Regulation of Jak2 tyrosine kinase by protein kinase C during macrophage differentiation of IL-3–dependent myeloid progenitor cells

32Dcl3 cells expressing PKC-α, PKC-δ, and PKC-ε isoforms have been described previously.22,36 Ba/F3 cells were a kind gift from Dr M. Heim. The cDNA constructs for different PKC isoforms and HA-epitope–tagged Jak2, Stat1, and Stat5a have been described previously.22,37,38 All transfections were carried out by electroporation. For Jak2 overexpression, parental and 32D-δ cells were transfected with wild-type Jak2 in pBabe-Puro expression vector. Twenty-four hours after electroporation, 2.5 μg/mL puromycin was added to the cells, and after 2 weeks, growing cells were dilution cloned. The Jak2 expression of different clones was analyzed by Western blotting from total cell lysates.

If not otherwise stated, cells were starved for 18 hours (0.5% fetal calf serum [FCS]), pretreated with 100 ng/mL TPA or vehicle alone for 30 minutes, and stimulated with 10 ng/mL IL-3 for 15 minutes. For immunoprecipitation, cells were lysed in Triton lysis buffer (50 mmol/L Tris-HCl, pH 7.5, 10% glycerol, 150 mmol/L NaCl, 1 mmol/L EDTA, 1% Tritonx100, 50 mmol/L NaF, and 1 mmol/L Na₃VO₄) supplemented with protease inhibitors or in 1% NP40 lysis buffer (as Triton lysis buffer but with Nonidet P40 as detergent) for coimmunoprecipitation studies. Immunoprecipitations were carried out as previously described,37 and immunodetection was performed with the ECL system (Amersham Pharmacia Biotech, Buckinghamshire, UK). The antibodies used were anti-phosphotyrosine-antibody (clone 4G10; Upstate Biotechnology, Lake Placid, NY), polyclonal anti-Jak2 antibody18 (kind gift from Dr Ihle), polyclonal anti–PKC-δ, -α, -ε, or -ζ (Gibco BRL Life Technologies, Gaithersburg, MD), and anti–IL-3Rβ (Transduction Laboratories, Lexington, KY). Anti-influenza virus hemagglutinin (HA)-epitope antibody (clone 12CA5) was purchased from Berkeley Antibody (Richmond, CA).

Cell lysates were prepared in WCE lysis buffer as previously described,37 and the lysates were analyzed using γ-³²P ATP-labeled GAS oligonucleotide from the murine IRF-1 gene 5′-CTAGAGCCTGATTTCCCCGAAATGATGAG-3′. The reactions were separated in 4.5% TBE-PAGE (2.2 × TBE) followed by autoradiography.

Jak2 substrate kinase assay was performed from immunoprecipitated proteins as described using synthetic peptide corresponding to the tyrosine phosphorylation site of Stat1 (GPKGTGYIKTELISVS).38 For PKC-δ kinase assay, the immunocomplexes were suspended in kinase reaction buffer (0.04 mg/mL phosphatidyl-L-serine, 10% glycerol, 0.1 mmol/L CaCl₂, 0.02% Triton-X-100, 10 mmol/L MgCl₂, 20 mmol/L HEPES, pH7.4) in the presence of γ-ATP (0.25 μCi/mL), and the reaction was allowed to proceed for 30 minutes at 30°C. The proteins were resolved in SDS-PAGE, and the gels were vacuum dried and exposed for autoradiography. Phospho-amino acid analysis was performed according to standard procedures using 2-dimensional thin-layer electrophoresis.39 

Growing cells were washed twice and stimulated as indicated, and 18 hours later ³H-thymidine was added to the cultures for 6 hours (1 μCi/well), after which the cells were harvested. Apoptotic cells were detected by analyzing DNA fragmentation. The TUNEL assay was performed using the In Situ Cell Death Detection Kit (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer's instructions. Alternatively, fragmented DNA was isolated from cells, electrophoresed in 2% agarose gels, and stained with ethidium bromide.

Treatment of 32D cells stably expressing PKC-α and PKC-δ (32D-α and 32D-δb, respectively) with TPA induces macrophage differentiation and cessation of cell proliferation.22 To investigate whether the differentiation-induced changes in 32D cells involved regulation of IL-3 signaling, the effect of PKC activation on cellular protein tyrosine phosphorylation was studied in 32D cell lines stably expressing different PKC isoforms.22Cells were treated with TPA for the indicated times before IL-3 stimulation. In parental (32D-wt) or in 32D-ε cells (32D cells expressing PKC-ε), TPA treatment did not affect the IL-3–induced tyrosine phosphorylation, whereas in 32D-α and 32D-δ cells, a marked inhibition in tyrosine phosphorylation was observed, particularly among the high molecular weight proteins (130-140 kd) (Figure 1). The inhibition was already evident after 5 minutes of TPA treatment and lasted for at least 6 hours. In 32D-δ cells, TPA also induced tyrosine phosphorylation of the 78-kd PKC-δ as previously reported.36 

