The effect of IL-3 on the B lymphoid potential of human hemopoietic stem cells is controversial. Murine studies suggest that B cell differentiation from uncommitted progenitors is completely prevented after short-term exposure to IL-3. We studied B lymphopoiesis after IL-3 stimulation of uncommitted human CD34+CD38− cells, using the stromal cell line S17 to assay the B lymphoid potential of stimulated cells. In contrast to the murine studies, production of CD19+ B cells from human CD34+CD38− cells was significantly increased by a 3-day exposure to IL-3 (p < 0.001). IL-3, however, did not increase B lymphopoiesis from more mature progenitors (CD34+CD38+ cells) or from committed CD34−CD19+ B cells. B cell production was increased whether CD34+CD38− cells were stimulated with IL-3 during cocultivation on S17 stroma, on fibronectin, or in suspension. IL-3Rα expression was studied in CD34+ populations by RT-PCR and FACS. High IL-3Rα protein expression was largely restricted to myeloid progenitors. CD34+CD38− cells had low to undetectable levels of IL-3Rα by FACS. IL-3-responsive B lymphopoiesis was specifically found in CD34+ cells with low or undetectable IL-3Rα protein expression. IL-3 acted directly on progenitor cells; single cell analysis showed that short-term exposure of CD34+CD38− cells to IL-3 increased the subsequent cloning efficiency of B lymphoid and B lymphomyeloid progenitors. We conclude that short-term exposure to IL-3 significantly increases human B cell production by inducing proliferation and/or maintaining the survival of primitive human progenitors with B lymphoid potential.
Interleukin-3 is commonly included in ex vivo expansion and gene transfer protocols that attempt to induce cycling of human hemopoietic stem cells (1, 2, 3, 4). The incorporation of IL-3 in many cytokine combinations was prompted by the early observations that IL-3 is a positive regulator of early myeloerythropoiesis causing rapid cell proliferation, increased cell survival, and increased retroviral mediated transduction of myeloid progenitors (5, 6, 7, 8, 9, 10, 11, 12, 13). IL-3 acts synergistically with several other cytokines such as steel factor (SF)3 (14), Flt3 ligand (FL) (15), IL-6 (16, 17, 18), IL-11 (19), thrombopoietin (20), and G-CSF (21) on committed and uncommitted myeloid progenitors (22). However, the effect of IL-3 on lymphoid production from hemopoietic stem cells has received little attention and remains controversial.
In murine in vitro studies, exposure to IL-3 abrogated the B lymphoid potential of lymphohemopoietic progenitors (23, 24, 25). B lymphoid potential of murine stem cells was rapidly lost when cells were exposed to IL-3 in combination with either SF, IL-6, IL-11, or G-CSF during culture in semisolid medium (23). Later studies showed that T lymphoid (26) and NK cell (27) potential were also lost after pluripotent murine progenitors were exposed to IL-3.
In vivo studies, however, suggest that murine B lymphopoiesis is not completely abrogated after bone marrow is exposed to IL-3. For example, murine bone marrow transduced with retroviral vectors in the presence of IL-3 is able to fully reconstitute hemopoiesis and B lymphopoiesis in ablated recipients (28, 29). Similarly, human B lymphoid cells can be generated in immunodeficient mice engrafted with human CD34+ progenitors that have been exposed ex vivo to IL-3 (30, 31, 32, 33, 34). However, each of the in vivo studies transplanted large numbers of functionally heterogeneous cells, making it difficult to determine the effects of IL-3 on the lymphoid potential of specific progenitor and stem cell populations within the graft.
Until recently, in vitro studies of IL-3 on primitive human progenitors have not been possible because assays of early human B lymphopoiesis were not available. Recently, three similar assays that permit human B lymphopoiesis from primitive progenitors have been developed; each assay uses cocultivation of human cells on one of three murine bone marrow stromal lines (Sys1, MS5, or S17) and generates B lymphoid cells that are predominantly CD34−CD19+CD10+CD20− surface Ig− (35, 36, 37, 38). Fluckiger et al. (39) further developed the S17 stroma-based assay to produce surface Ig+ mature B cells and measurable quantities of soluble IgM by adding CD40 ligand during a second phase of culture. Further modification of the S17 stromal assay by our group has allowed the identification of B lymphoid, myeloid, NK, and dendritic progeny in clones derived from single CD34+CD38− cord blood cells (Ref. 40 ; unpublished observations).
