INTRODUCTION

Protein kinase C (PKC)¹ is a family of at least 12 structurally related phospholipid-dependent serine/threonine protein kinases that are directly involved in the transmission of a wide variety of extracellular signals (1). The PKC enzyme family can be divided into three subgroups: the calcium-dependent or cPKCs (alpha (), beta I (_I), beta II (_II), and gamma ()); the novel or nPKCs (delta (), epsilon (), eta (), theta (), and mu (µ)); and the atypical or aPKCs (zeta (), iota (), and lambda ()) (reviewed in Ref. 2). These groupings are based on the presence or absence of functional domains that confer isotype-specific co-factor and activator requirements. Biochemical and immunologic studies indicate that multiple PKC isotypes are expressed in virtually all cell and tissue types (3, 4), suggesting a universal role in cellular function. The expression of individual PKC isotypes is developmentally regulated (5, 6) and is responsive to the differentiation state of many cell and tissue types (7, 8). For these reasons, it is thought that PKC isotypes fulfill distinct, nonredundant functions within the cell.

In human leukemia cells, we have provided direct evidence for PKC isotype specific function in the control of cellular proliferation and differentiation (8). The two conventional PKC isotypes expressed in these cells, PKC  and PKC _II play distinct roles in these cellular processes (8). PKC  is involved in cellular differentiation and overexpression of PKC  leads to gene dose-dependent cytostasis and increased sensitivity to differentiating agents such as phorbol myristate acetate (PMA) (8-10). In contrast, the levels of PKC _II correlate with the proliferative state of these cells, being lost when cells differentiate in response to PMA (8). Furthermore, cellular proliferation is blocked when PKC _II expression is specifically inhibited by antisense oligonucleotides directed against PKC _II, demonstrating that it is required for leukemia cell proliferation (8).

In the present report, we identify and determine the function of the third major PKC isotype expressed in K562 cells, the atypical PKC . Using isotype-specific antibodies to atypical PKC  and PKC  we demonstrate that K562 cells express PKC , but no detectable PKC . We further demonstrate that PKC  regulates cellular susceptibility to drug-induced apoptosis. Our results indicate that PKC  plays an important role in leukemia cell survival and suggest that PKC  may be an attractive target for development of new chemotherapeutic agents in the treatment of leukemia.

EXPERIMENTAL PROCEDURES

Human erythroleukemia (K562) cells were maintained in suspension culture and induced to undergo cytostasis and megakaryocytic differentiation in response to PMA as described previously (8). Apoptosis was induced in cells plated at 4 × 10⁵ cells/ml in growth medium containing okadaic acid (OA; Alexis Biochemicals) or taxol at the concentrations and for the times indicated in the figure legends. Parental K562 cells and transfectants overexpressing PKC isotypes (see below) were monitored for proliferation rate by daily cell counting using a hemacytometer. Megakaryocytic differentiation was monitored by assessing expression of the megakaryocytic cell surface marker glycophorin IIIa/IIb (gpIIIa/IIb) by fluorescence activated cell sorting of cells stained with fluorescein-labeled monoclonal antibody to gpIIIa/IIb (Dako) as described previously (11). Cell viability was determined by trypan blue exclusion and was >95% in all untreated cultures.

Transfection and Expression of the Human PKC  and PKC  cDNAs in K562 Cells

The full-length human PKC  cDNA (kindly provided by Dr. T. Biden, Garvan Institute of Biomedical Research, Sydney, Australia) was excised from pAXNeoRX using (5) SalI and (3) BamHI and cloned into the (5) XhoI and (3) BamHI sites within the multiple cloning site of pREP4 and pREP10 to achieve sense and antisense orientations, respectively. The full-length human PKC  cDNA (kindly provided by Sphinx Pharmaceutical Co.) was excised from pBluescript using (5) SpeI and (3) KpnI and cloned into the (5) NheI and (3) KpnI sites of pREP10 in sense orientation. K562 cells were transfected with plasmid containing PKC  or PKC  cDNA or with control plasmid and transfectants were selected and maintained as described previously (8). Transfectants were screened for PKC isotype expression by immunoblotting as described below.

