Cancer Therapy Vol 1, 275-281, 2003.

Protein kinase C-d and its downstream effectors as potential targets for cancer therapy

Review Article

Jihua Liu, David Durrant and Ray M. Lee*

Huntsman Cancer Institute and Section of Oncology, Department of Internal Medicine, University of Utah, 2000 Circle of Hope, Salt Lake City, UT, 84112

__________________________________________________________________________________

*Correspondence: Ray M. Lee; Phone: 801-585-0611; Fax: 801-585-0900; e-mail: ray.lee@hci.utah.edu

Key Words: Protein kinase C-d, cancer therapy, apoptosis

Abbreviations: Protein kinase C (PKC); nerve-growth factor, (NGF); Cytarabine, (ara-C); stress-activator protein kinase, (SAPK); mitogen-activated protein kinase, (MAPK); phospholipid scramblase 3, (PLS3); phospholipid scramblase 1, (PLS1); non-small cell lung cancer, (NSCLC); phosphatidylserine, (PS); chronic myelocytic leukemia, (CML)

Received: 23 October 2003; Revised: 4 December 2003;

Accepted: 4 December 2003; electronically published: December 2003

Summary

The Protein kinase C (PKC) family of serine/threonine protein kinases are important in many biologic processes. The PKC-d isoform is particularly involved in the regulation of cell death. During programmed cell death, PKC-d translocates from the cytoplasm to the nucleus, mitochondria, Golgi and plasma membrane. Several key substrates of PKC-d in the nucleus and mitochondria have been identified, and the linkage of these PKC-d targets to regulation of DNA damage checkpoints and mitochondrial apoptosis helps to elucidate the mechanism of PKC-d-induced apoptosis. Here we review the apoptotic effects of PKC-d and its substrates in various organelles, and discuss the possibility of using these targets to develop novel approaches for cancer therapy.

I. Introduction

In the post genomic era, the trend for developing novel cancer therapy is based on identification of molecular targets. This kind of mechanism-based drug development requires a comprehensive understanding of the function of molecular targets, their upstream activators and downstream effectors. In addition to utilizing kinases or other signal transduction mediators as targets, members of apoptotic pathways also serve as potentially useful targets due to the possibility of directly inducing cell death and minimizing the development of drug resistance. Here we review the biology of the apoptotic pathway involving PKC-d and the possibility of using PKC-d and its downstream effectors as potential targets for novel cancer therapy.

II. Biological functions of the protein kinase C family

Members of the Protein kinase C (PKC) family play diverse roles in many biological processes (Parker, 1997). Isoforms of the PKC family can be divided into three groups based on their interaction with and dependence on calcium and diacylglycerol. The classic PKCs, including isoforms a, bI, bII and g, require both calcium and diacylglycerol for activation. The novel PKCs, including d, e, h, and q, are independent of calcium but require diacylglycerol for activation. The atypical PKCs, including l and z, are independent of both calcium and diacylglycerol. Each isoform plays different roles in cell growth, proliferation, differentiation or apoptosis (Ohno, 1997). Due to their important roles in many different cancers, PKCs could be potential targets for developing novel cancer therapies. For example, treatments utilizing antisense oligonucleotides directed against PKC-a are in clinical development for several cancers (Wang et al, 1999; Roychowdhury and Lahn, 2003). Bryostatin, a compound that inhibits PKC-a, has gone through phase I clinical trials (Hofmann, 2001; Marshall et al, 2002; Swannie and Kaye, 2002). However, these two approaches target PKC-a, the most well-characterized isoform. Utilization of other PKC isoforms has not yet been realized.

III. Protein kinase C isoforms in apoptosis

The balance of survival and apoptotic signals determines cancer cell death, the ultimate goal of cancer therapy (Cory and Adams, 1998; Reed, 1999; Adams and Cory, 2001; Green and Evan, 2002). Although therapeutic approaches that block survival signals may tip the balance toward cell death, blocking such survival signals does not necessarily mean that cancer cells will automatically undergo apoptosis. In contrast, induction of apoptosis through modulation of key factors in the apoptotic pathway would have a direct and dominant effect on cells. Thus direct activation of the cell death response may be a better approach for cancer therapies. In this review, we focus on the PKC isoform best characterized in triggering apoptosis, PKC-d (Brodie and Blumberg, 2003). Compared with PKC-a, which provides a survival and proliferation signal, PKC-d provides a more direct target to enhance apoptosis (Mandil et al, 2001).

