|Home | About | Journals | Submit | Contact Us | Français|
IL-7 is critical for B cell production in adult mice, however its role in human B lymphopoiesis is controversial. One challenge was the inability to differentiate human cord blood (CB) or adult bone marrow (BM) hematopoietic stem cells (HSCs) without murine stroma. Here, we examine the role of IL-7 in human B cell development using a novel, human-only model based on co-culturing human HSCs on primary human BM stroma. In this model, IL-7 increases human B cell production by >60-fold from both CB and adult BM HSCs. IL-7-induced increases are dose-dependent and specific to CD19+ cells. STAT5 phosphorylation and expression of the Ki-67 proliferation antigen, indicate that IL-7 acts directly on CD19+ cells to increase proliferation at the CD34+ and CD34− pro-B cell stages. Without IL-7, HSCS in CB, but not BM give rise to a small but consistent population of CD19LO B lineage cells that express EBF and PAX-5 and respond to subsequent IL-7 stimulation. Flt3 ligand, but not thymic stromal-derived lymhopoietin (TSLP), was required for the IL-7-independent production of human B lineage cells. As compared to CB, adult BM shows a reduction of in vitro generative capacity that is progressively more profound in developmentally sequential populations, resulting in a ~50-fold reduction in IL-7-dependent B lineage generative capacity. These data provide evidence that IL-7 is essential for human B cell production from adult BM and that IL-7-induced expansion of the pro-B compartment is increasingly critical for human B cell production during the progression of ontogeny.
Murine studies have provided evidence that B cell development is controlled by a network of transcription factors and cytokine signaling that coordinates stage-specific expression of B lineage genes (1, 2). In mice, signaling through two Cytokine receptors, Flt3 (FMS-like tyrsosine kinase receptor) and the IL-7R, are critical for B cell development. The combined loss of signaling through both of these receptors completely blocks fetal and adult B cell development (3, Vosshenrich, 2003 #856) (4). Flt3 signaling upregulates IL-7Rα expression (5). In adult mice, signaling through the IL-7Rα upregulates expression of Early B cell factor (EBF). EBF in turn regulates expression of a cascade of B cell specific genes (1, 2) required for the transition from the common lymphoid progenitor to pro-B cells (6-8). During fetal and neonatal life, IL-7-independent expression of EBF allows early B lineage differentiation, but not proliferation, unless IL-7 is present (9).
Human B cell development is thought to differ from that in mouse with respect to the requirement for IL-7. Early evidence for the importance of IL-7 in murine B cell production came from in vitro studies showing that B cell precursors derived from adult BM increase by approximately 50 fold with the addition of IL-7 (10). Further evidence came from reports of blocked B cell development in the initial in vivo studies that characterized mutant mice with defects in IL-7 (11), or in components of the IL-7 receptor (IL-7R) (12, 13). Parallel in vitro studies of human B cell development were performed either using fetal hematopoietic sources (14-20) or using hybrid co-cultures with human progenitors and murine stromal cell lines (21-23). In those experiments IL-7 either had no effect on B cell production or resulted in IL-7-induced increases that were 10-fold less than those reported in the mouse. In vivo human data from clinical reports indicated that patients having severe-combined-immunodeficiency due to defects in IL-7/IL-7R signaling lack T cells, but have peripheral blood B cells (24-26). These data have been interpreted to mean that IL-7 is essential for normal murine, but not human, B cell development (27-31).
More detailed studies in the mouse revealed a dramatic change in B cell precursor requirements during ontogeny. IL-7 knockout mice show a block in adult B lymphopoiesis, but not in the production of B cells during the fetal or neonatal period (32). More recently it has been shown that IL-7-independent expression of EBF in the mouse allows B lineage differentiation early in ontogeny resulting in a “developmental switch” in murine stem cells with IL-7-independent B cell production occurring during fetal development and through the first two weeks of neonatal life, but not in the adult (9). Thus, it is not surprising that in vitro IL-7 effects on fetally-derived human B cell precursors (14-16, 20, 33, 34) were minor when compared to those observed for adult B lymphopoiesis in the mouse (10). Similarly, clinical data from patients with IL-7/IL-7R signaling defects was obtained early in life and is consistent with observations that fetal/neonatal B cell production is intact in mice with IL-7 defects.
It is also possible that residual, low level IL-7R signaling plays a role in B cell production in patients with defects in the IL-7R or its signaling pathway. Phosphorylation of STAT5 is a signaling event downstream of IL-7R ligation (35). Studies that compare B cell development in mice that are STAT5 null with that in STAT deficient mice that express a truncated form of STAT5 reveal that even low levels of residual IL-7R signaling can maintain B cell production in adult mice (36, 37). Experiments to assess for IL-7 function in human patients with defects in components of the IL-7R signaling pathway are limited by the number, type, and quantity of specimens available for study. Thus, in vitro studies of human B cell development are essential for determining the role of IL-7 in human B cell development and whether humans exhibit the “developmental switch” that results in the loss of IL-7 independent B cell production observed in the mouse.
The association between ontogeny and IL-7 requirements has not been rigorously examined in humans. In vitro studies that address the role of IL-7 in human B cell production during the neonatal period and through adulthood have been performed using hybrid human-mouse co-culture models (21-23). The limitations of this model were highlighted by recent reports demonstrating that murine IL-7 can stimulate signaling through the human IL-7Rα (22, 23, 38). Thus, the production of murine IL-7 by stroma in hybrid human-mouse co-cultures has rendered it difficult to evaluate the precise contribution of human IL-7, and the magnitude of IL-7-dependent effects in human B cell production from HSCs in CB and BM.
