Interleukin-3 (IL-3) inhibits erythropoietin-induced differentiation in Ba/F3 cells via the IL-3 receptor alpha subunit.

Introduction of erythropoietin receptors (EpoRs) into the interleukin-3 (IL-3)-dependent murine hemopoietic cell line, Ba/F3, enables these cells to not only proliferate, after an initial lag in G1, but also to increase β-globin mRNA levels in response to erythropoietin (Epo). With IL-3 and Epo costimulation, IL-3-induced signaling appears to be dominant since no increase in β-globin mRNA occurs. Differentiation and proliferation signals may be uncoupled since EpoRs lacking all eight intracellular tyrosines were compromised in proliferative signaling but retained erythroid differentiation ability. Intriguingly, a chimeric receptor of the extracellular domain of the EpoR and the transmembrane and intracellular domains of IL-3RβIL-3 chain (EpoR/IL-3RβIL-3) was capable of Epo-induced proliferative and differentiating signaling, suggesting either the existence of a second EpoR subunit responsible for differentiation or that the α subunit of the IL-3 receptor (IL-3R) prevents it. Arguing against the former, a truncated EpoR lacking an intracellular domain was incapable of promoting proliferation or differentiation. An EpoR/IL-3Rα chimera, in contrast, was capable of transmitting a weak Epo-induced proliferative signal but failed to stimulate accumulation of β-globin mRNA. Most significantly, coexpression of the EpoR/IL-3Rα chimera with either EpoR/IL-3Rβ or wild-type EpoRs suppressed Epo-induced β-globin mRNA accumulation. Taken together, these results suggest an active role for the IL-3Rα subunit in inhibiting EpoR-specific differentiating signals.


Introduction of erythropoietin receptors (EpoRs) into the interleukin-3 (IL-3)-dependent murine hemopoietic
cell line, Ba/F3, enables these cells to not only proliferate, after an initial lag in G 1 , but also to increase ␤-globin mRNA levels in response to erythropoietin (Epo). With IL-3 and Epo costimulation, IL-3-induced signaling appears to be dominant since no increase in ␤-globin mRNA occurs. Differentiation and proliferation signals may be uncoupled since EpoRs lacking all eight intracellular tyrosines were compromised in proliferative signaling but retained erythroid differentiation ability.

Intriguingly, a chimeric receptor of the extracellular domain of the EpoR and the transmembrane and intracellular domains of IL-3R␤ IL-3 chain (EpoR/IL-3R␤ IL-3 )
was capable of Epo-induced proliferative and differentiating signaling, suggesting either the existence of a second EpoR subunit responsible for differentiation or that the ␣ subunit of the IL-3 receptor (IL-3R) prevents it. Arguing against the former, a truncated EpoR lacking an intracellular domain was incapable of promoting proliferation or differentiation. An EpoR/IL-3R␣ chimera, in contrast, was capable of transmitting a weak Epo-induced proliferative signal but failed to stimulate accumulation of ␤-globin mRNA. Most significantly, coexpression of the EpoR/IL-3R␣ chimera with either EpoR/IL-3R␤ or wild-type EpoRs suppressed Epo-induced ␤-globin mRNA accumulation. Taken together, these results suggest an active role for the IL-3R␣ subunit in inhibiting EpoR-specific differentiating signals.
Erythropoietin (Epo), 1 the major in vivo stimulator of mammalian erythropoiesis (1), exerts its action by binding to specific cell surface receptors on immature erythroid progenitor cells (2,3). These cell surface Epo receptors (EpoRs) belong to the hemopoietin receptor superfamily and do not possess intrinsic tyrosine kinase activity (4). Nonetheless, within minutes of binding Epo, the EpoR and several intracellular pro-teins become tyrosine phosphorylated (5-7) through the action of an EpoR-associated tyrosine kinase, Jak2 (8). These Epoinduced tyrosine phosphorylations have been shown to correlate with both the expression of immediate-early response genes, such as c-jun and c-fos, and with mitogenesis (9). Moreover, tyrosine phosphorylation of the EpoR itself appears to be critical for activation of Stat5 and initiation of Epo-induced proliferation at physiological concentrations of Epo (10). In addition to its role in stimulating proliferation, Epo may have roles in preventing apoptosis (11,12) and in stimulating erythroid differentiation (13)(14)(15).