Two critical IL-3 signaling proteins, Jak2 and IL-3Rβ, are rapidly tyrosine phosphorylated on IL-3 treatment, and they migrate at approximately 130-140 kd; therefore, they were chosen as targets for further investigation. Jak2 was efficiently tyrosine phosphorylated in all 32D-derived cell lines after IL-3 stimulation, but TPA treatment inhibited tyrosine phosphorylation of Jak2 in 32D-α and in 32D-δ cells (Figure 2A). The TPA-dependent inhibition of Jak2 was analyzed in 2 independent 32D-δ clones with identical results (data not shown). In contrast, TPA stimulation had no effect on tyrosine phosphorylation of Jak2 in 32D-wt or in 32D-ε cells (Figure 2A). To verify that the observed inhibition of Jak2 was not a unique feature of 32D cells, the effect of TPA treatment on activation of Jak2 was investigated in the IL-3–dependent Ba/F3 cell line. In Ba/F3 cells, TPA-induced activation of endogenous PKC led to the abrogation of Jak2 tyrosine phosphorylation and the inhibition of IL-3–induced thymidine incorporation (Figure 2A; data not shown). The predominantly activated PKC isoform in Ba/F3 cells was PKC-α (data not shown). The PKC-mediated inhibition of Jak2 was studied in more detail in 32D-δ cells, and the inhibition was found to be TPA-concentration dependent and was not caused by the altered kinetics of Jak2 activation (Figure 2B). This differed from the previously demonstrated SHP-1 tyrosine phosphatase-mediated inhibition of Jak2, which primarily affects the kinetics of Jak2 activation.8 Taken together, these results indicate that PKC-α and PKC-δ can mediate the rapid inhibition of Jak2 in hematopoietic cells.

Another potential target for the PKC-mediated inhibitory effect was IL-3Rβ. Tyrosine phosphorylation of IL-3Rβ was investigated from immunoprecipitated proteins in 32D-wt and in 32D-δ cells. IL-3 stimulation resulted in rapid phosphorylation of IL-3Rβ, which was strongly inhibited on TPA treatment in 32D-δ cells but not in parental cells (Figure 2C). IL-3Rβ expression levels remained unchanged, and the mobility of the IL-3Rβ in SDS-PAGE was unaffected by TPA treatment, thus suggesting that the reduction in tyrosine phosphorylation resulted from the inhibition of Jak2 and not from the shedding or degradation of the receptor.

Protein kinase C is known to regulate several cellular signaling pathways, and at least 2 possible mechanisms for PKC-mediated inhibition of Jak2 can be envisaged: PKC could activate tyrosine phophatase(s)8,16 or it could induce Ser/Thr phosphorylation of Jak2, either directly or through the activation of other kinase(s), such as the MEK-ERK pathway. Pretreatment of 32D-δ cells with either tyrosine phosphatase inhibitor Na₃VO₄ (1 mmol/L, 30 minutes) or with the MEK inhibitor PD09805950 (μmol/L, 30 minutes) did not reverse the TPA-dependent inhibition of Jak2 activation in response to IL-3 (data not shown). To investigate whether Jak2 is the primary target for PKC-mediated inhibition of IL-3 signaling, Jak2 was activated independently of the receptor by overexpression in COS7 cells. Coexpression of either PKC-α, -δ, -ε, or -ζ with Jak2 and Stat5 revealed striking specificity of PKC isoforms in the inhibition of Jak2. In accordance with the results obtained in 32D cells, only PKC-α and PKC-δ could inhibit the in vivo catalytic activity of Jak2 as measured by Jak2-induced tyrosine phosphorylation of Stat5 (Figure 3A). A similar degree of overexpression of the PKC isoforms was verified by immunoblotting from cell lysates (data not shown). PKC-α and PKC-δ also inhibited the Jak2-induced functional activation of Stat1 in electrophoretic mobility shift assay (Figure 3B). Coexpression of Jak2 with PKC-δ in COS7 cells resulted in the inhibition of Jak2 tyrosine phosphorylation (Figure 4A). Further proof for a direct PKC-mediated inhibition of Jak2 was obtained from in vitro kinase reactions. COS7 cells were transfected with either Jak2 alone or together with PKC-δ or PKC-ζ, and Jak2 was immunoprecipitated and subjected to in vitro substrate kinase reaction. Coexpression of PKC-δ resulted in the inhibition of the catalytic activity of Jak2, whereas PKC-ζ did not reduce Jak2 activity (Figure 4B). Collectively, these results indicated that Jak2 is the target for the PKC-δ–mediated inhibition of cytokine signaling.

