A tumor suppressor role for PP2A-B56α through negative regulation of c-Myc and other key oncoproteins

Abstract

Loss or inhibition of the serine/threonine protein phosphatase 2A (PP2A) has revealed a critical tumor suppressor function for PP2A. However, PP2A has also been shown to have important roles in cell cycle progression and survival. Therefore, PP2A is not a typical tumor suppressor. This is most likely due to the fact that PP2A represents a large number of different holoenzymes. Further understanding of PP2A function(s), and especially its tumor suppressor activity, will depend largely on our ability to determine specific targets for these different PP2A holoenzymes and to gain an understanding of how these targets confer tumor suppressor activity or contribute to cell cycle progression and cell survival. Recent work has identified c-Myc as a target of the PP2A holoenzyme, PP2A-B56α. This holoenzyme also negatively regulates β-catenin expression and modulates the anti-apoptotic activity of Bcl2, thus characterizing PP2A-B56α as a tumor suppressor PP2A holoenzyme. This review will focus on the role of PP2A-B56α in regulating c-Myc and will place this tumor suppressor activity of PP2A within the context of its other tumor suppressor functions. Finally, the mechanism(s) through which PP2A-B56α tumor suppressor activity may be lost in cancer will be discussed.

1 Introduction

A multitude of signaling pathways are utilized to govern virtually all aspects of cellular function thus enabling cells to respond to both their external and internal environment and to maintain “normal” cellular homeostasis. These signaling pathways are exquisitely regulated through a variety of mechanisms including the post-translational modification of proteins by reversible phosphorylation. Two key groups of proteins are involved in the reversible phosphorylation process, kinases that add phosphates to proteins and phosphatases that remove these phosphates from such targets. Interestingly, the number of published reports identifying and/or characterizing kinases exceeds those dealing with the activities of phosphatases by 15:1. Therefore, our understanding of phosphatases significantly lags behind our understanding of kinases.

Over the past 15 years the tumor suppressor function of certain phosphatases have begun to be appreciated. For example, mutation or loss of the tumor suppressor lipid phosphatase, PTEN occurs very frequently in a wide array of human cancers and is only surpassed by the frequency with which the function(s) of the classical p53 tumour suppressor are altered [1]. Additionally, the protein phosphatase, PP2A has received considerable attention as a result of the demonstration that it also possesses tumor suppressor activity [2–4]. However, a somewhat confusing aspect of PP2A function in cancer arises because PP2A also plays important roles in promoting cell cycle progression and cell survival and these are functions which are usually associated with tumor initiation and progression rather than tumor suppression [5–9]. These apparently contradictory functions of PP2A were initially demonstrated following the global inhibition of PP2A catalytic activity using okadaic acid. This can result in either increased or decreased levels of cellular transformation depending in large part upon the biological system being utilized [10–12]. These divergent functions of PP2A are most likely due to the large number of PP2A holoenzymes, all of which are sensitive to inhibition by this marine toxin. Consequently, to further understand the role of PP2A in cancer, the characterization of specific PP2A holoenzymes and their targets is critical. In this review we will discuss work focused on characterizing a subset of PP2A holoenzymes containing the regulatory B subunit B′/PR61/B56/PPP2R5 family members that generally have tumor suppressor activity. In particular, we will focus on the important role of the PP2A-B56α holoenzyme in negatively regulating several key oncoproteins, specifically Cellular Myc (c-Myc), as well as β-catenin and Bcl2. We will conclude with an exploration of the mechanisms by which this PP2A holoenzyme may become disrupted in human cancers.

