (TGF-beta), and tumor necrosis factor-alpha (TNF-alpha), on the cytokine-induced proliferation of three different CD34+ cell subpopulations derived from cord blood and on total CD34+ cells derived from fetal liver. In cultures of cord blood cells, addition of MIP-1 alpha inhibited the numerical expansion of primitive CD34+ cells (CD34+ CD45RAlow CD71low cells) without inhibiting the proliferation of more mature subpopulations enriched for myeloid (CD34+ CD45RA+ CD71low cells) or erythroid (CD34+ CD45RAlow CD71+ cells) progenitors. TGF-beta significantly reduced the proliferation of all three subpopulations, although its effects were more pronounced on cells of the erythroid lineage, particularly immature erythroid progenitors. Similarly, TNF-alpha preferentially inhibited total nucleated and CD34+ cell production in the subpopulation enriched for erythroid cells. However, in contrast to TGF-beta, TNF-alpha preferentially inhibited the proliferation of more mature erythroid progenitors. In a separate set of experiments, MIP-1 alpha, TGF-beta, and TNF-alpha were added to cultures of total CD34+ cells purified from fetal liver. In keeping with the fact that the majority of the progenitors contained in these cells were erythroid progenitors, the inhibitory effects of the three cytokines were similar to those observed in cultures of CD34+ CD45RAlow CD71+ cord blood cells. The results of the present study demonstrate that MIP-1 alpha, TGF-beta, and TNF-alpha have the capacity to modulate cytokine-induced proliferation of cord blood and fetal liver progenitors. The differential effects of these three cytokines confirm their pleiotropic nature as regulators of hematopoiesis.

The observation that colony-forming unit-spleen (CFU-S) cells can be found in the yolk sac, fetal liver, and various hematopoietic tissues in adult mice has contributed to the concept that "stem cells" represent a single type of immortal cell that is present at variable numbers in tissues at different stages of development. However, more recent studies have shown dramatic changes in various functional properties of purified "candidate" stem cells during ontogeny. Together with previous studies documenting developmental changes in early hematopoietic cells and observations showing a decrease in the mean telomere length of chromosomes in hematopoietic cells during ontogeny, these findings are difficult to reconcile with the concept of a stem cell as an immortal and invariable cell type. In this minireview, some developmental aspects of stem cell biology are discussed in the context of stem cell definitions and transplantation biology.

Mobilized peripheral blood is increasingly used as a source of cells to provide hematopoietic reconstitution in marrow ablated patients. There is controversy as to how the engraftment potential of these preparations should be evaluated. The long-term culture-initiating cell is a very primitive hematopoietic with unique phenotypic feature. The potential role and significance of these cells in peripheral blood cell autografts is discussed.

Alternate splicing and glycosylation produce multiple CD45 isoforms which are selectively expressed on the surface of cells of the haemopoietic system. The expression of CD45RA, CD45RB and CD45RO on CD34+ and CD34- haemopoietic cells from umbilical cord blood, bone marrow and fetal liver were studied by flow cytometry. CD34+ subpopulations defined by CD45 isoform expression were sorted from bone marrow and tested in long-term culture assays. By combining results of functional studies with phenotypic data and previously published information, the following pattern of CD45 isoform expression on early haemopoietic cells was established. The most primitive CD34+ cells are CD45RO+ CD45RB+ and express low or undetectable levels of CD45RA. Upon erythroid differentiation, CD34+ cells remain CD45RO+ CD45RB+, whereas commitment into the myeloid and lymphoid lineages coincides with down-regulation of CD45RO and up-regulation of CD45RA. As a result, the majority of CD34+ cells can be divided into two mutually exclusive populations of cells which express either CD45RO or CD45RA. This notion was confirmed in this study by three-colour immunofluorescence. The alternative expression of various CD45 isoforms on functionally distinct haemopoietic cells suggests an important role for these molecules in the proliferation and differentiation of haemopoietic cells.

Incubation of chronic myeloid leukemia (CML) marrow for 10 days in vitro causes a marked and selective loss of very primitive Philadelphia chromosome (Ph)+ as compared with Ph- progenitors. We have autografted 22 patients with CML (16 in first chronic phase [group 1] and 6 with more advanced disease [group 2]) with marrow treated in this way to facilitate restoration of Ph- hematopoiesis after intensive therapy. Hematologic recovery to greater than 0.5 x 10(9)/L neutrophils occurred in 16 patients, and to greater than 20 x 10(9)/L platelets in 15 of 21 evaluable patients at a median of 29 and 48 days postautograft, respectively. Regenerating marrow cells were 100% Ph- in 13 patients and 75% to 94% Ph- in 3. Between 4 and 36 months (median 12) postautograft, Ph+ cells became detectable in all but 1 (who died in remission) of the 13 patients who achieved complete cytogenetic remission. Four of 7 evaluable patients treated with low-dose interferon alpha were returned to complete cytogenetic remission. Thirteen group 1 patients (81%) are alive 1.0 to 5.7 years (median 2.6) after autografting: 4 in complete cytogenetic remission, 2 in hematologic remission, 6 in chronic phase, and 1 in myeloid blast phase. Three group 2 patients (50%) are alive at 2.6, 3.8, and 4.3 years after autografting: 1 in partial cytogenetic remission, 1 in chronic phase, and 1 in accelerated phase. Thus, autografts of cultured marrow can result in prolonged restoration of Ph- hematopoiesis for some patients with CML.

