Identification of signaling pathways and proteins involved in melanoma progression and drug resistance has the potential to lead to the development of novel, more effective targeted therapies. Two of these pathways are the PI3K (phosphoinositide-3-kinase) and MAP (mitogen-activated protein) kinase signaling cascades involved in melanoma development by relaying extra-cellular signals to regulate diverse cellular processes including proliferation, survival, invasion, and angiogenesis (Chudnovsky et al., 2005; Hsu et al., 2002; Smalley and Herlyn, 2005). It is generally accepted in the melanoma research and clinical communities that therapeutic agents targeting these two pathways will be effective for the long-term treatment of advanced-stage patients but which members of the cascades to target remain to be established. This review provides an overview of the PI3 kinase pathway in melanoma development detailing mechanisms leading to deregulation of PTEN and Akt3 in this cascade, cellular processes regulated by these proteins and therapeutic implications of targeting these proteins to treat this deadly disease. Finally, key issues that remain to be answered concerning this important signaling cascade are discussed.
Introduction to PTEN as a cancer suppressor gene
The phosphatase and tensin homologue deleted from chromosome 10 (PTEN) gene, which is also known as mutated in multiple advanced cancers (MMAC1) and transforming growth factor beta (TGF-β) regulated and epithelial cell-enriched phosphatase (TEP1), is an important tumor suppressor gene located at 10q23–24 (Robertson et al., 1998a; Wu et al., 2003). PTEN is a unique 55 kDa dual specificity phosphatase dephosphorylating phosphoserine and phosphotyrosine residues in proteins (Lee et al., 1999; Maehama and Dixon, 1999; Waite and Eng, 2002). It is also a lipid phosphatase hydrolyzing the secondary messenger phosphatidylinositol 3,4,5-trisphosphate (PIP3) (Gericke et al., 2006; Maehama and Dixon, 1999). Structurally, it is the N-terminal phosphatase domain of PTEN from amino acids 22–185 that catalyzes the hydrolysis of phospholipids. The central C2 lipid binding domain from amino acids 190–351 and C-terminal PSD95, DlgA, Zo-1 (PDZ) ligand sequence from amino acids 401–403 are associated with binding to the lipid (Gericke et al., 2006; Miller et al., 2002; Tamguney and Stokoe, 2007). PTEN also contains the characteristic protein phosphatase signature sequence “HCKAGKGR” spanning amino acids 123–130 (Figure 1A) (Maehama et al., 2004). This signature sequence also contains two lysine residues that are key for establishing interactions with negatively charged PIPs (Maehama et al., 2004).
Although, PTEN exhibits both protein and lipid phosphatase activities, it is the lipid phosphatase activity that is generally thought to be the major function of this phosphatase (Maehama and Dixon, 1999; Waite and Eng, 2002). The PTEN active site is structurally designed to facilitate electrostatic interactions with PIP3 for binding and hydrolysis (Lee et al., 1999). As a protein phosphatase, PTEN can dephosphorylate focal adhesion kinase, regulating cell-to-cell adhesion and the Src homologous and collagen gene (an adaptor protein) to disrupt cell spreading and migration. PTEN can also suppress MAPK signaling activity to modulate cell-to-cell adhesion and migration (Li and Ross, 2007; Waite and Eng, 2002; Wu et al., 2003). In its role as a lipid phosphatase, PTEN removes a phosphate group at the third position from cellular PIP3 and phosphatidylinositol 3,4-bisphosphate (PIP2) thereby decreasing the activities of several proteins or pathways whose activities are regulated by this lipid thereby modulating cell proliferation, survival, and apoptosis (Mikhail et al., 2005; Wu et al., 2003).