Studies in both our laboratory (16) and others (17)(18)(19) have demonstrated that Ba/F3 cells engineered to express the EpoR rapidly accumulate ␤-globin mRNA upon exposure to Epo. Interestingly, the tyrosine kinase inhibitor genistein blocks both Epo-induced proliferation and ␤-globin mRNA accumulation in this model system. In contrast, inhibition of protein kinase C by Compound 3 suppresses only Epo-induced differentiation without affecting proliferation (16). These observations suggest that protein phosphorylation events play a critical role in Epo-induced differentiation and that the proliferative and differentiating functions of the EpoR can be uncoupled.
Interestingly, both the extra-and intracellular domains of the EpoR have been implicated in Epo-induced differentiation. Maruyama et al. (20), for example, showed that a chimeric receptor consisting of the extracellular and transmembrane domain of the epidermal growth factor receptor and the intracellular domain of the EpoR could induce hemoglobin synthesis in TSA-8 cells, consistent with a role for the cytoplasmic domain of the EpoR in differentiating signaling. In apparent contradiction to this, chimeric receptors consisting of the extracellular domain of the EpoR and the transmembrane and intracellular domain of the interleukin-3 receptor ␤ subunit (EpoR/IL-3R␤) were also found to be capable of increasing ␤-globin mRNA levels in Ba/F3 cells (18,21). This latter result was particularly surprising in view of the observation that IL-3 inhibited the Epo-induced accumulation of ␤-globin mRNA in Ba/F3 cells expressing EpoRs.
To examine in more detail the regions within the EpoR that might be responsible for eliciting differentiation-specific signals, we have monitored the effects of various mutant and chimeric EpoRs on the induction of ␤-globin mRNA accumulation in Ba/F3 cells. Our results suggest that the cytoplasmic domains of the EpoR and the IL-3R␤ subunit are equivalent in mediating ␤-globin mRNA induction. Our results further suggest that the IL-3R␣ subunit is responsible for the IL-3-induced suppression of ␤-globin mRNA accumulation and that, at least in this model system, late stage erythroid differentiation is contingent upon both EpoR-mediated signaling and the absence of IL-3R-mediated signaling.

Generation of EpoR Mutant and Chimeric cDNAs-The mutant and
chimeric EpoRs used in this study are depicted in Fig. 1. The null EpoR mutant in which all cytoplasmic tyrosines were replaced with phenylalanines was constructed using site-directed mutagenesis as described by Damen et al. (10). The EpoR/IL-3R␤ IL-3 and EpoR/IL-3R␣ chimeras and the C-terminal truncated EpoR(Ϫ230) contain the EpoR cDNA-derived extracellular domain encoded by a KpnI-NheI fragment of pXM EpoR (provided by A. D'Andrea, Harvard Medical School, Boston, MA). The EpoR/IL-3R␤ IL-3 encompassing the extracellular domain of the EpoR and the transmembrane and cytoplasmic domains of the IL-3R␤ IL-3 subunit was constructed by ligating a KpnI-NheI fragment of the EpoR to a fragment spanning the transmembrane region of the IL-3R␤ IL-3 subunit, generated by polymerase chain reaction amplification, and an NdeI-NotI fragment of pAIC2-26 (provided by A. Miyajima, DNAX Research Institute, Palo Alto, CA), encoding the intracellular domain of the IL-3R␤ IL-3 subunit. To generate the hybrid EpoR/IL-3R␣ gene, a 255-bp fragment encoding the transmembrane and intracellular domains of the IL-3R␣ subunit was amplified by polymerase chain reaction using pSUT-1 (provided by A. Miyajima, DNAX Research Institute) as a template and ligated to the KpnI-NheI fragment of the EpoR cDNA. The C-terminal truncated EpoR(Ϫ230) encompassing the extracellular and transmembrane domains of the EpoR was constructed using a cDNA-derived extracellular domain and a polymerase chain reaction-amplified transmembrane region.
