Protein Kinase C as a Therapeutic Target

The PKC family is important in normal keratinocyte biology and in skin diseases, especially skin cancer (12). Keratinocyte apoptosis induced by UV radiation is a major protective mechanism from photocarcinogenesis. The constitutively active catalytic domain of PKCδ is an apoptotic effector generated by caspase-3 cleavage of full-length PKCδ in response to a wide variety of apoptotic stimuli including UV radiation (13). The activated PKCδ catalytic domain triggers the redistribution and activation of Bax that can directly induce cytochrome c release. PKCδ also activates upstream components of the death effector pathway to insure apoptosis. The cleavage and activation of PKCδ are critical components of UV-induced apoptosis in human keratinocytes. Inactivation of PKCδ can promote the survival of keratinocytes exposed to UV radiation (14). Recently, Aziz et al. (15) showed that PKCδ overexpression in transgenic mice failed to suppress the induction of squamous cell carcinoma developed by repeated exposure to UV radiation.

Transgenic mice that overexpress PKCε protein in basal epidermal cells and cells of the hair follicle develop papilloma-independent metastatic squamous cell carcinoma elicited by 7,12-dimethylbenz(a)anthracene-initiation and 12-O-tetradecanoylphorbol-13-acetate promotion. Chronic exposure of PKCε transgenic mice to UV radiation indicates that PKCε signals tumor necrosis factor-α release, which is linked to photosensitivity in these mice (16). The overexpression of PKCε in the epidermis of mice may lead to the induction of specific cytokines that perturb normal hematopoiesis in bone marrow, resulting in a granulocytic skew toward neutrophils and eosinophils (17, 18). Both the transgenic overexpression of PKCε in mouse epidermis and human squamous cell carcinoma indicate that high expression of PKCε in the epidermis may lead to a microenvironment suitable for the development of squamous cell carcinoma by a paracrine mechanism.

PKC activity is higher in actively proliferating colonic epithelial cells than in their quiescent counterparts, suggesting a role for PKC in proliferation. Colonic epithelial cells express multiple PKC isozymes that are differentially modulated during colon carcinogenesis, suggesting a direct connection between PKC and colon carcinogenesis (10, 11).

PKCδ expression is decreased in azoxymethane-induced rat and sporadic human colon tumors (19). Human CaCo-2 gastric cancer cells stably transfected with PKCδ showed that increased expression of PKCδ induced cellular differentiation and limited survival. Elevated expression of PKCβII is an early event in colon carcinogenesis. Transgenic expression of PKCβII in mouse colon led to hyperproliferation and increased susceptibility to colon carcinogenesis that correlated with decreased expression of transforming growth factor-β type II receptor. Yu et al. (20) showed that transfection with PKCβII induced the expression of cyclooxygenase-2 in rat intestinal epithelial cells in culture and in transgenic PKCβII mice. Transgenic mice expressing PKCβII in colonic epithelium show that high expression of PKCβII is both necessary and sufficient to confer susceptibility to azoxymethane-induced colon carcinogenesis (21). PKCβII regulated its own expression in rat intestinal epithelial cells in culture and in animals through a mitogen-activated protein/extracellular signal-regulated kinase kinase–dependent signaling pathway. In the colon epithelium, elevated expression of PKCβII seems to be procarcinogenic, acting through both cyclooxygenase-2 and transforming growth factor-β signaling pathways.

PKCι is implicated as a critical downstream effector of oncogenic Ras in the colonic epithelium (11). Using the synthetic ether lipid, 1-O-hexadecyl-2-O-methyl-rac-glycerol, which inhibits PKC activation by competing with the natural PKC activator, diacylglycerol, Liou et al. (22) showed that human tumor cell lines with oncogenic p21ras mutations were much more sensitive to PKC inhibition by 1-O-hexadecyl-2-O-methyl-rac-glycerol than tumor cell lines with normal K-ras alleles. In transgenic mice expressing constitutively active PKCι in the colon epithelium, PKCι activity was required for oncogenic Ras- and carcinogen-mediated colon carcinogenesis. PKC isoforms may act in concert to promote malignancy in the colon because transgenic expression of PKCβII in mouse colon induces hyperproliferation and increased susceptibility to colon cancer through pathways that include the activation of Ras, PKCι/Rac1, and Mek leading to invasion in intestinal epithelial cells (23). PKC phosphorylation promotes rapid dissociation of K-Ras from the plasma membrane and association with intracellular membranes, including the outer membrane of mitochondria, in which phosphor-K-Ras interacts with Bcl-XL (24). Interestingly, gastric adenocarcinoma SNU-1 cells respond to dietary phytoalexin resveratrol by loss of membrane-associated PKCδ and an increase in cytosolic PKCα prior to the up-regulation of antiproliferative and proapoptotic signals (25).

