Involvement of Protein Kinase Cε in the Negative Regulation of Akt Activation Stimulated by Granulocyte Colony-Stimulating Factor¹

The serine/threonine kinase Akt, also termed protein kinase B, plays an important role in cell survival and proliferation. Akt is activated in cells in response to diverse stimuli such as growth factors, cytokines, and hormones through the PI3K pathway (1, 2, 3, 4, 5). PI3K generates phosphoinositides PI(3,4,5)P3 and PI(3,4)P2 that target Akt via its pleckstrin homology domain to the plasma membrane, where it is phosphorylated at Thr³⁰⁸ and Ser⁴⁷³ by phosphoinositide-dependent kinase (PDK)⁴ 1 and a yet to be identified kinase, respectively. Membrane targeting and subsequent phosphorylation result in Akt activation. Akt phosphorylates and regulates the functions of many cellular proteins implicated in apoptosis and proliferation. The activity of Akt is tightly controlled in normal cells, but frequently constitutively activated in many types of human cancer (1, 2, 3, 4, 5).

G-CSF plays a major role in granulopoiesis. Treatment of cells with G-CSF has been shown to activate Akt in a PI3K- and Src family kinase-dependent manner (6, 7). Akt is involved in the regulation of the survival, proliferation, and differentiation of myeloid cells in response to G-CSF (6, 7, 8). Notably, truncation of the C-terminal region of the G-CSF receptor, as seen in patients with acute myelogenous leukemia (AML) evolving from severe congenital neutropenia (SCN), results in markedly prolonged Akt activation (6). Apart from its role in normal granulopoiesis, Akt is constitutively activated in leukemic cells and plays a critical role in cellular transformation by leukemogenic proteins Bcr-Abl, Tel-Jak2, and Flt3/ITD (9, 10, 11, 12, 13).

PKC isozymes comprise a family of related serine-threonine kinases that can be grouped into three categories on the basis of their structural and biochemical properties (14, 15). The classical PKCs are composed of isoforms α, β_I, β_II, and γ, which are calcium dependent and activated by diacylglycerol (DAG) or phorbol ester. The novel PKCs include isoforms δ, ε, η, θ, and μ, which are calcium unresponsive and activated by DAG/phorbol ester. Atypical PKCs consist of isoforms ιλ and ζ, which are unresponsive to both calcium and DAG/phorbol ester. Specific PKC isoforms play pivotal roles in the regulation of myeloid, erythroid, and megakaryocytic development (16, 17, 18, 19). Expressions of various PKC isoforms are strictly regulated during hemopoietic development (20, 21). Although PKC pathway has been implicated in G-CSF receptor signaling (22), its biological significance remains to be determined.

Despite significant progress in our understanding of the molecular mechanisms by which Akt is activated, relatively less is known about the signaling events that negatively regulate Akt activation. In this work, we show that treatment with PMA, a well-known PKC activator, resulted in complete inhibition of Akt activation by G-CSF in myeloid 32D. The negative effect of PMA was abrogated by preincubation of cells with specific PKC inhibitors. We further show that PKCε, whose expression was down-regulated during G-CSF-induced granulocytic differentiation, was involved in mediating the inhibitory effect of PMA. Our data provide the first evidence that PKCε negatively regulates Akt activation, and add to the understanding of the complex regulatory mechanism that controls Akt activation in myeloid cells.

The 32D cells stably transfected with the wild-type and truncated (D715) forms of the G-CSF receptor have been described (23). The 32D cells used in this study did not express the endogenous G-CSF receptor. L-G cells (24) were provided by T. Honjo (Kyoto University, Kyoto, Japan). The 32D and L-G cells were grown in RPMI 1640 medium supplemented with 10% FBS and 10% WEHI-3B cell-conditioned medium as a crude source of IL-3, 100 μg/ml penicillin, and 100 μg/ml streptomycin.

