p38 Gamma MAPK
However, p38γ can be selectively activated under certain conditions after exposure to γ-radiation (Wang et al. 2000), cisplatin and UV (Pillaire et al. 2000), hypoxia (Conrad et al. 1999), DNA topoisomerase II inhibitors (Qi et al. 2011), antiestrogens (Qi et al. 2012), and oncogenes (Sakabe et al. 2002; Tang et al. 2005). p38γ is insensitive to p38α/p38β inhibitor SB203580 but is inhibited by pirfenidone (PFD) (Ono and Han 2000; Cuenda and Rousseau 2007; Ozes et al. 2008; Hou et al. 2012; Qi et al. 2014; Qi et al. 2015; Yin et al. 2016). Early studies showed that p38γ RNA is predominantly expressed in muscle tissues and is involved in muscle differentiation (Lechner et al. 1996; Li et al. 1996; Ono and Han 2000; Tortorella et al. 2003; Cuenda and Rousseau 2007; Perdiguero et al. 2007). Recent research found that p38γ protein is detectable in a variety of tissues (and cells), is upregulated in several types of human cancers, and promotes oncogenesis through interaction with several proteins (Loesch and Chen 2008; Hou et al. 2010a; Hou et al. 2010b; Meng et al. 2011) (Fig. 1). This review will focus on recent discoveries of new functions of p38γ MAPK through PDZ-dependent and independent interactions with its partner proteins.
p38γ Signals to and from Its Partners Through PDZ Binding
p38γ C-terminal PDZ-binding motif can mediate its specific interaction with PDZ-domain containing proteins, which may be the structural basis for p38γ-specific biological activities (Hou et al. 2010a).
The two-hybrid screening of human colon cDNAs showed that PDZ mediates p38γ interaction with a protein tyrosine phosphatase H1 (PTPH1) through which p38γ is dephosphorylated by PTPH1 (Hou et al. 2010b). PTPH1 is a nonmembrane tyrosine phosphatase containing a single PDZ domain (Tonks 2006). p38γ RNA and protein levels are upregulated by Ras oncogene in epithelial cells, and resultant p38γ is in turn required for Ras transforming and invasive activity (Tang et al. 2005; Qi et al. 2007; Hou et al. 2010b; Loesch et al. 2010). Further analysis showed that both K-Ras oncogene and p38γ overexpression increases PTPH1 protein expression, and the PDZ-mediated p38γ/PTPH1 interaction is required for p38γ as well as PTPH1 to potentiate K-Ras-induced transformation (Hou et al. 2010b). Although transient transfection studies showed that phosphorylated p38γ is decreased by K-Ras oncogene coexpression in rat intestinal epithelial cells (Tang et al. 2005), levels of phosphorylated p38γ proteins are increased in K-Ras mutated human colon cancer cells as compared to those without K-Ras mutation independent of PTPH1 (Hou et al. 2012). Moreover, stable expression of a dominant negative p38γ/AGF has a diminished potentiating effect on K-Ras transformation as compared to a wild-type p38γ (Hou et al. 2012), whereas stably expressed p38γ-MKK6 fusion protein blocks stress-induced death independent of phosphorylation (Qi et al. 2007). Together, these results indicate that p38γ can signal to increase K-Ras oncogenesis and/or to inhibit stress-induced cell death dependent and independent of its phosphorylation and of its substrate PTPH1 (Hou et al. 2012).