PTEN-null embryonic fibroblasts have elevated levels of PIP3 leading to constitutive Akt activity, which demonstrates that PTEN regulates the Akt pathway (Dahia et al., 1999; Wu et al., 1998a). PTEN is essential for murine embryonic development beyond day 7.5 (Di Cristofano et al., 1998) and loss of a single PTEN allele leads to hyperplasia and dysplasia in the skin, gastrointestinal tract, and prostate, as well as tumor formation (Di Cristofano et al., 1998). PTEN protein is also absent or functionally inactivated in many different tumor types causing increased Akt activation (Dahia et al., 1999; Wu et al., 1998b). PTEN expression in cells can suppress colony formation, growth in soft agar, and tumor formation in nude mice (Cheney et al., 1998).
Regulation of PTEN activity
PTEN expression and activity are regulated at transcriptional and post-translational levels (Gericke et al., 2006; Mirmohammadsadegh et al., 2006; Robertson et al., 1998a). Alterations in PTEN promoter activity have been shown to modulate cellular protein levels in response to positive and negative regulators (Gericke et al., 2006; Maehama et al., 2004). Some of the positive regulators of PTEN promoter activity include EGR-1 (early growth regulated transcription factor), PPARγ (peroxisome proliferator-activated receptors) and p53 (Gericke et al., 2006). Negative regulators such as mitogen-activated protein kinase-4 (MKK-4), NFκB (nuclear factor kappa-light-chain-enhancer of activated B-cells), TGF-β, and c-JUN have been shown to inhibit PTEN protein expression in several cancers (Gericke et al., 2006; Lopez-Bergami et al., 2008; Wu et al., 2003). Although, these pathways have been demonstrated to regulate melanoma development, detailed mechanisms by which they inhibit PTEN activity remain to be established.
PTEN activity is also modulated by interaction with lipids and proteins. Recently, protein interacting with the C-tail-1 (PICT-1) has been shown to interact with the C-tail of PTEN thereby regulating its turnover (Gericke et al., 2006; Tamguney and Stokoe, 2007). SiRNA-mediated down regulation of PICT-1 protein levels led to destabilization of PTEN and decreased protein levels. Aberrant PICT-1 mRNA levels have been reported in human gliomas, neuroblastomas and ovarian carcinomas (Tamguney and Stokoe, 2007). However, its expression and activity patterns in melanomas are yet to be studied.
Post-translational modifications such as phosphorylation, oxidation, acetylation, and ubiquitination can also regulate the activity of this phosphatase (Figure 1A) (Gericke et al., 2006; Miller et al., 2002; Tamguney and Stokoe, 2007). The oxidation of cysteine residues C71 and C124 by reactive oxygen species inhibits the phosphatase activity of PTEN. Acetylation is another mechanism by which PTEN activity is negatively regulated in many cancer cell types. The histone acetyltransferase p300/CBP-associated factor (PCAF) acetylates lysine residues 125 and 128 (located within the catalytic cleft) of PTEN thereby decreasing affinity towards PIP3’s. PCAF regulates melanoma development by acetylation, thereby stabilizing NFκB in the nucleus (Ueda et al., 2007). Regulation of PTEN activity by PCAF in melanomas warrants investigation.
Phosphorylation of the C-terminal tail of PTEN has been found to regulate PTEN activity. The C-terminal tail of PTEN contains multiple serine and threonine residues (S362, T366, S370, S380, T382, T383, and S385), which can be phosphorylated (Figure 1A) (Miller et al., 2002; Tamguney and Stokoe, 2007). Amino acids S370 and S385 are phosphorylated by casein kinase-2 (CK2). Recently, other kinases such as LKB1, Src, glycogen synthase kinase GSK3β have also been shown to play a role in PTEN phosphorylation (Gericke et al., 2006; Tamguney and Stokoe, 2007). Phosphorylated PTEN is stable but less active compared with unphosphorylated PTEN, which is less stable (Dahia, 2000; Gericke et al., 2006). C-tail mutants of PTEN, including deletions and missense mutations, tend to be short lived. A deletion of the C-terminal 18 amino acids has been shown to completely destabilize PTEN protein. Furthermore, amino acid substitutions at L345 or T348 to L345Q and T348I, respectively, produce highly unstable proteins, contributing to ∼20% of known tumor-associated PTEN mutations (Dahia, 2000; Maehama et al., 2004; Tamguney and Stokoe, 2007). It is uncertain whether these occur in melanoma.