K562 cell transfectants were screened for PKC isotype expression by immunoblot analysis using isotype-specific PKC antibodies. Antibodies specific to PKC  and PKC _II were produced against peptides corresponding to the carboxyl terminus of each isotype and were characterized previously (8). Antibody to PKC / was produced against peptide corresponding to the carboxyl terminus (amino acids 576-592) of rat PKC  (12). Isotype-specific antibodies to PKC  and PKC  (Santa Cruz Biotechnology, Inc.) were generated against peptides corresponding to amino acids 2-21 within the divergent amino terminus of human PKC  and PKC , respectively.

Following drug treatments, cells were harvested and DNA isolated as described previously (13). Briefly, cells were lysed at 1 × 10⁶ cells/100 µl in lysis buffer containing 0.5% SDS and 0.2 mg/ml proteinase K and incubated at 50 °C for 3 h. RNase A was added to a concentration of 0.2 mg/ml and incubated at 37 °C overnight. Cell lysates were extracted with phenol/chloroform (1:1, v/v) and precipitated with ethanol. The dried pellet was resuspended in 10 mM Tris, pH 8.0, 1 mM EDTA and DNA concentration determined spectrophotometrically. 2 µg of DNA/lane was subjected to electrophoresis in 1.5% agarose and visualized by staining with ethidium bromide.

Cellular DNA morphology was assessed by staining with 4,6-diamidino-2-phenylindole (DAPI) and observation under phase-fluorescence optics as described previously (14). Images were captured using a Hamamatsu video camera and processed using Photoshop 5.0.

RESULTS AND DISCUSSION

K562 Cells Express the Atypical PKC Isoform, PKC

In previous studies, we determined that the two calcium-dependent PKC isotypes expressed in human K562 leukemia cells, PKC  and PKC _II, play distinct functional roles in K562 cell physiology (8). PKC  is involved in induction of cytostasis and differentiation, while PKC _II is required for cellular proliferation (8). In addition to these two classical PKC isotypes, we detected expression of an atypical PKC, which we identified as PKC  based on immunoreactivity with an antipeptide antibody against the carboxyl terminus of rat PKC  (8). Subsequent to our report, a second atypical PKC isotype was identified and cloned, termed PKC  in mouse and PKC  in human, that has high sequence homology to PKC  (15, 16). Comparison of the carboxyl-terminal regions of human PKC  and PKC  reveals extensive homology in the peptide region used to generate the PKC  antibody, raising the possibility that this antibody may cross-react with PKC . These findings led us to test the specificity of this antibody and to reexamine the expression of atypical PKC isotypes in K562 cells using isotype-specific antibodies to PKC  and PKC , generated against the divergent amino-terminal regions of these isotypes (Fig. 1).

K562 cells express PKC  but not PKC .

[View Larger Version of this Image (25K GIF file)]

As observed previously (8), the COOH-terminal anti-rat PKC  antibody recognizes a single band of molecular mass 74 kDa in K562 cell lysates (Fig. 1A, lane 1). In addition, this antibody recognizes human PKC  when overexpressed in K562 cells (Fig. 1A, lane 3) and purified recombinant baculovirus-expressed human PKC  (Fig. 1A, lane 4). However this antibody also recognized a 74-kDa band of increased intensity in lysates from K562 cells transfected with the human PKC  cDNA (Fig. 1A, lane 2), indicating that it also recognizes PKC . In contrast, an isotype-specific PKC  antibody recognizes PKC  overexpressed in K562 cells and recombinant purified PKC  (Fig. 1B, lanes 3 and 4) but fails to recognize a band in either parental K562 cells or cell transfectants overexpressing PKC  (Fig. 1B, lanes 1 and 2). Conversely, a PKC -specific antibody recognizes a 74-kDa band in K562 cells (Fig. 1C, lane 1) whose intensity is increased in PKC -overexpressing cells (Fig. 1C, lane 2). This PKC  antibody does not recognize PKC  when expressed in K562 cells or recombinant baculovirus-expressed human PKC  (Fig. 1C, lanes 3 and 4), demonstrating its specificity for PKC . From these results we conclude that the carboxyl-terminal rat PKC  antibody used in our previous studies (8) recognizes both human PKC  and PKC , whereas the amino-terminal PKC  and PKC  antibodies recognize their respective antigens specifically. Furthermore, our results demonstrate that K562 cells express PKC , but no detectable PKC .

