Phosphorylation is the switch that turns PEA-15 from tumor suppressor to tumor promoter

Abstract

Abnormal ERK signaling is implicated in many human diseases including cancer. This signaling cascade is a good target for the therapy of certain malignancies because of its important role in regulating cell proliferation and survival. The small phosphoprotein PEA-15 is a potent regulator of the ERK signaling cascade, and, by acting on this pathway, has been described to have both tumor-suppressor and tumor-promoter functions. However, the exact mechanism by which PEA-15 drives the outcome one way or the other remains unclear. We propose that the cellular environment is crucial in determining PEA-15 protein function by affecting the protein’s phosphorylation state. We hypothesize that only unphosphorylated PEA-15 can act as a tumor-suppressor and that phosphorylation alters the interaction with binding partners to promote tumor development. In order to use PEA-15 as a prognostic marker or therapeutic target it is therefore important to evaluate its phosphorylation status.

MAPK/ERK Signaling—A Powerful Pathway Driving Tumorigenesis

The mitogen-activated protein kinase/extracellular signal regulated kinase (MAPK/ERK) cascade is one of the major signaling pathways regulating a cell’s proliferation and survival. It is therefore not surprising that mutations leading to abnormal signaling in this pathway are associated with a variety of malignancies. Aberrations in the epidermal growth factor receptor (EGFR) or the small GTPase Ras are two examples that lead to changes in MAPK/ERK signaling with pathogenic effects.¹ Ras is mutated in about 30% of human cancers and constitutively active Ras isoforms have been shown to transform both fibroblasts and epithelial cells.^2-4 The H-Ras isoform in particular is mutated in about 10% of all bladder and kidney cancers.⁵ The MAPK/ERK pathway is a prime target for cancer therapy because of its important role in tumor development and progression. This has led to the development of therapeutic drugs inhibiting members of the signaling cascade like B-RAF and MEK1/2.^1,2 However, inhibitors of these upstream kinases shut down the entire pathway thereby affecting a whole spectrum of downstream targets potentially leading to unwanted side effects. Therefore, alternative targets, that block only certain functions of the pathway, are desirable.

The cells physiological regulation of the MAPK/ERK pathway consists of a variety of feedback loops orchestrating the multiplicity of outcomes.⁶ Of special interest among the regulators of MAPK/ERK signaling are protein scaffolds that enhance the transduced signal by targeting it to a certain downstream effector and protein anchors that restrict components of the pathway to specific locations. PEA-15 (phosphoprotein enriched in astrocytes, 15 kDa) behaves as both a protein scaffold and ERK anchor. It was first described as a substrate for protein kinase C (PKC) involved in the functional regulation of murine astrocytes.⁷ The protein consists of an N-terminal canonical death effector domain (DED) and a C-terminal tail containing two phosphorylation sites (Ser104 and Ser116).⁸ Phosphorylation at these sites determines what binding partner PEA-15 interacts with and what signaling pathway it affects.⁹ PEA-15’s regulation of the MAPK/ERK signaling cascade is one of its most prominent functions. PEA-15 can bind ERK and sequester it in the cytoplasm.¹⁰ It thereby prevents ERK mediated phosphorylation of nuclear substrates like transcription factors, leading to a decrease in proliferation.¹¹ At the same time PEA-15 does not block ERK activity itself and does not inhibit activation of cytoplasmic targets like the ribosomal S6 kinase isoform 2 (RSK2). It in fact mediates ERK driven RSK2 activation by acting as a scaffold to bring both proteins together.^12,13

PEA-15 Regulation—Implications for Cancer

PEA-15’s regulation of the MAPK/ERK pathway has been associated with anti-tumor outcomes. PEA-15 blocked tumor development in a breast cancer xenograft model and it has been shown to induce cellular senescence in human fibroblasts thereby preventing oncogenic transformation.^14,15 PEA-15 has been proposed to be a good prognostic marker in ovarian cancer, where higher expression levels correlated with prolonged patient survival, and in astrocytoma where PEA-15 expression levels were inversely correlated with the stage of the tumor.^16,17 In addition, PEA-15 has been shown to block tumor cell motility and predicted good prognosis in neuroblastoma.^18,19 Based on these findings PEA-15 is considered a potential good prognostic marker in these cancers, and has been proposed to have therapeutic value as a tumor suppressor. However, our recent findings question the simplicity of this hypothesis.

