PerspectiveCancer Biology

P-REX2a Driving Tumorigenesis by PTEN Inhibition

See allHide authors and affiliations

Science Signaling  27 Oct 2009:
Vol. 2, Issue 94, pp. pe68
DOI: 10.1126/scisignal.294pe68


The phosphatase PTEN (phosphatase and tensin homolog deleted on chromosome 10) antagonizes phosphoinositide 3-kinase (PI3K) signaling and is one of the most frequently mutated tumor suppressors in human cancers. Its regulation appears complex and is of great potential clinical importance. The protein P-REX2a (phosphatidylinositol 3,4,5-trisphosphate Rac exchanger 2a), better known as a regulator of the small guanosine triphosphatase Rac, has been identified as a direct regulator of PTEN activity and as a potential oncoprotein. P-REX2a can stimulate cell proliferation by inhibiting PTEN and stimulating downstream PI3K-dependent signaling. This suggests that aberrant control of PTEN by P-REX2a may represent a key tumorigenic mechanism, in agreement with recent studies supporting the pathological relevance of several other proposed PTEN regulators.

Cancer develops through the accumulation in cells of diverse genetic and epigenetic changes that pervert the behavior of these cells from their normal physiological program toward aberrant proliferation and malfunction (1). The key targets for genetic mutation in cancer are generally classified into two groups: oncogenes, which have cellular functions that drive cancer when activated, and tumor suppressors, which have normal cellular functions that drive the development of cancer when lost.

The application of genome-scale analyses to cancer research has shown that the number of genetic mutations in tumors is higher than had been previously anticipated, with perhaps 10 or more mutations that are key contributors to the development of any particular tumor and an even larger number of less pathologically important “passenger” mutations (2). Such analyses confirm that many tumor suppressors and oncogenes cluster in a relatively small number of important functional pathways and that the coexistence in one cancer of multiple mutations in the same pathway (such as the p53 and retinoblastoma pathways) is very rare (3). However, the coexistence of multiple mutations is not uncommon within the broader phosphoinositide 3-kinase (PI3K) signaling network (3, 4).

PTEN (phosphatase and tensin homolog deleted on chromosome 10) is one of the most frequently mutated tumor suppressors in human cancer, with genetic or epigenetic loss of expression and function contributing to the pathogenesis of a diverse range of tumor types (5, 6). The principal mechanism of action of the PTEN protein in cells is the dephosphorylation of phosphatidylinositol 3,4,5-trisphosphate (PIP3), the lipid product of PI3K (710). Reducing PIP3 concentrations inhibits the downstream components of the PI3K signaling pathway, such as the oncogenic protein kinase Akt, and suppresses cell growth, survival, and migration (6, 10).

Studies from many labs have shown that the activity of the PTEN protein can be regulated by several mechanisms, including phosphorylation, oxidation, ubiquitination, and localization (11). However, it is currently unclear whether these mechanisms contribute in a meaningful manner to tumor pathology. A report from Fine, Parsons, and co-workers has identified a physical interaction between PTEN and P-REX2a (phosphatidylinositol 3,4,5-trisphosphate Rac exchanger 2a), a protein previously recognized as a regulator of the small guanosine triphosphatase (GTPase) Rac (12). Their data suggest that this interaction may not only be important for inhibiting PTEN function, but also that P-REX2a function may be elevated in some tumors.

Small GTPases such as Rac take up an active conformation when they are in a GTP-bound state and undergo a shift to an inactive conformation when they hydrolyze their bound GTP to GDP. These signaling molecules are activated by guanine nucleotide exchange factors (GEFs) that promote the dissociation of GDP and its replacement with GTP, which is present in the cell at higher concentrations. P-REX2a is a GEF for the small GTPase Rac and appears to activate Rac when both PI3K and heterotrimeric GTP-binding protein (G protein)–coupled receptor signaling pathways are activated (1315) (Fig. 1).

Fig. 1

P-REX2a, PTEN, and PI3K signaling. Isoforms of PI3K, which phosphorylate PI(4,5)P2 to convert it to PIP3, can be activated by stimulation of receptor tyrosine kinases (RTKs) or G protein–coupled receptors (GPCRs). PIP3 acts through downstream targets, including Akt, P-REX2a, other GEFS, and indirectly, Rac, to promote cell growth, survival, and motility. PTEN suppresses downstream signaling by converting PIP3 back to PI(4,5)P2. P-REX2a appears to play two roles in the pathway: first, as an activator of Rac GTPase signaling in response to PIP3 and Gβγ subunits released after GPCR activation, and second, as a direct inhibitor of PTEN.

