Reassessment of interactions between hematopoietic receptors using common beta-chain and interleukin-3–specific receptor beta-chain–null cells: no evidence of functional interactions with receptors for erythropoietin, granulocyte colony-stimulating factor, or stem cell factor

Interleukin (IL)-3 has numerous effects on hematopoietic cells including actions on precursors and mature cells.1-4 The receptor for IL-3 consists of a unique specific α-chain, IL-3Rα, which binds IL-3 with low affinity,5,6 and the common β-chain (β_c) which is also used by granulocyte macrophage–colony-stimulating factor (GM-CSF) and IL-5. Following the binding of IL-3 to IL-3Rα, β_c converts the interaction to one of high affinity.7 In mouse cells, but not human cells, an additional IL-3–specific β-chain (β_IL3) is used in preference to β_c for signaling by IL-3⁸ and, unlike β_c, uses low affinity to directly bind IL-3.

Mice lacking β_c (β_cnull mice), have an eosinopenia and, like mice deficient in GM-CSF, develop lung disease reminiscent of human pulmonary alveolar proteinosis.9-12 Cells from these mice lack high-affinity binding for GM-CSF and IL-5. Mice that lack β_IL-3(β_IL-3 null mice)8,10 show decreased biological responsiveness of cells to IL-3 (via intact β_csignaling).8 Ablation of β_IL-3 explained the conflicting results observed for hierarchical receptor interactions in mouse cells versus human cells. Although GM-CSF was not able to “transmodulate” IL-3 receptors and alter IL-3 binding in wild-type murine cells (because of the availability to IL-3 of β_IL-3), it was able to trans-down-modulate IL-3 binding in β_IL-3 null cells, as a result of competition between GM-CSF and IL-3 for binding to βc chains.8 

Accumulated evidence suggests a role for β_c in modulating signaling of other hematopoietic cytokines including G-CSF, erythropoietin (EPO), and stem cell factor (SCF). Both GM-CSF and IL-3 transmodulated binding of G-CSF and M-CSF in normal cells.13 IL-3 also showed this effect in both β_c null and in β_IL3null cells.8 However, transmodulation by GM-CSF required a functional β_c receptor. Based on these results we proposed that β_c and β_IL-3 interacted with the G-CSF receptor (G-CSFR) and M-CSFR and/or that the GM-CSF or IL-3 activation of cellular signaling pathways modified G-CSFR and M-CSFR, perhaps resulting in their internalization. However, this phenomenon remains unexplained.8 

Interaction between β_c and the EPO receptor (EPOR) has been demonstrated. EPO stimulated tyrosine phosphorylation of β_c in the UT-7 erythroleukemia cell line,14 although neither GM-CSF14 nor IL-315 stimulated tyrosine phosphorylation of EPOR. It was therefore suggested that EPO might activate the GM-CSF signaling pathway by phosphorylating β_c.14 Functional and physical interactions between β_c and EPOR were demonstrated using the murine IL-3–dependent cell line (Ba/F3), which expresses IL-3Rα, β_c, and β_IL-3. The Ba/F3 cells transfected with murine EPOR acquired responsiveness to EPO, and increased expression of murine β_c resulting in heightened responsiveness to EPO. Conversely, inhibition of murine β_c function in Ba/F3/EPOR cells inhibited both IL-3–dependent and EPO-dependent cell growth. Moreover, an EPO-independent physical interaction between β_c and EPOR was demonstrated by coimmunoprecipitation.16 We sought to address interactions involving the β_c/β_IL-3R system using cells from β_c/β_IL-3null mice.

The B_c and β_IL-3loci are closely linked on mouse chromosome 15.17 To generate mice with a mutation in both loci, the embryonic stem (ES) cell line 3.15, which contains a targeted mutation of one allele of theBIL3 locus, was electroporated with a linearized targeting construct for the B_c locus. This construct was as previously described,9 except that a cassette containing the hygromycin-resistance gene18 was inserted in exon 7, and a thymidine kinase cassette19 was ligated to the 5′-end of the construct. Selection and screening of hygromycin and FIAU-resistant ES cell clones were performed as previously described. To detect homologous recombinants, BamHI-digested DNA was hybridized with probe A (Figure1) and a 3′-probe. Correctly targeted clones were further analyzed by Southern blot analysis for the presence of the BIL3 mutation.8 We used 3 ES cell lines to create chimeric mice, which were screened for cosegregation of theBIL3 and B_c mutations. Control mice were C57BL/6 or 129/Sv, and the experiments were conducted on mice 6 to 14 weeks of age.