Reverse transcription-polymerase chain reaction (RT-PCR) was used to detect interleukin-1 beta (IL-1 beta) mRNA in candidate human hematopoietic stem cells. The cells, obtained from adult bone marrow (BM) or umbilical cord blood, had a CD34+ CD45RAlo CD71lo phenotype and were further fractionated into CD38+ and CD38- or Thy-1+ and Thy-1- subpopulations. The purity of these fractions was always more than 99%. IL-1 beta and CD34 mRNA were detected in pools of 30 BM-derived CD34+ CD45RAlo CD71lo cells. To further exclude any contribution by contaminating cells, individual cells were analyzed for CD34 and IL-1 beta mRNA. Positive results were obtained with 2 of 5 individual BM-derived CD34+ CD45RAlo CD71lo CD38+ cells isolated by micromanipulation after overnight culture in serum-free medium without any exogenous cytokines, and 1 of 10 individual CD34+ CD45RAlo CD71lo CD38- cells isolated immediately after sorting. Moreover, of 10 pools of three BM-derived CD34+ CD45RAlo CD71lo cells cultured overnight in the presence of a mixture of various cytokines (Steel factor, IL-3, IL-6, macrophage colony-stimulating factor [M-CSF], erythropoietin, and IL-3/granulocyte-macrophage colony-stimulating factor [GM-CSF] fusion protein), 5 were positive for IL-1 beta mRNA. This result was compatible with more than 20% (95% confidence limit 0.06-0.61) of the BM cells with the CD34+ CD45RAlo CD71lo phenotype expressing IL-1 beta mRNA. IL-1 beta expression was also consistently observed from day 0 to day 9 in liquid cultures of cord-blood-derived CD34+ CD45RAlo CD71lo Thy-1+ or Thy-1- cells. The cultures contained the same combination of cytokines and resulted in an expansion of cell numbers of up to 400-fold. GM-CSF mRNA was not detected in the equivalent of 75 cells at any day, even though it could be detected with high sensitivity in control stromal cells. Because IL-1 beta is a powerful and pleiotropic biomodulator of cytokines and adhesion molecules, our observations suggest that at least some primitive hematopoietic cells do not merely respond passively to signals from their environment, but may themselves regulate the paracrine production of cytokines from neighboring stromal cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Peripheral blood cells (PBCs) collected by leukapheresis after progenitor mobilization with chemotherapy and growth factors have been used successfully to replace marrow autografts in protocols requiring stem-cell support. Moreover, such transplants are often associated with more rapid recovery of blood cell counts than is routinely achieved with bone marrow. While conditions that mobilize colony-forming cells (CFCs) into the circulation are becoming increasingly well characterized, little information is available as to how these or other mobilizing treatments may influence the release of more primitive cells into the peripheral blood. To quantitate the peripheral blood content of such cells, we used the long-term culture-initiating cell (LTC-IC) assay, which detects a cell type that is able to produce progeny CFCs after a minimum of 5 weeks in cultures containing marrow fibroblasts. In this report, we present the findings on 21 patients who were transplanted over a 7-year period at our institution with PBCs alone. PBCs were collected in steady-state (n = 6) or during the recovery phase after high-dose cyclophosphamide (Cy; n = 15, nine with and six without additional growth factor administration). PBCs collected from another 11 patients given granulocyte colony-stimulating factor (G-CSF) were transplanted together with autologous marrow. Time-course studies of nine patients after Cy +/- granulocyte-macrophage CSF (GM-CSF) showed that CD34+ cells, CFCs, and LTC-ICs fell from normal to undetectable levels after Cy, and increased at the time of white blood cell (WBC) recovery: LTC-ICs to a mean of sixfold and CFCs to a mean of 26-fold higher than normal. The mean number of CD34+ cells, CFCs, and LTC-ICs present in the PBC harvest was twofold to 10-fold higher after mobilization than in steady-state collections; however, more than 2-log interpatient variability was observed. After PBC transplantation, the median time to a WBC count more than 10(9)/L was 12 days; polymorphonuclear leukocyte (PMN) count more than 0.5 x 10(9)/L, 15 days; and platelet count more than 20 x 10(9)/L, 17 days, although patients who received fewer than 1.5 x 10(5) CFCs/kg had a more than 50% chance of delayed count recovery (> 28 days). Patients who received Cy + GM-CSF-stimulated PBCs had more rapid and consistent platelet recoveries as compared with other groups receiving Cy mobilized or steady-state PBCs alone, and a rapid WBC recovery after Cy predicted a rapid WBC recovery after transplantation.(ABSTRACT TRUNCATED AT 400 WORDS)