PKCα and PKCδ promote apoptosis in androgen-dependent prostate cancer cells (26). Stewart and O'Brian (27) found 12-O-tetradecanoylphorbol-13-acetate–induced PKCα activation in epidermal growth factor receptor transactivation and signaling to Erk1/2 activation in PC-3 cells. The PKC-selective kinase inhibitors, GF109203X and Go6983, blocked phorbol ester–induced epidermal growth factor receptor transactivation, indicating a requirement for PKC, and supporting PKCα inhibition as a therapeutic approach for prostate cancer.

PKCε expression in LNCaP cells was sufficient to confer a significant resistance to phorbol ester induction of Bax oligomerization suggesting the association of PKCε with Bax and the neutralization of apoptotic mitochondrial death-signaling pathways (28). Failure of hormonal therapy frequently involves an outgrowth of PKCε-positive cells in recurrent prostate cancer. Wu et al. (29) found a functional interplay among the integrin receptors, PKCε and protein kinase B/Akt, in recurrent CWR-R1 prostate cancer, suggesting that sustained activation of these β1 integrins induced an androgen independent–like behavior in these cells.

Atypical PKCι is critical in the establishment and maintenance of epithelial cell polarity and is genomically amplified and overexpressed in serous epithelial ovarian cancers (30). PKCι is markedly increased and mislocalized in all serous ovarian cancers. In nonserous ovarian cancers, increased PKCι is associated with decreased overall survival. Thus, PKCι may be an oncogene in ovarian cancer regulating epithelial cell polarity and proliferation. Compared with normal tissues, endometrial cancers also exhibit alterations in the expression profile or in the activities of the PKC family, and aberrant levels or activation of specific PKC isoforms are postulated to contribute to endometrial neoplasia and transformation. Lahn et al. (31) found that PKCα increased in breast cancers with low or negative estrogen receptor levels. In breast cancer cells, the indolocarbazole PKC inhibitor, Go6976, abrogated S and G₂ cell cycle arrest (32). In p53-mutated human SKOV3 ovarian cancer xenograft tumors, the administration of Go6976 and caffeine was able to restore the response to cisplatin (33).

Atypical PKCι is highly expressed in human non–small cell lung carcinoma (NSCLC) lines, whereas the closely related PKC isozyme, PKCζ, is undetectable in these cells (34). PKCι is a critical lung cancer gene that activates a Rac1 → Pak → Mek1,2 → Erk1,2 signaling pathway required for transformed growth. PKCι may be an oncogene in NSCLC because: (a) PKCι is overexpressed in most NSCLCs, (b) PKCι expression predicts poor survival in patients with NSCLC, (c) the PKCι gene is frequently amplified in NSCLC lines and primary NSCLC, (d) gene amplification drives PKCι expression in NSCLC lines and primary tumors, and (e) PKCι disruption signaling with a dominant-negative PKCι allele blocks the transformed growth of human NSCLC cells harboring PKCι gene amplification (34, 35). PKCι is implicated in carcinogen-induced survival of malignant cells (36). Stallings-Mann et al. (37) found that aurothioglucose and aurothiomalate were inhibitors of PKCι-Par6 interactions and slowed the growth of human A549 NSCLC xenograft tumors.

Clark et al. (38) found PKCδ to be antiapoptotic in NSCLC cells. Ding et al. (39) identified PKCε as critically involved in drug-resistance in lung cancer cells by preventing cell from undergoing apoptosis. PKCα and PKCε work cooperatively in lung cancer cellular response to various stimuli (40). In cocultures of normal pulmonary fibroblasts and NSCLC cells, PKCα and PKCε, but not PKCδ, were critical to stromelysin-3 (MMP-11) expression, indicating the involvement of both classical and novel PKC in the regulation of stromelysin-3 expression (41).