In our study, we show that in an environment where H-Ras is constitutively active PEA-15 does not block but rather enhances Ras-MAPK/ERK signaling and thereby potentiates H-Ras driven transformation of kidney epithelial cells.²⁰ We show that this effect is at least in part mediated by PEA-15’s interaction with phospholipase D1 (PLD1). PEA-15 causes an increase in PLD1 activity, likely through stabilization of the PLD1 protein levels. This interplay of PEA-15 and PLD1 has been proposed before to have functional significance for the activation of ERK.²¹ We propose that higher PLD1 activity in turn results in higher production of the second messenger phosphatidic acid (PA). As shown before, PA may promote the recruitment of either SOS, a guanine nucleotide exchange factor that activates Ras, or the downstream kinase Raf to the plasma membrane.^22,23 Both events may lead to elevated Ras-MAPK/ERK signaling causing the observed increase in cell transformation.

Importantly, Ras transformation leads to phosphorylation of PEA-15 at Ser116, and thereby prevents PEA-15 restriction of ERK to the cytoplasm. Phosphorylation at this site doesn’t affect PEA-15 stabilization and activation of PLD1. Thus in this signaling context, PEA-15 acts to enhance ERK activation via PLD1, but no longer binds ERK in the cytoplasm. The net effect is increased ERK signaling and enhanced Ras induced transformation. Our proposed mechanism is in agreement with earlier reports showing PEA-15’s ability to promote ERK activation.²⁴ In this study activation of ERK by PEA-15 did not require Grb2 binding to SOS, leaving PEA-15’s point of action downstream of Grb2. The next signaling molecule downstream of Grb2 is SOS, a guanine nucleotide exchange factor that binds to Grb2 and promotes Ras activation through GTP-loading. Interestingly, isoform 2 of PLD (PLD2) has been proposed to be crucial in recruiting SOS to the plasma membrane and placing it between the EGFR/Grb2 and Ras.²⁵ PEA-15 can also bind PLD2.²¹ It is therefore possible that PEA-15 may enhance activation of Ras/ERK signaling by increasing PLD2 activity and thereby PA recruitment of SOS to the plasma membrane.

Our study is not the first to describe tumor-promoting effects of PEA-15. The protein’s ability to interfere with apoptotic signaling through its death effector domain suggested an oncogenic role of PEA-15 early on. PEA-15 can bind to FADD and prevent recruitment of initiator caspases in death receptor activated apoptosis. Through interference with the programmed cell death PEA-15 has been shown to increase susceptibility to chemically induced skin cancer in transgenic mice and to mediate chemo-resistance in human breast cancer cells.^26,27 Furthermore, PEA-15 protected human glioblastoma cells from glucose deprivation-induced apoptosis. In this way it acted as a survival factor for cancer cells in microenvironments with low glucose supplies.²⁸

Phosphorylation—The Key to PEA-15 Function

PEA-15 seems to play a role as either a tumor-promoter or tumor-suppressor depending on the cellular environment. This raises the question of how to assess PEA-15’s role as a prognostic marker or even a therapeutic target in specific cancer subsets. High levels of PEA-15 expression have been described as a good prognostic marker in ovarian cancer, neuroblastoma and astrocytoma.^16,17,19 On the other hand PEA-15 is highly expressed in certain human glioblastomas and human malignant pleural mesotheliomas, both late stage malignancies with a poor prognosis.^28,29 The key to evaluating PEA-15’s function on a molecular level therefore lies beyond its expression levels and has to include other post-translational factors that determine its effects on tumor growth and metastasis.

One of these factors is the phosphorylation status of PEA-15. The serine phosphorylation sites at Ser104 and Ser116 can be phosphorylated by protein kinase C (PKC) and calcium calmodulin kinase II (CamK II) respectively.⁷ Protein kinase B/AKT has also been shown to phosphorylate PEA-15 at Ser116.³⁰ Phosphorylation at these sites on PEA-15 modulates its binding specificity and impacts how PEA-15 regulates cell-signaling processes. However, the biochemical mechanism by which PEA-15 phosphorylation affects the interaction with binding partners remains elusive. Callaway and colleagues reported that phosphorylation of PEA-15 does not affect its affinity for ERK per se, however phosphorylation seems to promote the binding of other proteins like FADD that can compete with ERK for the binding of PEA-15.³¹ Competitive inhibition might therefore be the cause for the observed decrease in ERK-PEA-15 interaction upon phosphorylation of PEA-15.

