MIG-6 and SPRY2 in the Regulation of Receptor Tyrosine Kinase Signaling: Balancing Act via Negative Feedback Loops

2.1. MIG-6

MIG-6(also known as gene 33, ERRFI1orRALT) is a unique and immediate early response gene that is not present in relatively simple organisms like Drosophilaand C. elegans. It emerges in the more complex, higher-order species [9, 11], underlying its importance in evolution. It encodes a 58 kDa nonkinase protein that resides in the cytoplasm and functions as a scaffolding adaptor for modulating signal transduction. Structurally, MIG-6 has several functional motifs/domains that are crucial for interaction with other signaling molecules [9, 12]. The Cdc42/Rac-interaction and binding (CRIB) domain of MIG-6 (Figure 1A) shares consensus sequences with many other proteins that associate with Cdc42 or Rac small GTPases, which are important regulators of actin cytoskeleton remodeling and signal transduction [11, 13]. CRIB domain mediates the binding of MIG-6 to active (GTP-bound) Cdc42 and negatively regulates HGF-induced Cdc42 activation and cell migration [12, 14]. This domain has also been shown to play a role in regulating transactivation of nuclear factor κB (NFκB) by sequestering the inhibitor of κBα (IκBα) [15, 16]. Within its middle region, MIG-6 has several proline-rich motifs that likely mediate its binding to various Src homology-3 (SH3] domain-containing proteins such as GRB2, Src, PI3K p85 and PLC-γ [14, 17]. MIG-6 interacts with 14-3-3ζ via the 14-3-3 protein binding motif [12]. MIG-6 also possesses two PEST sequences, and is targeted by ubiquitination and proteasome degradation [18]. The ErbB-binding region (EBR), a large portion of the carboxyl terminus in MIG-6, is required for physical interaction with EGFR family receptors, which resulted in attenuation of EGFR/ErbBs signaling [17, 19, 20]. The EBR domain shares a high homology with the activated Cdc42-associated kinase 1 (ACK1), a non-receptor tyrosine kinase that also interacts with and regulates EGFR [21].

MIG-6is expressed in many tissues/organs, with high expression in liver and kidney and low to moderate expression in brain, lung, heart, and other tissues [22, 23]. Its expression can be induced by diverse factors ranging from hormones and growth factors, to chemical agents, to stress stimuli [9]. The induction of MIG-6 expression by growth factors is mainly mediated by the RAS-MEK-MAPK/ERK pathway, while other inducers may involve other pathways such as PI3K [9]. MIG-6 has also been reported to be a G-actin-regulated target gene, because the actin-MAL-serum response factor (SRF) cascade mediates MIG-6 induction by serum or lipid agonists such as lysophosphatidic acid (LPA) or sphingosine 1-phosphate (S1P) [24]. MIG-6 may play a crucial role in patho-physiological conditions such as myocardial ischemic injury and infarction [25], liver regeneration [26-28], joint mechanical injury [29], and diabetic nephropathy and hypertension [12]. Its activity is required for skin morphogenesis [30, 31] and lung development in mice [32], and it plays an important role in the maintenance of tissue homeostasis in joints, the lungs and the uterus [32-35].

2.2. SPRY2

In term of evolution, the Sprygene emerged far earlier than Mig-6. SPRY2is the mammalian homolog of Drosophila melanogaster Spry(dSpry), and is one of the four SPRYgenes in the human genome [10]. The dSpry protein is 63 kDa, while its mammalian counterparts are 32–34 kDa, but they all contain a functional cysteine-rich region in their C-terminus (designated the SPRY domain) and an SH2-binding motif carrying a conserved tyrosine at the N-terminus (Tyr55 residue in human SPRY2) (Figure 1B). These conserved regions are essential for SPRY2 to fully execute its inhibitory function in the regulation of RTK signaling [10, 36, 37]. The SPRY domain is also found in the SPRED (Sprouty-related EVH1 domain-containing protein) family, which like SPRY proteins inhibits RTK signaling upon stimulation by growth factors [10, 38]. The SPRY domain mediates the binding of SPRY2 to many signaling molecules including GAP1, FRS2, SHP2, RAF, PKCδ, TESK1 and caveolin1 [10, 39]. The SPRY domain is also required for translocation of SPRY2 to the membrane during its activation. The Tyr55 residue is essential for the SPRY2 protein’s interaction with CBL, PP2A and GRB2 [10, 39]. A cryptic SH3-binding motif (PxxPxR) in the C-terminal end of SPRY2 (but not present in other SPRY members) has been shown to mediate GRB2 interaction as well [39-41]. The dual-specificity tyrosine phosphorylation–regulated kinase (DYRK1A) interacts with and phosphorylates Thr75 on SPRY2 [42]. SPRY2 is targeted for ubiquitination and proteasome degradation by at least two ubiquitin E3 ligases: CBL-mediated Tyr55 phosphorylation-dependent ubiquitination and SIAH2-mediated Tyr55 phosphorylation-independent ubiquitination [43, 44]. On the other hand, SPRY2 protein can be stabilized by phosphorylation of its serines 112 and 121 residues by the mitogen-activated protein kinase-interacting kinase 1 (Mnk1], thereby decreasing growth factor–induced degradation [45].