2 PP2A is a heterotrimeric substrate specific serine/threonine protein phosphatase

Protein phosphatases are grouped into two general families based on their ability to either dephosphorylate Tyrosine or Serine/Threonine (S/T) residues. PP2A is a member of the S/T family of protein phosphatases, which also includes PP1, PP2B (also called calcineurin), PP2C, PP3, PP4, PP5, PP6, PP7, PP8, and possibly others, which are yet to be identified. Of the S/T family of protein phosphatases, PP2A is considered to be one of the most prevalent members constituting 30–50% of cellular S/T protein phosphatase activity depending upon cell and tissue type. PP2A is a heterotrimeric protein comprised of a structural A subunit, catalytic C subunit, and highly variable, regulatory B subunit (Fig. 1, top). The structural A subunit is found in two isoforms, α or β and serves as a scaffold protein comprised of 15 non-identical huntingtin-elongation-A subunits-TOR-like (HEAT) repeats that facilitate the binding of the B and C subunits [13, 14]. Similar to the A subunit, the C subunit also exists in two isoforms, α or β and is the catalytically active phosphatase subunit of PP2A holoenzymes. The regulatory B subunits fall into four unrelated families called B/B55/PPP2R2, B′/PR61/B56/PPP2R5 (referred to as B56 throughout the remainder of this review), B″, and B‴ (Fig. 1) that bind the A subunit in a mutually exclusive manner to form distinct PP2A holoenzymes. Importantly, the regulatory B subunits direct PP2A holoenzyme activity to specific substrates as well as acting to target PP2A holoenzymes to various subcellular locations. Altogether, there are 25 reported regulatory B subunits, which allows for ~100 different PP2A holoenzymes to be formed through the various combinations of A and C isoforms along with the different B subunits. However, it is unlikely that all PP2A holoenzymes are formed in all cell types as the regulatory B subunits display differential expression depending upon cell type, tissue type and developmental stage.

Fig. 1.

Schematic of heterotrimeric PP2A holoenzyme, which is comprised of a structural A subunit, catalytic C subunit and a variable regulatory B subunit from one of four families of regulatory B subunits. The B56/PPP2R5 family of regulatory B subunits are encoded by five genes and the γ and δ genes encode several splice variants. A Subunit Binding Domain 1 and 2 (ASBD1 and 2) are regions of higher amino acid sequence homology that facilitate the association of the B subunits with the A subunit. Known targets of specific B56 family members are shown with indication of suppression or activation by a particular PP2A-B56 holoenzyme

3 Summary of the B56 subunit family in tumorigenesis

Considerable research activity has been focused on the tumor suppressor activity of PP2A following the demonstration that global inhibition of PP2A can contribute to the transformation of primary human cells [15–19]. However, there is now growing evidence that PP2A also possesses potentially oncogenic roles [7, 20–22]. As mentioned above, much of the ambiguity regarding PP2A tumor suppressor verses oncogenic activity stems from the fact that PP2A represents a large number of distinct holoenzymes. Consequently, significant contributions to our understanding of PP2A function in tumorigenesis have resulted from the identification of specific PP2A holoenzyme/target protein interactions. Of particular interest for the purposes of this review are the B56 regulatory B subunits: α, β, γ, δ, and ε, which have been shown to have specific targets that are important in tumorigenesis (Fig. 1). Since their identification in 1995, most of the B56 subunits have been shown to have tumor suppressor activity, detailed in Fig. 1.