Here we describe a novel, entirely human model of in vitro B cell development that is based on co-culturing human HSCs on primary human BM stroma. Using this human-only model we show that human IL-7 increases the production of human B cell precursors by more than 60 fold, both from HSCs in adult BM, and from the HSCs in CB that give rise to B cells during the neonatal period. IL-7 effects were mediated through the direct action of human IL-7 on human CD19+ pro-B cells to increase proliferation at both the CD34+ and CD34− stages of B cell development. Adult BM, as compared to CB, shows a reduction of in vitro generative capacity that is progressively more profound in developmentally sequential populations, resulting in a ~50-fold reduction in B lineage generative capacity. In the absence of IL-7, HSCS in CB, but not BM give rise to a small but consistent population of CD19LO B lineage cells that express EBF, PAX-5 and respond to IL-7 with increased cell size and upregulation of CD19. These data suggest that in humans the role of IL-7 in B cell production becomes increasingly critical as ontogeny progresses, and that it is an essential factor for B cell production in adults.
CB was collected according to guidelines approved by the Institutional Review Boards at Childrens Hospital Los Angeles and Loma Linda University. Adult human BM was purchased from AllCells, LLC, Berkeley, CA; the National Disease Research Interchange (NDRI), Philadelphia, PA; or Poietics Cell Systems of Lonza Walkersville, Inc., Gaithersburg, MD. The leukopheresis product from a patient with pre-B ALL (> 90% CD19+ cells) served as a source of primary human pre-B ALL cells. The use of all human tissues was reviewed and approved by the Institutional Review Boards at Childrens Hospital Los Angeles and/or Loma Linda University. Primary murine BM was obtained from normal adult control mice immediately after they were sacrificed as a part of other studies approved by the Institutional Animal Care and Use Committees at Childrens Hospital Los Angeles or Loma Linda University.
Lineage marker negative (Lin−) CD34+ cells were isolated by FACS sorting from CD34-enriched CB and BM mononuclear cells as previously described (39). Primary stromal cells from adult human or murine BM were obtained by plating whole BM in Dexter’s original medium as described previously (40). Non-adherent cells were removed on the second day and adherent cells were used through passage six.
The murine S17 (41) stromal cell line is a generous gift from Dr. Kenneth Dorshkind, University of California Los Angeles and the murine MS-5 (42) stromal cell line is a generous gift from Dr. Kazuhiro Mori.
Human CB and adult BM HSCs (typically ~10,000 and ~15,000 respectively) were seeded onto murine stromal cells pre-plated at 3,500 cells/well (hybrid cultures) or onto sublethally irradiated (2000 rads) primary human BM stromal cells plated at 7,000 cells/well (human-only cultures) in 96-well tissue culture plates in B cell medium: RPMI medium (Irvine Scientific, Santa Ana, CA) containing penicillin/streptomycin, L-glutamine, 50 μM 2-mercaptoethanol, and 5% FBS (Omega Scientific, Inc. Tarzana, CA) lot tested for ability to support B cell production. Sublethal irradiation assured that culture progeny did not arise from residual human HSCs that might be associated with the BM-derived stroma. Control cultures were supplemented with IL-3 at 1 ng/ml (first week only) and Flt-3 ligand at 1 ng/ml or 50 ng/ml (R&D Systems, Minneapolis, MN) two Cytokines that have previously been shown to support the production of human B lineage cells in vitro (19, 43-45). Additional supplementation with recombinant human cytokines and/or antibodies is described in fig. legends or as follows: IL-7 at 5 ng/ml, stem cell factor at 25 ng/ml (R&D Systems) neutralizing anti-human IL-7 antibody, or control antibody at 10 ng/ml (BD Pharmingen, San Diego, CA), neutralizing anti-mouse IL-7 antibody (Abcam, Inc. Cambridge, MA) or control antibody (Jackson ImmunoResearch, West Grove, PA), neutralizing anti-human TSLP antibody (R&D) or control antibody (Jackson ImmunoResearch) at 1 ug/ml, neutralizing anti-human Flt-3 ligand or control antibody (R&D) at .6 μg/ml. The ND50 for rabbit anti-mouse IL-7 neutralizing antibody was 0.012 to 0.02 μg/ml of antibody for mouse IL-7 at 600 pg/ml. The ND50 for rat anti-human IL-7 neutralizing antibody was .0001 to .001 μg/ml of antibody for human IL-7 at 167 pg/ml. The ND50 for sheep anti-human TSLP was .05 to .25 μg/ml for .5 ng/ml human TSLP. The ND50 for mouse anti-human Flt-3 ligand was .02 to .06 μg/ml for 5 ng/ml Flt-3 ligand.
For cultures described in Fig. 4A each well contained 3 antibodies: either the neutralizing antibody to Flt-3 ligand, IL-7, and/or TSLP, or the corresponding control antibody at an identical concentration as neutralizing antibody. Wells to which exogenous cytokine were added received the appropriated control antibody for that cytokine. This assured that antibody isotypes and concentrations were the same in all wells. To simplify, only neutralizing antibodies (and not control antibodies) are indicated in the legend to Fig. 4A.
CD19+ progeny from human-only cultures were enriched for apoptosis assays using the Human B cell enrichment kit (StemCell Technologies, Vancouver, BC, Canada) and for assays of STAT5 phosphorylation using the Miltenyi Biotech CD19 microbead kit (Miltenyi Biotec, Auburn, CA).
For flow cytometry cells were surface stained as described previously, (46) and resuspended in PBS for FACS sorting using the FACSVantage or FACSAria cell sorter (Becton Dickinson Immunocytometry Systems, [BDIS] San Jose, CA) or in PBS or 1% freshly prepared formaldehyde for analysis using the FACSCalibur flow cytometer. The following antibodies were used: kappa FITC, lambda FITC, IgM FITC, IgM PE-Cy5 (Cychrome), CD19 APC, CD34 PerCP (BD Pharmingen), and IL-7Rα PE (Beckman-Coulter Immunotech).
To assess and cytoplasmic mu (cμ) and Ki-67 expression, cells were first stained for surface markers then stained with anti-IgM antibodies to detect intracellular cμ with anti-Ki-67, or with an isotype-matched control antibody (BD Pharmingen) using the Fix and Perm Cell Permeabilization Kit (Caltag Laboratories, Burlingame, CA) as per manufacturer’s instructions.