Using the S17 stromal assay, we studied the effects of IL-3 on B lymphoid production from specific human hemopoietic progenitor subpopulations. No inhibition of B lymphoid cell production from CD34+CD38− human progenitors exposed short-term to IL-3 was seen. On the contrary, brief exposure to IL-3 significantly increased B lymphoid production from CD34+CD38− cells by inducing proliferation of B lymphoid and B lymphomyeloid progenitors. IL-3-responsive B lymphopoiesis was seen predominantly in the most immunophenotypically primitive progenitors that express very low levels of IL-3-Rα. We conclude that the incorporation of IL-3 into protocols using short-term ex vivo manipulation does not damage the B lymphoid potential of human hemopoietic stem cells.
Materials and Methods
Isolation of hemopoietic cells
Cord blood and bone marrow samples were obtained under guidelines of the Childrens Hospital Los Angeles’ Committee of Clinical Investigations and processed within 24 h of collection. Bone marrow samples were obtained from the filter screens of bone marrow harvests from the posterior iliac crests of two healthy donors (7 and 4 yr old). After Ficoll-Hypaque (Pharmacia, Piscataway, NJ) centrifugation, CD34+ cells were enriched from the mononuclear population with the MiniMACS device (Miltenyi Biotec, Auburn, CA) using manufacturer’s guidelines. CD34+-enriched cells were then incubated with CD34-FITC (HPCA2; Becton Dickinson Immunocytometry Systems (BDIS), San Jose, CA) and CD38-PE (Leu-17; BDIS). Isotype controls were used to set positive and negative quadrants and CD34+CD38+ and CD34+CD38− cells were isolated by FACSvantage using an argon laser. Gates for cell sorting have been previously described (41); the gate for CD34+CD38− sorting is shown as R2 in Fig. 4 C. Cell purity checked by reanalysis following isolation was 99.6%.
Initial stimulation conditions
Following isolation, cells were stimulated in 96-well plates (Becton Dickinson Labware, Franklin, NJ) for the first 3 days in B cell medium (RPMI 1640 (Irvine Scientific, Santa Ana, CA), 5% FCS (screened for B cell cultures), 50 μmol/L 2-ME (Sigma, St. Louis, MO), penicillin/streptomycin (Gemini Bio Products, Calabasas, CA), and glutamine (Gemini Bio Products)) containing combinations of the following cytokines: IL-3 (10 ng/ml), FL (100 U/ml, a gift from Dr. Charles Hannum, DNAX, Palo Alto, CA), IL-7 (10 ng/ml). In most experiments, the 3-day stimulation took place during cocultivation on S17 stroma (a gift from Dr. Kenneth Dorshkind, UCLA). In certain experiments, cells were stimulated instead either on the recombinant fibronectin fragment CH-296 (Takara Shuzo, Otsu, Shiga, Japan) or in suspension (i.e., in tissue culture plates). Following stimulation, at day 3, approximately one-half the culture medium was changed to fresh B cell medium as below and, if not already present, S17 stroma was added to begin B cell culture.
Culture and analysis of human cells on S17 stroma
Following 3 days of stimulation with or without IL-3, cells were cultured in B cell conditions, i.e., RPMI, 5% FCS, 2-ME, penicillin/streptomycin, and glutamine with FL (100 U/ml) on S17 stroma (40). Half-medium changes were performed every 7 days. At 7–14-day intervals, cultures were harvested, and cells were stained with trypan blue and counted to determine the fold increase in total viable cells since day 0. Aliquots were then analyzed by FACS to determine the proportion of B lymphoid (using CD19-FITC; BDIS) and myeloid (using CD33-PE; BDIS) cells. In all experiments, cultured cells were used as negative controls (IgG-FITC and IgG-PE; BDIS) to set parameters for positive Ag expression (thus allowing for autofluorescence and nonspecific binding to Ab commonly seen when analyzing cultured cells). The remaining cells were replated to continue cultures. CD19+ cell numbers were compared across experiments by standardizing data to an input day 0 cell number of 1000. The relative number of CD19+ cells produced in culture was thus calculated using the following formula: (% CD19+ of the total population/100) multiplied by (fold increase in total viable cells from day 0 × 1000). Clonal analysis was performed as previously described (40). In brief, single cord blood CD34+CD38− cells were plated by the Automated Cell Deposition Unit on the FACSvantage onto S17 stroma in B cell medium in individual wells of 96-well plates and cultured for 3 days with or without IL-3. On day 3 and again on day 7, one-half the medium was removed and replaced with B cell medium containing FL only. Timing of appearance of each clone was recorded. Clones large enough for analysis (>1000 cells) were then analyzed by FACS for CD19 expression to measure the presence of B cells. Aliquots of each clone (approximately 50%) were also replated into methylcellulose culture to detect CFU and thus prove the presence of myeloid progenitors (40). Clones in which at least 10% of cells were CD19+ and which also produced CFU in methylcellulose culture were recorded as B lymphomyeloid. Clones in which at least 10% of cells were CD19+ but which did not produce CFU were recorded as B lymphoid.