Of the B56 regulatory B subunits studied to date, B56α and B56γ are perhaps the best characterized in terms of both their cellular targets and tumor suppressor functions. We will discuss B56α targets later in more detail and in particular its role in negatively regulating c-Myc. There are four splice variants of B56γ (Fig. 1). Studies using anti-sense knockdown of multiple B56γ variants have shown that loss of B56γ expression contributes to the transformation of human kidney epithelial cells that also express Simian virus 40 (SV40)-Large-T-antigen, H-Ras, and hTERT [16]. It has also been shown in cultured human cells that both B56γ1 (variant 3) and B56γ3 (variant 1) can associate with and dephosphorylate p53 at T55, an event which stabilizes p53 in response to DNA damage [23]. Moreover, RNAi knockdown of B56γ led to a decrease in p53 stability, reduced expression of Bax, and reduced apoptosis. B56γ has also been shown to promote the degradation of the transcriptional co-activator, p300 in response to valproic-acid treatment of human cultured cells [24]. The biological importance of B56γ regulation of p300 is yet to be fully appreciated. However, p300 has been shown to promote cellular proliferation and conversely induce apoptosis depending upon cellular context through its roles as a transcriptional co-activator and its ability to acetylate p53, respectively [25]. In a somewhat contradictory role, B56γ could help to resolve the DNA damage response as it was shown to negatively regulate Chk2 activity in response to genotoxic stress in cultured human cells [26, 27]. In addition, a truncated form of B56γ1, called ΔB56γ1, was identified in a mouse melanoma cell line that demonstrated increased metastatic potential [28]. It was further demonstrated that paxillin is a direct target of full-length B56γ and expression of ΔB56γ1 could disrupt this regulation leading to increased paxillin phosphorylation and cell motility, suggesting that ΔB56γ1 functions as a dominant negative form of B56γ. Following up on this initial study, the same group demonstrated that ΔB56γ1 expression led to increased polyploidy in gamma-irradiated BL6 tumor cells [29]. More recent work has shown that ΔB56γ1 expression can disrupt the phosphorylation of Mdm2 in cultured cells in response to irradiation [30]. The phosphorylation of Mdm2 is an important event in activating p53 as phosphorylation of Mdm2 inhibits its ability to target p53 for degradation. Altogether, these studies highlight important roles of B56γ in activating p53 signaling and apoptosis, inhibiting proliferation and cell motility, and regulating genomic stability.

Two other B56 family members that have been characterized with specific targets are B56β and B56δ, with a single transcript reported for B56β and three variants reported for B56δ (Fig. 1). Recent work has demonstrated that B56β specifically interacts with and negatively regulates protein levels of the Pim-1 kinase in 293 cells [31]. This finding suggests a tumor suppressor role for B56β, since elevated Pim-1 kinase levels have been shown to cooperate with other oncoproteins such as c-Myc in cellular proliferation and lymphomagenesis [32]. A role for B56δ in dephosphorylating Cdc25 at T138 and thus preventing the release of 14-3-3 from Cdc25 and inhibiting cell cycle progression through mitosis was shown in Xenopus egg extract studies [33]. Other B56δ targets include the Heart and Neural Crest Derivatives Expressed 1 (HAND1) transcription factor, which is important in development, and the Dopamine- and cAMP-Regulated Phosphoprotein, (DARPP-32), which is critical for dopaminergic neurotransmission in striatal neurons [34, 35].

Finally, B56ε has been shown to be required for Wnt signaling in Xenopus embryogenesis [22]. Although no specific targets were identified for B56ε in this study, it was demonstrated that B56ε is required upstream of Dishevelled (DVL), but downstream of Wnt ligand. Having a positive role in Wnt signaling allows for a potentially oncogenic role for B56ε as Wnt signaling results in increased proliferation. In summary, for all B56 family members, it will be essential to understand what cellular contexts are important for their respective tumor suppressor or oncogenic activities. In particular, understanding the relative importance of specific PP2A holoenzymes for the biology of different cancers, the mechanisms that govern the regulation of different PP2A holoenzymes, and how these mechanisms are perturbed in cancer will give us critical insight into PP2A's role in tumorigenesis.