Phosphorylation of PEA-15 at Ser104 by PKC blocks ERK binding both in vitro and in vivo. We showed that this phosphorylation subsequently impaired ERK’s ability to activate the cytoplasmic target RSK2.^9,13,32 Ser104 phosphorylation also abolished PEA-15 mediated inhibition of HeLa cell invasion, likely due to its loss of ERK binding capacity.¹⁸ In addition it has been reported that phosphorylation at this site is necessary for PEA-15’s regulation of glucose metabolism.³³

Phosphorylation at Ser116 promotes PEA-15 binding to the apoptotic adaptor protein FADD and this phosphorylation is required for PEA-15 recruitment to the death inducing signaling complex (DISC) to inhibit apoptotic signaling.⁹ This phosphorylation was shown to be required for blocking apoptosis in glucose-deprived glioblastoma cells.²⁸ In our recent study investigating its involvement in H-Ras driven cell transformation we found that PEA-15 was phosphorylated at Ser116 when co-expressed with a constitutively active H-Ras isoform. H-Ras activity was responsible for the phosphorylation, since cells lacking mutated H-Ras did not express PEA-15 phosphorylated at this site. Note that oncogenic H-Ras activates the AKT pathway among others and it is likely this activity that leads to the Ser116 phosphorylation. As expected, in these cells phosphorylated PEA-15 did not sequester ERK in the cytoplasm.²⁰

Our study shows, that PEA-15’s function not only depends on the cell type, but can change due to the subcellular signaling environment. We previously described that stimulation of Jurkat T cells through the CD3/28 T-cell co-receptors led to increased phosphorylation of PEA-15 at Ser104, likely through PKC. In this scenario PEA-15 may act as T-cell regulator by inhibiting cell proliferation in the resting T cell. Activation of the T cell through the T-cell receptor also starts signaling pathways leading to the phosphorylation of PEA-15 causing it to release ERK, which in turn induces the transcription of nuclear targets resulting in T-cell proliferation.¹¹ We also previously described that stimulation of Cos-7 cells and thymocytes with phorbol esters like PMA induced PEA-15 phosphorylation at Ser104.⁹ It is the activity of pathways leading to PEA-15 phosphorylation that ultimately determines the cellular outcome by regulating PEA-15’s interaction with distinct binding partners. Although we know the proteins that phosphorylate PEA-15, the activity of these pathways and the timing of this activity in the cancers in which PEA-15 is expressed are not uniformly addressed.

Trencia and colleagues reported that AKT phosphorylates PEA-15 at Ser116 in Hek293 cells thereby stabilizing PEA-15 protein levels. When AKT activity was inhibited in these cells, PEA-15 lost its ability to block apoptosis. A similar increase in sensitivity to TRAIL-induced apoptosis was observed in U-373 MG glioblastoma cells.³⁰ Interestingly we found that in Cos-7 cells stimulation with epidermal growth factor (EGF) can induce MAPK-independent phosphorylation of PEA-15 at Ser104 (Fig. 1). PEA-15 phosphorylation at Ser116 did not change upon EGF treatment, however basal phosphorylation at this site was already very high. Active EGF receptor (EGFR) can therefore induce PEA-15 phosphorylation at Ser104. These findings are intriguing since both EGFR and PI3K/AKT signaling has been shown to be involved in tumor development and progression. A high number of malignancies show mutations in either one of the two proteins causing them to be overly active. Abnormal EGFR signaling has been related to a variety of malignancies and is often connected to poor patient survival rates.³⁴ Likewise, PI3K/AKT signaling has been implicated in cancer and the AKT isoforms have been related to functions like tumor cell invasion and hormone independence.³⁵ Thus, we propose that in cells with increased EGFR or AKT activity, phosphorylated PEA-15 is likely to be pro-tumorigenic, potentially through effects on PLD activity, blockade of apoptosis or both.