For phospho STAT5 staining, cells were washed, fixed by incubating for 10 min with 1.5% paraformaldehyde at room temperature, pelleted by centrifugation and then paraformaldehyde was decanted off. To permeabilize, cells were vortexed, resuspended in 100% methanol at 4°C, and incubated for 10 min at 4° C or at -20° C overnight. Cells were pelleted by centrifugation, washed in PBS with 10% FBS and .01% sodium azide, then incubated with anti-phospho STAT5 PE or mouse IgG1 PE isotype control antibody (BD Pharmingen) for 30 min at room temperature Cells were then washed and resuspended in staining buffer for flow cytometry.
To assess apoptosis, harvested cells were incubated with CD19 APC and the apoptosis indicator Annexin V FITC (BD Pharmingen) and prepared for flow cytometry according to manufacturers instructions.
Fluorescence intensity on all flow cytometry dot plots and histograms is shown on a log scale. Forward scatter (FSC) and percent maximum on histograms is shown in a linear scale.
Primary murine and human BM stromal cells (non-irradiated, 3.5 × 105 cells or sublethally irradiated [2000 rads], 7.0 × 105 cells) and murine MS5 and S17 stromal cells (non-irradiated, 3.5 × 105 cells) were plated in 200 μl of B cell medium (described above) without cytokines in individual wells of a 96-well plate. At weekly intervals 100 μl of culture medium was harvested, filtered with a 0.2 μm filter, and 100 μl of new media added to each well. Harvested supernatants were frozen, thawed, and IL-7 levels were detected using the Mouse IL-7 or Human IL-7 ELISA kit (R&D Systems).
cDNA was prepared as described previously. Reverse-transcription (RT)-PCR to detect expression of IL-7 and β2 microglobulin was performed as described previously (39). Intron-spanning PCR primers for mouse and human IL-7 and B2 microglobulin were as follows: human β2 microglobulin, 5’–CTC GCG CTA CTC TCT CTT TC–3’ and 5’ –CAT GTC TCG ATC CCA CTT AAC–3’ (330 base product); human IL-7, 5’–CTC CCC TGA TCC TTG TTC TG–3’ and 5’–TCA TTA TTC AGG CAA TTG CTA CC–3’ (151 base product); mouse β2 microglobulin, 5’–TGC TAT CCA GAA AAC CCC TC–3’ and 5’–GGC ATG CTT AAC TCT GCA GG –3’ (259 base product); and mouse IL-7, 5’ –TGG AAT TCC TCC ACT GAT CC–3’ and 5’ –TGG TTC ATT ATT CGG GCA AT–3’ (160 base product): (Invitrogen Life Technologies, Paisley, PA). PCR conditions for human IL-7 and B2 were as follows: 10 min 94° C (1 cycle); 94° C for 1 min, 50° C for 45 s, 72° C for 1 min and 15 s (5 cycles); 94° C for 1 min, 55° C for 30 s, 72° C for 1 min and 15 s (29 cycles); 72° C for 5 min (1 cycle). PCR conditions for mouse IL-7 and B2 were as follows: 15 min 95° C (1 cycle); 94° C for 30 s, 55° C for 30 s, 72° C for 30 s (35 cycles); 72° C for 10 min (1 cycle).
cDNA was amplified for detection of CD19, EBF, Pax-5, and GAPDH (housekeeping gene) using the LightCycler 1.0 system real-time thermal cycler and the LightCycler FastStart DNA MasterPLUS SYBR Green kit (Roche Applied Science, Indianapolis, IN). The small numbers of B lineage cells produced in the absence of IL-7 activity precluded quantitative or semi-quantitative assessment of mRNA levels. The assessment of samples as positive for expression was based on 1) observing reproducible CT values in 2 replicates of the target gene where the GAPDH housekeeping gene showing a CT value of <30 cycles and 2) melting curve analysis showing superimposable melting curves across samples (including pre-B ALL cells as a positive control) and replicates for a given target gene in the absence of a superimposable product in negative control samples. Gene specific primer pairs were designed to span introns using Roche’s Assay Design Center at www.roche-applied-science.com. Primers were as follows: GAPDH left 5’- GAG TCC ACT GGC GTC TTC AC; GAPDH right 5’- GTT CAC ACC CAT GAC GAA CA; CD19 left 5’- TCC ACC TGG AGA TCA CTG CT; CD19 right 5’- GAC CTT CCA GCC ACC AGT C; Pax5 left 5’- GTC CCA GCT TCC AGT CAC AG; Pax5 right 5’- GTG CTC ACC GAG GAC ACC; EBF left 5’-AGC TGC CAA CTC CCC CTA T; EBF right Lower 5’ GGG AGG CTT GTG GAG GAG. Cycling parameters for all products: initial denaturation of 15 min at 95 degrees followed by 50 cycles of 10 s at 95°C (denaturing); 5 s at 60 °C (annealing); 15 s at 72 °C (elongation).
In flow cytometry analysis, CD19+ B lineage cells and CD34+ HSC and progenitor populations were identified based on light scatter within a lymphocyte/progenitor light scatter gate and by expression of indicated surface markers. CD19−CD34− non-B lineage cells were identified based on light scatter within a living cell light scatter gate. Gates for determining percentages of cells positive for surface or intracellular markers were set on similarly gated CD19+ B lineage and CD19− cells in samples co-stained for CD19 and isotype-matched control antibodies.
Absolute numbers of culture progeny were determined by multiplying the fraction of cells within a particular subset (gated based on surface immunophenotype: CD19−, CD19+, CD19+CD34+IL-7Rα+, etc.) as a percentage of the cells within living cell light scatter (assessed by flow cytometry) by the total count of living cells generated in the culture (determined by hemocytometer count of cells negative for trypan blue staining or by flow cytometry using Caltag Counting Beads, [Caltag Laboratories]). Relative numbers of cells were determined by dividing the absolute number of cells of a particular type generated in each experimental condition by the absolute numbers of the same type of cells generated under control conditions (Fig. 1, ,2)2) or under control conditions plus anti-human IL-7 neutralizing antibodies (Fig. 3, ,8)8) in each independent experiment.