Generation of Retroviral Vectors Encoding Chimeric and Mutant EpoRs-To generate the JZenEpoR/IL-3R␤ IL-3 retroviral vector, a 2279-bp KpnI-NotI fragment encompassing the coding region of the chimeric receptor was isolated from pBS-EpoR/IL-3R␤ IL-3 and inserted by blunt end ligation into the XhoI site of JZen TK neo upstream of the neo r gene. Expression of the hybrid gene was thus under control of the myeloproliferative sarcoma virus long terminal repeat. To create MSCV-EpoR/IL-3R␣, a 1010-bp KpnI-NotI fragment encompassing the hybrid EpoR/IL-3R␣ gene was isolated from the pBS-EpoR/IL-3R␣ and inserted by blunt end ligation into the HpaI site of the MSCV PGK pac vector upstream of the PGK puro r cassette. A MSCV-EpoR(Ϫ230) PGK neo retroviral vector was generated by subcloning a blunted 843-bp KpnI-BamHI fragment of pBS-EpoR(Ϫ230), encoding the truncated EpoR, into the HpaI site of the MSCV PGK neo vector upstream of the PGK neo r cassette. The JZen null EpoR TK neo vector was constructed as reported previously (10).
Cell Lines-The ecotropic GPϩE-86 retrovirus packaging cell line (22) was obtained from Dr. A. Banks (Columbia University, New York) and was maintained in Dulbecco's modified Eagle's medium with 4500 mg/liter glucose and 10% heat-inactivated newborn calf serum, supplemented with 15 g/ml hypoxanthine, 250 g/ml xanthine, and 25 g/ml mycophenolic acid (HXM selective medium). GPϩE-86 subclones transfected with recombinant retroviral vectors were selected and maintained in HXM medium containing 1 mg/ml G418 (Canadian Life Technologies, Burlington, Ontario, Canada) or 2 g/ml puromycin (Sigma), as appropriate for selection of the virus-encoded selectable marker. The IL-3-dependent murine cell line, Ba/F3, was kindly provided by Dr. A. Miyajima (DNAX Research Insitute) and was maintained in RPMI 1640 medium with 10% heat-inactivated fetal calf serum and 3 nmol/liter COS cell-derived murine IL-3. The retrovirally infected cells were maintained in the same medium, supplemented with 1.8 mg/ml G418 and/or 2 g/ml puromycin. All media were obtained from StemCell Technologies Inc. (Vancouver, British Columbia, Canada).
Viral Production and Infection of Ba/F3 Cells-The generation of GPϩE-86 clones producing recombinant retrovirus and the infection of IL-3-dependent Ba/F3 cells were as described previously (10,16). Clones expressing various chimeric and mutant EpoRs were selected by plating the infected cells in standard methylcellulose (StemCell Technologies Inc.) supplemented with IL-3 and 1.8 mg/ml G418 or 2 g/ml puromycin. Selection of clones coexpressing WT EpoR or EpoR/IL-3R␤ IL-3 with the EpoR/IL-3R␣ chimera was carried out in the presence of 1.8 mg/ml G418 and 2 g/ml puromycin.
Proliferation Assays-Cells were grown for 2-4 days in the absence of G418, washed, and then deprived of IL-3 and fetal calf serum for 6 h in RPMI 1640 supplemented with 0.1% BSA. Cells were then washed, resuspended in RPMI 1640 containing 0.1% BSA, and aliquoted into 96-well U-shaped microtiter plates at 2.5 ϫ 10 4 cells/well, and growth factors were added to a final volume of 0.1 ml/well. Following 22 h of incubation at 37°C in a humidified atmosphere containing 5% CO 2 , cells were pulsed with 1 Ci of [ 3 H]thymidine ( 3 H-Tdr, 2 Ci/mmol, DuPont NEN) for 2 h. The cells were then harvested onto filter mats using an LKB 1295-001 Skatron cell harvester, and 3 H-Tdr incorpora-tion was determined in an LKB 1205 Betaplate liquid scintillation counter (LKB Wallac, Turku, Finland).