Figure 1. PEA-15 is Phosphorylated at Serine 104 Upon EGF Stimulation. Cos-7 cells were transfected with PEA-15 plasmid DNA using Lipofectamine (Life Technologies, NY) and cultured in DMEM containing 10% fetal bovine serum. After 24h the cells were serum-starved for 18hrs before treating them with the MEK inhibitor U0126 (10μM, 15min pre-treatment) and/or EGF (10ng/ml) for the indicated amount of time. Cell lysates were prepared using a M2 lysis buffer (0.5% NP-40, 20mM TRIS-HCl, pH 7.6, 250mM NaCl, 5mM EDTA, 3mM EGTA, 20mM NaPO₄, 20mM NaPP_i, 3mM β-glycerophosphate and protease/phosphatase inhibitors). Cell lysates were resolved using SDS-PAGE, followed by immunoblotting and protein expression was detected with specific antibodies against PEA-15 phospho-Ser104, phospho-Ser116, total PEA-15, phospho-ERK1/2 and total ERK1/2.

Although a functional connection between PEA-15 phosphorylation and pathways activated in cancer such as EGFR, PKC and AKT is likely, little direct evidence has been established. Some recent studies have begun to more completely examine this. In ovarian clear cell carcinoma, Ser116 phosphorylation renders cells more sensitive to selumetinib, a selective inhibitor of the MEK1/2 kinase.³⁶ Selumetinib treatment upregulated PEA-15 phosphorylation at Ser116 in RMG-I, but not in OVTOKO ovarian clear cell carcinoma lines. Interestingly, both selumetinib sensitivity and PEA-15 phosphorylation correlated with EGFR expression in these cells. The low EGFR expressing cell line OVTOKO showed resistance to selumetinib treatment and no upregulation of PEA-15 phosphorylation was detected. Thus in ovarian clear cell carcinoma, PEA-15 phosphorylation correlates with EGFR expression. Importantly, in this cancer there is a direct connection between PEA-15 phosphorylation and cancer progression as well as sensitivity to anti-cancer therapeutics. More work is necessary to determine if there are similar correlations between mutational activation of signaling pathways and PEA-15 phosphorylation and function in other cancer models.

Concluding Remarks

We summarize phosphorylation dependent functions of PEA-15 in Figure 2. To utilize PEA-15 as a prognostic marker it is necessary to evaluate not only PEA-15 expression levels, but also its phosphorylation status. This is especially important in cancer types that have known mutations in pathways that alter PEA-15 phosphorylation including EGFR, Ras or AKT. So far only unphosphorylated PEA-15 has been described to act as tumor-suppressor. Studies with phosphorylated PEA-15, including our recent report on PEA-15’s involvement in H-Ras driven cell transformation, describe a potent tumor-promoter function. Therefore, to use PEA-15 as a therapeutic target it is important to determine the signaling pathways active in the specific tumor and the resultant phosphorylation status of PEA-15. Thus far the data indicates that only non-phosphorylatable PEA-15 constructs are likely to be effective as anti-cancer therapeutics.

Figure 2. PEA-15 function is determined by phosphorylation. Only unphosphorylated PEA-15 has been associated with tumor-suppressor functions so far. In this state PEA-15 interacts with ERK preventing ERK nuclear translocation and activation of nuclear targets. Phosphorylation at Ser104 by protein kinase C (PKC), which can be induced by phorbol esters (PMA) or epidermal growth factor (EGF), blocks ERK binding of PEA-15. This reverses the effects PEA-15 has on MAPK/ERK signaling, resulting in elevated transcription and cell proliferation. Phosphorylation at Ser116 by calcium calmodulin kinase II (CamKII) or AKT promotes PEA-15 binding to FADD. Thereby it blocks death receptor activated apoptosis. For this function PEA-15 phosphorylation at Ser104 is not required (*). Ser116 phosphorylated PEA-15 retains its ability to activate PLD1 and can thereby enhance ERK pathway activation. The phosphorylated states of PEA-15 have been implicated in tumor progression and have been found to induce higher tumor cell invasion and increased tumor cell survival.

Sulzmaier FJ, Valmiki MK, Nelson DA, Caliva MJ, Geerts D, Matter ML, White EP, Ramos JW. PEA-15 potentiates H-Ras-mediated epithelial cell transformation through phospholipase D. Oncogene. 2011 doi: 10.1038/onc.2011.514.