The oncogenic role of PDZ-mediated p38γ-PTPH1 complex is further supported by the fact that p38γ phosphorylates PTPH1 at S459 in vitro and in vivo, which is important for their potentiation of K-Ras transformation and antiapoptotic activities (Hou et al. 2012). Moreover, disruption of this complex by a p38γ C-terminal peptide decreases the growth of colon cancer cells (Hou et al. 2010b), and application of a specific p38γ pharmacological inhibitor pirfenidone (PFD) inhibits the malignant growth and reduces levels of p-PTPH1/S459 protein expression (Hou et al. 2012; Qi et al. 2014). Thus, PDZ-mediated p38γ-PTPH1 complex and resultant PTPH1 phosphorylation/activation plays an important role in K-Ras oncogenesis, albeit p38γ may additionally promote proliferation and/or oncogenesis independent of its own phosphorylation and/or PTPH1. Recent crystal-structure analysis of the p38γ/PTPH1 complex further suggests an essential role of p38γ in PTPH1 activity (Chen et al. 2014). In addition, phosphorylation of PTPH1 by p38γ is independent of stress-induced activation of conventional p38α and JNK pathway activities (Hou et al. 2012). Overall, PDZ-coupled p38γ/PTPH1 complex may play an important role in Ras oncogenesis through increasing cell growth and/or decreasing cell death. Because p38γ depends on its phosphorylation status to bind PTPH1, the PDZ-complex is most likely driven by p38γ phosphorylating PTPH1, rather by PTPH1 dephosphorylating of p38γ, to promote K-Ras oncogenesis (Fig. 1).
α1-Syntrophin and SAPs
Two-hybrid screening using human brain cDNA library revealed that p38γ, through its C-terminal PDZ-binding motif, binds and phosphorylates α1-syntrophin (Hasegawa et al. 1999). α1-syntrophin, a PDZ domain protein, functions as a modular adapter to recruit signaling proteins and is highly expressed in skeletal muscle (Peters et al. 1997). Although this PDZ-coupled interaction is essential for p38γ phosphorylation of α1-syntrophin, biological consequence of this interaction remains unclear (Hasegawa et al. 1999) (Fig. 1). A similar PDZ-dependent interaction was also demonstrated between p38γ and SAP90 (synapase-associated protein 90)/PSD95 (postsynaptic density protein 95) (Sabio et al. 2004) and between p38γ and SAP97 (hDlg) (Sabio et al. 2005), through which both proteins are phosphorylated with the biological consequences also unknown. Recent studies showed that the oncogenic protein β-catenin interacts with SAP97 through its PDZ motif to maintain the integrity of tight junctions (Gujral et al. 2013). Since p38γ also binds β-catenin (Yin et al. 2016), it would be interesting to explore if PDZ-coupled multiple-protein complex plays a role in maintaining epithelial polarity (Chen and Macara 2005). The interaction of SAP97/hDlg with p38γ can further regulate its interaction with polypyrimidine tract-binding protein-associated-splicing factor (PSF) and thereby causes the hDlg-RNA dissociation in stress response (Sabio et al. 2010). While these effects appear specific for p38γ and are activated in response to stress (such as sorbitol and UV radiation), cellular effects of blocking p38γ interaction with these proteins have not been demonstrated (Sabio et al. 2004; Sabio et al. 2005; Sabio et al. 2010). Since PDZ binding is a central module of scaffold proteins in maintaining barrier function (Fanning and Anderson 1999; Schlieker et al. 2004; Smock and Gierasch 2009; Monteriro et al. 2013), p38γ may play an active role in epithelial integrity through interaction with PDZ-domain containing proteins.
p38γ Binds and Activates Its Partner Proteins
In addition to PDZ-mediated interactions, p38γ can bind to other proteins by a yet unknown mechanism to impact life-important events. These interactions typically lead to its binding partner phosphorylation, stabilization, or degradation, leading to an oncogenic response. Moreover, the ability of p38γ to interact with a transcription factor will enable its recruitment to target gene promoters and thereby directly regulate gene transcription as a coactivator.