Northern Blot Analysis-Ba/F3 cells expressing chimeric or mutant EpoRs were incubated for 4 h in RPMI 1640 supplemented with 0.1% BSA. Cells were then washed with RPMI 1640 and resuspended in RPMI 1640 containing 0.1% BSA, or in the same medium supplemented with Epo (0.05 or 0.5 unit/ml) or 3 nmol/liter COS cell-derived IL-3, or a combination of 0.5 unit/ml Epo and 3 nmol/liter IL-3. After 18 -20 h of incubation at 37°C in a humidified atmosphere with 5% CO 2 , cells were lysed in RNAzol (Canadian Life Technologies), and the total cellular RNA was isolated as recommended by the manufacturer. Northern blot analysis was performed as described previously (16). Probes used for hybridization were a 295-bp Sau3AI-AccI fragment encompassing the first exon and intron of the murine ␤-major globin gene (provided by Dr. P. Lebouche, MIT, Boston, MA) and either a 1.3-kilobase pair PstI fragment of rat glyceraldehyde-3-phosphate dehydrogenase cDNA (provided by P. Jeanteur, Centre Paul Lamarque, Montpellier, France) or a 1.6-kilobase pair PstI fragment of chicken ␤-actin cDNA used to test for loading equivalence.

RESULTS
To examine the regions within the EpoR that might be inducing ␤-globin mRNA, we engineered retroviral vectors carrying coding regions for various mutant and chimeric EpoRs that could be compared with the WT EpoR. The forms of the various receptors studied are illustrated in Fig. 1. These included a full-length EpoR in which all eight intracellular tyrosines were exchanged for phenylalanines (null EpoR) (10), two chimeric EpoRs containing the extracellular domain of the EpoR and the transmembrane and cytoplasmic domains of the IL-3R␤ or IL-3R␣ subunit (termed EpoR/IL-3R␤ or EpoR/IL-3R␣, respectively), and a C-terminal truncated EpoR (EpoR(Ϫ230)) in which the entire cytoplasmic domain was replaced with two primer-derived arginine residues. Following retrovrial infection, IL-3-responsive Ba/F3 cells expressing the various EpoR forms were then selected for assessment of Epo- stimulated proliferation and differentiation.
EpoR Tyrosine Phosphorylation Is Not Required for EpoRmediated ␤-Globin Gene Induction in Ba/F3 Cells-To determine the importance of EpoR tyrosine phosphorylation to Epoinduced ␤-globin gene expression, we tested the mutant EpoR in which all eight cytoplasmic tyrosines were substituted with phenylalanines (null EpoR). Several independent Ba/F3 clones were obtained following retroviral mediated gene transfer and assessed for expression of cell surface null EpoRs by both flow cytometry using biotin-labeled Epo and by Scatchard analysis using 125 I-labeled Epo. Following selection of clones expressing similar numbers of cell surface WT or null EpoRs (approximately 3000/cell, as determined by Scatchard analysis), the Epo-induced proliferation and induction of ␤-globin message were compared. Null EpoR-expressing Ba/F3 cells proliferated in response to Epo as determined by 3 H-Tdr incorporation assays but required approximately 5-fold higher concentrations of Epo to achieve levels of 3 H-Tdr incorporation comparable with those obtained with WT EpoR-expressing cells (Fig. 2A). The null and WT EpoR-expressing cells accumulated comparable levels of ␤-globin mRNA upon stimulation with Epo, and for both, no induction of ␤-globin mRNA could be detected in response to IL-3 or to IL-3 plus Epo (Fig. 2B, representative Northern blot analysis). Interestingly, this Epo-induced differentiating response could be detected in both cell types at concentrations of Epo that stimulated proliferation of WT EpoR cells but were markedly less effective in promoting proliferation of null EpoR cells, consistent with recently published observations that Epo-induced differentiation of Ba/F3 cells can occur in the absence of proliferation (16,19). Our results further suggest that tyrosine phosphorylation of the EpoR itself is not required for induction of ␤-globin mRNA.