An antisense agent designed to inhibit the production of PKCα has undergone clinical trials in NSCLCs (42). In preclinical studies, the antisense oligonucleotide, LY900003 (ISIS3521), inhibited PKCα mRNA and protein expression and had antitumor activity. In early clinical studies, LY900003 had single-agent activity and was promising in combination, particularly in patients with NSCLC, which led to a phase III clinical trial of LY900003 in combination with either cisplatin and gemcitabine or carboplatin and paclitaxel. However, in the phase III setting, LY900003 did not add significantly to the chemotherapy regimens.

In rat hippocampal H19-7 cells and rat pheochromocytoma PC12 cells, the extracellular signal-regulated kinase activation by the neurogenic agents, fibroblast growth factor and nerve growth factor, is PKCδ-dependent, whereas extracellular signal-regulated kinase activation in response to epidermal growth factor is PKCδ-independent. Inhibition of PKCδ by antisense or rottlerin, blocked neurite outgrowth induced by fibroblast growth factor and nerve growth factor in PC12 cells and by activated Raf in H19-7 cells, indicating the involvement of PKCδ in the neurogenic effects of fibroblast growth factor, nerve growth factor, and Raf. Epidermal growth factor activates mitogen-activated protein kinase through a PKCζ-dependent pathway involving phosphatidylinositol-3-kinase and PDK1 in H19-7 cells (43). PKC regulates the activation of Raf-1 signaling from growth factors by phosphorylating the Raf kinase inhibitory protein, thereby relieving the inhibition of the Raf/mitogen-activated protein kinase signaling cascade (44).

Phorbol 12-myristate 13-acetate–stimulated activation of PKCε in U-251 glioblastoma cells resulted in the activation of both Akt and the mammalian target of rapamycin signaling pathways as well as an increase in cell proliferation. Increased PKCη expression correlated with increased phorbol 12-myristate 13-acetate–induced proliferation in U-1242 MG cells (45). In the glioblastoma cell lines (U-1242 MG and U-87 MG), phorbol 12-myristate 13-acetate–induced epidermal growth factor receptor phosphorylation was blocked by bisindolylmaleimide, a PKC inhibitor, and rottlerin, a PKCδ-specific inhibitor (46). Thus, phorbol 12-myristate 13-acetate transactivates the epidermal growth factor receptor and increases cell proliferation by activating the PKCδ/c-Src pathway in glioblastomas.

PKC is directly involved in modulating apoptosis hematopoietic neoplasms. PKCδ, PKCι, PKCμ, and PKCζ expression was detected in multiple myeloma cell lines (U266, RPMI-8226, and K620; ref. 47). Using immunohistochemistry, PKCδ expression was confirmed in plasma cells from 11 patients. Alkan et al. (48) investigated PKC isoform expressions in chronic lymphocytic leukemia cells from seven patients and found the expression of PKCβ, PKCγ, PKCδ, and PKCζ by Western blot in all of the samples.

Several PKC isoforms including α, βII, ε, ι, and η seem to be pro-oncogenic in the skin, colon, prostate, ovarian, breast, lung, and brain. PKCδ seems to have some possible tumor-inhibitory characteristics except in the brain, where it is more pro-oncogenic. Upstream and downstream pathways from PKCα and PKCδ are shown in Fig. 2. Many factors related to proliferation and invasion have been connected with PKCα; however, the pathways downstream from PKCδ form a less clear pattern. Several additional PKC isoforms have been identified in malignant tissues; however, no active role in the disease process has been elucidated thus far.

Small molecule approaches to inhibiting PKC have been directed toward the ATP-binding site or the diacylglycerol binding site. Varied levels of isoform selectivity can be shown for many naturally occurring and synthetic PKC inhibitors. Conformationally constrained analogues of diacylglycerol-lactones with approximately the same binding affinity for PKC as phorbol esters were explored (49, 50). Safingol, the l-threo enantiomer of dihydrosphingosine, acts as a competitive inhibitor at the diacylglycerol-binding domain of the enzyme. Balanol analogues with benzophenone subunits were inhibitors of PKC and PKA (51). The advent of staurosporine-like bisindolylmaleimide structures opened up the field of inhibitors such as CGP41251, Go-6850, Ro-31-8220, and UCN-01, at the ATP-binding site of the enzyme (52, 53). The National Cancer Institute 60-cell line identified the PKC inhibitor, UCN-01, 7-hydroxy-staurosporine. UCN-01 was shown to inhibit the growth of many tumor types in vitro and in vivo (54). UCN-01 has undergone phase I and II clinical trials alone and in combination regimens (55, 56).