Anti-Akt Ab, phospho-specific Abs against Akt, PDK1, STAT3, STAT5 Erk1/2, JNK, and p38; and Akt kinase assay kit were purchased from Cell Signaling Technology. Anti-PKCε Ab and anti-G-CSF receptor Ab were obtained from Santa Cruz Biotechnology and BD Biosciences, respectively. Anti-FLAG (M2) and anti-β-actin Abs were purchased from Sigma-Aldrich. ECL kit and GelCode blue stain reagent were purchased from Pierce Biotechnology.

The expression constructs for the wild-type, constitutively active (A159E), and kinase-dead (K437R) forms of PKCε have been described (25). The 32D and L-G cells were transfected by electroporation and selected in medium containing G418 (0.8 mg/ml) 24 h after transfection. Individual clones were expanded and examined for expression of transfected proteins by Western blotting. Two independent clones for each PKCε form were used in subsequent experiments.

Cells were washed with ice-cold PBS and resuspended in lysis buffer (50 mM Tris (pH 7.5), 150 mM NaCl, 10 mM NaF, 0.5 mM DTT, 1% Triton X-100, 1 mM PMSF, and 1 mM vanadate). After incubation on ice for 20 min, lysates were cleared by centrifugation at 12,000 rpm for 30 min at 4°C. The preparation of membrane and cytosolic extracts was essentially as described (26). Cells were resuspended in hypotonic buffer (20 mM Tris (pH 7.5), 10 mM EGTA, and 1 mM PMSF) and lysed by 50 strokes in a glass Dounce homogenizer. After centrifugation, the supernatants were collected as cytoplasmic fraction. The membrane pellets were solubilized in hypotonic buffer containing 1% Triton X-100 and centrifuged at 12,000 rpm for 30 min. The resulting supernatants were collected as membrane fractions.

Cell extracts were boiled in SDS sample buffer and resolved by SDS-PAGE before transfer to Immobilon membranes. The membranes were incubated with the appropriate Abs, and the reactive proteins were visualized by ECL.

The kinase activity of Akt was analyzed using the Akt kinase assay kit, according to the manufacturer’s protocol. Cells were starved in serum-free medium for 4 h and stimulated with G-CSF for 10 min with or without PMA pretreatment. Akt was immunoprecipitated with an immobilized Akt Ab from equal amounts of whole cell extracts and incubated with a glycogen synthase kinase-3 (GSK-3) fusion protein in the presence of ATP. The reaction mixtures were separated by SDS-PAGE and transferred to Immobilon membranes. The membranes were incubated with a phospho GSK-3αβ (Ser^21/9) Ab supplied in the kit. GSK3 phosphorylation was visualized by chemiluminescence.

To investigate the intracellular events that negatively regulate Akt activation, we examined the effect of PMA on G-CSF-stimulated activation of Akt in myeloid 32D cells stably transfected with the wild-type G-CSF receptor (32D/WT) (23). Cells were incubated with PMA for different times before stimulation with G-CSF for 15 min. Whole cell extracts were prepared and examined for Akt phosphorylation by Western blot analysis using Abs that recognize Akt only when phosphorylated on Ser⁴⁷⁶ or Thr³⁰⁸ (see below). As shown in Fig. 1 A, PMA treatment of cells for only 5 min effectively blocked G-CSF-stimulated activation of Akt. PMA significantly attenuated Akt activation even when added with G-CSF simultaneously or 5 min later. In contrast to Akt, G-CSF-stimulated activation of other signaling pathways, including STAT3, STAT5, p38, JNK, and Erk1/2, was not significantly affected by PMA treatment. Thus, it appeared that PMA treatment specifically inhibited Akt activation by G-CSF.

We then determined the duration of PMA-mediated inhibitory effect. Cells were pretreated with PMA for 15 min, washed extensively to remove PMA, and incubated in serum-free medium for different times before G-CSF stimulation. As shown in Fig. 1 B, Akt was not activated by G-CSF during the first 60 min post-PMA treatment, but reactivated by G-CSF after 120 min.