p38γ is required for Cot- and RhoA-induced activation of c-Jun promoter (Chiariello et al. 2000; Marinissen et al. 2001). Moreover, p38γ overexpression alone is sufficient to stimulate c-Jun promoter activity through AP-1 and MEF2 binding sites (Marinissen et al. 1999; Loesch et al. 2010). The work by Loesch et al. further showed that p38γ depends on both its C-terminal PDZ motif and phosphorylation to bind and to trans-activate c-Jun, which is essential for basal AP-1 transcription activity (Loesch et al. 2010). Further studies showed that p38γ is required for MKK6 stimulation of MMP9 promoter activity (Simon et al. 2001). Since c-Jun is positively autoregulated by its own product through AP-1 on its promoter (Angel et al. 1988), p38γ interaction with c-Jun may directly link its activity to increase AP-1 target gene expression. Indeed, c-Jun can act as a carrier to recruit p38γ onto the MMP9 and Nanog promoters, and these events are critical for p38γ-induced invasive response and cancer stem-like cell (CSC) expansion (Loesch et al. 2010; Qi et al. 2015). Since there is no PDZ domain in c-Jun protein, this interaction may not directly involve PDZ binding.
Estrogen Receptor α (ER)
p38γ interacts with the nuclear receptor ER (estrogen receptor), which plays an important role for ER to antagonize nuclear p38γ activity (Qi et al. 2006). Further, p38γ phosphorylates ER/S118 and forms a complex with ER and c-Jun on cyclin D1 promoter (Qi et al. 2012). ER requires both T311 and S118 to bind p38γ, which is important for p38γ invasive activity (Qi et al. 2006) and for p38γ increasing hormone sensitivity in breast cancer cells (Qi et al. 2012). However, PTPH1 can catalyze ER/Y537 dephosphorylation and thereby increase ER nuclear translocation and breast cancer hormonal sensitivity (Suresh et al. 2014). Although S459 is important for the PTPH1 phosphatase catalytic activity toward EGFR (epidermal growth factor receptor) (Hou et al. 2012; Ma et al. 2015), whether p38γ phosphorylates PTPH1/S459 in breast cancer has not been demonstrated. It is possible, however, that p38γ may increase breast cancer hormonal sensitivity directly by increasing p-ER/S118 levels and indirectly by decreasing p-ER/Y537 (through phosphorylating PTPH1/S459). This is because increased levels of p-ER/S118 (Kok et al. 2009) and p-ER/Y537 (Skliris et al. 2010) in clinical breast cancer are good and worse biomarkers for antiestrogen therapy, respectively. Furthermore, p38γ binds both ER and c-Jun in breast cancer cells, and treatment with tamoxifen stimulates this ternary-complex formation, which may also be critical for p38γ potentiation of breast cancer sensitivity to antiestrogens (Qi et al. 2012). In addition, p38γ is a breast cancer metastasis gene in triple-negative breast cancer (Qi et al. 2006; Meng et al. 2011; Rosenthal et al. 2011; Lee et al. 2013; Qi et al. 2015). Overall, p38γ may promote malignant progression in ER-negative breast cancer and increase hormonal sensitivity in ER-positive breast tumors.
Heat Shock Protein 90 (Hsp90)
Proteomic analysis of p38γ precipitates identified a mutant K-Ras-dependent interaction of p38γ with Hsp90 (heat shock protein 90) (Qi et al. 2014). Importantly, this complex contains mutated K-Ras protein, and p38γ protects the oncoprotein from degradation by phosphorylating Hsp90 at S595 (Qi et al. 2014). Hsp90/S595 is phosphorylated by p38γ, but not its family member p38α, which is important for stabilizing mutated (but not wild-type) K-Ras protein against proteasome-dependent degradation (Qi et al. 2014). Significantly, high levels of p38γ proteins in K-Ras mutant colon cancer cells are required to maintain the endogenous oncoprotein levels, and targeting p38γ by shRNA or pharmacological inhibitor PFD selectively inhibits K-Ras mutant colon cancer growth in vitro and/or in vivo (Qi et al. 2014). Although Hsp90 activity was previously shown to be important for K-Ras mutant cancers (Sos et al. 2009; Azoitei et al. 2012), the mutant K-Ras-specific binding of p38γ together with Hsp90 and the resultant Hsp90/S595 phosphorylation reveal a novel mechanism that can be explored for targeting mutant K-Ras protein.