Each independent co-culture experiment represents an independent FACS sort of Lin−CD34+ BM or CB cells. In most cases adult BM in independent experiments is from different donors and in all cases CB in each independent experiment is pooled from multiple different donors. The hematopoietic cells for all conditions in each independent experiment are aliquoted from a pool of cells FACS sorted at the same time from the same donor(s).
Flow cytometry analysis was performed using FlowJo software (Tree Star, Inc. Ashland, OR). Statistical analysis was performed using InStat3 software (GraphPad Software, Inc., San Diego, CA).
We developed a novel, human-only in vitro model of B cell development based on co-culturing nonfetal human hematopoietic progenitors with primary human stroma from adult BM (Fig. 1A). Initial experiments used CB as a source of HSCs. While CB contains fetally-generated B cells, the stem cells present in CB are of interest because they give rise to B cells in the neonatal period and are increasingly used clinically as a source of HSCs for transplant.
B cell production in human-only cultures was compared to that in an established hybrid co-culture model (45, 47-49) consisting of human HSCs and murine S17 stroma. (Fig. 1A). The emergence of human B lineage cells in human-only cultures paralleled that observed in hybrid cultures with murine S17 stroma. CD19+ B lineage cells were undetectable at one week and rare at two weeks (data not shown). At three weeks, CD19+ B lineage cells were detectable in both human-only and hybrid S17 cultures under control conditions. However the frequency of B lineage cells was low in human-only cultures (Fig. 1B, middle panel). With the addition of human IL-7, human-only cultures produced a substantial population of CD19+ cells (Fig. 1C, middle panel). At 3 weeks, the CD19+ B lineage cells present in human-only cultures, like those in hybrid cultures, were B cell precursors that did not express the cytoplasmic mu (cμ) Ig heavy chain (Fig. 1B, C, middle and left panels) or surface immunoglobulin (Ig) as evidenced by the absence of kappa and lambda Ig light chains (data not shown). At 5 weeks, surface IgM+ cells were observed in both human-only (Fig. 1B, C, right panels) and hybrid co-cultures (data not shown), and while the production of cμ+ and IgM+ B cells in human-only cultures was inefficient (~11 to 16% of CD19+ cells), it was consistently higher than that observed in hybrid co-cultures (~3% of CD19+ cells). These data provide evidence that co-cultures with primary human BM stromal cells support the production of human B cells and establish human-only cultures as an in vitro model for studying early stages of human B lymphopoiesis.
Using human-only and hybrid S17 cultures we characterized the effect of IL-7 on the production of human B cell precursors from CB HSCs. The addition of IL-7 to human-only, but not hybrid S17 cultures, significantly increased the production of CD19+ B lineage cells as compared to cultures without IL-7 (Fig. 1B, C, D). Stem cell factor (SCF) had no significant effect on human B cell production, either alone or in combination with IL-7 in human-only or hybrid S17 cultures (Fig. 1D). Thus, human-only cultures revealed specific, IL-7 effects on human B cell production.
A titration of IL-7 into human-only cultures showed that IL-7-induced increases in the production of B lineage cells were dose-dependent and specific to CD19+ cells, achieving maximal effects (43-fold) at 10 ng/ml of IL-7 (Fig. 1E). In contrast, IL-7-induced increases observed in hybrid S17 cultures at 3 weeks never averaged more than 3-fold at any concentration of IL-7 (Fig. 1E). These data provide evidence that IL-7 induces a lineage-specific increase in the in vitro production of human B cell precursors that is not detectable in hybrid cultures with the murine S17 stromal cell line.
To determine if the inability to support IL-7-induced increases in human B cell production is specific to murine S17 stroma, we examined IL-7 effects in hybrid cultures with the MS-5 murine stromal cell line, another established model (42, 50, 51) for in vitro human B lymphopoiesis. As with S17 stroma, the addition of IL-7 did not significantly increase the production of human B cell precursors in hybrid MS-5 cultures (data not shown). However, substantial numbers of human B lineage cells were produced in the absence of exogenous IL-7 in hybrid cultures with murine S17 and MS-5 stroma (Fig. 1B and data not shown) suggesting that another factor replaces human IL-7 in hybrid human-mouse co-cultures.
While early studies suggested that mouse IL-7 does not act on human cells (52, 53), more recent reports indicate that this is not the case (22, 38). A titration of anti-mouse IL-7 neutralizing antibodies, but not control antibodies, into hybrid cultures with murine stroma gave a dose-dependent reduction in human B cell production. At the highest concentrations, anti-mouse IL-7 neutralizing antibodies decreased B cell production from CB HSCs in hybrid S17 and MS-5 cultures by 10 and 17 fold, respectively, as compared to co-cultures with no neutralizing antibody (Fig. 2A). The production of non-B lineage (CD19−) cells was not affected by anti-mouse IL-7 neutralizing antibodies (Fig. 2B). These data provide evidence that endogenously produced murine IL-7 is critical for the production of human B cell precursors in hybrid co-cultures with murine S17 and MS-5 stroma.
To determine if stromal cells can serve as a source of IL-7 for in vivo human B cell development we examined the ability of primary human BM stroma to produce IL-7. IL-7 mRNA was detected in primary adult human BM stroma as well as primary murine BM stroma and murine stromal cell lines (Fig. 2C). Next we compared the levels of IL-7 protein produced by murine and human stroma grown as in co-cultures. Weekly assessment of IL-7 by ELISA showed that primary adult human BM stroma produce low levels of IL-7 in vitro (ranging from .15 to 1.02 pg/ml during a five week culture period) and that IL-7 production is not affected by sub-lethal irradiation prior to plating for co-culture. In contrast, the murine S17 and MS-5 stromal cell lines produced levels of IL-7 that were at least 40 fold higher than that of primary human BM stroma (Fig. 2D). However, in vitro IL-7 production by primary adult murine BM stroma, either with or without sub-lethal irradiation, was below the threshold of detection by the ELISA for mouse IL-7. Thus, the high level of in vitro IL-7 production by murine S17 and MS-5 murine stromal cell lines does not reflect that of primary murine BM stroma that support in vivo IL-7-dependent B cell development in the mouse.