It has been shown that truncation of the C-terminal region of the G-CSF receptor, as seen in patients with AML/SCN, results in prolonged activation of Akt upon G-CSF stimulation (6, 7). We were interested to know whether PKC shortened the duration of Akt activation mediated by the truncated G-CSF receptor. Cells expressing a carboxy truncated G-CSF receptor, which was expressed in AML/SCN patients (27, 28, 29), were incubated with G-CSF for 15 min before addition of PMA. PMA reduced Akt phosphorylation to nearly basal levels within 15 min after it was added to the culture (Fig. 2). In the absence of PMA, Akt remained strongly phosphorylated for at least 90 min. Thus, PMA suppressed sustained Akt activation mediated by the carboxy truncated G-CSF receptor.

To evaluate the role of PKC pathway in G-CSF-stimulated Akt activation, we investigated whether PKC inhibitors abrogated the negative effect of PMA. The 32D/WT cells were preincubated with different PKC inhibitors before addition of PMA. PMA-mediated inhibition of Akt activation was completely reversed by GF109203X, an inhibitor of PKCα, β_I, β_II, γ, δ, and ε isoforms, and PKCε inhibitor Ro318220 (30), whereas Go6976, an inhibitor of PKCα and β_I (31), and PKCδ inhibitor rottlerin (32) had no significant effect (Fig. 3,A). GF109203X and Ro318220 also blocked the inhibitory effect of PMA on the kinase activity of Akt that was stimulated by G-CSF (Fig. 3 B).

The results obtained with the different PKC inhibitor suggested that the classical and/or novel PKC isoforms, including PKCε, might be involved in the negative regulation of Akt activation. We further investigated whether PKCε was activated by PMA treatment by examining its membrane translocation. As shown in Fig. 3 C, PKCε was translocated to plasma membrane upon PMA treatment in 32D/WT cells as well as in 32D/WT cells overexpressing PKCε.

PDK1 is involved in Akt activation by phosphorylating Akt at Thr³⁰⁸. PDK1 is activated by phosphorylation of Ser²⁴¹ located in the activation loop of its kinase domain (33), which correlates with Akt activation (34). To assess whether PMA treatment influenced PDK1 activation, we examined PDK1 phosphorylation at Ser²⁴¹ by Western blot analysis. PDK1 phosphorylation was not altered by PMA treatment (Fig. 4,A). Because PDK1 activation may also involve its translocation to the plasma membrane (35, 36), we investigated whether PMA treatment altered PDK1 membrane localization. As shown in Fig. 4 B, treatment of 32D cells with PMA did not significantly affect the membrane localization of PDK1.

PP2A has been shown to dephosphorylate Akt and negatively regulate its activity (37, 38, 39, 40). To assess the potential involvement of PP2A in PMA-mediated inhibition of Akt activation, we examined whether the inhibitory effect of PMA was abolished by pretreatment of cells with okadaic acid, a PP2A inhibitor. Preincubation of 32D/WT cells with okadaic acid at concentrations up to 50 μM had no significant effect on PMA-mediated inhibition of Akt activation (Fig. 4 C and data not shown). Okadaic acid alone did not affect Akt phosphorylation induced by G-CSF. Together, the results suggested that PP2A might not play a major role in the regulation of Akt activation by G-CSF in 32D cells.

To directly demonstrate the involvement of PKCε in the negative regulation of Akt activation, we examined whether overexpression of PKCε and the constitutively active PKCε AE (25, 41) affected Akt activation. When overexpressed, a portion of PKCε was constitutively translocated to the membrane in 32D/WT cells (Fig. 3,C). G-CSF-stimulated activation of Akt was markedly attenuated upon overexpression of PKCε or PKCε AE, but not the kinase-dead PKCε KR (Fig. 5,A). Notably, treatment of cells with PKC inhibitors GF109203X and Ro318220, but not with Go6976 and rottlerin, restored Akt activation (Fig. 5 B). GF109203X and Ro318220 appeared less effective in abrogating the inhibitory effect of PKCε AE, suggesting that the two inhibitors may be less efficient in suppressing the activity of PKCε AE.