DNA Topoisomerase IIα (Topo IIα)
Topo IIα is an important therapeutic target for cancer chemotherapy, and application of Topo II inhibitors (such as Adriamycin: ADR; etoposide: VP16) is a standard therapeutic strategy for many types of human cancer (Chen and Liu 1994). However, determinants for therapeutic response to Topo II drugs are largely unknown (Pritchard et al. 2008). Studies showed that treatment of breast cancer cells with Topo II inhibitors, but not with the antimicrotubule drug taxol, increases p38γ (but not p38α) phosphorylation, which is associated with increased sensitivity of breast cancer cells to Topo II drugs (Qi et al. 2011). Topo IIα is a nuclear protein (Chen and Liu 1994), and in contrast to p38α, phosphorylated p38γ is also mostly accumulated in the nucleus (Qi et al. 2007; Sabio et al. 2010). These findings suggest that p38γ activation may be required for the positive feedback loop between Topo II and its inhibitors, through which p38γ phosphorylates and thereby activates Topo II to increase its therapeutic target activity. Indeed, p38γ binds, phosphorylates Topo IIα/S1542, and increases the Topo II catalytic activity (Qi et al. 2011). In addition, p38γ increases Topo IIα protein stability dependent of phosphorylation, and elevated p38γ expression in breast cancer specimens correlates with increased Topo IIα levels (Qi et al. 2011). Furthermore, Ras oncogene stimulates Topo IIα (Chen et al. 1999) and p38γ expression (Tang et al. 2005), and transformed cells are more sensitive to the Topo II inhibitor VP-16 as compared to their nontransformed counterparts (Chen et al. 1997). These results together indicate that increased p38γ expression in malignant cells may be a good marker for their sensitivity to Topo II inhibitors.
β-catenin is a central component of Wnt signaling and plays a critical role in colon cancer development and progression by stimulating Wnt transcription activity (Clevers 2006). Conditional p38γ knockout from intestinal epithelial cells (IECs) decreases β-catenin expression, inhibits Wnt activity, and attenuates colon tumorigenesis in an azoxymethane(AOM)/dextran sodium sulfate (DSS) mouse model (Yin et al. 2016). Further analysis showed that p38γ binds β-catenin and increases its protein stability by stimulating its S605 phosphorylation and thereby decreasing its proteasome-dependent degradation (Yin et al. 2016). Moreover, inflammation stimulates p38γ and β-catenin phosphorylation, and β-catenin/S605 is required for p38γ stimulating of Wnt transcriptional activity and for colon cancer growth (Yin et al. 2016). The role of p38γ-phosphorylating β-catenin/S605 in p38γ promoting K-Ras oncogenesis and inflammation-induced colon cancer, however, remains to be explored further.
p38γ Phosphorylates Other Proteins
Studies also found that p38γ can phosphorylate Tau and p53 proteins. Although this may result from their interactions, experimental evidence has not been demonstrated.
Tau is a microtubule-associated protein that is hyperphosphorylated in Alzheimer’s disease. Studies showed that p38γ and its family member p38δ can directly phosphorylate Tau at several serine and threonine residues (Goedert et al. 1997). Moreover, p38γ appears to be the most potent MAPK to phosphorylate Tau in intact cells as compared to its family proteins (Buee-Scherrer and Goedert 2002). However, biological consequences of tau phosphorylation by p38γ remain unknown.
p38γ and p38α also phosphorylate the tumor suppressor p53 at Ser33 (Kwong et al. 2009). Although p38γ is required for Ras-induced senescence as well as induction of p21 (a p53 target) in fibroblasts, whether p53/S33 phosphorylation by p38γ plays a role in this process is unclear (Kwong et al. 2009).