Involvement of PKC-d in apoptosis was first demonstrated by activation of PKC-d in cells treated with a variety of apoptotic stimuli, including H₂O₂ (Konishi et al, 2001; Majumder et al, 2001), TNF-a (Emoto et al, 1995), the Fas ligand (Scheel-Toellner et al, 1999), UV and g irradiation (Denning et al, 1998; Yuan et al, 1998), and etoposide treatment (Reyland et al, 1999; Blass et al, 2002). Inhibition of PKC-d activity by a PKC-d-specific inhibitor, rottlerin, or by a dominant negative mutant resulted in suppression of the apoptotic response (Li et al, 1999; Majumder et al, 2000). Another important clue was shown in PKC-d-deficient mice, which had an increased B cell population and formed numerous germinal centers in the absence of stimulation (Mecklenbrauker et al, 2002; Miyamoto et al, 2002). The observed abnormal B cell proliferation was associated with enhanced autoimmunity due to the persistence of self antigen-recognizing B cells that failed to undergo apoptosis during positive selection. Analysis of these knockout mice thus established a role for PKC-d in controlling B-cell apoptosis in the regulation of B cell tolerance.

IV. Regulation of the activity of PKC-d

Regulation of PKC-d activity is mediated by at least three mechanisms. The regulatory C1 domain has an inhibitory effect on the catalytic domain found at the carboxyl terminus (Ohno, 1997; Parker, 1997). One way to release inhibition is by interaction of diacylglycerol or phorbol esters with the C1 domain, which triggers a conformational change. A second mechanism is mediated by cleavage of the catalytic domain from the C1 regulatory domain, which is achieved during apoptosis by activated caspase 3 (Ghayur et al, 1996; Denning et al, 2002). Tyrosine phosphorylation of PKC-d at Tyr64 and 187 is essential for the cleavage and the apoptotic effect of PKC-d (Blass et al, 2002). Tyr311 phosphorylation by Lck kinase after H₂O₂ treatment enhances basal PKC-d activity and elevates its maximal activity in the presence of diacylglycerol (Konishi et al, 2001). Finally, activated PKC-d undergoes ubiquitination and degradation through the proteasome pathway, which prevents a persistent effect of PKC-d (Lu et al, 1998).

V. Translocation of PKC-d during apoptotic responses

Depending on the cell types and the apoptotic stimuli, PKC-d has been reported to translocate to nearly all subcellular organelles, including nuclei, mitochondria, the Golgi complex, endoplasmic reticulum (ER) and the plasma membrane (Brodie and Blumberg, 2003; Roychowdhury and Lahn, 2003). At each subcellular organelle, PKC-d phosphorylates different substrates, inducing various responses that eventually lead to cell death. Identification of the substrates is critical to understanding the mechanism of PKC-d, but it has been very challenging to identify physiologic substrates in each organelle. We define three criteria that must be met to convincingly claim any protein as a physiologic substrate of PKC-d; (1) there must be evidence that PKC-d phosphorylates the protein, (2) there should be evidence for their interaction, and (3) most importantly, deletion or inactivation of the substrate must lead to at least a partial loss of PKC-d-induced response. Here we review several known substrates of PKC-d with a goal of connecting these substrates to the apoptotic pathway so that we can understand how PKC-d induces apoptosis.

VI. PKC-d substrates in the nucleus

Translocation of PKC-d to the nucleus has been established in T cells and C6 glioma cells (Scheel-Toellner et al, 1999; Blass et al, 2002). A putative nuclear localization signal has been identified at the carboxyl terminus of the catalytic domain of PKC-d (DeVries et al, 2002). Previously, nucleolin, which is required for nerve-growth factor (NGF)-induced differentiation of pheochromocytoma cells PC12, was identified as a substrate of PKC-z (Zhou et al, 1997). However, neither PKC-a nor PKC-d can phosphorylate nucleolin, and nucleolin is not involved in the apoptotic response.