4 PP2A-B56α negatively regulates the potent oncoprotein c-Myc

As previously mentioned, B56α is one of the more extensively characterized regulatory B subunits within the B56 family. Here we will examine in more detail recent work that has identified c-Myc as a B56α target, followed by a discussion of other B56α targets and its tumor suppressor functions. C-Myc is a transcription factor that regulates the expression of numerous genes involved in cellular proliferation, cellular growth, apoptosis, angiogenesis, and differentiation [36–40]. Over the past 20 years c-Myc has been intensely studied due to both the requirement of c-myc for normal development in mice as well as the high prevalence of elevated c-Myc expression in a wide array of human cancers [41, 42]. Moreover, expression of c-Myc in mice using inducible c-myc transgenes results in the development of neoplastic pre-malignant and malignant phenotypes that often spontaneously regress when c-myc expression is shut off [43, 44]. In addition, c-Myc has been shown to be one of four critical factors for the self-renewal of stem cells and could be important in cancer stem cell biology [45, 46]. It is therefore critical to maintain tight control of c-Myc expression in order to maintain normal cellular function. Regulation of c-Myc expression occurs at several levels through transcriptional, translational, and post-translational mechanisms [47–52]. The regulation of c-Myc at the post-transcriptional level appears to be particularly critical since elevated expression of c-Myc protein is observed in ~70% of human tumors, but only ~20% of these show amplification of the c-myc gene or translocation of the gene to areas of high transcriptional activity [42].

Work from a number of labs including ours has focused on the post-translational regulation of c-Myc through a series of sequential, reversible phosphorylation events on two highly conserved residues, Threonine 58 (T58) and Serine 62 (S62) (Fig. 2(a)). Phosphorylation of T58 and S62 has opposing affects on c-Myc protein stability with S62 phosphorylation stabilizing and T58 phosphorylation destabilizing the c-Myc protein [52]. The PP2A holoenzyme, PP2A-B56α, plays an important role in the regulation of these reversible phosphorylation events [47, 53]. The regulated phosphorylation of c-Myc on T58 and S62 is summarized in Fig. 2(a). As shown, in response to mitogenic stimulation, c-Myc can be phosphorylated on S62 by ERK through the Ras/Raf/MEK/ERK kinase cascade [54, 55]. It has also been shown that other MAPK and cyclin dependent kinases can phosphorylate c-Myc on S62 [56, 57]. Simultaneous activation of the PI3K/Akt pathway by Ras can result in the phosphorylation and inactivation of GSK3β [58]. These two Ras regulated pathways keep c-Myc phosphorylated on S62, which stabilizes c-Myc protein. When Ras activity falls, inhibition of GSK3β activity is relieved, allowing GSK3β to phosphorylate c-Myc on T58 in a processive manner, requiring previous phosphorylation on S62 [55, 57]. Following T58 phosphorylation by GSK3β, the phosphorylation-directed prolylisomerase, Pin1 associates with c-Myc and mediates isomerization to facilitate dephosphorylation of S62 by PP2A, which is conformation sensitive, dephosphorylating Serine or Threonine residues followed by Proline when the peptidyl prolyl bond is in trans [47]. In the case of c-Myc, PP2A-B56α is the specific PP2A holoenzyme that dephosphorylates S62 leaving c-Myc singly phosphorylated at T58 within this motif [53]. Knockdown of B56α was shown to increase c-Myc S62 phosphorylation and protein stability, while ectopic expression of B56α in combination with additional PP2A-C subunit expression substantially reduced c-Myc protein levels [53]. C-Myc phosphorylated at T58 and lacking the stabilizing S62 phosphate is unstable, as it is recognized by the SCF^Fbw7 ubiquitin machinery, which multi-ubiquitinates c-Myc marking it for degradation by the 26S proteosome [59, 60]. Very recent experiments in our lab have found that multiple components of the degradation pathway for c-Myc are recruited into a complex by the scaffold protein, Axin1 (Fig. 2(a)), which will be discussed below. Altogether, c-Myc protein stability is tightly regulated by these hierarchical, reversible phosphorylation events in which PP2A-B56α plays an essential role.

Fig. 2.