Overexpression and Antisense Inhibition of Expression of PKC

Having identified the atypical PKC isotype expressed in K562 cells as PKC , we wished to evaluate its functional role. For this purpose, we stably transfected K562 cells with expression plasmids containing the full-length cDNA for human PKC  in either sense or antisense orientation or with cDNA for human PKC . Immunoblot analysis with the PKC -specific antibody demonstrated that these cell lines express either enhanced (Fig. 2A, lane 2) or reduced (Fig. 2A, lane 3) levels of PKC , respectively, when compared with control vector-transfected cells (Fig. 2A, lane 1) or those expressing PKC  (Fig. 2A, lane 4). Importantly, the changes in the expression level of PKC  or PKC  seen in these transfectants had no effect on the level of expression of PKC  and PKC _II (Fig. 2, B and C). Therefore, the changes in PKC  and PKC  expression do not lead to compensatory changes in the levels of other PKC isotypes.

Overexpression and antisense inhibition of expression of PKC  and PKC  in K562 cells.

[View Larger Version of this Image (50K GIF file)]

Effect of PKC  Expression on Cellular Proliferation and Susceptibility to PMA-induced Cytostasis

Having established cell lines selectively overexpressing or inhibited from expressing PKC , we next investigated the role of PKC  in K562 cell physiology. First, we assessed whether, like PKC _II (8), PKC  was involved in K562 cell proliferation (Fig. 3A). Neither overexpression or antisense inhibition of expression of PKC  had a significant effect on the proliferation rate of K562 cells, suggesting that PKC  functions in signaling pathways that are not rate-limiting for normal K562 cell proliferation. It is possible that the very low level of PKC  expressed in antisense PKC  cells is sufficient to support cellular proliferation; however, these results are clearly distinct from our previous findings with PKC _II, in which antisense inhibition of PKC _II expression led to dramatic inhibition of K562 cell proliferation (8).

Changes in cellular PKC  levels have no effect on K562 cell proliferation or differentiation.

[View Larger Version of this Image (24K GIF file)]

We next determined whether PKC  is important for K562 cell differentiation. Treatment of K562 cells with PMA induces cytostasis and megakaryocytic differentiation (8). We demonstrated previously that PMA-induced cytostasis is mediated, at least in part, by PKC  (8). PMA induces increased expression of PKC , and overexpression of PKC  confers a gene dose-dependent reduction in proliferative capacity and increased susceptibility to PMA-induced cytostasis (8, 10). The expression of PKC  is also induced by PMA treatment (8), suggesting that PKC  may also participate in PMA-induced cytostasis and differentiation. Therefore, we evaluated the effect of overexpression and antisense inhibition of expression of PKC  on PMA-induced cytostasis and megakaryocytic differentiation (Fig. 3B). Cells expressing increased or decreased levels of PKC  showed the same susceptibility to PMA-induced cytostasis as cells carrying control vector. Furthermore, expression of the megakaryocytic marker gpIIIa/IIb was unaffected by changes in PKC  expression (Fig. 3C). These results indicate that PKC does not play a determinant role in PMA-induced cytostasis or megakaryocytic differentiation.

Effect of PKC  Expression on Drug-induced Apoptosis

We next determined whether PKC  plays a role in K562 cell survival and drug-induced apoptosis. K562 cells are resistant to many apoptotic agents (17, 18) but can be induced to undergo apoptosis in response to the protein phosphatase inhibitor OA (19). Therefore, we determined whether K562 cells overexpressing or inhibited from expressing PKC  differ in their sensitivity to OA-induced apoptosis. First, parental K562 cells were treated with increasing concentrations of OA to establish conditions that induce apoptosis as measured by a standard interchromosomal DNA fragmentation assay (Fig. 4A). OA induces dose-dependent apoptosis when cells are treated for 24 h, with a maximal effective dose of about 45 nM. These results are consistent with the reported IC₅₀ of 10 nM for OA-induced cytotoxicity in K562 cells (19).

Changes in cellular PKC  levels affect susceptibility of K562 cells to drug-induced apoptosis.