We have previously shown that the most primitive human hematopoietic cells are included within a cell subpopulation expressing high levels of CD34 and low or undetectable levels of CD45RA and CD71. In this study, cord blood cells with this phenotype were sorted and further separated based on their expression on the Thy-1 antigen. The proliferation and differentiation of the purified cell fractions in response to a mixture of hematopoietic cytokines was analyzed in serum- and stroma-free liquid cultures. Thy-1+ cells (25% of CD34+ CD45RAlo CD71lo cells) were particularly enriched for high proliferative potential colony-forming cells (HPP-CFC; up to 45% of the clonogenic cells), whereas Thy-1- cells were enriched for multipotential colony-forming cells (CFU-MIX; up to 46% of the clonogenic cells). When both subpopulations were cultured in serum-free liquid cultures supplemented with a cytokine mixture that included steel factor, interleukin-6 (IL-6), granulocyte-macrophage colony-stimulating factor (GM-CSF)/IL-3 fusion protein, M-CSF, G-CSF, and erythropoietin, Thy-1+ cells showed a much higher numerical expansion of CD34+ cells (30,000-fold) and colony-forming cells (4,700-fold) than was observed in cultures initiated with Thy-1- cells (900-fold increase in CD34+ cell numbers and 241-fold increase in CFC numbers). Cells coexpressing CD34 and Thy-1 were only transiently expanded (up to 29-fold) and were not detected after day 22 of culture. When CD34+ CD45RAlo CD71lo Thy-1+ cells were cultured, either in semi-solid or liquid cultures, in the presence of anti-Thy-1 antibody, a significant reduction in progenitor cell numbers (particularly HPP-CFC) was observed. In contrast, CD34+ CD45RAlo CD71lo Thy-1- cells were not affected by anti-Thy-1. The results of this study indicate that Thy-1 is expressed on primitive cord blood progenitors with the highest in vitro proliferative potential, and further suggest that Thy-1 is involved in hematopoietic cell development, possibly by mediating a negative signal that results in inhibition of primitive cell proliferation.

In long-term cultures (LTC) of human haematopoietic cells, primitive progenitors termed LTC-initiating cells can be maintained for several months and will differentiate to produce clonogenic cells and mature granulocytes and macrophages when provided with a supportive feeder layer of adherent mesenchymal cells. Primitive haematopoietic cells become associated with this feeder layer and their proliferative status and differentiation are regulated by their interaction with these feeder cells and the growth factors they produce. Both positive and negative regulators are generated in LTC and the balance between these diverse factors is readily manipulated by both direct and indirect mechanisms which appear to operate in a localized fashion. These features parallel those believed to characterize the mechanisms that regulate haematopoiesis in the bone marrow microenvironment in vivo and suggest that further analysis of the LTC system will be useful in delineating the full mystery of this process.

Normal murine bone marrow (BM) cells were sorted on the basis of low forward and orthogonal light scatter properties, Sca-1 expression (Sca-1+), lack of staining with a cocktail of mature hematopoietic lineage markers (Lin-), and binding of wheat germ agglutinin (WGA+). This approach allowed the reproducible isolation of a very small subpopulation (0.037% +/- 0.023% of all nucleated BM cells) that was approximately 400-fold enriched in cells capable of reconstituting both lymphoid and myeloid lineages in lethally irradiated recipients. Transplantation of 30 or 10 of these Sca-1+Lin-WGA+ cells resulted in > or = to 20% donor-derived nucleated peripheral blood cells 3 months posttransplantation in 100% and 22% of the recipients, respectively. When Sca-1+Lin-WGA+ cells were cultured in serum-free medium supplemented with Steel factor, interleukin-6 (IL-6), and erythropoietin (with or without IL-3), a large increase in total cell number, including cells with an Sca-1+Lin-WGA+ phenotype was observed. Single cell cultures showed that 90% to 95% of the input cells underwent at least one division during the first 2 weeks and the remainder died. Interestingly, this proliferative response was not accompanied by a parallel increase in the number of cells with both lymphoid and myeloid repopulating potential in vivo, as quantitation of these by limiting dilution analysis showed they had decreased slightly (1.3-fold) but not significantly below the number initially present. These results demonstrate that Sca-1+Lin-WGA+ cells with long-term repopulating potential can be maintained for 2 weeks in a serum- and stroma cell-free culture, providing a simple in vitro system to study their behavior under well-defined conditions. The observed expansion of Sca-1+Lin-WGA+ cells in vitro without a concomitant increase in reconstituting cells also shows that extensive functional heterogeneity exists within populations of cells with this surface phenotype.

Recent studies in our laboratory have shown striking differences in the functional properties of candidate hematopoietic stem cells purified from fetal, neonatal, and adult human tissues. These differences include the ability to produce CD34+ cells, the turnover rate, and the fraction of cells that respond to a mixture of cytokines. All these parameters decrease with the age of the cell donor, and some of these observations are summarized here. Extensive qualitative changes in hematopoietic cells from various stages of development should be taken into account in the design of novel therapeutic strategies.

We have developed a reliable, purely immunological means of quantitatively labeling cells with magnetic colloidal dextran iron. Labeled cells can be efficiently separated with a high gradient magnetic filter inside a magnetic field. Separation conditions and filter design can be changed to accommodate large and small scale positive or negative selection. This separation technique has been applied to tumor cell depletions, stem cell enrichment, and large scale preclinical T-cell depletions.