(a) Schematic showing the sequential reversible phosphorylation pathway that regulates c-Myc protein stability and turnover via the 26S proteosome. Several of the proteins involved in this pathway are coordinated into a degradation complex for c-Myc by the scaffold protein Axin1. (b) Schematic of complexes that PP2A-B56α has been shown to be a part of

Since post-translational regulation of c-Myc expression is likely a key level of regulation to prevent tumorigenic levels of c-Myc expression, our lab has been analyzing cancer cell-lines and patient samples for possible deregulation of c-Myc expression at the post-translational level. Thus far, observations in several breast cancer and leukemia cell-lines and patient samples have shown c-Myc protein stability to be significantly increased when compared to respective control cell-lines or normal samples with unstable c-Myc ([61] and Sears lab unpublished). Moreover, the c-Myc in these cancer cell-lines shows significantly increased levels of S62 phosphorylation when compared with control cell-lines, indicating that PP2A-B56α function with respect to c-Myc may be impaired in these cancer cell-lines. In support of this hypothesis, a more oncogenic point mutant of c-Myc, c-Myc^T58A, is resistant to PP2A-mediated S62 dephosphorylation and is more stable with higher levels of S62 phosphorylation than wild-type c-Myc [47, 52]. Moreover, this c-Myc^T58A mutant can also transform cultured primary human fibroblasts without any requirement for the presence of SV40 small-T antigen, which has been shown to disrupt PP2A-B56α as well as other PP2A holoenzymes [16, 47]. In contrast, SV40 small-T expression was required for wild-type c-Myc to transform these same cells. Additionally, preliminary analyses demonstrate that specific RNAi-mediated knockdown of B56α can also facilitate transformation of these cells without the need for global inhibition of PP2A activity by SV40 small-T antigen (Sears lab, unpublished). Altogether, these findings strongly suggest that the maintenance of S62 in the unphosphorylated state by PP2A-B56α is critical for the negative regulation of c-Myc expression and inhibition of its oncogenic function. Consequently, understanding the mechanism(s) through which B56α activity may be lost in tumorigenesis is critical and some possible mechanisms are discussed below.

5 PP2A-B56α associates with the scaffold protein Axin1

The general view of PP2A localization and regulation has been largely reported to be intrinsic, with the regulatory B subunits acting to both localize and regulate PP2A holoenzyme activity. Moreover, post-translational modification of the C subunit by phosphorylation and methylation regulate C subunit activity as well as the binding of particular regulatory B subunits [62–68]. Another level of PP2A regulation now being uncovered is the association of PP2A holoenzymes with adaptor and scaffold proteins (Fig. 2(b)). These proteins may act to both coordinate PP2A with its targets and help to localize PP2A holoenzymes. This mode of regulation was initially revealed by the recruitment of PP2A holoenzyme containing B56 family members to dephosphorylate Mdm2 through association with Cyclin G in cell culture experiments [69]. It was also shown that B56α can bind the adaptor protein, ankyrin-B and through this interaction B56α was correctly localized in primary cardiomyocytes (Fig. 2(b)) [70]. Of particular interest in the context of this review, PP2A-B56α holoenzymes have also been shown to associate with the scaffold protein, Axin1. This was initially demonstrated in Xenopus egg extract experiments [71]. A well characterized function of Axin1 is the coordination of APC, GSK3β and now PP2A-B56α into a degradation complex for β-catenin (Fig. 2(b)) [71–74]. However, the exact role of PP2A-B56α within this Axin1-mediated degradation complex is not clear. Recent work in our lab has now shown that Axin1 also associates with c-Myc and regulates its stability by coordinating a degradation complex containing GSK3β, PP2A-B56α, and Pin1 (Sears lab, unpublished). Within the Axin1-mediated degradation complex for c-Myc, one role for PP2A-B56α would be the direct dephosphorylation of c-Myc at S62. However, it is likely that PP2A-B56α targets other components as well. Uncovering the regulation of PP2A by adaptor and scaffold proteins will greatly expand our understanding of PP2A biology.