The Intracellular Domain of the IL-3R␤ IL-3 Subunit Induces ␤-Globin Gene Expression-IL-3 was previously reported to inhibit the Epo-induced increase in ␤-globin mRNA seen in Ba/F3 cells engineered to express the normal, wild-type EpoR (16,19). To identify domains within the EpoR and/or the IL-3R that might be involved in regulating ␤-globin mRNA induction in this model system, we first examined a chimeric receptor (EpoR/IL-3R␤), consisting of the extracellular domain of the EpoR and the transmembrane and cytoplasmic domains of the major signal transducing component of the IL-3R complex, the IL-3R␤ IL-3 subunit. Ba/F3 cells engineered to express this chimeric receptor proliferated as well as cells expressing similar   Fig. 2A. B, the Epo-induced accumulation of ␤-globin mRNA by cells of two representative Ba/F3 clones expressing EpoR/IL-3R␤ IL-3 chimera. Growth factor-deprived cells were incubated for 20 h in RPMI 1640 supplemented with 0.1% BSA in the absence of growth factors (lane 0) or in the the presence of 3 nmol/liter of IL-3 or IL-3 and 0.5 unit of Epo/ml or Epo alone (0.5 or 0.05 unit/ml). Lanes contained approximately 10 g of total cellular RNA. Hybridization probes are listed on the right. Clone 4 was subsequently used to assess the effect of coexpressing the EpoR/IL-3R␣ (see Fig. 6). GAPDH, glyceraldehyde-3-phosphate dehydrogenase. numbers of WT EpoRs (3000/cell as determined by Scatchard analysis) in response to Epo (Fig. 3A); moreover, the ␤-globin message was induced in these cells in response to Epo (Fig. 3B) but not in response to IL-3 or IL-3 plus Epo, consistent with results obtained by other groups (18,21).
One possibility suggested by these observations is that a second as yet unidentified subunit associates with the extracellular domain of the EpoR and provides the differentiating signal. Another intriguing possibility is that the cytoplasmic domain of the IL-3␤ subunit on its own is permissive for differentiation, but signaling through the intact IL-3R suppresses differentiation, thus pointing to a specific inhibitory role for the ␣ subunit of the IL-3R. To discriminate between these two possibilities, we tested a C-terminal truncated EpoR (EpoR(Ϫ230)) possessing only 2 amino acids within the cytoplasmic domain (Fig. 1). The differentiating and proliferative capacity of this truncated EpoR was examined in several independent EpoR(Ϫ230)-transduced Ba/F3 clones, expressing between 8000 and 12,000 cell surface EpoRs (determined by flow cytometry and Scatchard analyses, data not shown). Viability of these cells decreased in Epo-supplemented medium within 24 h to approximately 25-30%, and no viable cells could be detected by 48 h (Fig. 4A). Moreover, no accumulation of ␤-globin mRNA by the EpoR(Ϫ230)-transduced cells could be de-tected in response to Epo or Epo plus IL-3 (Fig. 4B). This indicates that an EpoR lacking the intracellular domain is not capable of promoting the survival and differentiation of Ba/F3 cells in response to Epo and argues against differentiating signaling being activated through molecules associated with the extracellular domain of the EpoR, at least not in the absence of a functional cytoplasmic domain.
The ␣ Subunit of IL-3R Inhibits Epo-induced ␤-Globin Gene Expression-To test the hypothesis that the IL-3R␣ subunit can suppress Epo-induced differentiation, we examined the differentiating and proliferative capacities of a chimeric receptor composed of the extracellular domain of the EpoR and the transmembrane and cytoplasmic domains of the IL-3R␣ subunit (EpoR/IL-3R␣). Ba/F3 cells expressing this EpoR/IL-3R␣ chimera proliferated in Epo-supplemented medium but required approximately 5-fold higher concentrations of Epo to achieve 3 H-Tdr incorporation levels comparable with those obtained by WT EpoR-expressing cells (Fig. 5A). ␤-Globin mRNA levels were then examined in several independent EpoR/IL-3R␣ clones, and no accumulation of ␤-globin mRNA could be detected in a variety of experimental conditions, including different concentrations of Epo (Fig. 5B) and different times of exposure to Epo (data not shown), suggesting that this EpoR/ IL-3R␣ chimeric receptor was not capable of promoting differentiation.