Midostaurin (PKC 412). The N-benzoylated staurosporine analogue, midostaurin, was originally identified as a PKC inhibitor but has subsequently been found to inhibit both tyrosine kinase and serine/threonine kinases (57). Midostaurin is a potent radiation sensitizer both in cell culture and human tumor xenograft studies through a phosphoinositide-3-kinase/Akt pathway (58). In B16F10 murine melanoma, midostaurin was antimetastatic because of the inhibition of the invasive and/or platelet-aggregating activities of the melanoma cells (59).

In culture, midostaurin was more cytotoxic toward B cell chronic lymphocytic leukemia than toward normal B cells (60). Exposure to midostaurin sensitized human U1810 NSCLC cells to DNA damage by etoposide (61). Midostaurin suppressed Akt kinase activation and induced apoptosis in human multiple myeloma cells (62).

Midostaurin is a potent inhibitor of mutant FLT3 and of the abnormal fusion protein, ZNF198-fibroblast growth factor receptor 1, as well as mutant KIT tyrosine kinase activities (63, 64). Gleevec-resistant disease remains sensitive to midostaurin (65). In culture, exposure of human leukemia cells to rapamycin, along with midostaurin, resulted in synergistic inhibition of cells expressing midostaurin-sensitive or -resistant leukemogenic FLT3 mutants. Similarly, the exposure of human acute myelogenous leukemia cells with mutant FLT3 to 17-allylamino-demethoxygeldanamycin or histone deacetylase inhibitor, LAQ824, and midostaurin resulted in increased cell killing compared with either agent alone, and relatively increased cell killing compared with acute myelogenous leukemia cells expressing normal FLT3 (66, 67).

Midostaurin has undergone several phase I clinical trials as a PKC inhibitor and as a tyrosine kinase inhibitor. Phase I single agent studies showed that midostaurin can be safely administered by regular oral therapy. Phase I studies of midostaurin in combination with protracted continuous infusion of 5-fluorouracil in advanced solid tumors, in combination with paclitaxel and carboplatin, or gemcitabine and cisplatin in advanced NSCLC, established midostaurin doses that could be safely administered in these regimens (68, 69). Although phase II trials of midostaurin alone and along with imatinib in mutant FLT3 acute myeloid leukemia or gastrointestinal stromal tumors have shown activity, a mutation in the FLT3 tyrosine kinase domain which confers resistance to midostaurin has been identified (70).

Enzastaurin (LY317615). Enzastaurin is a potent inhibitor of PKCβ (71). Endogenous PKCβII mRNA levels in HCT116 cells were significantly reduced by enzastaurin, consistent with the effect on PKCβ promoter activity (72). Exposure to enzastaurin (600 nmol/L, 72 hours) profoundly inhibited the proliferation of vascular endothelial growth factor (VEGF; 20 ng/mL)–stimulated human umbilical vascular endothelial cells. When human SW2 small cell lung carcinoma cells were exposed to enzastaurin for 72 hours, a potency differential on the malignant cells versus human umbilical vascular endothelial cell was apparent. Cell culture studies indicate that exposure to enzastaurin can cause growth inhibition and apoptosis in human multiple myeloma cells, diffuse large cell lymphoma, and mantle cell lymphoma (73–76).