The role of the catalytic activity of PKC in inhibition of Jak2 was studied in 32D cells constitutively expressing a kinase-inactive mutant of PKC-δ.36 In these cells, TPA treatment did not inhibit IL-3–induced tyrosine phosphorylation of Jak2, suggesting that the inhibition of Jak2 required PKC-δ–mediated phosphorylation (Figure5A). Identical results were obtained with 2 independent 32D clones expressing kinase-inactive PKC-δ (data not shown). To investigate whether Jak2 was a substrate for PKC, COS7 cells were transfected with HA-tagged Jak2 either alone or together with PKC-δ or PKC-ε, and the immunoprecipitated Jak2 was subjected to in vitro kinase assay under conditions that support the catalytic activity of PKC but not Jak2. In this assay, only Jak2, which was purified from PKC-δ–expressing cells but not from PKC-ε cells, became phosphorylated (Figure 5B). Phosphorylation of Jak2 by PKC-δ indicated a direct interaction between the proteins, and, in support of this notion, PKC-δ was found to coimmunoprecipitate with the HA-tagged Jak2 (Figure 5C). The band corresponding to the phosphorylated Jak2 was excised from the gel, and phospho-amino acid analysis demonstrated that Jak2 was phosphorylated by PKC-δ on both serine and threonine residues but not on tyrosine residues (Figure 5C).

To investigate the in vivo significance of Jak2 inhibition during macrophage differentiation, we wanted to revert the inhibitory effect of PKC by stably overexpressing Jak2 in 32D-δ cells (32D-δ–Jak2). In those cells, overexpression of Jak2 resulted in increased IL-3–dependent tyrosine phosphorylation of Jak2 and subsequently diminished PKC-δ–mediated inhibition compared with 32D-δ cells (Figure 6A). In thymidine incorporation assay, overexpression of Jak2 doubled the IL-3–induced increase in DNA synthesis (Figure 6B). As predicted, TPA treatment inhibited DNA synthesis only in PKC-δ–expressing cells but not in parental 32D cells. By contrast, the tyrosine kinase inhibitor AG490, which has been shown to inhibit various Jak kinases, inhibited cell proliferation both in parental and in 32D-δ cells. Furthermore, TPA-treated 32D-δ cells, like the IL-3 starved or the AG490-treated cells, similarly arrested in the G1 phase of the cell cycle, thus supporting the notion that Jak2-induced cell-cycle progression signal were abrogated by PKC-δ (data not shown). TPA induced early morphologic differentiation in 32D-δ and 32D-δ–Jak2 cells, but the sustained activation of Jak2 in differentiating 32D-δ–Jak2 cells promoted apoptotic cell death, as shown by the TUNEL analysis (Figure 6C). The overexpression of either Jak2 or PKC-δ did not cause apoptosis because the apoptotic DNA ladder was observed only in TPA-treated 32D-δ–Jak2 cells but not in similarly treated parental 32D cells overexpressing Jak2 at a comparable level or in 32D-δ cells (data not shown). This result was confirmed using several independent cell clones. Together these results suggested that the inhibition of Jak2 activity serves as an important mechanism for cell-cycle arrest during macrophage differentiation of IL-3–dependent 32D progenitor cells.

Development and maintenance of hematopoietic cells require coordinated regulation of cytokine signaling. Terminal differentiation of macrophages is accompanied by the cessation of cytokine-dependent cell growth, which is likely to be important for the successful accomplishment of the differentiation program.40 In cytokine-dependent progenitor cells, growth arrest could potentially involve the inhibition of cytokine receptor mitogenic signaling. The current study was aimed at characterizing the IL-3R signaling events during terminal differentiation of macrophages induced by PKC activation. Biologic responses to PKC activation are regulated by the differential expression of various PKC isoforms and by their functional specificities.27 In support of this notion we found that 2 PKC isoforms, PKC-α and PKC-δ, specifically inhibited the activation of Jak2. The PKC-α and PKC-δ isoforms mediate macrophage differentiation and growth inhibition of myeloid progenitor cells,21,22 and in 32D cells the differentiation-inducing PKC isoforms caused a rapid inhibition of IL-3R–induced activation of Jak2 that was closely correlated with growth arrest. The inhibition was not cell-type dependent, and the activation of endogenous PKC in Ba/F3 cells inhibited IL-3–stimulated activation of Jak2. During macrophage differentiation, the M-CSF receptor has been shown to inhibit IL-3 growth signal through a PLC-γ2–dependent mechanism.26,28,29 In light of our current results, the PKC-mediated inhibition of Jak2 could serve as a molecular mechanism for this response.