To analyze expression of IL-3-Rα by FACS analysis, CD34+-enriched cord blood cells were stained with combinations of CD34 (HPCA2)-FITC, CD38-APC (BDIS), CD19-FITC, and anti-human IL-3-Rα-PE (CDw123, clone 9F5, a nonblocking Ab; PharMingen, San Diego, CA). Anti-IL-3-Rα was used at 20 μl per 106 CD34+ cells (1/50 dilution). Cells were analyzed by FACSVantage using argon and HeNe lasers, and populations isolated by FACS were subjected to RT-PCR to detect IL-3Rα transcripts and to B lymphoid culture to assess B lymphoid potential after an initial 3-day stimulation with or without IL-3. RNA was extracted from 20,000 cells of each phenotype (CD34+ IL-3Rhigh, CD34+ IL-3Rdim, and CD34+ IL-3Rnegative) according to manufacturer’s guidelines using the RNA STAT-60 kit (Tel-Test, Friendswood, TX). One-half of each sample was subjected to RT-PCR (+RT) for cDNA production, and one-half was used as a negative control (−RT). The +RT and −RT products were then each divided in half for PCR detection of IL-3Rα and β2-microglobulin (used as a positive control for loading of cDNA). Primers for detection of IL-3Rα were designed based on the published cDNA and genomic sequences (42, 43), as follows: sense, CGT CGC TGC TGA TCG CGC, and antisense, CCC AGA CCA CCA GCT TGT CG. This primer pair amplified a 156-bp sequence (nucleotides 916-1072). No signal was detected in the absence of RT or when samples of genomic DNA were tested, confirming that the primers span at least one intron. Detection of β2-microglobulin cDNA used the following primers: sense, CTC GCG CTA CTC TCT CTT TC, and antisense, CAT GTC TCA ATC CCA CTT AAC. These primers were also confirmed to span at least one intron. PCR conditions for detection of IL-3Rα were as follows: 94°C (15 min for one cycle), 94°C (30 s), 58°C (30 s), 72°C (30 s) for 35 cycles, then 72°C (30 s for one cycle). Conditions for detection of β2-microglobulin were as follows: 94°C (15 min for one cycle), 94°C (1 min), 58°C (1 min), 72°C (2 min) for 33 cycles, then 72°C (10 min for one cycle). Gels were imaged using the Stratagene Eagle Eye system (La Jolla, CA).
Early addition of IL-3 increases CD19+ B cell output from CD34+CD38− cells
To determine whether short-term exposure of primitive human hemopoietic progenitors to IL-3 blocks their ability to subsequently differentiate into B cells, CD34+CD38− cells were initially cultured for 3 days with or without IL-3 on S17 stroma, and then assayed for subsequent B cell production during long-term culture on S17 stroma (in B cell medium without IL-3). A 3-day stimulation was chosen, as this is a commonly used period of ex vivo manipulation in many clinical hemopoietic gene transfer protocols. During the long-term B lymphoid culture, cell number was measured and cultures were serially analyzed by FACS for CD19 and CD33 expression. Early exposure to IL-3 did not inhibit, but instead increased B lymphoid production from both cord blood (n = 4) and bone marrow (n = 2) CD34+CD38− cells. Although the absolute level of B cell production varied from sample to sample, the effect of IL-3 was highly reproducible. The absolute number of CD19+ B cell progenitors was significantly increased in the S17 cultures initially exposed to IL-3 relative to those not exposed to IL-3 (n = 6, p < 0.001) (Fig. 1). The number of B lymphoid cells was higher after IL-3 exposure at all time points during subsequent long-term culture on S17 stroma. The proportion of CD33+ cells was not significantly changed by IL-3 stimulation. The increase in B cell numbers was accomplished by both a significant increase in total cell output (p < 0.001) and a significant increase in the purity (frequency) of CD19+ B cells in culture after IL-3 stimulation (p = 0.007). In cultures with cord blood, the number of CD19+ cells was 53.2 ± 26.5 (mean ± SEM)-fold greater with IL-3 stimulation than without IL-3, whereas total cell numbers in culture increased only 13.3 ± 2.7-fold with IL-3. Total cell and B cell proliferation, with or without IL-3 stimulation, was lower with bone marrow than with cord blood. Bone marrow total cell numbers were 4.9 ± 1.4-fold higher after short exposure to IL-3, and CD19+ cell numbers were 11.4 ± 4.3-fold higher after IL-3 exposure.