Bone marrow (BM) progenitor cells were assayed in clonal culture as previously described.9 Semisolid 1-mL agar cultures containing 5 × 10⁴ BM cells or 10⁵ spleen cells in 0.3% agar in Dulbecco's modified Eagle's medium (DMEM) with 20% newborn calf serum were plated in triplicate and stimulated by multiple combinations of purified recombinant growth factors. To determine cytokine responsiveness, BM cells fromβ_c/β_IL3 null mice and control mice were cultured in agar using serial dilutions of G-CSF (initial G-CSF concentration, 500 U/mL) or SCF (initial SCF concentration, 100 ng/mL) for 7 days of incubation at 37°C in a fully humidified atmosphere of 10% carbon dioxide (CO₂) in air. The colonies were enumerated using a dissection microscope. For colony-forming unit–E (CFU-E) assays, bone marrow cells were cultured using serial dilutions of EPO (initial EPO concentration, 4 U/mL) in methylcellulose cultures incubated for 2 days in 5% CO₂ in air. The colonies were enumerated using an inverted microscope. Cultures were scored by an investigator blinded to the genotype of the cells.

Baseline hematopoiesis in β_c/β_IL-3 null mice was no different than that seen in β_c null mice (C.L.S., L.R., and C. G. B., unpublished observations). This was in keeping with previous reports of mice lacking eitherβ_c,9,10β_IL-3,8,10 or the combination of β_c and β_IL-3.20 21Thus, in mouse cells and in spite of the presence of an additional β_c specific for IL-3, which might imply an important function, IL-3 did not have an essential role in steady-state hematopoiesis.

Hematopoietic progenitor cells fromβ_c/β_IL-3 null mice were examined. The lack of responsiveness to IL-3 and GM-CSF was confirmed. There was no proliferation when BM cells fromβ_c/β_IL3 null mice were stimulated by either cytokine. In contrast, colony formation in response to stimulation with other hematopoietic cytokines, including SCF and the combination of SCF, G-CSF, and IL-6, was normal. Analysis of erythroid colonies in methylcellulose cultures revealed that the number of erythroid progenitor cells (both BFU-E and CFU-E) was also normal (C.L.S., L.R., and C.G.B., unpublished observations).

We have previously demonstrated that IL-3 was able to transmodulate both G-CSFR and M-CSFR in the absence of eitherβ_c null or β_IL3 null cells.13 However, transmodulation of G-CSFR and M-CSFR by GM-CSF required the presence of β_c.8 These results predict that a biochemical analysis of transmodulation of G-CSFR and M-CSFR by IL-3 in the absence of both β chains would show a lack of transmodulation capacity by IL-3. Indeed, in keeping with the lack of high-affinity IL-3 binding that results from generating β_c/β_IL-3 null cells, we did not see transmodulation of G-CSF receptors by IL-3 on BM cells from β_c/β_IL-3 null mice, even at doses of 100-nmol/L IL-3. This was in contrast to the transmodulation of G-CSF receptors by IL-3 that was observed on normal BM cells in the same experiments (N.A.N., unpublished observations).

The corollary of the G-CSFR transmodulation results predicted that the response to G-CSF may, in part, be mediated by β_c or β_IL-3, although neither receptor was directly engaged by G-CSF. This issue was addressed usingβ_c/β_IL3 null cells. We would have predicted that the absence of B_crendered cells less sensitive to stimulation with G-CSF. However, as shown in Figure 2,β_c/β_IL3 null cells showed normal responsiveness to G-CSF. This indicated that biological responsiveness to G-CSF was not dependent on β_cor β_IL3, implying that the biochemical phenomenon of transmodulation was of no genuine biological significance.

Several different observations have suggested a role for β_c in signaling by EPO. In addition to the phosphorylation and physical interaction data described above, IL-3 and GM-CSF are known to cooperate with EPO in erythropoiesis in vitro,22,23 and common signal transduction pathways involving STAT5 are used by their receptors.24 In addition, β_c has been implicated in signaling by SCF via its receptor, c-kit, and SCF is able to induce serine/threonine phosphorylation of β_c.25 

To determine the physiological significance of these observations, responsiveness of β_c/β_IL-3 null BM cells to EPO and SCF was examined. However, there was no observed difference in responsiveness to EPO (Figure 2). Nor was there a difference in responsiveness to SCF for BM cells cultured from β_c/β_IL-3 null mice compared to wild-type mice (Figure 2).

We therefore conclude that previously documented biochemical phenomena, including receptor transmodulation and receptor transphosphorylation, are not physiologically relevant in the context of hematopoietic cell growth responses to individual cytokines. Moreover, even the demonstration of direct physical interaction involving the β_c/β_IL-3 receptor systems in cell lines did not extrapolate to an interaction of physiological significance in primary hematopoietic cells. This result is important for interpreting the significance of biochemical interactions between receptor molecules, particularly in studies in which cell lines are employed.

The authors thank Louise Barnett for her work in generating the ES cell line.