The subcellular distribution of PTEN is a further key to its activity and may play a role in cancer development (Gericke et al., 2006; Wu et al., 2003). C-tail phosphorylation can influence translocation of PTEN to the cell membrane. Kinases such as CK2 and GSK3β have been shown to phosphorylate PTEN thereby decreasing its association with the cell membrane (Al-Khouri et al., 2005; Gericke et al., 2006). Cytosolic PTEN is inactive but highly stable. PTEN has intrinsic membrane translocation signals in the phosphatase and C2 domains. Mutations in these regions impair its ability to translocate to the plasma membrane. Additionally, PTEN localization and stability were regulated by interaction with other proteins (Dahia, 2000; Gericke et al., 2006; Tamguney and Stokoe, 2007). For example, monoubiquitination of PTEN promotes nuclear localization whereas polyubiquitination causes the protein to remain in the cytosol. Ubiquitination of PTEN is catalyzed by neural precursor cell expressed, developmentally down-regulated 4 (NEDD4) and NEDD1 (Kim et al., 2008; Wang et al., 2007b). However, whether similar ubiquitination patterns occur in melanomas needs investigation.
Finally, genetic and epigenetic mechanisms such as promoter methylation, mutations, chromosomal deletions, and expression of endogenous microRNAs have also been shown to modulate the expression and activity of this important phosphatase (Gericke et al., 2006; Waite and Eng, 2002). Genetic mechanisms regulating PTEN expression are: (i) loss of the whole chromosome, (ii) homozygous deletion, (iii) frameshift or point mutations, and (iv) inframe deletion and truncation (Li and Ross, 2007; Yin and Shen, 2008). Loss of chromosome 10, which contains the PTEN gene, has been reported in several cancers, including ∼30–60% of non-inherited melanomas (Stahl et al., 2003). Furthermore, mutations in the PTEN gene have been shown to be associated with ∼28% of glioblastomas, ∼34% endometrial carcinomas, ∼11.8% prostate cancer and ∼12% melanomas (Yin and Shen, 2008). Mechanistically, loss or inactivation of PTEN leads to Akt activation, which upon induction leads to the phosphorylation and monoubiquitination of DNA damage checkpoint kinase (Chk1), causing genomic instability, double stranded DNA breaks and ultimately, cancer development (Yin and Shen, 2008). Epigenetic modifications of PTEN arise by intrinsic (development, differentiation and cancer) or extrinsic (environmental) factors and can be reversed by chemical or environmental treatment modalities. CpG-rich regions of DNA, also known as CpG islands, which are distributed throughout the whole genome, have been shown to undergo methylation in several cancers including melanoma (Costello et al., 2000; Furuta et al., 2004; Wu et al., 2003). DNA methylation is a well-studied epigenetic mechanism for silencing of PTEN in melanoma (Furuta et al., 2004; Liu et al., 2008; Mirmohammadsadegh et al., 2006).
MicroRNAs are small non-coding RNAs of ∼22 nucleotides long (Wang and Lee, 2009). Recent studies have shown that microRNAs can negatively regulate gene expression at the post-transcriptional and/or translational level in different cell types (Wang and Lee, 2009). MicroRNAs targeting PTEN gene expression have been identified in several cancers (Huang et al., 2009; Pezzolesi et al., 2008; Yang et al., 2008). miR-214 has been shown to regulate PTEN translation by binding to its 3′-UTR, which in turn leads to Akt activation and chemoresistance in human ovarian cancers (Yang et al., 2008). In a different study, miR-21 has been demonstrated to cause hepato-cellular carcinoma by inhibiting PTEN expression (Meng et al., 2007). However, microRNAs regulating PTEN expression in melanoma have not yet been identified. As melanomas are highly resistant to chemotherapies and currently no treatment options are available for combating the advanced stage disease, identification of microRNAs regulating PTEN expression has significant potential to lead to novel therapeutic interventions.