IL-3 in combination with either FL or IL-7 further enhances CD19+ B cell output from CD34+CD38− cells
We have previously noted that FL and IL-7 are able to increase B cell production from CD34+CD38− cells cultured on S17 stroma (37 ; unpublished data). We therefore next studied whether early exposure to IL-3 would further enhance B cell production from CD34+CD38− cells cultured in FL or IL-7. Cord blood cells were cultured on S17 stroma, either with FL or IL-7 each in the presence or absence of IL-3 during the first 3 days after isolation. As seen in Fig. 2, IL-3 increased CD19+ B cell production when added to either FL (p = 0.01, n = 4) or IL-7 (p = 0.004, n = 3). Adding IL-3 to FL or to IL-7 also slightly increased CD33+ myeloid cell production, but this effect did not reach statistical significance.
IL-3 stimulates uncommitted and primitive progenitors, and not more mature B lymphoid cells
The above experiments all studied the effect of a brief exposure to IL-3 early in culture. In contrast, B lymphoid cells were not produced from cord blood CD34+CD38− cells cultured on S17 stroma during continuous exposure to IL-3 (data not shown). Similarly, when IL-3 was added late to established B lymphoid cultures (day 14), myeloid cells became predominant and B lymphoid cells rapidly disappeared. We thus reasoned that IL-3 may increase B cell production from early or uncommitted progenitors, but inhibit growth of more mature CD19+ B cells produced during S17 cocultivation.
To determine more specifically at which stage of lymphopoiesis IL-3 acts to increase B lymphoid production, we studied fresh cells isolated at different stages of B lymphoid differentiation. Early exposure of cord blood and bone marrow CD34+CD38+ cells (a mixture of committed lymphoid and myeloid progenitors) to IL-3 gave variable results with no consistent increase or decrease in B lymphoid production (n = 9); B cell production from these mature progenitors, however, was not prevented by IL-3. In contrast, myeloid cells (measured by % CD33 expression) were consistently increased in S17 stroma cultures of CD34+CD38+ cells after an initial 3-day stimulation with IL-3.
Freshly isolated cord blood CD34+CD19+ pro-B cells cultured on S17 stroma grew poorly relative to more primitive progenitors, whether or not they were stimulated by IL-3. However, early exposure to IL-3 slightly increased CD19+ cell numbers short-term; at day 16, the number of IL-3-stimulated CD19+ cells increased 1.94-fold compared with a decrease to 5% of the starting number of CD19+ cells without IL-3 stimulation. When more mature B cells (freshly isolated CD34−CD19+ cells) were exposed to IL-3, cultures rapidly died; by day 15, no cells were detectable in culture. Thus, IL-3 increased B cell production predominantly from primitive, immunophenotypically uncommitted progenitors. B cell production from committed progenitors was not blocked, but was barely if at all increased. Mature B cells were not maintained when exposed briefly to IL-3.
IL-3 effects in the absence of S17 stroma
As human IL-3 does not act on murine cells (44), it is unlikely that the results seen in the above experiments were due to an indirect action of human IL-3 mediated through murine S17 stroma. To exclude this possibility, however, we studied the effect of IL-3 stimulation on B cell production in the absence of S17 stroma. Cord blood (n = 3) and bone marrow (n = 1) CD34+CD38− cells were cultured for 3 days either in suspension or on fibronectin in the absence or presence of IL-3. Cells were then washed and replated onto S17 stroma to assay B lymphoid production. The presence of IL-3 during 3-day culture, either in suspension or on fibronectin, increased subsequent CD19+ B cell production (suspension, p = 0.06, n = 3; fibronectin, p = 0.02, n = 4) (Fig. 3). Thus, the effect of IL-3 on human B cell production was not mediated through the murine S17 stromal cells.