Recently, Yoshida et al, (2003) reported that PKC-d is responsible for constitutive and DNA damage-induced phosphorylation of Rad9, a key factor involved in checkpoint regulation of the DNA damage response (al-Khodairy et al, 1994). The authors also demonstrated an interaction between PKC-d and Rad9, and showed that PKC-d phosphorylated Rad9 both in vitro and in cells treated with Cytarabine (ara-C) or g-irradiation (Yoshida et al, 2003). Nuclear Rad9 forms a critical heterotrimeric complex with Hus1 and Rad1, the 9-1-1 complex that is involved in DNA damage checkpoint control (Volkmer and Karnitz, 1999). PKC-d, which translocated to the nucleus during apoptosis, enhanced the phosphorylation of Rad9 and the formation of the Rad9-Hus1-Rad1 complex (Yoshida et al, 2003). Interestingly, Rad9 is also phosphorylated by ATM (Chen et al, 2001) and c-Abl (Yoshida et al, 2002). The latter kinase also interacts with PKC-d (Sun et al, 2000a, b). Using ATM siRNA to down-regulate the level of ATM, a diminished nuclear targeting of PKC-d was observed, suggesting that ATM is required for nuclear targeting of PKC-d and is functionally upstream of PKC-d (Yoshida et al, 2003). These studies provide a direct linkage between PKC-d and DNA damage-induced checkpoint regulation, and form the basis for future studies of the mechanism of PKC-d-induced apoptosis.

Another downstream effector of PKC-d in DNA damage response in cells treated with ara-C is stress-activator protein kinase (SAPK/JNK) (Yoshida et al, 2002) (Figure 1). DNA damage-induced SAPK/JNK activation was attenuated by rottlerin, a dominant negative mutant of PKC-d, and PKC-d siRNA. PKC-d did not directly phosphorylate SAPK/JNK, rather SAPK/JNK was indirectly phosphorylated through the mitogen-activated protein kinase (MAPK) pathway, PKC-d ® MEKK1 ® MKK7 ® SAPK/JNK (Yoshida et al, 2002). The finding that SAPK/JNK is a downstream effector of PKC-d provides another mechanism of PKC-d-induced apoptosis. Interestingly, SAPK/JNK was shown to be the substrate of PKC-b and to translocate to mitochondria after phosphorylation to induce cytochrome c release (Ito et al, 2001a).

VII. PKC-d substrates in the mitochondria

Translocation of PKC-d to mitochondria was shown in U937 myeloid leukemia cells and keratinocytes (Li et al, 1999; Majumder et al, 2000). The translocation can be induced by phorbol ester (Denning et al, 1998) and the oxidative stress (Majumder et al, 2001). With UV irradiation, mitochondrial targeted PKC-d was cleaved by caspase 3 to generate the active catalytic fragment of PKC-d (Denning et al, 2002). One known substrate of PKC-d is c-Abl kinase (Sun et al, 2000a). It has been demonstrated that PKC-d interacts with c-Abl, and that the phosphorylation of c-Abl results in activation of c-Abl kinase. Cells treated with H₂O₂ had an increase in c-Abl activity, which was attenuated by the PKC-d inhibitor, rottlerin, and by overexpression of the regulatory domain of PKC-d (Sun et al, 2000a). In the unstimulated condition, c-Abl localized to the nucleus, ER and cytoplasm. On ER stress caused by calcium ionophore A23187, brefeldin A or tunicamycin treatment, c-Abl translocated to mitochondria (Ito et al, 2001b).

The second mitochondrial target of PKC-d is the phospholipid scramblase 3 (PLS3) (Liu et al, 2003a), a member of the scramblase family that is responsible for bidirectional movement of phospholipids in the lipid bilayer. Unlike PLS1, which is localized in the plasma membrane (Zhou et al, 1997; Sims and Wiedmer, 2001), PLS3 is found exclusively in the mitochondria (Liu et al, 2003a, b). The function of PLS3 is currently unclear, but is likely involved in translocation of cardiolipin from the mitochondrial inner membrane to the outer membrane during apoptosis. Mitochondria with expression of an inactive mutant of PLS3 have a low level of cardiolipin and poor respiration (Liu et al, 2003b). They also display a unique morphology, being larger in size, fewer in number, and with tightly packed cristae, consistent with the notion that PLS3 moves phospholipids from the inner membrane to the outer membrane (Liu et al, 2003b).