While the production of IL-7 from primary adult human BM stroma was detectable with the much more sensitive ELISA assay for human IL-7, those levels would have been below the sensitivity of the mouse IL-7 ELISA assay. These data provide evidence that primary human adult BM stroma produce IL-7 and that the level of in vitro IL-7 production by these cells is consistent with that observed for mouse, an animal model where the physiological relevance of IL-7 in B cell development has been demonstrated in vivo (11, 32)
To determine the extent to which B cell production is dependent on IL-7 and how this might change during ontogeny, we compared the production of B cell precursors from CB and adult BM HSCs in human-only cultures supplemented with anti-human IL-7 neutralizing antibodies or with IL-7 (Fig. 3A). A small population of CD19+ B lineage cells were generated from CB HSCs in cultures where IL-7 activity was neutralized. In contrast, B lineage cells derived from adult human BM were almost totally absent from cultures with anti-IL-7 antibodies, (Fig. 3A). These data suggest that HSCs in CB but not adult BM have the potential for IL-7-independent human B lymphopoiesis.
A comparison of human-only cultures initiated with CB and adult BM HSCs showed that IL-7-induced increases in both cases were similar in magnitude and specific to CD19+ cells (Fig. 3B). However, the fraction of culture progeny comprised of B lineage cells was significantly less in cultures with HSCs from adult BM (Fig. 3A). Therefore, we compared the generative capacities of CB and BM HSCs with respect to B lymphoid (CD19+) and non-B lymphoid (CD19−) progeny and the effects of IL-7 on these capacities (Fig. 3C). For each Lin−CD34+ CB cell plated at culture initiation, an average of 5.7 CD19+ cells were generated at 3 weeks in human-only cultures with IL-7. Surprisingly, the average B lymphoid generative capacity of adult BM HSCs in identical conditions was only .11 CD19+ cells/HSC plated. The capacities of CB and adult BM HSCs to generate CD19− progeny were not significantly different and were unaffected by IL-7. Thus, the in vitro B lymphoid generative capacity of adult BM HSCs is more than 50 fold less than that of CB HSCs and this decreased generative capacity is specific to CD19+ cells. These data suggest that IL-7 is more critical in maintaining human B cell production during adult life than during the neonatal period and that this may be due in part to a selective decrease in the B lymphoid generative capacity of HSCs in adult BM as compared to CB.
Studies of mouse B cell development suggest that TSLP, a cytokine that signals through the IL-7Rα and/or Flt3 ligand support the production of fetal/neonatal B cells in the absence of IL-7 (4, 54). Therefore we examined the extent to which the IL-7-independent production of B lineage cells from CB HSCs in our human-only cultures (Fig. 3A left panel) is dependent on Flt3 ligand and TSLP. The production of CD19+ cells from CB HSCs in human-only cultures supplemented with Flt-ligand and anti-IL-7 was compared to that observed with neutralizing anti-Flt-ligand, neutralizing anti-TSLP, or both. As shown in Fig. 4A, the addition of anti-TSLP antibodies had no effect on IL-7 independent production of B lineage cells, while CD19+ B lineage cells were essentially absent in cultures that contained anti-Flt3 ligand. These data show that Flt3 ligand, but not TSLP, is required for the IL-7-independent production of B lineage cells from CB HSCs that we observed in human-only cultures.
IL-7 has been reported to increase the expression of CD19 on fetally-derived human B cell precursors (55) and in the presence of IL-7, IL-7Rα+ B cell precursors are larger in size than B cell precursors that lack the IL-7Rα (22). Consistent with this report, flow cytometry analysis showed that the B lineage cells generated from CB and BM HSCs in human-only cultures expressed high levels of CD19 in the presence of IL-7, while cultures supplemented with anti-IL-7 neutralizing antibodies expressed much lower levels of CD19 (Fig 3A). Human-only cultures supplemented with IL-7 also showed higher frequencies of B cells that were larger (as indicated by increased forward scatter (FSC) in flow cytometry analysis) as compared to those produced in cultures with anti-IL-7 neutralizing anti-bodies (see below).
To determine if the B lineage cells generated in the absence of IL-7 were capable of responding to IL-7 by upregulating CD19 and exhibiting increased cells size characteristic of cycling cells, we switched the culture progeny produced from CB HSCs in human-only cultures supplemented with anti-IL-7 neutralizing antibodies to human stromal cell cultures with IL-7 (anti-IL-7 to IL-7 switch cultures). After 3 days, anti-IL-7 to IL-7 switch cultures were compared to cultures that had been switched to human stroma and anti-IL-7 (anti-IL-7 to anti-IL-7 switch cultures) or to human-only cultures that had been both initiated with and switched to IL-7 supplementation (IL-7 to IL-7 switch cultures). The percentage of cells expressing high levels of CD19 (CD19HI) was increased in anti-IL-7 to IL-7 switch cultures (Fig. 4B middle panel), as compared to anti-IL-7 to anti-IL-7 switch cultures (Fig 4B top panel), however, the percentages were not as high as those observed among B lineage cells that were produced under continuous exposure to IL-7 (Fig 4B, bottom panel, and Fig 4D). Cultures with B cell precursors that were never exposed to IL-7 (anti-IL-7 to anti-IL-7 switch cultures) showed a low frequency of cells with the high levels of FSC characteristic of cycling cells. The switch from anti-IL-7 to IL-7 resulted in a significant increase in the frequency of cells that were FSCHI and an FSC histogram profile that was similar to that seen for IL-7 to IL-7 switch cultures.