IL-3R expression on progenitor cells
To explore further the issue of the type of progenitor able to respond to IL-3, we analyzed the expression of IL-3Rα on the surface of freshly isolated cord blood populations. CD19+ cells did not express IL-3Rα (data not shown). Most CD34+ cells also had no IL-3Rα expression detectable by FACS (Fig. 4,D). However, a small percentage of CD34+ cells (1.7%) expressed high levels of IL-3Rα (defined as CD34+ IL-3Rhigh) and approximately 12% of CD34+ expressed IL-3Rα levels near the threshold set by the isotype control (defined as CD34+ IL-3Rdim) (Figs. 4,D and 5A). CD34+ cells with high IL-3Rα expression had a more mature progenitor immunophenotype, i.e., they were CD38 positive and had low CD34 expression (Fig. 4,E). CD34+CD38− cells were either IL-3Rdim or IL-3Rnegative (Fig. 4,F). RNA expression of IL-3Rα was detectable by RT-PCR in all CD34+ populations, including CD34+ IL-3Rnegative cells (Fig. 5 B).
B lymphoid potential of progenitors isolated by IL-3Rα expression
To determine whether IL-3 responsiveness of B cell progenitors was similar in all progenitors that express detectable levels of IL-3Rα on the cell surface, CD34+CD38+ IL-3R+ cells and CD34+CD38− IL-3R+ cells were isolated by FACS, exposed to IL-3 for 3 days, and then studied in B lymphoid cultures. The IL-3R+ gate used for these studies included both IL-3Rhigh and IL-3Rdim cells. The CD34+CD38− IL-3R+ cells produced a significantly higher purity (p = 0.045) and absolute number (p = 0.016) of CD19+ cells than did CD34+CD38+ IL-3R+ cells after stimulation in IL-3 (n = 2 experiments), demonstrating again that IL-3 acts on B lymphopoiesis at a primitive progenitor level (Fig. 6).
As IL-3Rα is expressed at lower levels in CD34+CD38− than in CD34+CD38+ cells, we next determined whether the effect of IL-3 varies on populations expressing high, dim, and undetectable levels of IL-3Rα. CD34+ cells were isolated according to IL-3Rα expression, irrespective of CD38 expression, and cultured for 3 days in the presence or absence of IL-3. IL-3 exposure significantly increased the production of B lymphoid cells from both CD34+ IL-3Rdim (p = 0.001, n = 5) and CD34+ IL-3Rnegative cells (p = 0.033, n = 3) (Fig. 7). There was no significant difference between the B lymphoid capacity of CD34+ IL-3Rdim and CD34+ IL-3Rnegative cells either in the presence or absence of IL-3. CD34+ IL-3Rhigh cells, however, had little capacity for B lymphoid production, and IL-3 did not enhance B cell production. In the presence of IL-3, CD34+ IL-3Rhigh cells produced significantly lower numbers of CD19+ cells than did either CD34+ IL-3Rdim (p < 0.0001, n = 6) or CD34+ IL-3Rnegative cells (p = 0.004, n = 3) (Fig. 7). Thus, B lymphoid cells are produced almost entirely from the most primitive CD34+ progenitors with low or undetectable expression of IL-3Rα by flow cytometry. Despite the low expression of IL-3R, IL-3 significantly increases B lymphoid production from primitive human progenitors.