Enzastaurin at 10 mg/kg orally twice per day for 10 days post-surgical implant of VEGF-impregnated filters resulted in markedly decreased vascular growth to about one-half of the VEGF-stimulated controls, whereas 30 mg/kg decreased vascular growth to the level of surgical control (71). Enzastaurin at 30 mg/kg orally twice per day for 10 days post-surgical implantation of basic fibroblast growth factor resulted in decreased vascular growth to 26% of the basic fibroblast growth factor control. When nude mice bearing human tumor xenografts were treated with enzastaurin orally, twice daily for 2 weeks, the number of intratumoral vessels was decreased to one-half to one-quarter of the controls in treated animals (Fig. 3A; refs. 71, 77–80). Although some tumors responded to enzastaurin as an antiangiogenic agent, in no case was angiogenesis completely blocked as in the corneal micropocket neoangiogenesis model. The tumor growth delay did not correlate with intratumoral vessel decrease (Fig. 3B). In most tumor models, the tumor growth delay produced by enzastaurin as a single agent was not sufficient to predict single agent activity in the clinic. However, the combination regimens suggested high activity. VEGF plasma levels in mice bearing the human SW2 SCLC and Caki-1 renal cell carcinomas treated or untreated with enzastaurin were obtained every 3 days starting on day 7 post-implantation, throughout treatment, and after the termination of treatment. Plasma VEGF levels were similar between the treated and untreated groups up to day 20 (81, 82). Plasma VEGF levels in the control groups continued to increase throughout the study; however, even after termination of enzastaurin treatment, plasma VEGF levels in enzastaurin-treated mice were significantly decreased (83).

Simultaneous and sequential treatment regimens combining treatment with enzastaurin and 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) were studied in human T98G glioblastoma multiforme using tumor growth delay as the end point when the tumor was implanted s.c., and increase-in-life span as the end point when the tumor was implanted intracranially (80). For the simultaneous treatment regimen, enzastaurin was administered on days 4 to 18, and BCNU was administered on days 7 to 11 after tumor implantation. In this tumor model, treatment with enzastaurin produced tumor growth delays of 6.0 and 8.6 days at doses of 10 and 30 mg/kg, respectively (Fig. 4B). Administration of the combination of enzastaurin (30 mg/kg) with BCNU resulted in a tumor growth delay of 15.2 days.

Nude mice implanted with T98G (10⁴ cells) intracranially had a mean survival time of 37 days. Administration of enzastaurin (30 mg/kg) alone doubled the survival time, producing a mean survival time of 72 days (Fig. 4A). Administration of BCNU to mice bearing the intracranial tumor implant resulted in a mean survival time of 41 days. The simultaneous combination regimen of BCNU with enzastaurin (10 or 30 mg/kg) produced mean survival times of 90 and 102 days, respectively (80).

Enzastaurin efficacy in the sequential treatment regimen after completion of BCNU was examined over a dosage range from 3 to 30 mg/kg. Enzastaurin alone on days 12 to 30 produced tumor growth delays of 18.5 to 32.6 days, respectively, in the s.c. T98G tumor (Fig. 4D). BCNU treatment followed by enzastaurin (30 mg/kg) resulted in a tumor growth delay of 46.8 days. Delaying treatment with enzastaurin (3-30 mg/kg) from days 4 to 12 after intracranial T98G tumor implantation produced mean survival times of 57 to 72 days compared with 37 days for controls (Fig. 4C). BCNU treatment followed by enzastaurin (3-30 mg/kg) resulted in mean survival times of 63 to 74 days compared with 41 days for BCNU alone (80).

Phase I clinical studies for enzastaurin were completed in 2003 (84, 85). Monocytes collected from patients treated with enzastaurin showed decreased PKC activity after ex vivo stimulation, with phorbol ester providing a potential biomarker in a surrogate cell (86). Phase II studies have been conducted in patients with relapsed diffuse large B cell lymphoma and in patients with recurrent high-grade gliomas. Several patients with multiple relapsed diffuse large B cell lymphomas achieved prolonged periods of stable disease following enzastaurin treatment, although the objective tumor response rate was low (87). In patients with high-grade gliomas, enzastaurin was well tolerated and seemed to have promising antitumor activity in a significant percentage of highly pretreated patients (88, 89). In early combination studies, enzastaurin has proven safe and potentially efficacious along with capecitabine, pemetrexed, and gemcitabine and cisplatin (90–92). A phase III randomized, open label registration trial of enzastaurin in comparison with lomustine in glioblastoma has been initiated. A second phase III trial has been initiated in late-stage non–Hodgkin's lymphoma where enzastaurin will be evaluated as a maintenance therapy in patients with diffuse large B cell lymphoma who have achieved remission following first line therapy to prevent relapse.

In summary, among the kinases being targeted with small-molecule inhibitors, the PKC family is perhaps the most ubiquitously expressed and centrally involved in cellular metabolism. While the clinical trials progress, science will continue to elucidate the spectrum of kinases effected by these agents and useful therapeutics will emerge.