To test whether the absence of the Epo-induced accumulation of ␤-globin mRNA by EpoR/IL-3R␣ cells reflected an inability of this chimeric receptor to induce ␤-globin gene expression or an ability to actively suppress expression of the ␤-globin gene, two additional types of Ba/F3 clones were created. First, EpoR/IL-3R␤ cells, shown to accumulate high levels of ␤-globin mRNA in response to Epo (Fig. 3B, clone 4), were engineered to coexpress the EpoR/IL-3R␣ chimera, and several clones expressing 2-5-fold higher levels of cell surface EpoRs than the parental EpoR/IL-3R␤ cells were identified by Scatchard analysis. Ba/F3 clones coexpressing EpoR/IL-3R␤ and EpoR/IL-3R␣ proliferated in response to Epo (Fig. 6A) but failed to accumulate ␤-globin mRNA upon Epo stimulation (Fig. 6C). Regeneration of the IL-3R signaling complex thus prevented Epo-induced differentiation despite the capacity of the IL-3R␤ subunit alone to transduce a differentiating signal. Lastly, the capacity of the IL-3R␣ subunit to inhibit ␤-globin gene expression was examined in WT EpoR cells engineered to coexpress the EpoR/IL-3R␣ chimera. Several clones expressing 2-4-fold higher numbers of cell surface EpoRs (than the parental WT EpoR cells) were identified by Scatchard analysis (i.e. 3300 WT EpoRs/cell plus 7000 -12,000 EpoR/IL-3R␣/cell). These cells required 5-30-fold higher concentrations of Epo to achieve proliferation levels comparable with those obtained by the parental WT EpoR cells (Fig. 6B). More importantly, cells coexpressing WT EpoR and EpoR/IL-3R␣ ceased to accumulate ␤-globin mRNA in response to Epo (Fig. 6C) suggesting that the IL-3R␣ subunit inhibited the differentiating function of the EpoR.

DISCUSSION
In this study, we expressed various mutant and chimeric EpoRs in the IL-3-responsive cell line, Ba/F3, to examine the functional roles of the EpoR and IL-3R in regulating ␤-globin gene induction in Ba/F3 cells. Surprisingly, EpoRs totally lacking in potential tyrosine phosphorylation sites (null EpoR) were found to be capable of inducing ␤-globin mRNA as well as the normal EpoR. This finding is compatible with our previously published results showing that Epo-induced differentiation was blocked by genestein (16), since null EpoRs mediate tyrosine phosphorylation (and activation of Jak2) as well as WT EpoRs (10). Taken together, our data suggest that tyrosine phosphorylation of Jak2, but not the EpoR, is critical for Epoinduced differentiation. Null EpoRs, however, were severely compromised in their ability to promote proliferation of Ba/F3 cells, consistent with our previous results (10). It is conceivable that the reduced ability of the null EpoR to promote proliferation of Ba/F3 cells could create conditions permissive for differentiation. In this regard, Ba/F3 cells, although originally described as a pro-B cell line (23), express eythroid-specific transcription factors such as GATA-1 and NF-E2 (17) and very low levels of endogenous EpoRs (7) and may thus have evolved in culture toward an erythroid phenotype or have always been erythroid in nature. The Epo-induced delay in progression through G 1 of the cell cycle, shown previously for the WT (16,19) and currently for the null EpoR-expressing cells (data not shown), may simply be triggering a predetermined erythroid differentiating program, as seems to be the case for MEL cells engineered to express p53 (24). However, a simple delay in G 1 of parental Ba/F3 cells does not induce ␤-globin mRNA (16), suggesting that induction of this gene depends on EpoR-mediated signaling.