Our results indicated that PKC-α and PKC-δ isoforms displayed significant specificity for the inhibition of Jak2, and the inhibition involved direct PKC-δ–mediated phosphorylation of Jak2 on serine and threonine residues. Inhibition of Jak2 was not caused by the down-regulation of IL-3R, and the PKC-α and PKC-δ isoforms also specifically inhibited Jak2 in ligand-independent overexpression models. The possibility that the PKC-induced inhibition of Jak2 would be mediated indirectly either by tyrosine phosphatases8,16or by MEK-ERK kinases was argued against by experiments with tyrosine phosphatase and MEK kinase inhibitors, which did not affect the inhibition. Furthermore, MEK1 is activated similarly by all PKC isotypes, which makes it a less likely mediator of the PKC isoform-specific inhibition of Jak2 in these cells.41 On the other hand, the rapid kinetics of the inhibition does not support involvement in this response for the transcriptionally regulated Jak inhibitors SOCS/SSI/JAB.3-5 Recently, ERK2 kinase was shown to associate with and to inhibit Stat3.13,14 Interestingly, the effect of ERK2 was receptor specific because the inhibition was observed in IL-6 but not in interferon-α–induced activation of Stat3.13 Thus, it appears that PKC, or certain isoforms of PKC, can regulate cytokine signaling by 2 distinct mechanisms, directly through interaction with Jak2 and indirectly through activation of ERK2 in a receptor-specific manner.

Jak2 plays an indispensable role in IL-3R mitogenic signaling.18-20 PKC-mediated inhibition of Jak2 was closely correlated with the abrogation of cell growth in differentiating 32D cells. The inhibition of Jak2 either by PKC-δ, tyrosine kinase inhibitor AG490, or IL-3 deprivation all resulted in similar G1 arrest of the cell cycle. However, the differentiated phenotype was observed in only TPA-treated cells (data not shown). These results are in agreement that Jak2 activation is critical for cell-cycle progression and induction of short-term proliferation,20,42 and they suggest that in growth factor–dependent myeloid cells, the inhibition of Jak2 functions as a mechanism to control cell proliferation during differentiation. However, cell-cycle progression can be negatively regulated by other mechanisms. In growth factor–independent myeloid cell lines, differentiation-inducing agents have been shown to stimulate cyclin-dependent kinase inhibitor gene expression.43,44 The existence of parallel mechanisms for controlling cell proliferation emphasizes the importance of this regulation for successful execution of cellular differentiation. In support of this notion, the reversal of Jak2 inhibition by ectopic expression of Jak2 resulted in increased apoptosis among the differentiating cells. The molecular mechanism for the observed apoptosis in 32D-δ–Jak2 is unknown. However, c-myc has been shown to be directly regulated by Jak2,42 and constitutive c-myc expression has been shown to promote apoptosis under restrictive growth conditions.45,46 In our experiments, apoptosis was more pronounced in reduced serum concentrations, though it was observed also in optimal growth conditions. On the other hand, PKC-δ has been associated with the induction of apoptosis.47 Prolonged (24-hour) TPA treatment used in our differentiation experiments resulted in the down-regulation of PKC; thus, the observed apoptosis could have been caused by either PKC-δ–mediated signaling or the absence of PKC-δ. Proteolytic cleavage and activation of PKC-δ have been directly implicated in the onset of apoptosis in various cell types,47-49 possibly because of the PKC-δ–mediated inhibition of DNA-dependent protein kinase-mediated phosphorylation of p53.50 However, in 32D-δ cells, macrophage differentiation did not result in marked apoptosis. Therefore, it is likely that the activation of PKC-δ, rather than its down-regulation, with the simultaneous activation of Jak2 in 32D-δ–Jak2 cells, results in a proapoptotic effect that overrides the Bcl-2– and Bcl-x_L–mediated anti-apoptotic pathway from IL-3R and Jak2.20 

Inhibition of Jak2 by specific PKC isoforms represents a previously uncharacterized mechanism of negative regulation of cytokine signaling that appears to integrate different signaling pathways during hematopoietic cell differentiation. However, IL-3 deprivation or AG490 treatment does not lead to macrophage phenotype. Thus, Jak2 inhibition alone is not sufficient to cause terminal differentiation of 32D cells. This suggests that PKC mediates at least 2 distinct signals, both of which are required for proper execution of the differentiation process: one signal for the direct phosphorylation and inhibition of Jak2 associated with growth arrest, and another signal for the induction of differentiation.

We thank Paula Kosonen for technical assistance, Drs O. Jaakkola, H. Kankaanranta, and A. Lagerstedt for help with apoptosis assays and photography, Dr K. Saksela for discussions, and Drs K. Alitalo, L. Andersson, M. Hurme, and H. Jacobs for reading the manuscript.