6 PP2A-B56α negatively regulates β-catenin expression and the anti-apoptotic function of Bcl2

In addition to c-Myc, β-catenin and Bcl2 have been shown to be targets of PP2A-B56α activity (Fig. 1). β-catenin was the first identified target of B56α activity, as it was demonstrated that overexpression of B56α decreased β-catenin expression in mammalian cells and Xenopus embryo explants [75]. Additionally, it was also demonstrated that oncogenic point mutations in β-catenin could prevent its down-regulation by B56α. This has important parallels with the finding that the oncogenic c-Myc^T58A-mutant is resistant to PP2A activity [47, 52]. Intriguingly, although direct dephosphorylation of β-catenin by PP2A-B56α has not been demonstrated, PP2A-B56α was shown to be one component of a degradation complex for β-catenin mediated by the scaffold protein Axin1, as described above. PP2A-B56α is now thought to have targets within this Axin1-mediated degradation complex for β-catenin. These may include APC, since B56α and APC were found to interact in yeast-two hybrid experiments [75]. As part of the Axin1-mediated degradation complex for β-catenin, PP2A-B56α was found to inhibit Wnt signaling and to be critical for normal dorsal/ventral axis formation in Xenopus development [71]. Importantly, β-catenin and Wnt signaling are critical for stem-cell maintenance and cellular proliferation, and deregulation of β-catenin and Wnt signaling are often observed in human malignancies [76]. Therefore, as a negative regulator of β-catenin and Wnt signaling, B56α has been identified as a putative tumor suppressor [71, 75].

Following the identification of β-catenin as a B56α target, Bcl2 was shown to associate with B56α in cell culture experiments [77]. Furthermore, in this study it was demonstrated that in response to ceramide treatment, PP2A-B56α dephosphorylated Bcl2 at the mitochondrial membrane. These dephosphorylation events are thought to occur within the flexible loop domain (FLD) of Bcl2, including the key S70 residue, known to be important for Bcl2's anti-apoptotic function [78]. More specifically, phosphorylation of Bcl2 within the FLD increases its pro-survival activity by increasing its association with Bax and decreasing its association with p53 [79, 80]. Consequently, PP2A-B56α-mediated dephosphorylation of Bcl2 decreases its prosurvival activity, supporting a tumor suppressor role for PP2A-B56α. It is also interesting to note that unspecified PP2A holoenzyme(s) have been shown to promote apoptosis by increasing the pro-apoptotic activity of Bax and BAD [81, 82]. Although the specific B subunit responsible for directing PP2A activity to these targets has not been identified, it is becoming clear that PP2A plays an important role in regulating the induction of apoptosis. In particular, the identification of PP2A-B56α as a regulator of Bcl2 represents the beginning of our understanding of the roles of specific PP2A holoenzymes in controlling key proteins involved in apoptosis.