IL-7R-signaling, partially via upregulation of EBF, has been reported to be a regulator of the transcription factor network that controls B cell development in the mouse (2). Expression of EBF in the absence of IL-7 in fetal and neonatal mice allows IL-7-indepdent B cell differentiation at early points in ontogeny (9). Therefore we were interested in examining expression of EBF and its downstream target PAX-5 (56) in the early B lineage cells generated in our human-only cultures. CD19+ cells generated in switch cultures were FACS-sorted for RT-PCR. Real time RT-PCR showed that B lineage cells generated in all three switch culture conditions expressed mRNA for CD19, PAX-5, and EBF, (Supplemental Data) however the small number of cells generated in cultures initiated with anti-IL-7 precluded a quantitative comparison of mRNA levels. Taken together these data provide evidence that the CD19LO Cells generated from CB HSCs in the absence of IL-7 are B lineage cells and have the capacity to respond to subsequent IL-7 stimulation.
To identify the cellular targets of IL-7 we used flow cytometry to examine IL-7Rα expression among cells present during in vitro B lymphopoiesis (Fig. 5A). IL-7Rα was not detectable on primary human BM stromal cells used in co-cultures (Fig. 5B). Approximately 35% of CD19+ B cell precursors produced in co-culture expressed the IL-7Rα, regardless of whether they arose from HSCs in CB or adult BM or whether IL-7 was present (Fig. 5C and Fig. 5D right panel). We gated CD19− cells into a CD34− subset that included more mature non B-lineage cells and into a CD34+ subset that included HSCs and progenitors present in the 3-week human-only cultures. Of the CD19−CD34− non B lineage cells generated from CB and BM HSCs in human-only cultures, 3-12 % expressed the IL-7Rα (Fig. 5C left panel) while 4-14% of the of the CD19−CD34+ HSC and progenitor subset was IL-7Rα+ (Fig. 5C middle panel). The percentage of cells in each subset that expressed IL-7Rα was not significantly different in cultures initiated with CB and BM HSCs, and the presence or absence of IL-7 activity had no significant impact on the percentage of cells expressing the IL-7Rα in any of the subsets (Fig 5D). These data show that CD19+ B cell precursors represent the major IL-7Rα+ population in human-only cultures and are therefore a potential IL-7 target during in vitro B cell development.
To determine if IL-7 acts directly on B cell precursors we used flow cytometry to assess phosphorylation of STAT5, an event downstream of IL-7R signaling (35). CD19+ B cell precursors were isolated from human-only cultures and assessed by flow cytometry for purity (based on CD19 expression) and for IL-7Rα expression (Fig. 5E, left panel) and then for STAT5 phosphorylation following incubation with or without IL-7 (Fig. 5E, right panel). Phosphorylated STAT5 was detected in B cell precursors incubated with IL-7, but not in B cell precursors cultured without IL-7 (Fig. 5E right panel). The percentage of B cell precursors positive for phosphorylated STAT5 following incubation with IL-7 was virtually identical to the percentage of cells expressing IL-7Rα (Fig. 5E compare left and right panels). These data provide evidence that B cell precursors are directly targeted by IL-7 in human-only cultures.
To determine if increased cell survival is involved in the IL-7-induced increases observed in our co-cultures, we examined the effect of IL-7 on apoptosis among culture-generated human B cell precursors. Human-only cultures initiated with CB HSCs were harvested at 3 weeks and CD19+ cells (either enriched or unseparated from CD19− cells) were placed in co-culture with fresh human stroma and supplemented with anti-IL-7 neutralizing antibodies, control antibodies, or IL-7 Cells were harvested at daily time points and flow cytometric analysis of staining with the Annexin V apoptosis indicator was used to determine the percentages of CD19+ B cell precursors undergoing apoptosis. IL-7 had no effect on the percentages of CD19+ that were Annexin V+, whether they were in co-culture alone or together with CD19− cells (data not shown). Similarly IL-7 had no effect on apoptosis among in vivo-generated B cell precursors (CD19+K/L−) that were FACS sorted from adult BM and assessed in identical assays (data not shown). These results suggest that IL-7 does not mediate protection from apoptosis and that increased survival of CD19+ B cell precursors is unlikely to be a major mechanism by which IL-7 increases human B cell production.
To determine if increased proliferation is a mechanism for IL-7-induced increases in human B cell production, we used flow cytometry to examine expression of the Ki-67 proliferation antigen in cells harvested from human-only cultures supplemented with IL-7 or with anti-IL-7 neutralizing antibodies (Fig. 6A). IL-7 did not change the percentage of CD19− cells that were Ki-67+ in cultures of CB or BM HSCs (Fig. 6A, B). In contrast, in CB cultures the percentage of CD19+ cells expressing the Ki-67 was significantly higher with IL-7 supplementation (Fig. 6B). Cultures of BM HSCs generated too few B cell precursors to be assessed for Ki-67 when anti-IL-7 antibodies were added. However, the percentage of B cell precursors expressing Ki-67 was the same in CB and BM cultures with IL-7 (Fig. 6 A, B).
To verify that the increases in proliferation observed in cultures supplemented with IL-7 were due to the direct action of IL-7 on B cell precursors, we compared Ki-67 expression (Fig. 6C) and size (indicated by FSC, Fig. 6D) in IL-7Rα+ and IL-7Rα− B cell precursors generated in cultures with IL-7. For both CB and BM cultures, average cell size and the percentages of cells expressing Ki-67 was significantly higher for IL-7Rα+ than for L-7Rα− B cell precursors. These data provide further evidence that IL-7 increases the production of human B lineage cells through increased proliferation due to direct targeting of B cell precursors by IL-7.
To determine whether the IL-7 targets that we identified in our human-only cultures are reflective of human B cell development in vivo, we examined the expression of IL-7Rα in early B lineage cells that were generated in vivo in adult human bone marrow. The average frequency of IL-7Rα+ cells among in vivo-generated B cell precursors (CD19+K/L−) in adult human BM (Fig. 7A, Table 1) was similar (32%) to that observed for B cell precursors generated in vitro (35%) from HSCs in CB and adult BM in human-only cultures (Fig. 5C, D). The percentages of in vivo-generated CD19+K/L− B cell precursors that we assessed in human subjects (Table 1) exhibited a broad range in the percentage of cells that were IL-7Rα+ (17%-56%) and showed no obvious correlation with age, gender, or race, suggesting that there is substantial natural variation among adult individuals. Thus, the IL-7Rα+ population of B cell precursors observed in human-only cultures is reflective of in vivo B cell development in adult human BM and suggests a physiological role for IL-7 in human B cell development.