IL-3 increases cloning efficiency of B lymphoid progenitors and B lymphomyeloid progenitors
To determine whether IL-3 acts by increasing the number of B lymphoid progeny from each CD34+CD38− cell or by increasing cloning efficiency of normally quiescent progenitors with B lymphoid potential, the clonal behavior of single CD34+CD38− cells was studied. Individual CD34+CD38− cord blood cells were plated in each well of 96-well plates (containing S17 stroma) and cultured with or without IL-3 during the first 3 days. IL-3 exposure for 3 days increased the subsequent cloning efficiency of CD34+CD38− cells grown on S17 stroma (Fig. 8). It should be noted that the S17 culture system is not optimal to measure production of differentiated myeloid cells from progenitors (as demonstrated, for example, by CD33 expression). Coculture on S17 stroma does, however, allow preservation of myeloid progenitors in a relatively nondifferentiated state. The presence of myeloid potential is revealed when S17-cultured cells are switched to myeloid conditions (e.g., methylcellulose medium with IL-3, IL-6, SF, and erythropoietin) (40). Thus, clones were analyzed by FACS (for CD19+ B cell production) and by secondary CFU plating (to prove myeloid potential). Exposure of single CD34+CD38− cells to IL-3 led to an increase in the frequency of total B lymphoid clones (i.e., CD19+ clones with or without myeloid potential) and of the subset of clones with both B lymphoid and myeloid cells (bipotent progenitors) (Fig. 8). IL-3, however, did not affect the proportion of CD19+ cells within the clones analyzed; IL-3-stimulated B cell clones contained 30.2 ± 4.1% (mean ± SEM) CD19+ cells, and clones arising without prior IL-3 stimulation contained 30.4 ± 4.1% CD19+ cells. These data support the contention that IL-3 either stimulates proliferation or improves survival of primitive progenitors, including pluripotent B lymphomyeloid progenitors. Furthermore, these single cell studies show that IL-3 acts directly on progenitors with B lymphoid and B lymphomyeloid potential rather than indirectly through accessory cells in the culture.
These studies show that short-term exposure to IL-3 causes a consistent and marked increase in B lymphoid cell production from primitive human hemopoietic progenitors. IL-3 responsiveness was maximal in CD34+CD38− cells, the most immature progenitors studied. IL-3 increased the cloning efficiency of both B lymphoid and B lymphomyeloid progenitors, but did not increase the proportion of CD19+ cells in each clone. Thus, IL-3 increased B lymphoid production by either inducing proliferation and/or promoting survival of primitive and normally quiescent progenitors with B lymphoid or lymphomyeloid potential, rather than forcing B lymphoid differentiation. IL-3 was also synergistic with FL and IL-7, growth factors known to act on stem cells and/or primitive B lymphoid progenitors (45, 46). The proliferative effect of IL-3 was limited to primitive progenitors. No proliferation or increased survival of CD34−CD19+ B cells was seen. IL-3 consistently increased myeloid cell production from CD34+CD38+ cells, but results on B lymphopoiesis from this heterogeneous and lineage committed population were inconsistent.
The results were unexpected in view of studies reporting the effects of IL-3 on murine lymphohemopoietic progenitors using in vitro clonal assays. Stimulation of murine stem cell populations (e.g., murine Linnegative Sca-1+ c-kit+ cells) in IL-3 for more than 6 days has been reported to completely prevent the subsequent production of B lymphoid cells (23, 24, 25, 47). Certain important differences in the nature of the in vitro assays used for the murine studies and our own should be noted. The murine studies used semisolid medium during both the primary culture (IL-3 stimulation) and the secondary culture (B cell assay). However, in vitro identification of human B lymphopoiesis requires contact of progenitors with an adherent stromal layer. In most cases, we used S17 stroma layers during IL-3 stimulation and always in subsequent B cell assay. Nevertheless, IL-3 stimulation in the absence of S17 stroma (on fibronectin or in suspension) also increased subsequent B cell production. Thus, the actions of human IL-3 on B lymphopoiesis were not mediated through S17 stroma.
IL-3 exposure during primary culture was slightly longer in the murine studies than in our own (7–10 days vs 3 days). However, in our own experience, after culturing CD34+CD38− cells on primary human stroma continuously in IL-3 for at least 1 mo, B cells can still be generated when cultures are switched to S17 stroma without IL-3 (unpublished data). Thus, long-term exposure to IL-3 does not prevent subsequent B lymphopoiesis. It is possible, however, that a combination of the above differences in experimental design could have produced the different results in the murine and human studies. For example, the combination of the longer period of IL-3 stimulation in the absence of stroma in the murine studies may have led to apoptosis of cycling progenitors with B lymphoid potential.