The Epo-induced accumulation of ␤-globin mRNA in cells expressing the EpoR/IL-3R␤ chimera is consistent with the previously published findings of Carroll et al. (18) and Chiba et al. (21). This observation pointed to the possibility that Epo-  3R␣ chimera (B). 3 H-Tdr incorporation assay was performed as described for Fig. 2A. Results are representative of three independent experiments. C, Northern blot analysis of ␤-globin mRNA levels upon IL-3 or Epo stimulation of Ba/F3 cells expressing WT EpoR versus a representative clone coexpressing EpoR/IL-3R␣ chimera, and Ba/F3 cells expressing the EpoR/IL-3R␤ alone (Fig. 3, clone 4) versus that clone coexpressing the EpoR/IL-3R␣ chimera. Growth factor-deprived cells were stimulated with 3 nmol/liter of IL-3 or 0.5 unit of Epo/ml in RPMI 1640 containing 0.1% BSA for 1 day (WT EpoR and EpoR/IL-3R␤) or for 1-3 days as indicated for cells coexpressing EpoR/ IL-3R␣. Each lane contains approximately 10 g of total cellular RNA. Hybridization probes are listed on the right. specific signaling might depend on the interactions of the extracellular domain of the EpoR (25) with a second as yet unidentified subunit of the EpoR complex. However, cells expressing high levels of a truncated EpoR(Ϫ230) lacking the cytoplasmic domain neither survived nor accumulated ␤-globin mRNA upon Epo stimulation, suggesting that the cytoplasmic region of the EpoR is indispensable for EpoR function. The differentiating capacity of the EpoR/IL-3R␤ chimera, however, also suggested that the cytoplasmic domains of the EpoR and the ␤ IL-3 subunit of the IL-3R were interchangeable in providing for the differentiating signal, which argues against the existence of a differentiation-specific domain within the cytoplasmic region of the EpoR and points to a permissive rather than an instructive role for the EpoR in Epo-induced differentiation.
Our finding that the EpoR/IL-3R␣ chimera was capable of supporting proliferation of Ba/F3 cells was somewhat surprising since Kitamura and Miyajima (26) reported that the human IL-3R␣ subunit alone was unable to support proliferation of IL-2-dependent CTLL-2 cells. These cells proliferated in response to human IL-3 only when engineered to coexpress human IL-3R␣ and murine IL-3R␤ c suggesting that IL-3-induced mitogenic signaling depends on interactions between ␣ and ␤ subunits of the IL-3R complex likely mediated by their extracellular domains (27). It seems unlikely that a similar association occurs between the extracellular domains of IL-3R␤ and EpoR. Our results are consistent with a steadily growing body of data suggesting that the membrane proximal region conserved among ␣ subunits of receptors for IL-3, IL-5, and granulocyte macrophage colony-stimulating factor is essential for mitogenic signaling (28 -30) and Jak2 activation (31,32). It is possible that the Epo-induced dimerization of EpoR/IL-3R␣ chimeras results in activation of a mitogenic signal through the cytoplasmic domain of the IL-3R␣ subunit that is not compatible with differentiation.
Several possible mechanisms could account for the observed inhibitory effect of the EpoR/IL-3R␣ on the partial Epo-induced differentiation of Ba/F3 cells. First, EpoR/IL-3R␣ chimeras could be forming unproductive heterodimers with EpoR/IL-3R␤ IL-3 or WT EpoRs. This mechanism seems unlikely since EpoR/IL-3R␣ also inhibited the Epo-induced accumulation of ␤-globin mRNA in a clone expressing 3300 WT EpoRs and 7000 EpoR/IL-3R␣. Random dimerization of cell surface EpoRs in this clone would be expected to yield 9% or approximately 1000 WT EpoR dimers/cell. Second, EpoR/IL-3R␣ chimeras, when overexpressed, could shorten the duration of the G 1 phase of cell cycle required for induction of ␤-globin mRNA. This was, however, not the case since exposure to Epo of all EpoR/IL-3R␣expressing Ba/F3 clones led to a marked delay in G 1 to S progression (data not shown). Thus, a delay in G 1 is not sufficient for ␤-globin mRNA induction. Third, the Epo-induced dior oligomerization of EpoR/IL-3R␣ chimeras could activate IL-3-specific pathways involved in inhibition of ␤-globin gene expression. Our results favor this last possibility and are consistent with the concept that the ␣ subunits of the IL-3 (27), , and granulocyte macrophage colony-stimulating factor (33) receptors initiate distinct ligand-induced events.
The data presented in this study suggest both a permissive role for Epo in inducing ␤-globin mRNA in Ba/F3 cells expressing EpoRs and an active role for the IL-3R␣ subunit in the IL-3-induced inhibition of Epo-induced differentiation. Based on these findings, ongoing studies are aimed at identifying effectors that are, through interaction with the IL-3R␣ subunit, involved in suppression of ␤-globin gene expression in Ba/F3 cells.