A substantial portion of the B cell precursors generated in human-only cultures initiated with CB and BM HSCs were IL-7Rα−. This was also the case for CD19+ K/L−cells in adult BM. To determine if the expression of IL-7Rα correlated with the expression of developmental markers during in vivo B cell differentiation we examined IL-7Rα expression among developmental subsets of CD19+ K/L− B cell precursors identified based on expression of CD34+ or cμ (Fig 7B). The percentages of IL-7Rα+ cells were similar for all developmental subsets (cμ+, cμ−, CD34−, and CD34+) of CD19+K/L− cells. Thus, the expression of IL-7Rα among B cell precursors produced in vivo does not appear to correlate with particular stages of development that can be identified based on expression of CD34 or with successful rearrangement of the Ig heavy chain as indicated by the expression of cμ. (Fig 7B).
In the mouse, IL-7R signaling has been associated with distinct differentiation and proliferation signals. Thus, while the expression of IL-7Rα may not correlate with markers of development in CD19+ B cell precursors, it is possible that IL-7 may directly target distinct developmental populations of IL-7Rα+ B lineage precursors for proliferation or for differentiation. To identify the developmental targets of IL-7-induced proliferation we examined IL-7-induced increases and expression of the Ki-67 proliferation antigen in four subsets that include developmentally sequential populations of B cell precursors. FACS analysis of the cells present at 3 weeks in human-only cultures was used to identify four populations (Fig 8A): 1) The most primitive IL-7R−HSCs and progenitors (HSC) through which IL-7 effects might be indirectly mediated, 2) IL-7Rα+ progenitors (IL-7R+ P) that would include the earliest CD19− lymphoid-committed progenitors that express the IL-7Rα (57) (58) ; 3) CD34+ pro-B cells (CD34+ B) and 4) CD34− pro-B cells (CD34− B).
Human-only cultures initiated with Lin−CD34+ cells from CB and BM and supplemented with IL-7 showed significantly increased numbers of cells in the CD34+ B and CD34− B subsets at three weeks when compared to cultures with anti-IL-7. The relative numbers of CD34+ B cells increased by only 5 to 14-fold (CB and BM cultures respectively) when IL-7 was present while CD34− B cells went up by 194 to 90-fold (CB and BM cultures respectively). In contrast, the relative numbers of HSCs, and IL-7R+ P were not significantly increased in cultures with IL-7 (Fig. 8B). Consistent with the IL-7 effects that we observed on cell numbers, CB cultures supplemented with IL-7 showed significantly higher percentages of Ki-67+ cells among the CD34+ B and CD34− B, but not among the HSC or IL-7R+ P subsets (Fig 8C). The percentage of Ki-67+ B cells among CD34− B cells in BM cultures was similar to that observed in CB cultures, however there were too few CD19+ cells to assess in Ki-67 expression in any BM cultures with anti-IL-7 or in the CD34+ B cells in BM cultures where IL-7 was present (Fig 8C right panel). Taken together these data provide evidence that IL-7 targets both CD34+ and CD34− pro-B cells for increased proliferation. In contrast, in our culture model we were unable to detect any significant changes in proliferation due to IL-7 effects among developmental subsets that would include CD19− B cell precursors.
In Fig. 3 we showed that IL-7 is particularly critical for the production of B lineage cells from adult BM because the capacity of adult BM to generate B lineage cells is greatly reduced as compared to CB. To get a better understanding of the dynamics and potential mechanisms that might be responsible for the changes in B cell development that occur during ontogeny, we compared the ability of CB and adult BM to generate the developmental subsets observed in vitro in human-only cultures. In cultures with adult BM the HSC generative capacity was about half, and the IL-7R+ P generative capacity about 10 fold less than that observed for CB cultures (Fig 8D, Note the 10X scale reduction in the right panel). In cultures with IL-7, the adult BM generative capacities for cells in the CD34+ B and CD34− B subsets were ~40-fold and ~50-fold less, respectively than those observed for CB. While BM cultures generated detectable CD19−subsets, and the capacity to generate these cells was unaffected by IL-7, B lineage precursors that reached the CD19+ stage of development were essentially absent in cultures with anti-IL-7. These data suggest that there is a progressive loss of generative capacity during the course of B lineage differentiation in adult BM, as compared to CB, and that as a consequence of this reduced B lineage generative capacity, IL-7 is more critical for B cell development from HSCs in adult BM.
Defining the role of IL-7 in human B cell development has been difficult for several reasons. First, most early studies of IL-7 and human B cell development used fetal hematopoietic sources. This is particularly relevant as subsequent murine studies revealed that IL-7-dependent B lymphopoiesis is linked with ontogeny–only in the adult mouse does B cell development become dependent on IL-7. Thus, differences in fetal and adult B cell development have been interpreted as fundamental differences in mouse and human B lymphopoiesis. Second, studies of humans that are defective in components of the IL-7R signaling pathway are hampered by limitations in the number, type, and quantity of specimens available for assessment of IL-7 function. Third, studies of nonfetal B lymphopoiesis were hampered by the lack of an in vitro model of human B cell development that reconstitutes the human BM stroma microenvironment without the confounding factors introduced by murine stromal cell lines. Here we describe a novel human-only model of in vitro B cell development based on co-culturing HSCs on primary human BM stroma. We used this model to examine the role of IL-7 in human B lymphopoiesis from HSCs in adult BM and from the HSCs in CB that give rise to B cells during the neonatal period.