Fine and co-workers identified P-REX2a in a screen for PTEN-interacting proteins using protein extracts from a glioblastoma cell line that lacks PTEN and an immobilized PTEN column to pull out binding partners. After verification of the interaction using endogenous proteins, the authors showed that purified P-REX2a inhibited PTEN lipid phosphatase activity in vitro when the two proteins were present at equal concentration. This agreed with data from cell lines, in which knockdown or overexpression of P-REX2a respectively inhibited or increased both Akt phosphorylation and proliferation. Moreover, these effects were observed only in cell lines in which wild-type PTEN was present, not in cell lines lacking the phosphatase, unless PTEN had been reexpressed.

The authors provide evidence that P-REX2a may display elevated activity in tumors. First, the PREX2 gene at 8q13.2 lies in a region that is frequently amplified in breast, prostate, colorectal, and ovarian cancers (12, 16, 17). Accordingly, analysis of published gene expression data using Oncomine (18) shows that the mRNA abundance of P-REX2a is increased in several tumor data sets relative to normal tissue, in breast, prostate, and ovarian tumors, as well as in gliomas and pancreatic cancer. Finally, current data suggest that P-REX2a may be one of the most frequently mutated GEFs in human cancers, because nine different P-REX2a point mutations have been described in a range of cancers, corresponding to 3% of the tumors analyzed (12).

A critical part of the Fine et al. data supports the role of the P-REX2a−PTEN interaction in mediating the functional effects of P-REX2a. PTEN was required for the effects of P-REX2a on Akt phosphorylation and correlated with the effects of P-REX2a on cultured cell proliferation. Similarly, P-REX2a mRNA abundance was significantly higher in breast cancer samples that contained PTEN than in those lacking the phosphatase, and also correlated with the presence of mutant active PI3K. This agreed with experiments in cultured MCF10A cells, in which P-REX2a and mutant PI3K cooperated in cellular transformation (12).

The inhibitory activity of P-REX2a on PTEN in vitro and in cells appeared to require only the GEF active regions of P-REX2a: the pleckstrin homology (PH) and Dbl homology (DH) domains. However, GEF activity itself did not seem to be required (12). Therefore, the data suggest a model in which P-REX2a binds directly to PTEN through its GEF domains (DH-PH) to inhibit PTEN activity and activate downstream PI3K-dependent signaling. A prediction of this analysis is that the mutations of P-REX2a identified in tumors should enhance or positively deregulate the inhibitory activity of P-REX2a toward PTEN. This touches upon regulation of the P-REX2a protein, its inhibition of PTEN, and whether PTEN binding in turn affects P-REX2a GEF activity. Is the ability of P-REX2a to inhibit PTEN controlled by P-REX2a conformation and perhaps modulated by PIP3 and G protein βγ subunits, like its GEF activity? The apparently equal ability of full-length P-REX2 and the DH-PH region to inhibit PTEN in vitro and in cells argues against this type of regulation, yet the presence of P-REX2 mutations in tumors would seem to favor it, assuming that the mutations affect P-REX2a function. It is also possible that the mutation of PREX2 in tumors drives tumorigenesis by a PTEN-independent mechanism. There is evidence that other Rac GEFs show elevated activity in tumors (1921) and that P-REX1 can promote prostate cancer metastasis through its Rac GEF activity (22). Therefore, it seems quite plausible that some effects of P-REX2a activity in tumors might act through the activation of Rac signaling in addition to effects on PTEN.

Although PTEN is present in almost all normal human cell types, the distribution of P-REX2a is not so ubiquitous (13, 14, 23). It is currently unknown whether the closely related P-REX1 protein or other, more distantly related exchange factors might also regulate PTEN; however, P-REX1 protein is also detected in many, but not all tissues. Mice lacking both P-REX proteins are overtly healthy and fertile, suggesting that these proteins are dispensable for viability. However, these mice suffer from morphological defects in the cerebellum, including thinner main dendrites on Purkinje cells and related motor coordination problems (23).

The principle of important tumor suppressors and oncogenes acting through a dominant downstream tumor suppressor has been well established with p53. In that system, either expression of oncoproteins, such as MDM2 or several viral proteins, or mutations in the ARF tumor suppressor promote carcinogenesis indirectly through suppression of p53 abundance and function (24). In some settings, partial loss of PTEN function is sufficient to drive tumor development (6, 2527); thus, binding partners that can control PTEN activity represent potential oncogenes and tumor suppressors. Over the past few years, several such proteins have been proposed, including the tyrosine kinase RAK (28), GLTSCR2 [glioma tumor suppressor candidate region gene 2; also known as protein interacting with carboxyl-terminus 1 (PICT-1)] (29, 30), the E3 ubiquitin ligase NEDD4.1, (31), the hydrogen peroxide–metabolizing enzyme peroxiredoxin 1 (PRDX1) (32, 33), and now P-REX2a. Future work should tell us whether PTEN frequently acts in tumors as a downstream mediator of changes in other oncogenes and tumor suppressors, including P-REX2a, and whether identifying these changes can be used successfully to match patients with the drugs that are becoming available to target components of the PI3K signaling network.


View Abstract

Navigate This Article