Other evidence supports our contention that IL-3 does not abrogate human B lymphopoiesis. A number of investigators have stimulated human CD34+ or CD34+CD38− cells in cytokine combinations, which include IL-3 for up to 8 days and nevertheless demonstrated robust human B lymphoid production in vivo after transplantation into immunodeficient mice (30, 31, 32, 33, 34, 48). Previous reports suggest that IL-3 may stimulate B lymphopoiesis from human pro-B cells and mature B cells (49, 50, 51, 52). In one report, brief initial exposure to IL-3 (in combination with other cytokines) appeared to increase the frequency of B lymphoid clones arising from single CD34+ lin− CD38− cells during subsequent culture on the murine stromal line AFT024 (53). In the same study, however, no increase in B lymphopoiesis was seen with IL-3 containing combinations using bulk cultures of CD34+ lin− Dr− cells, possibly because overgrowth of other cell lineages affected B cell growth (53). The effect of IL-3 used in isolation and the requirement of AFT024 stroma during IL-3 stimulation were not studied.
One explanation for the contrasting results is that intrinsic differences in regulation of B lymphopoiesis may exist between the species. The different requirements for IL-7 and for stromal contact in postnatal in vitro human and murine B lymphopoiesis support this possibility (54, 55). The structures of murine and human IL-3 and their receptors differ significantly. Only 29% amino acid homology exists between murine and human IL-3 (44, 56). IL-3R from both species are heterodimers consisting of α and β subunits (42, 57). The α subunit confers low affinity, IL-3-specific binding; the murine and human forms of IL-3Rα have the same low homology (30%), as do their ligands (58). The β subunit in both species confers high affinity binding and mediates signal transduction. In the human, only one β subunit exists (βc), which is shared with the receptors for GM-CSF and IL-5 (42, 57). The murine IL-3R has two β subunits, βc and βIL-3, which have significant sequence homology (91% at amino acid level). Either of the murine β subunits can combine with IL-3Rα to form high affinity receptors (58), and some functional redundancy exists between murine βc and murine βIL-3. For example, either murine β subunit can transduce the negative regulatory signals of IL-3; murine B lymphopoiesis is abrogated when uncommitted progenitors from knockout mice deficient in either βc or βIL-3 are exposed to IL-3 (25). The murine β subunits, however, are not completely interchangeable. For example, although murine βc is shared by the receptors for murine GM-CSF and IL-5, βIL-3 is not. Thus, differences both in ligand-receptor interactions and signal transduction pathways may result in critical differences in the responses of murine and human progenitors to IL-3.
A recent report by Brown et al. (59) argues against the hypothesis that IL-3 responsiveness is restricted to human B lymphopoiesis. In this study, IL-3 administration in vivo was found to partially restore both T and B lymphopoiesis in Jak3-deficient mice, suggesting that IL-3 also has a positive regulatory role in primitive murine lymphopoiesis.
IL-3Rα expression on murine and human progenitors has been previously reported (60, 61, 62). Of interest in the current studies was the unexpected finding that IL-3 stimulation of B cell production was most impressive in cells with low or undetectable cell surface IL-3Rα expression. PCR revealed that message for IL-3Rα was present even in cells with expression undetectable by flow cytometry. IL-3Rα is the only subunit of the receptor that binds to IL-3; it thus seems that a very low receptor number is sufficient for the proliferative or survival effects on primitive progenitors with B lymphoid potential. In contrast, high expression of IL-3Rα was largely restricted to a mature, committed myeloid progenitor population.
Murine studies have also shown that culture in IL-3 causes loss of long-term reconstituting cells (63, 64, 65). Many investigators are now using ex vivo expansion and gene transfer protocols that avoid IL-3 because of concern that terminal differentiation and/or impairment of engrafting ability of hemopoietic stem cells occur with IL-3-stimulated proliferation. The studies reported in this work were not designed to demonstrate whether IL-3 stimulates self-renewing vs nonrenewing cell divisions of long-term reconstituting cells. However, the results do show that loss of B lymphoid potential should not be of concern during short-term ex vivo manipulation of pluripotent human hemopoietic stem cells in IL-3.
We thank Dr. Kimberly Payne for careful review of the manuscript. Thanks also go to the Labor and Delivery staff of Kaiser Sunset Permanente Hospital for their generous assistance in collection of cord blood samples.
This work was supported by grants from National Institutes of Health, RO1 DK54567-01 and P50 HL54850, and by a Translational Research Grant from The Leukemia and Lymphoma Society. G.M.C. is a Scholar of The Leukemia and Lymphoma Society.
Abbreviations used in this paper: SF, steel factor; FL, Flt 3 ligand.