p38γ Promotes Pro-Inflammatory Reactions
p38 MAPKs are activated by inflammation and regulate expression of cytokines, inflammatory mediators, and survival genes (Kumar et al. 2003). Among the 4 family proteins, p38γ and p38α are frequently coactivated in response to inflammation (Abdollahi et al. 2003; Korb et al. 2006; Long and Loeser 2010; Tian et al. 2013). Experiments with IEC-specific p38α knockout showed that p38α is inhibitory to inflammation-induced colon tumorigenesis (Otsuka et al. 2010; Wakeman et al. 2012; Gupta et al. 2014). On the contrary, studies with a whole body (Del Reino et al. 2014) and IEC-specific p38γ knockout (Yin et al. 2016) demonstrated a promoting role of p38γ in colon cancer development in the AOM/DSS inflammation mouse model. The required role of p38γ in inflammation-induced tumorigenesis was further demonstrated in a mouse skin cancer model (Zur et al. 2015). Critically, a systemic application of the p38γ inhibitor PFD only blocks the AOM/DSS induced pro-inflammatory cytokine expression and colon-tumorigenesis in wild-type, but not IEC-specific p38γ knockout, mice (Yin et al. 2016). In addition, one recent study showed that neutrophils in myeloid-specific compound p38γ and p38δ double knockout mice are deficient in migration and infiltration in response to the liver metabolism reprograming (Gonzalez-Teran et al. 2016). These results together indicate a critical role of p38γ in inflammation and in inflammation-induced tumorigenesis.
Mechanisms by which p38γ is required for inflammation-induced tumorigenesis are multiple and may be tissue-, cell-, and even stage-specific. p38γ is activated in cells in response to TNF (Cuenda et al. 1997) and is also required for TNF and IL-1 induced activation of NF-kB (Tian et al. 2013). Further, levels of pro-inflammatory cytokines (IL-6, IL-1β, and TNF) are decreased in bone-marrow macrophages derived from whole body p38γ knockout mice after LPS (Risco et al. 2012) and in intestinal tissues of IEC-specific p38γ knockout mice in response to DSS (Yin et al. 2016). Also, treatment of mice with the inflammation stimulus DSS specifically activate p38γ, but not p38α, together with activation of IL-1β, IL-6, and TNF (Yin et al. 2016), and treatment with IL-1β activates both p38α and p38γ in chondrocytes (Long and Loeser 2010). These results indicate an integrated role of p38γ and p38α in feedback loops of pro-inflammatory cytokine signaling. Furthermore, treatment of mice with the p38γ inhibitor PFD reduces TNF-induced shock and decreases IL-6 expression (Cain et al. 1998) and T cell activation (Visner et al. 2009). These results together suggest that p38γ may be important for pro-inflammatory cytokine expression and/or secretion and may collaborate with p38α in fine-tuning coordinated inflammatory response. Targeting p38γ with PFD may have application potentials in prevention of inflammation-induced tumorigenesis.
Other properties of p38γ may also contribute its pro-inflammatory activity. At cellular levels, p38γ is proliferative and/or antiapoptotic (Qi et al. 2007; Loesch et al. 2010; Wu et al. 2010; Kukkonen-Macchi et al. 2011; Hou et al. 2012). Moreover, p38γ is invasive and metastatic (Qi et al. 2006; Loesch et al. 2010; Meng et al. 2011; Rosenthal et al. 2011) and increases CSC expansion (Qi et al. 2015). Furthermore, p38γ promotes differentiation in fibroblasts (Lechner et al. 1996; Gillespie et al. 2009; Zhang et al. 2011) and facilitates glucose transport (Ho et al. 2003), metabolic adaption (Pogozelski et al. 2009), and cell-cycle progression through G2/M phase (Wang et al. 2000). These properties may play a critical role for p38γ pro-inflammatory activity. It should be noted that p38α is a tumor suppressor (Chen et al. 2000; Bulavin and Fornace 2004; Qi et al. 2004; Dolado et al. 2007; Kennedy et al. 2007) and antagonizes p38γ activity (Qi et al. 2007; Loesch and Chen 2008; Lassar 2009). Because p38α and p38γ are coactivated in response to inflammation (Abdollahi et al. 2003; Korb et al. 2006; Long and Loeser 2010; Tian et al. 2013), the resultant p38γ/p38α activity ratio may determine if the integrated response is pro-inflammatory versus anti-inflammatory.