Different models have been proposed to explain lineage commitment in hemopoiesis. Some suggest that lineage commitment occurs in a stochastic manner without the direct influence of extracellular factors; others postulate that cytokines determine whether multipotent cells will become erythroid or granulocyte/macrophage progenitors. In the present study, the patterns of proliferation and differentiation of individually sorted human cord blood-derived primitive hemopoietic cells (highly enriched for multipotent progenitors) were analyzed in a serum-free culture system supplemented with different cytokine combinations. In a first set of experiments, the response of individual cells to different cytokine combinations was compared, whereas in a second set of experiments, single cells were allowed to undergo one division after which the two daughter cells were physically separated and cultured in either the same or different cytokine combinations. Proliferation of progenitor cells was absolutely dependent on cytokines, and the combination of mast cell growth factor plus interleukin 6 was sufficient to induce mitosis. When cytokine combinations favoring erythropoiesis and/or myelopoiesis were added to the cultures, a more vigorous proliferative response of the sorted primitive progenitors was observed. Interestingly, the relative proportions of granulocyte/macrophage, erythroid, and multipotent progenitors remained more or less the same regardless of the cytokine combination used, indicating a permissive rather than an instructive role for cytokines in hemopoietic differentiation. Asymmetric cell divisions, defined as a division that yields two daughter cells with distinct functional properties, were observed in 3-17% of the progenitor cells capable of forming colonies under our experimental conditions. In the rest, symmetric divisions involving multipotent and lineage-committed progenitors were observed. The results of this study demonstrate that the asymmetric cell divisions that occur in the early stages of hemopoiesis at the level of multipotent progenitors cannot be skewed by the addition of specific cytokine combinations. These findings support the hypothesis that lineage commitment in hemopoiesis occurs in a stochastic manner by mechanisms that remain to be elucidated.

We have previously identified, based on the expression of the CD45RA and CD71 antigens, three major subpopulations of CD34+ cells derived from human umbilical cord blood: CD34+ CD45RAloCD71lo cells (up to 42% multipotent progenitors), CD34+ CD45RA+ CD71lo cells (90% myeloid progenitors), and CD34+ CD45RAloCD71+ cells (70% erythroid progenitors). In the present study, we have investigated the long-term proliferation and differentiation of these subpopulations in response to hematopoietic cytokines. Cells from each subpopulation were cultured for 38 days in serum- and stroma-free liquid cultures supplemented with cytokine combinations that favor either erythropoiesis or myelopoiesis. In keeping with their high content of primitive progenitors, CD34+ CD45RAloCD71lo cells showed the highest CD34+ cell expansion (up to 532-fold) throughout the culture period, followed by CD34+ CD45RA+ CD71lo (130-fold) and CD34+ CD45RAloCD71+ (28-fold) cells. Interestingly, the cytokine combination favoring myelopoiesis was always more efficient in inducing CD34+ cell expansion than the one favoring erythropoiesis. In all but one of the cultures, a predominance of myelopoiesis was observed after 2 weeks, even in those supplemented with the cytokine mixture that favors erythropoiesis. Only when CD34+ CD45RAloCD71+ cells were cultured in the presence of erythroid cytokine mixture, erythropoiesis was evident at all time points. However, such cultures could be sustained for only 29 days. The results of this study demonstrate that the cord blood-derived CD34+ cell compartment consists of functionally distinct cell subpopulations that possess different proliferative capacities in vitro. Our results also show that the cytokine combinations used here were able to modulate proliferation and, to a much lesser extent, differentiation of such subpopulations, probably by favoring the expansion of committed progenitors rather than by acting on uncommitted cells.

Long-term culture initiating cells (LTC-IC) are primitive hemopoietic progenitors that give rise to clonogenic cells when provided with a supportive feeder layer of mesenchymal cells. These LTC-IC possess many of the characteristics expected of marrow-repopulating "stem cells" including high proliferative and multilineage-differentiative capacity and resistance to 4-hydroperoxy-cyclophosphamide (4-HC) killing. In addition, stem cells are known to persist and may proliferate in murine LTC, and human marrow grown in LTC has been successfully used as hemopoietic support for myeloablative therapy. LTC-IC, as well as clonogenic precursors, circulate in normal peripheral blood, and the concentration of both progenitor types can be increased by cytotoxic chemotherapy and/or growth factors. When mobilized peripheral blood cells are used as hemopoietic support for high-dose chemo/radiotherapy, engraftment has often been more rapid than that achieved with autologous marrow. Thus, primitive hemopoietic cells circulate in human blood, which can enable hemopoietic reconstitution following aggressive therapy for malignant disease.

When chronic myeloid leukemia (CML) marrow is set up in long-term culture (LTC), Philadelphia chromosome (Ph)-positive (Ph+) cells typically decline and Ph-negative (Ph-) hematopoietic cells often become detectable. In 1987, we initiated a study to evaluate the feasibility of using 10-day cultured marrow autografts to allow intensive treatment of CML. Patients were selected on the basis of a previous assessment of the frequencies of normal and leukemic LTC-initiating cells (LTC-IC) remaining in their marrow after 10 days of LTC. Of the 87 patients evaluated, 36 (41%) were considered eligible, and 22 (15 in first chronic phase [CP], Group 1; and 7 with more advanced disease, Group 2) were autografted with 10-day cultured marrow after intensive therapy. Satisfactory hematological recovery occurred in 16 patients, and of these, only Ph- cells were detected in 13 (nine in Group 1), with 76-94% Ph- cells in the other three (two in Group 1). Ph+ cells reappeared between 4 and 36 months post-autograft in all but one of the 13 patients in whom complete (morphological and cytogenetic) remission had been achieved; the remaining patient died in remission. Nine of these twelve patients were then treated with alpha-interferon (IFN-alpha) 1-3 x 10(6) units/m2, 3-7 days/week; four returned to complete remission, three developed increasing numbers of Ph+ cells, and two are still too early to evaluate. Fifteen patients (12 in Group 1) remain alive and well, nine in hematological remission (eight in Group 1), 9 to 64 months (median 28) post-autograft.(ABSTRACT TRUNCATED AT 250 WORDS)