7 Interplay between PP2A-B56α targets

An interesting aspect of the identification of these PP2A-B56α targets arises from the transcriptional interplay and cooperation in tumorigenesis between c-Myc, β-catenin, and Bcl2, summarized in Fig. 3. The c-myc gene has been shown to be a transcriptional target of β-catenin/TCF/Lef and observations from different human tumors show elevated co-expression of c-Myc and β-catenin [83–85]. Interestingly, there are also reports that β-catenin and c-Myc expression do not correlate in human tumors and it is suggested from these observations that multiple levels of regulation are responsible for the coordination of β-catenin and c-Myc expression [86, 87]. β-catenin/TCF/Lef transcriptional activity has also been shown to induce luciferase activity from a Bcl2 promoter-luciferase reporter construct [88]. Of note, the induction of luciferase activity in this assay could be suppressed by RNAi knockdown of c-Myc, indicating that β-catenin transcriptional regulation of Bcl2 is likely dependent upon its ability to induce the expression of c-Myc. In contrast, elevated c-Myc expression has also been shown to repress expression of Bcl2 by interfering with Bcl2 gene transcription activated by the Myc Interacting Zinc-finger 1 (MIZ1) (Fig. 3), resulting in the apoptosis of primary human cultured cells [89–91]. The ability of c-Myc to repress Bcl2 expression through this mechanism is thought to inhibit the transformation of cells due to elevated c-Myc expression. Importantly, this protective mechanism against elevated c-Myc expression is frequently lost in tumors as c-Myc and Bcl2 often cooperate in tumorigenesis [92, 93]. Altogether, the expression of c-Myc, β-catenin, and Bcl2 have been shown to affect one another at the transcriptional level, yet discrepancies in a number of reports strongly suggest that other levels of control are also important in the coordinated regulation of these oncoproteins. Therefore, the regulation of c-Myc, β-catenin, and Bcl2 by the same PP2A holoenzyme, PP2A-B56α at the post-translational level could offer insight into new mechanisms through which these oncoproteins may influence one another. Furthermore, loss of PP2A-B56α activity could be disastrous as it may result in the simultaneous deregulation of β-catenin, c-Myc and Bcl2.

Fig. 3.

Schematic of known B56α targets and target sequences if available. Also shown is the transcriptional cross-talk connection between known B56α targets and the biological function of these targets

8 Disruption of PP2A-B56α holoenzyme function

Since B56α plays a critical role in regulating multiple oncoproteins, an important question that arises is whether loss of B56α function occurs in cancer. Furthermore, identifying mechanisms through which loss of B56α function can occur is a critical aspect to understanding the importance of B56α function in tumorigenesis. To date, several mechanisms have been reported that could disrupt PP2A-B56α function and these are summarized in Fig. 4.

Fig. 4.

Schematic representations of mechanism through which B56α function could be lost

One of the first reported mechanisms shown to disrupt PP2A holoenzyme function is the ability of viral proteins such as Polyoma small and middle T antigens and SV40-small T antigen to bind the structural A subunit and prevent the association of regulatory B subunits with the A subunit (Fig. 4) [94–96]. Specifically, expression of SV40 small T antigen has been shown to disrupt PP2A-B56α and other B56 holoenzymes in cultured human primary cells [16]. Moreover, expression of these viral proteins has been shown to facilitate the transformation of cultured primary human cells [97–99]. Unfortunately, it is not possible to conclude from these experiments which PP2A holoenzymes are critical in preventing transformation since multiple holoenzymes are simultaneously inhibited by the viral T antigens.

There are a growing number of reports that have identified mutations and mis-expression of PP2A in various human cancers. In particular, mutations in the Aα and Aβ subunits have been reported in human melanoma, breast, lung, colon and colorectal cancers, as well as B-cell chronic lymphocytic leukaemia [86, 100–102]. Moreover, some of the mutations identified in the A subunit from breast tumors have been shown to specifically disrupt its interaction with B56 family members [103]. It is important to note that screening by DNA sequence analysis of PP2A-Aα/β in some other human cancers such as gliomas, cervical, and Wilms tumor revealed no mutations [104–106]. However, in a study of human gliomas by Colella et al., expression of the PP2A-Aα subunit was found to be reduced by 10 fold in 43% of the tumors screened. This suggests that even without mutations in PP2A-Aα/β, down-regulation of the PP2A-Aα subunit may still play a role in tumorigenesis. Analogous to the mechanisms of viral disruption of PP2A holoenzyme function, mutations or mis-expression of the A subunit, as described above, will most likely affect multiple regulatory B subunits. Consequently, it is again difficult to identify specific PP2A holoenzymes that have tumor suppressor activity from current reports.