Studies showed several specific properties of p38γ MAPK. Structurally, p38γ is a nonclassical p38 MAPK family member and is the only MAPK that has a unique C-terminal PDZ-binding motif (Cuenda and Rousseau 2007; Hou et al. 2010a). In response to stimuli, p38γ is activated both by elevated RNA and/or protein expression and by increased phosphorylation (Cuenda et al. 1997; Boppart et al. 2000; Chiariello et al. 2000; Franco et al. 2002; Tang et al. 2005; Qi et al. 2006; Qi et al. 2007; Ding et al. 2009; Pogozelski et al. 2009; Qi et al. 2011; Hou et al. 2012; Qi et al. 2012). This property may enable p38γ to have a sustained effect in various biological responses. Furthermore, p38γ is the only MAPK so far that has its own specific phosphatase PTPH1 through PDZ binding (Hou et al. 2010b). This complex may promote K-Ras oncogenesis through a dynamic p38γ-induced PTPH1 phosphorylation and PTPH1-induced p38γ dephosphorylation by a time-, site-, and microenvironment-specific mechanism (Hou et al. 2010b; Hou et al. 2012; Kolch et al. 2015). Although p38γ can signal through several substrates and/or partners, it may only need some of them to trigger a proliferative and/or pro-inflammatory response and to promote inflammation-induced oncogenesis through their integrated activities (Fig. 1). It is important to mention that the p38γ-specific pharmacological inhibitor PFD would be a potent tool for further verification of systemic p38γ activities (Ozes et al. 2008; Moran 2011; Qi et al. 2014; Qi et al. 2015; Yin et al. 2016). Most importantly, PFD is nontoxic and PDA-approved for the treatment of lung fibrosis in clinic (Noble et al. 2011; Richeldi et al. 2011; Schaefer et al. 2011; King et al. 2014). Targeting p38γ by PFD may be a novel strategy for prevention and treatment of inflammation-induced cancer.
The work in Chen lab was supported by grants from National Institutes of Health, Department of Veterans Affair (VA), and Department of Defense (DoD). We would like to acknowledge the former lab members Drs. Jung Tang, Rocky Pramanik, Song-Wang Hou, Mathew Loesch, Adrienne Lepp, Padmanaban S. Suresh, Shao Ma, and Ning Yin for their contributions.
- Ho RC, Alcazar O, Fujii N, Hirshman MF, Goodyear LJ. p38γ MAPK regulation of glucose transporter expression and glucose uptake in Ly myotubes and mouse skeletal muscle. Am J Phys Regul Integr Comp Phys. 2003;286:R342–R9.Google Scholar
- Marinissen MJ, Chiariello M, Pallante M, Gutkind JS. A network of mitogen-activated protein kinases links G protein-coupled receptors to the c-jun promoter: a role for c-Jun NH2-terminal kinase, p38s, and extracellular signal-regulated kinase 5. Mol Cell Biol. 1999;19:4289–301.PubMedPubMedCentralCrossRefGoogle Scholar
- Ozes O, Blatt LM, Seiwert SD. Use of pirfenidone in therapeutic regimens. United States Patent-US 7,407,973 B2. 2008;Aug. 5th:1–46.Google Scholar
- Qi X, Zhi H, Lepp A, Wang P, Huang J, Basir Z, et al. p38γ mitogen-activated protein kinase (MAPK) confers breast cancer hormone sensitivity by switching estrogen receptor (ER) signaling from classical to nonclassical pathway via stimulating ER phosphorylation and c-Jun transcription. J Biol Chem. 2012;287:14681–91.PubMedPubMedCentralCrossRefGoogle Scholar
- Risco A, Fresno C, Mambol A, Alsina-Beauchamp D, MacKenzie KF, Yang HA. p38γ and p38δ kinases regulate the toll-like receptor 4 (TLR4)-induced cytokine production by controlling ERK1/2 protein kinase pathway activation. Proc Natl Acad Sci U S A. 2012;109:11200–5.PubMedPubMedCentralCrossRefGoogle Scholar