Serum-free cultures supplemented with IL-6, IL-3, steel factor, and erythropoietin support extensive production of erythroid cells from purified bone marrow stem cell candidates which are themselves maintained in number in such cultures. In this study, the mechanism responsible for the observed maintenance of primitive hematopoietic cells in rapidly proliferating suspension cultures was examined. The following models were considered: (1) proliferating cells represent a small minority of cells at start of culture (large quiescent pool), (2) self-renewal of primitive cells (balanced by loss through differentiation and death), and (3) asymmetrical divisions (each division of a primitive cell yielding one equally primitive daughter cell and one less primitive, i.e., committed daughter cell). To discriminate between these various models, the proliferative behavior of purified CD34+ CD71 cells was studied at a population level using the fluorescent membrane dye PKH26 and at a single cell level by studying cultures of individually sorted CD34+ CD45RAloCD71lo bone marrow stem cell candidates. The results from these experiments indicate that the majority of purified stem cell candidates do not rapidly proliferate in response to the combination of growth factors used and that limited production of CD34+ cells compensates for the loss of such cells in culture. These observations have to be taken into account in the development of clinical useful strategies for the expansion of primitive hematopoietic bone marrow cells ex vivo.

High gradient magnetic separators (HGMS) create magnetic field gradients that can be used to attract much smaller and less magnetic particles than those required for conventional magnetic separation techniques. As a result cells can be labeled with submicron magnetic particles and still be separated using an HGMS filter. Typically, HGMS filters consist of random arrays of wire such as stainless steel wool. Wire elements arranged regularly in a filter should allow more efficient separation of cells. Filters were constructed containing ordered wire arrays composed of 430 series stainless steel wire mesh with wire diameters of 50, 100, or 150 microns. The ability of these filters to separate T cells from peripheral blood mononuclear cell suspensions was tested and found superior to random arrays of 302 series stainless steel wire (Thomas et al, 1992). Target cells recognized by OKT5 monoclonal antibody were cross-linked to dextran-iron particles of approximately 20 nm in diameter. Separation conditions were optimized and after one passage through the filter 88% of the OKT5+ cells were recovered in the enriched fraction with 85% purity (%OKT5+). Multiple passages (3 times) could achieve 99% purity with 68% recovery. Variations in separation flow rate had a large effect on the balance between purity and recovery. Optimum separation efficiencies were achieved only when > 10(8) cells were processed. The primarily cause of nonspecific entrapment of CD8- cells was not nonspecific magnetic labeling of cells but the physical (nonmagnetic) characteristics of the filter/filter chamber.

We previously described that cells with a CD34+CD71lo phenotype from adult human bone marrow are maintained at constant numbers in long-term suspension cultures supplemented with interleukin-6 (IL-6), IL-3, mast growth factor (MGF) (a c-kit ligand), and erythropoietin (Epo). In view of the large increase in cell numbers in such cultures (for example, > 10(6)-fold per cell), this was an unexpected finding. The following models for the observed maintenance of CD34+CD71lo cells in our cultures were considered: (1) survival of non-dividing cells; (2) self-renewal balanced by loss of cells; (3) asymmetrical divisions; and (4) combinations of the above. Two experimental strategies were explored to discriminate between these models. In the first, sorted CD34+CD45RAloCD71lo cells were labeled with the flourescent tracking dye PKH26, followed by analysis of PKH26 fluorescence of CD34+CD71lo and other cells present in the cultures at various times (up to 11 weeks). In the second approach, single CD34+CD45RAloCD71lo cells were directly sorted into individual wells, and growing cells were then analyzed by flow cytometry. Results from these experiments indicated a considerable variability in (1) the number of surviving input cells (ranging from 30 to 80%); (2) the proportion of cells that contributed significantly to the total cell production measured at day 20 (ranging from 1 to 5%); and (3) the number of CD34+ cells present in individual clones. Taken together, the observed maintenance of primitive CD34+ cells in our cultures apparently involves a combination of survival of CD34+CD71lo cells with a vary low turnover together with a very limited production of CD34+ cells. Clonal heterogeneity, differences in cell cycle kinetics between CD34+ and CD34- cells, and observations that the majority of bone marrow-derived CD34+CD45RAloCD71lo cells do not show a rapid proliferative response to a mixture of IL-6, IL-3, MGF, and Epo will have to be taken into account in the development of experimental strategies aimed at clinically useful expansion of primitive hematopoietic cells ex vivo.

Blood cells originate from hematopoietic stem cells that are located at different sites during ontogeny. Production of human stem cells and their progeny in culture is expected to have important implications for experimental therapeutic strategies involving gene transfer and transplantation. Here we report striking differences between primitive hematopoietic cells purified from adult bone marrow, umbilical cord blood, and fetal liver in cytokine-supplemented, serum-free cultures. In such cultures both the fraction of responding cells and their ability to produce CD34+ progenitor cells decreased markedly with the age of the cell donor. These results document extensive, ontogeny-related functional differences between primitive hematopoietic cells.