An alternative mechanism that could lead to disruption of PP2A-B56α holoenzyme activity is a “substrate trap” mechanism by which B56α and/or PP2A-B56α holoenzymes can be abnormally sequestered. This mechanism was first suggested by the observation that ΔNp63, an oncogenic form of p63, interacted with B56α resulting in increased β-catenin protein levels and TCF/Lef-mediated transcription in cultured cells [107]. More recently, we reported that c-Myc^T58A, a highly oncogenic form of c-Myc, has a significantly increased association with B56α when compared to c-Myc^WT [53]. These results suggest that oncogenes such as ΔNp63 and c-Myc^T58A can potentially act as “substrate traps” for B56α. Consistent with this, data also suggests that the amount of B56α is limiting [53, 108]. Consequently, these oncoproteins could interfere with the ability of B56α to regulate its other targets, thus resulting in the deregulation of multiple oncoproteins, such c-Myc, β-catenin, and/or Bcl2 at the post-translational level.

Lastly, the association of PP2A-B56α with adaptor and scaffold proteins opens the door to new mechanisms by which B56α function could be disrupted. It was recently demonstrated in primary cardiomyocytes that B56α associated with the adaptor protein, ankyrin-B [70]. Furthermore, the reduced levels of ankyrin-B in primary cardiomyocytes derived from heterozygous ankyrin-B knockout mice resulted in disorganized localization of B56α and this could be rescued by exogenous expression of ankyrin-B. PP2A-B56α has also been shown to associate with the scaffold protein Axin1 [71] and several studies have identified mutations in Axin1 from a wide array of human tumors that potentially affect the association of PP2A-B56α with this scaffold [109]. Clearly, the molecular and biological consequences of these mutations in Axin1 will need to be characterized in order to determine the importance of Axin1/PP2A-B56α association and its regulation of PP2A-B56α function in tumorigenesis. Altogether, there are several potential mechanisms that could disrupt normal PP2A-B56α function resulting in loss of its tumor suppressor activity in cancer (Fig. 4).

9 Conclusions

The results of numerous studies now support the idea that PP2A has important tumor suppressor activities. However, it should also be noted that PP2A has been reported to possess oncogenic potential. An increased understanding of PP2A tumor suppressor function will depend largely on the identification of specific target/PP2A holoenzyme interactions. Recent work discussed in this review has now identified the PP2A holoenzyme containing the regulatory B subunit B56α as possessing critical tumor suppressor activity via the regulation of several potent oncoproteins. These include, β-catenin, Bcl2, and c-Myc. Furthermore, potential loss of PP2A-B56α function(s) can occur via several mechanisms resulting in the simultaneous deregulation of these oncoproteins, and may be a critical event in the progression of certain human tumors. Clearly, our current knowledge needs to be greatly expanded and important questions such as the underlying mechanisms and frequency of loss or disruption of B56α function need to be answered before we can fully understand the importance of PP2A-B56α tumor suppressor activity. However, the mechanism of PP2A's tumor suppressor activity is now beginning to be revealed one B subunit at a time.

Abbreviations

APC

adenomatous polyposis coli

Bcl2

B-cell lymphoma 2

Cdc25

cell division cycle 25

c-Myc

cellular Myelocytomatosis

DARPP-32

dopamine- and cAMP-regulated phosphoprotein 32 kD

DVL

dishevelled

ERK

extracellular receptor kinase

FLD

flexible loop domain

GSK3β

glycogen synthase kinase 3β

HAND1

heart and neural crest derivatives expressed 1

HEAT

huntingtin-elongation-A subunits-TOR-like

hTERT

human telomerase reverse transcriptase

MAPK

mitogen activated protein kinase

Mdm2

mouse double minute 2

MIZ1

Myc interacting zinc-finger 1

PI3K

phosphoinositide-3-kinase

Pin1

protein (peptidylprolyl cis/trans isomerase) NIMA-interacting 1

PP2A

protein phosphatase 2A

PTEN

phosphatase and tensin homolog

S

serine

SCF

Skp/Cullin/F-box

SV40

Simian virus 40

T

threonine

TCF

T cell specific factor

Wnt

wingless/Int