Figure 1. Diagram of PKC-d and its downstream effectors in DNA damage-induced apoptosis. PKC-d is localized in the cytoplasm before the induction of apoptosis. When DNA is damaged by apoptotic stimuli, PKC-d translocates to the nucleus, where it phosphorylates Rad9 and c-Abl and activates MEKK1. PKC-d also translocates to mitochondria, where it phosphorylates PLS3 and c-Abl, and to the Golgi, where it phosphorylates c-Abl and SAPK/JNK. In the plasma membrane, PKC-d phosphorylates PLS1. The pathway from DNA damage to checkpoint activation and apoptosis is outlined.

Cardiolipin translocation is directly tied to the sensitivity of Cardiolipin translocation is directly tied to the sensitivity of mitochondria to tBid-induced cytochrome c release. Because tBid targeting to the mitochondria is mediated by cardiolipin (Lutter et al, 2000), the translocation of cardiolipin from the inner membrane to outer membrane facilitates the recruitment of tBid (Liu et al, 2003b). This idea was confirmed by the finding that mitochondria overexpressing PLS3 were more sensitive to tBid-induced cytochrome c release, whereas those expressing inactive mutant PLS3 were more resistant (Liu et al, 2003b).

PLS3 fulfills the three criteria we defined for a physiological substrate of PKC-d. PLS3 can interact with PKC-d and be phosphorylated by PKC-d in vitro (Liu et al, 2003a). HeLa cells expressing PLS3 become more positive in TUNEL studies when they were treated with the phorbol ester, PMA. Expression of mitochondria-targeted PKC-d in cells resulted in apoptosis, and overexpression of PLS3 enhanced this effect. In contrast, overexpression of the inactive PLS3 mutant did not generate this response (Liu et al, 2003a). These data support the view that PLS3 is a mitochondrial target of PKC-d-induced apoptosis.

VIII. PKC-d substrates in the plasma Another member of PLS family, PLS1, has been shown to be a target of PKC-d in the plasma membrane (Frasch et al, 2000). PLS1 is phosphorylated by PKC-d at a PKC phosphorylation consensus site, Thr161. Co-expression of PKC-d and PLS1 significantly increased the activity of scramblase following PMA treatment (Frasch et al, 2000). In contrast, co-expression of PKC-d and a T161A mutant of PLS1 showed no increase in scramblase activity, indicating that phosphorylation of Thr161 by PKC-d is important for scramblase function (Frasch et al, 2000). In addition, PLS1 can be phosphorylated by c-Abl, a kinase known to interact with PKC-d in other organelles (Sun et al, 2001).

Although there is solid evidence that PLS1 is activated during apoptosis (Zhao et al, 1998 ; Frasch et al, 2000), a direct link between PLS1 and apoptosis has not been fully established. During apoptosis, phosphatidylserine (PS) translocates from the inner leaflet to the outer leaflet of the plasma membrane. The regulation of transbilayer movement of phospholipids is controlled by at least three enzymes. One is aminophospholipid translocase, or flippase, which moves phospholipids inwards. One is the phospholipid scramblase (PLS1) that moves phospholipids bidirectionally. The third is a less well-characterized outward-directed floppase (Bevers et al, 1999). Probably due to the complexity of the regulation of phospholipid topology in lipid bilayers, cells from mice with homozygous for a deletion of PLS1 still maintain their ability to translocate PS to the surface (Zhou et al, 2002). This could presumably be due to compensation by aminophospholipid translocase activity. Therefore the mechanism of surface translocation of PS remains unclear. It has been hypothesized that apoptosis is associated with inactivation of aminophospholipid translocase and activation of the scramblase (Bevers et al, 1998; 1999), but definitive proof of this hypothesis has not yet been materialized.

A second plasma membrane target of PKC-d is Fyn kinase, found in the plasma membrane of platelets (Crosby and Poole, 2003)