Using the human-only culture model we show that IL-7 increases the production of human B cell precursors from HSCs in CB and adult BM by more than 60-fold as compared to cultures with anti-IL-7 neutralizing antibodies (Fig. 3). IL-7-induced increases were dose-dependent (Fig. 1E), specific to CD19+ cells (Fig. 1E and and5),5), and mediated through increased proliferation (Fig. 6), but not increased cell survival (data not shown). Expression of the IL-7Rα and the induction of STAT5 phosphorylation following stimulation with IL-7 (Fig. 5) indicate that IL-7 acts directly on CD19+cμ−cells present in 3-week human-only cultures (Fig. 1) to expand both the CD34+ and CD34− pro-B cell compartments (Fig. 8).
The IL-7-induced increases in B cell precursors that we observed in human-only cultures were not detectable in hybrid cultures where human HSCs were co-cultured on murine S17 or MS-5 stromal cell lines (Fig. 1, ,2).2). The absence of IL-7 effects that we and others (21) (51) have observed in hybrid cultures is likely due to high IL-7 production by MS-5 and S17 stromal cell lines (>40-fold that of primary human or murine BM stroma [Fig. 2]) since human B cell production was reduced up to 17 fold (Fig. 2A) by high levels of anti-mouse IL-7 neutralizing antibodies. Thus, the human-only model revealed IL-7 effects on human B cell development that were not easily detectable in hybrid human-mouse co-cultures.
A critical role for IL-7 in adult B cell development has been demonstrated in vivo in the mouse model (11)-(12, 13). IL-7 expression by primary murine BM stroma (Fig. 2) and the level of IL-7 that elicits in vitro B cell precursor responses (1 ng/ml) in the mouse (59) are consistent with what we see during in vitro human B cell development (Fig. 1, ,2).2). The expression of IL-7Rα by B cell precursors (Fig. 7) and of IL-7 by stromal cells (Fig. 5) in primary human BM suggests that the IL-7 effects we observe in vitro are physiologically relevant for in vivo human B cell development. Indirect evidence for the in vivo role of IL-7 in human B cell development comes from patients with T cell lymphopenia in whom endogenous IL-7 is upregulated and expansion of the immature B cell pool correlates with increased serum levels of IL-7 (60), (61).
Data obtained using our human-only culture model provide the first evidence that human B cell production from HSCs in adult BM is dependent on IL-7. Robust IL-7-independent B cell production has been reported in human fetal BM (62). In the absence of IL-7, we observed that HSCS in CB, but not BM consistently give rise to a small population of CD19LO B lineage cells (Fig. 3) that express EBF and PAX-5 at three weeks and differentiate to generate very small numbers of IgM+ B cells at 5 weeks (Fig. 1). When switched to IL-7 at 3 weeks, CD19LO Cells generated in the absence of IL-7 respond by upregulating CD19 and exhibiting increased cell size that is characteristic of cycling cells (Fig. 4B-D). Studies of B cell development in the mouse model suggest that IL-7-independent expression of EBF allows early B lineage differentiation (but not expansion unless IL-7 is present) during the fetal and neonatal period. This results in what has been described as a “developmental switch” in stem cells with IL-7-independent B cell production occurring through the first two weeks of neonatal life, but not in the adult (9). Our data showing that human B cell development from HSCs in adult BM is IL-7-dependent, coupled with our characterization of B lineage cells generated from CB in the absence of IL-7, provide the first evidence that that such a “developmental switch” may also occur in human B lymphopoiesis.
We found that adult BM, as compared to CB, shows a reduction of in vitro generative capacity that is progressively more profound in developmentally sequential populations (Fig 3, ,8).8). The ability of adult BM to generate HSCs, IL-7R+ progenitors, and pro-B cells after 3 weeks in human-only cultures with IL-7 was reduced by half, 10-and 50-fold, respectively, of that observed for CB (Fig. 8D). Indeed the CD19+ generative capacity of BM HSCs in the presence of IL-7 was similar to that of CB HSCs in the absence of IL-7 activity (Fig. 3). The B lymphoid-specific decrease in generative capacity that we observed for human HSCs in adult BM, as compared to CB, is consistent with a selective loss of B lymphoid generative capacity observed in xenograft transplant models and in the clinical transplant setting (63, 64).
The decline in IL-7-dependent B lymphoid generative capacity that we have observed in adult versus fetal B cell development suggests, that a mechanism other than the “developmental switch” observed in the mouse model may contribute to the decline in IL-7-independent B cell production that occurs during the progression to adulthood. It is possible that IL-7-induced expansion of the pro-B compartment (absent in the IL-7-indepdent B cell production observed by Kikuchi et al.) is essential for B lymphopoiesis from adult HSCs because of their reduced generative capacity. Further studies will be needed to determine the extent to which the decline in B cell generative capacity contributes to the loss of IL-7-independent B lymphopoiesis in adult humans and mice and whether B cells arising from IL-7-dependent and IL-7-independent pathways in humans are functionally distinct.
In summary, data obtained using our human-only culture model provide the first evidence that B cell development from HSCs in adult BM is dependent on IL-7 and characterize the progressive decline in developmentally sequential populations of B cell precursors that correlates with B cell production that is the dependent on IL-7. These studies have implications for immune reconstitution following stem cell transplant with CB versus BM in the clinical setting.
We thank Labor and Delivery at Kaiser Permanente-Sunset, Los Angeles, CA for CB collection. We are grateful for FACS sorting services provided by the FACS core of the Division of Research Immunology and BM Transplantation, Childrens Hospital Los Angeles, and by Barbara N. Walter and the Genomics Core Flow Cytometry Facility of the University of California, Riverside Institute for Integrative Genome Biology.
1This work was supported by the National Institutes of Health grants: K22 CA 111392 (S.D.), 5K01 DK066163 (K.J.P), a Research Career Development Award from the Saban Research Institute at Childrens Hospital Los Angeles (K.J.P), and the Center for Health Disparities and Molecular Medicine and the Dept. of Pathology and Human Anatomy, Loma Linda University (K.J.P).