In addition to G-CSF, IL-3 has also been shown to stimulate Akt activation. We investigated whether IL-3-stimulated Akt activation was affected by PKCε overexpression. As shown in Fig. 5 C, Akt was activated by IL-3 in 32D/WT cells transfected with the empty vector, but not in cells overexpressing PKCε and PKCε AE. Thus, PKCε inhibited Akt activation that was stimulated by G-CSF and IL-3.

We further examined G-CSF- and IL-3-dependent proliferation and survival of 32D/WT cells overexpressing PKCε, PKCε AE, or PKCε KR. Cells were cultured in IL-3- or G-CSF-containing medium. Cell numbers and viability were determined on different days. The 32D/WT cells transfected with PKCε and PKCε AE barely grew and lost viability rapidly when cultured in G-CSF-containing medium (Fig. 6). Few living cells could be seen after culture in G-CSF for 5 or 6 days. The 32D/WT cells transfected with the empty vector or PKCε KR showed comparable growth and survival. Notably, although PKCε inhibited IL-3-stimulated Akt activation, overexpression of PKCε and PKCε AE had no or only minimal effect on the proliferation and survival of 32D/WT cells cultured in IL-3. Thus, PKCε specifically inhibited cell proliferation and survival stimulated by G-CSF.

In addition to 32D cells, Akt activation in myeloid L-G cells was also inhibited by PMA treatment (data not shown). L-G cells expressed the endogenous G-CSF receptor and underwent terminal granulocytic differentiation in response to G-CSF (24). To examine whether PKCε overexpression in L-G cells also inhibited G-CSF-dependent proliferation and survival, we stably transfected L-G cells with the expression constructs for PKCε and PKCε AE. The expression of PKCε and PKCε AE was confirmed by Western blot analysis (Fig. 7,A). L-G cells showed transient and modest growth in G-CSF-containing medium (Fig. 7 B). Overexpression of PKCε and PKCε AE in L-G cells resulted in rapid loss of viability, and essentially no living cells were seen after culture in G-CSF for 3 days.

G-CSF induced the terminal granulocytic differentiation of 32D/WT and L-G cells (24, 27). We investigated whether the expression of PKCε altered during G-CSF-induced granulocytic differentiation. Expression of PKCε was examined by Western blot analysis. As shown in Fig. 8, PKCε protein decreased gradually when 32D/WT and L-G cells were induced to differentiate, and was barely detectable after culture of the cells in G-CSF for 5 or 6 days.

Akt plays an important role in the survival, proliferation, and differentiation of myeloid cells in response to G-CSF (6, 7, 8). In this work, we have shown that PKC activator PMA specifically inhibits G-CSF-stimulated activation of Akt, but not the other signaling pathways in myeloid 32D. PMA also inhibits Akt activation in myeloid L-G cells and pro-B Ba/F3 cells stably transfected with the G-CSF receptor (data not shown). The inhibitory effect of PMA is abrogated by preincubation of cells with specific PKC inhibitors, indicating that the PKC pathway negatively regulates Akt activation by G-CSF. We have further demonstrated that the inhibitory effect of PMA is mediated, at least in part, by PKCε. Notably, expression of PKCε is down-regulated during G-CSF-induced granulocytic differentiation, and overexpression of PKCε suppresses the survival and proliferation of 32D and L-G cells stimulated by G-CSF.