To study the role of different cytokine combinations on the proliferation and differentiation of highly purified primitive progenitor cells, a serum-free liquid culture system was used in combination with phenotypic and functional analysis of the cells produced in culture. CD34+ CD45RAlo CD71lo cells, purified from umbilical cord blood by flow cytometry and cell sorting, were selected for this study because of their high content of clonogenic cells (34%), particularly multipotent progenitors (CFU-MIX, 12% of all cells). Four cytokine combinations were tested: (1) mast cell growth factor (MGF; a c-kit ligand) and interleukin-6 (IL-6); (2) MGF, IL-6, IL-3, and erythropoietin (Epo); (3) MGF, IL-6, granulocyte-macrophage colony-stimulating factor (GM-CSF)/IL-3 fusion protein (FP), macrophage colony-stimulating factor (M-CSF), and granulocyte-CSF (G-CSF); and (4) MGF, IL-6, FP, M-CSF, G-CSF, and Epo. Maximum numbers of erythroid progenitors (BFU-E, up to 55-fold increase) and mature erythroid cells were observed in the presence of MGF, IL-6, IL-3, and Epo, whereas maximum levels of myeloid progenitors (CFU-C, up to 70-fold increase) and mature myeloid cells were found in cultures supplemented with MGF, IL-6, FP, M-CSF, and G-CSF. When MGF, IL-6, FP, M-CSF, G-CSF, and Epo were present, maximum levels of both erythroid and myeloid progenitors and their progeny were observed. These results indicate that specific cytokine combinations can act directly on primitive hematopoietic cells resulting in significant expansion of progenitor cell numbers and influencing their overall patterns of proliferation and differentiation. Furthermore, the observations presented in this study suggest that the cytokine combinations used were unable to bias lineage commitment of multipotent progenitors, but rather had a permissive effect on the development of lineage-restricted clonogenic cells.

Major hurdles in studies of stem cell biology include the low frequency and heterogeneity of human hematopoietic precursor cells in bone marrow and the difficulty of directly studying the effect of various culture conditions and growth factors on such cells. We have adapted the cell analyzer imaging system for monitoring and recording the morphology of limited numbers of cells under various culture conditions. Hematopoietic progenitor cells with a CD34+ CD45RAlo CD71lo phenotype were purified from previously frozen organ donor bone marrow by fluorescence activated cell sorting. Cultures of such cells were analyzed with the imaging system composed of an inverted microscope contained in an incubator, a video camera, an optical memory disk recorder and a computer-controlled motorized microscope XYZ precision stage. Fully computer-controlled video images at defined XYZ positions were captured at selected time intervals and recorded at a predetermined sequence on an optical memory disk. In this study, the cell analyzer system was used to obtain descriptions and measurements of hematopoietic cell behavior, like cell motility, cell interactions, cell shape, cell division, cell cycle time and cell size changes under different culture conditions.

Expression of Thy-1 on hematopoietic cells from human fetal liver (FL), cord blood (CB), and bone marrow (BM) was studied with a novel anti-Thy-1 antibody, 5E10. Specificity of 5E10 for human Thy-1 was demonstrated by immunoprecipitation of a 25-35-kD molecule, and the sequence of a cDNA that was cloned by immunoselection of COS cells transfected with a cDNA library derived from a 5E10+ cell line. Two- and three-color immunofluorescence staining experiments revealed that the Thy-1 expression is restricted to, an average, 1-4% of FL, CB, and BM cells, and binding to these cell types is essentially restricted to a very small subset of lymphoid cells and approximately 25% of CD34+ cells. Thy-1+ CD34+ cells were further characterized as CD38lo/CD45RO+/CD45RA-/CD71lo/c-kit(lo) and rhodamine 123dull. When CD34+ cells were sorted on the basis of Thy-1 expression, the majority of clonogenic cells were recovered in the CD34+Thy-1- fraction, whereas the majority of cells capable of producing myeloid colonies after 5-8 wk of long-term culture (long-term culture initiating cells) were recovered in the Thy-1+CD34+ fraction. In addition to CD34+ cells, Thy-1 was found to be expressed on a variable, very small number (< 1%) of CD34- mononuclear cells in BM, CB, and peripheral blood that were further characterized as CD3+ CD4+ lymphocytes. The restricted expression of Thy-1 on primitive hematopoietic cells is in agreement with a previous report (Baum et al., 1992. Proc. Natl. Acad. Sci. USA. 89:2804) in which Thy-1 expression was used to enrich for primitive hematopoietic cells from fetal tissue. Compared with those previous studies, we found Thy-1 expression on a larger proportion of CD34+ cells (25% in our study vs. 5% in Baum et al.) and furthermore performed studies on Thy-1 expression on CD34+ cells from CB, FL, and BM in relation to markers that are known to be differentially expressed on hematopoietic cells. Taken together our results indicate that Thy-1-specific antibody 5E10 is an attractive tool for further studies on the biology and purification of human stem cells.