[View Larger Version of this Image (83K GIF file)]

Based on these data, we chose to use 30 nM OA, a dose that gave demonstrable, but submaximal, DNA fragmentation, to assess the effect of selective overexpression and antisense inhibition of expression of PKC  on OA-induced apoptosis (Fig. 4B). Overexpression of PKC  leads to enhanced resistance to OA-induced DNA fragmentation when compared with control cells (Fig. 4B, compare lanes 4 and 2). In contrast, antisense inhibition of PKC  leads to increased sensitivity to OA-induced apoptosis (Fig. 4B, compare lanes 8 and 2). The protective effect of PKC  is isotype-selective, since overexpression of the related atypical PKC isotype, PKC  to similar levels gave no demonstrable protection from OA-induced apoptosis (Fig. 4B, compare lanes 6 and 2). Interestingly, overexpression of PKC  also protects K562 cells from apoptosis induced by taxol (Fig. 4C, lane 4), whereas inhibition of expression of PKC  enhances taxol-induced apoptosis (Fig. 4C, lane 6). These results demonstrate that the protective effects of PKC  are not specific for OA-induced apoptosis.

As an independent measure of apoptosis, we next examined nuclear DNA morphology of control and OA-treated cells. Cells undergoing apoptosis exhibit apoptotic nuclear DNA morphology characterized by highly condensed chromosomal masses, nuclear swelling, and formation of apoptotic bodies (20, 21). Therefore, control cells and those expressing increased or decreased levels of PKC  were treated with 30 nM OA and nuclear morphology was visualized using the fluorescent DNA dye DAPI (Fig. 4D). The nuclear morphology of cells expressing enhanced or reduced levels of PKC  was identical to control cells in the absence of OA, indicating that changes in PKC  expression have no direct effect on DNA morphology (data not shown). However, control and PKC -overexpressing cells treated with 30 nM OA (Fig. 4D, panels 2 and 3, respectively) displayed mitotic chromosomal condensation, consistent with the known mitotic arrest induced by OA in K562 cells (22). In contrast, cells expressing reduced levels of PKC  exhibit a more pronounced apoptotic effect on DNA morphology, including the presence of apoptotic bodies (Fig. 4D, panel 4, arrowhead). These results are consistent with the DNA fragmentation data (Fig. 3), and support the conclusion that the level of expression of PKC  is a key determinant of cellular susceptibility to drug-induced apoptosis.

PKC  is a member of the atypical PKC subfamily. These enzymes lack a calcium-binding domain and contain only one cysteine-rich zinc finger (15) and, as such, are not calcium-dependent and are not activated by phorbol esters or diacylglycerol (15). Relatively little is known about the in vivo activators and cofactor requirements of the atypical PKCs. Recent studies indicate that the atypical PKCs can be activated by phosphatidylinositol phosphates, specifically those phosphorylated in the 3 position by phosphatidylinositol 3-kinase (23), and by ceramide generated by sphingomyelinase (24). These second messengers have been linked to a number of cellular functions including mitogenesis and cell survival. Early reports suggested that PKC / is critical for Xenopus maturation and cellular proliferation in mouse fibroblasts (25, 26). Other studies have placed PKC / downstream of epidermal growth factor and platelet-derived growth factor in the activation of AP-1 (27, 28). In addition, recent reports suggest a role for PKC / in cell survival and UV irradiation-induced signal transduction (29, 30). Our present data demonstrate that PKC  plays a critical role in human leukemia cell survival, particularly in the presence of apoptotic stimuli, and indicate that the atypical PKC isotypes may be involved in cell survival in response to many types of cellular stress. Further studies will be required to elucidate the specific pathways in which PKC  participates to mediate its anti-apoptotic effects. Future studies will focus on the role of cellular phospholipids, including phosphatidylinositol phosphates and ceramides, and of the recently identified atypical PKC binding proteins (30-32), in regulating PKC  function and human leukemia survival. The recent finding that expression of the atypical PKC-binding protein, prostate apoptosis regulator 4 (PAR-4), is induced during apoptosis (33) provides an intriguing potential mechanism by which atypical PKC activity, and thereby cell survival, may be regulated.