PMA has previously been shown to inhibit Akt activation in different cell types (42, 43, 44, 45). Interestingly, the PKC isoforms involved in mediating PMA effect were different in these studies. For instance, activation of PKCα by PMA was shown to inhibit insulin-induced Akt phosphorylation in vascular smooth muscle cells, whereas PKCδ was suggested to be involved in PMA-induced inhibition of Akt phosphorylation in a prostate cancer cell line (44, 45). In 32D cells, however, the negative effect of PMA was not abolished by preincubation of cells with PKCαβ_I inhibitor Go6976 (31) and PKCδ inhibitor rottlerin (32), suggesting that PKCα, β_I, and δ are unlikely the major players in PMA-induced inhibition of Akt activity. Our data implicate PKCε as an important mediator of the inhibitory effect of PMA. However, it is of note that G-CSF-stimulated Akt activation was not always completely blocked by overexpression of PKCε and PKCε AE in 32D cells (Fig. 5 A). Further studies are needed to investigate whether other PKC isoforms such as PKCβ_II and γ are also involved in PMA-induced inhibition of Akt activation.

The mechanism by which PKCε negatively regulates Akt activation is still speculative. PMA treatment had no significant effect on PDK1 phosphorylation and membrane localization. The inhibitory effect of PMA was not abolished by PP2A inhibitor okadaic acid. These results suggest that PDK1 and PP2A may not be the targets of regulation by PMA. Akt activation by G-CSF is PI3K and Src family kinase dependent (6, 7). The Src family kinases Lyn and Hck are activated by G-CSF and are required for G-CSF-stimulated cell proliferation and survival (46, 47). Notably, PMA has been shown to inhibit the activation of PI3K and Lyn stimulated by GM-CSF in human neutrophils (48). Thus, it is possible that PMA-mediated inhibition of Akt activation may result from the negative effect of PMA on PI3K and Lyn activation. However, the possibility cannot be excluded that PKCε might target other components of the G-CSF receptor/Akt-signaling pathway.

Our results are in contrast to several recent studies that implicate PKCε as a positive regulator of Akt activation in nonhematopoietic cells (49, 50, 51). Akt appeared to act downstream of PI3K and PDK1, as suggested by the fact that a kinase-dead PKCε mutant inhibited insulin-induced activation of Akt, but not PI3K and PDK1, in Chinese hamster ovary cells (49). Furthermore, PKCε was shown to interact with Akt in prostate cancer cells and mouse heart, and to directly phosphorylate and activate Akt in vitro (50, 51). Thus, it is plausible that the role of PKCε in Akt activation may depend on the types of cells and stimuli.

The biological significance of PKCε-mediated inhibition of Akt activation in granulopoiesis remains to be explored. Overexpression of PKCε, which results in its constitutive membrane translocation (see Fig. 3,C), inhibits the survival and proliferation of myeloid 32D and L-G cells stimulated by G-CSF, consistent with its negative role in Akt activation. Although PKCε also inhibits IL-3-stimulated Akt activation, overexpression of PKCε or PKCε AE has no significant effect on the proliferation and survival of 32D and L-G cells cultured in IL-3 (Figs. 6 and 7, and data not shown), indicating that PKCε specifically suppresses G-CSF-dependent cell proliferation and survival. Notably, expression of PKCε was down-regulated in 32D and L-G cells that were induced with G-CSF to undergo terminal granulocytic differentiation. Myeloid progenitor cells gradually lose the abilities for proliferation and survival with terminal granulocytic differentiation. It is possible that PKCε down-regulation may be necessary for the survival of myeloid progenitor cells in order for them to complete the differentiation process. It is also of note that the sustained Akt activation mediated by the carboxy truncated G-CSF receptor, which is associated with AML development in patients with SCN, is rapidly attenuated by PMA treatment in 32D cells. Akt is frequently constitutively activated in leukemia and may contribute to leukemogenesis (9, 10, 11, 12, 13). Given its negative role in Akt activation and in the survival and proliferation of myeloid cells, it would be interesting to examine PKCε expression in leukemic cells and to investigate the effect of stimulating PKCε activity on leukemogenic transformation.

The authors have no financial conflict of interest.