The total number of clonogenic cells present in 5-week-old long-term cultures (LTC) initiated by seeding normal human marrow cells on competent adherent cell feeder layers allows for the quantitation of a more primitive hematopoietic input precursor cell type referred to as an LTC-initiating cell (LTC-IC). Previous studies have suggested that LTC-IC also circulate because production of clonogenic cells continues for many weeks when cells from the light-density (< 1.077 g/mL), T-cell-depleted fraction of normal blood are maintained on irradiated, marrow-derived feeder layers in LTC medium. We now show that the number of clonogenic cells present in such reconstructed LTC after 5 weeks is linearly related to the input number of peripheral blood (PB) cells over a wide range of cell concentrations, thereby permitting the quantitation of circulating LTC-IC by limiting dilution analysis. Using this approach, we have found the concentration of LTC-IC in the circulation of normal adults to be 2.9 +/- 0.5/mL. This is approximately 75-fold lower than the concentration of circulating clonogenic cells (ie, burst-forming units-erythroid plus colony-forming units [CFU] granulocyte-macrophage plus CFU-granulocyte, erythroid, monocyte, megakaryocyte) and represents a frequency of LTC-IC relative to all nucleated cells that is approximately 100-fold lower than that measured in normal marrow aspirate samples. Characterization studies showed most circulating LTC-IC to be small (low forward light scatter and side scatter), CD34+, Rh-123dull, HLA-DR-, and 4-hydroperoxycyclophosphamide-resistant cells, with differentiative and proliferative potentialities indistinguishable from LTC-IC in normal marrow. Isolation of the light-density, T-cell-depleted, CD34+, and either HLA-DR(low) or Rh-123(dull) fraction of normal blood yielded a highly enriched population of cells that were 0.5% to 1% LTC-IC (approximately 1,500-fold enriched beyond the light-density, T-cell-depletion step), a purity comparable to the most enriched populations of human marrow LTC-IC reported to date. However, purification of PB LTC-IC on the basis of these properties did not allow them to be physically separated from a substantial proportion (> 30%) of the clonogenic cells in the same samples, in contrast to previous findings for LTC-IC and clonogenic cells in marrow. These studies show the presence in the blood of normal adults of a relatively small but readily detectable population of functionally defined, primitive hematopoietic cells that share properties with marrow LTC-IC, a cell type thought to have in vivo reconstituting potential.(ABSTRACT TRUNCATED AT 400 WORDS)

The peripheral blood of chronic myeloid leukemia (CML) patients with chronic-phase disease and elevated white blood cell (WBC) counts typically contains markedly increased numbers of a variety of neoplastic pluripotent and lineage-restricted hematopoietic progenitors. These include cells detected in standard colony assays as well as their more primitive precursors. The latter are referred to as long-term culture-initiating cells (LTC-IC) because of their ability to generate clonogenic cell progeny detectable after a minimum of 5 weeks incubation on competent fibroblast feeder layers. In this study, we have investigated a number of the properties of the LTC-IC and clonogenic cells present in the blood of such CML patients with high WBC counts. This included an analysis of the light scattering properties of these progenitors, as well as their expression of CD34 and HLA-DR, Rhodamine-123 staining, and in vitro sensitivity to 4-hydroperoxycyclophosphamide. In the case of LTC-IC, the production of different types of lineage-restricted and multipotent progeny was also analyzed. Most of the circulating LTC-IC and clonogenic cells in the CML patients studied (on average approximately 70% and approximately 90%, respectively) showed features of proliferating or activated cells. This is in marked contrast to the majority of progenitors in the blood of normal individuals and most of the LTC-IC in normal marrow, all of which exhibit a phenotype expected of quiescent cells. Interestingly, a significant proportion of the circulating clonogenic cells and LTC-IC in the CML samples studied (on average approximately 10% and approximately 30%, respectively) appeared to be phenotypically similar to normal circulating progenitors, although their absolute numbers were indicative of a neoplastic origin. Both phenotypes of circulating CML clonogenic cells and LTC-IC could be obtained at approximately 10% to 20% purity by differential multiparameter sorting. These findings suggest that expansion of the Philadelphia chromosome-positive clone at the level of the earliest types of hematopoietic cells results from the activation of mechanisms that enable some, but not all, signals that block the cycling of normal stem cells to be bypassed or overcome. In addition, they provide strategies for purifying these primitive leukemic cells that should facilitate further analysis of the mechanisms underlying their abnormal proliferative behavior.

The possibility of separating cells on the basis of levels of antigen expression was explored in a model system using fixed erythrocytes and high gradient magnetic separation (HGMS). Fixed human erythrocytes were labelled to varying degrees with tetrameric monoclonal antibody complexes specific for both dextran and glycophorin A-M. The cells were then mixed and incubated with dextran iron particles prior to magnetic separation. The small size of the dextran iron particles (less than 0.2 microns) resulted in quantitative magnetic labelling of cells as shown using fluoresceinated anti-dextran antibodies and flow cytometry. The relationships between the initial percentage of labelled cells, cell recovery, non-specific entrapment of unlabelled cells, the purity of the removed fraction, the degree of antigen expression and separation conditions (flow rate and field strength) were determined and used to establish separation conditions that allowed recovery of cells that differ only in the degree of antibody labelling.