The Akt family as oncogenes
The Akt kinase family consists of three protein kinases Akt1 (PKBα), Akt2 (PKBβ), and Akt3 (PKBγ) (Brazil et al., 2002; Nicholson and Anderson, 2002). Akt isoforms share >80% amino acid homology, containing three structurally distinct functional domains (Bellacosa et al., 2004; Brazil and Hemmings, 2001; Brazil et al., 2002, 2004; Datta et al., 1999; Nicholson and Anderson, 2002; Scheid and Woodgett, 2001, 2003; Testa and Bellacosa, 2001) (Akt3 shown in Figure 1B). The N-terminal pleckstrin homology domain spans amino acids 1–107, mediating protein–protein and protein–lipid interactions (Ferguson et al., 2000; Lietzke et al., 2000). The central catalytic domain contains a threonine residue (T305) whose phosphorylation is essential for activating this protein (Andjelkovic et al., 1995; Jones et al., 1991). The carboxy terminal tail region, known as the regulatory domain, contains a second phosphorylation site serine (S472). Phosphorylation of this second site is required for complete Akt3 activation (Figure 1B). In response to growth factors or other extracellular stimuli both sites become phosphorylated, resulting in the complete activation of Akt isoforms (Alessi et al., 1996). Other possible phosphorylation sites may also be important and research in this area continues (Alessi et al., 1996). Splice variants of Akt3 lacking serine 472 have been identified but the significance of this form remains unknown (Brodbeck et al., 2001; Konishi et al., 1995).
Regulation of Akt activity
Activity of the Akt isoforms is regulated through similar mechanisms, but activity of a particular isoform is cell type dependent (Bellacosa et al., 2004; Brazil and Hemmings, 2001; Brazil et al., 2002, 2004; Datta et al., 1999; Nicholson and Anderson, 2002; Scheid and Woodgett, 2001, 2003; Testa and Bellacosa, 2001). In normal cells, growth factors bind to a cell surface receptor or through G-protein-coupled receptors, activating PI3K as shown in Figure 2 (Burgering and Coffer, 1995; Franke et al., 1995). Active PI3K then phosphorylates PIP2 on the 3-OH group generating the second messenger PIP3 (Vanhaesebroeck and Alessi, 2000). PIP3 does not activate Akt directly, but rather promotes Akt translocation to the plasma membrane, where altered conformation of the protein mediated by PIP3 binding allows subsequent phosphorylation by phosphoinositide-dependent kinase-1 (PDK-1).
Akt3 activity is regulated by phosphorylation at threonine 305 and serine 472 (Figures 1B and 2). Phosphorylation at threonine 305, which is mediated by PDK1, partially activates Akt3, but full activation also requires phosphorylation of serine 472 (Alessi et al., 1996). Mechanism mediating serine 472 phosphorylation on Akt3, though controversial, is suspected to involve a second PDK, called PDK2; the identity of this factor remains unclear (Chan and Tsichlis, 2001). Integrin-linked kinase (ILK) and mammalian target of rapamycin complex-2 (mTORC2) can phosphorylate serine 472. However, studies have shown that ILK only acts as a facilitator; leaving mTORC2 as the PDK2 candidate (Alessi et al., 1996; Delcommenne et al., 1998). PDK2 activity of the mTOR complex has been recently established but in melanomas its role in maintaining Akt3 phosphorylation has not been studied (Feldman et al., 2009; Huang and Manning, 2009). Akt autophosphorylation might also play a role in the activation process (Bellacosa et al., 2004; Brazil and Hemmings, 2001; Brazil et al., 2002, 2004; Datta et al., 1999; Nicholson and Anderson, 2002; Scheid and Woodgett, 2001, 2003; Testa and Bellacosa, 2001). Thus, basal Akt activity in a cell depends upon a balance between positive signals from elevated PIP3 levels and negative signals leading to Akt dephosphorylation and inactivation. As this review focuses on Akt signaling in melanoma, subsequent sections discuss what is known about Akt3 deregulation in melanomas and the signals that regulate its activity.