Clinical uses of gene transfer to bone marrow transplants require the establishment of a reproducible method for infecting large numbers of very primitive hematopoietic cells at high efficiency using cell-free retrovirus-containing media. In this study we report the results of experiments with preparations of a high-titer (2-5 x 10(7)/ml) helper-free recombinant neo(r) retrovirus that indicate this goal can now be achieved based on measurements of gene transfer efficiencies to cells referred to as long-term culture initiating cells (LTC-IC) because they give rise to clonogenic cells after greater than or equal to 5 wk in long-term culture (LTC). Intermittent, repeated exposure of normal human marrow mononuclear cells to virus-containing supernatant over a 3-d period of cell maintenance on an IL-3/granulocyte colony-stimulating factor (G-CSF) producing stromal layer resulted in gene transfer efficiencies to LTC-IC of 41%; a level previously obtainable only using co-cultivation infection techniques. Marrow cells enriched greater than or equal to 500-fold for LTC-IC (1-2% pure) by flow cytometry showed gene transfer efficiencies of 27% when infected in a similar fashion over a shorter period (24 h), but in the presence of added soluble IL-3 and G-CSF without stromal feeders, and this increased to 61% when Steel factor was also present during the infection period. By using a less highly enriched population of LTC-IC obtained by a bulk immunoselection technique applicable to large-scale clinical marrow harvests, gene transfer efficiencies to LTC-IC of 40% were achieved and this was increased to 60% by short-term preselection in G418. Southern analysis of DNA from the nonadherent cells produced by these LTC over a 6-wk period provided evidence of clonal evolution of LTC-IC in vitro. Leukemic chronic myelogenous leukemia LTC-IC were also infected at high efficiency using the same supernatant infection strategy with growth factor supplementation. These data demonstrate the feasibility of using cell-free virus preparations for infecting clinical marrow samples suitable for transplantation, as well as for further analysis of human marrow stem cell dynamics in vitro.

To directly study the biological properties of purified hematopoietic colony-forming cell precursors, cells with a CD34+ CD45RAlo CD71lo phenotype were purified from human bone marrow using density separation and fluorescence-activated cell sorting, and were cultured in serum-free culture medium supplemented with various cytokines. In the presence of interleukin 3 (IL-3), IL-6, erythropoietin, and mast cell growth factor (a c-kit ligand), cell numbers increased approximately 10(6)-fold over a period of 4 wk, and the percentage of cells that expressed transferrin receptors (CD71) increased from less than 0.1% at day 0 to greater than 99% at day 14. Interestingly, the absolute number of CD34+ CD71lo cells did not change during culture. When CD34+ CD71lo cells were sorted from expanded cultures and recultured, extensive cell production was repeated, again without significant changes in the absolute number of cells with the CD34+ CD71lo phenotype that were used to initiate the (sub)cultures. These results document that primitive hematopoietic cells can generate progeny without an apparent decrease in the size of a precursor cell pool.

The binding of lectins to nucleated cells from human bone marrow was studied in a search for markers that can be used to subdivide further immature hemopoietic cells that are characterized by their expression of CD34. Low-density bone marrow cells were indirectly labeled with biotinylated lectins and streptavidin-R-phycoerythrin (SA-RPE) together with FITC-labeled monoclonal anti-CD34. Four-parameter flow cytometric analysis was then performed and list mode data analyzed. Of the 21 lectins tested, only a few showed differential staining of CD34+ versus CD34- cells. These include soybean agglutinin (SBA) and Ulex europaeus agglutinin I (UE). Lycopersicon esculentum (LE) and Erythrina cristigalli (EC) reacted preferentially with, respectively, CD34+ and CD34- cells, suggesting their usefulness in some method to enrich for CD34+ cells. This possibility was tested by passing cells labeled with biotinylated lectins over a column containing streptavidin-coated beads. CD34+ cells could be enriched > 10-fold by competitive (sugar) elution of LE-labeled cells from the column. Similarly, depletion of biotinylated EC-labeled cells by passage through the streptavidin column enriched CD34+ cells several fold. The results of these studies document the reactivity of a large panel of lectins with subpopulations of nucleated bone marrow cells and indicate that certain lectins could possibly be used for development of cell separation procedures aimed at the selective enrichment of cells that express CD34.

Erythropoietin-receptor (EpR) expression on bone marrow cells from normal individuals and from patients with chronic myeloid leukemia (CML) was examined by multiparameter flow cytometry after stepwise amplified immunostaining with biotin-labeled Ep, streptavidin-conjugated R-phycoerythrin, and biotinylated monoclonal anti-R-phycoerythrin. This approach allowed the detection of EpR-positive cells in all bone marrow samples studied. Most of the EpR-positive cells in normal bone marrow were found to be CD45-dull, CD34-negative, transferrin-receptor-positive and glycophorin-A-intermediate to -positive. This phenotype is characteristic of relatively mature erythroid precursors, ie, colony-forming units-erythroid and erythroblasts recognizable by classic staining procedures. Approximately 5% of normal EpR-positive cells displayed an intermediate expression of CD45, suggesting that these represented precursors of the CD45-dull EpR-positive cells. Some EpR-positive cells in chronic myeloid leukemia (CML) bone marrow had a phenotype similar to the major EpR-positive phenotype in normal bone marrow, ie, CD34-negative and CD45-dull. However, there was a disproportionate increase in the relative number of EpR-positive/CD45-intermediate cells in CML bone marrow. Even more striking differences between normal individuals and CML patients were observed when EpR-expression on CD34-positive marrow cells was analyzed. Very few EpR-positive cells were found in the CD34-positive fraction of normal bone marrow, whereas a significant fraction of the CD34-positive marrow cells from five of five CML patients expressed readily detectable EpR. These findings suggest that control of EpR expression is perturbed in the neoplastic clone of cells present in patients with CML. This may be related to the inadequate output of mature red blood cells typical of CML patients and may also be part of a more generalized perturbation in expression and/or functional integrity of other growth factor receptors on CML cells.