In Drosophila, dSpryexpression is detected at the tips of branching lung buds and is induced by branchless, the DrosophilaFgf. Losing dSpryleads to excessive tracheal branching as a result of increased Fgf signaling activity [46]. TheXenopushomolog,xSpry2, is expressed in a pattern resembling that of XenopusFgf8 and inhibits Fgf-mediated gastrulation [47]. During mammalian embryogenesis, Spry2 expression tends to localize closest to the sites where Fgf activity is needed for organ/tissue development, underlying the importance of this molecule as an intrinsic Fgf signaling regulator in organogenesis [48-50]. Mouse Spry2is highly expressed in the terminal buds of peripheral mesenchyme in the embryonic lung, adjacent to that of Fgf10, a key mouse lung-branching morphogen [50]. Ectopic expression of Spry2 in the mouse pulmonary epithelium results in decreased branching; exogenous Fgf10 produces greater lung branching and higher Spry2 expression [50]. Spry2 also plays a role in mouse kidney development; its ectopic expression in the ureteric bud leads to postnatal kidney failure due to deficiency in ureteric branching [51]. Moreover, Spry2 deficiency in mice results in defects of the auditory sensory epithelium development in the inner ear, and leads to enteric neuronal hyperplasia and esophageal achalasia [52, 53]. The lack of phenotypes in other Spry2-deficient tissues is likely due to compensatory roles of other Spry family members, because there are overlapping expressions of Spry1, 2 and 4 in many tissues during the development [54]. In adult mice, Spry2 expression is abundant in the brain, lung, heart, kidney, skeletal muscle and mammary glands [48, 55].

Beyond being an intrinsic inhibitor for Fgf signaling, Spry also regulates signaling driven by other RTKs like Egfr. The dSprygene is required for eye and wing development in Drosophila, antagonizing Egfr signaling for neuronal induction in the retina and for vein formation in the wings [56, 57]. Loss of dSpryresults in excess photoreceptors, cone cells and pigment cells in the retina, while its overexpression leads to phenotypes that mimic loss of Egfr signaling [56, 57].

3.1. Regulation of RTK pathways by MIG-6

Many growth factors can induce MIG-6 expression, including EGF, FGF, and HGF [9]. Upon induction, MIG-6 proteins rapidly and transiently accumulate in the cytoplasm, where they feed back to inhibit the activated RTK signaling (Figure 2]. The inhibition of EGFR/ErbB signaling by MIG-6 can occur at two molecular levels: one on the receptor itself, and another on the signaling molecules downstream of the receptor (Figure 2). Through its C-terminal EBR domain, MIG-6 directly binds to EGFR and other ErbB members [17, 19, 20, 58]. The interaction involves the kinase domain of EGFR or ErbB2 and requires their catalytic activities, but does not involve their C-terminal regions in which there are tyrosine residues essential for activating downstream signaling [20, 58]. Crystal structures reveal that the MIG-6 EBR domain binds to the distal surface of the carboxy-terminal lobe (C-lobe) in the kinase domain of EGFR [59]. The C-lobe is crucial in asymmetric EGFR dimer formation [60]; binding of MIG-6 to the C-lobe blocks the dimer interface thereby preventing EGFR activation [59]. MIG-6 coupling also promotes clathrin-mediated endocytosis of EGFR [61], an important mechanism for timely attenuation of ligand-induced EGFR activation [62, 63]. The region responsible for promoting EGFR endocytosis has been mapped to the endocytic domain (ED) of MIG-6 (see Figure 1A) [61], which mediates the binding of MIG-6 to the AP2 adaptor complex, a key component in forming clathrin-coated pits during endocytosis [63]. Interestingly, the non-receptor tyrosine kinase ACK1 also binds to EGFR upon EGF stimulation through a region sharing high homology with the MIG-6 EBR domain [21]. This interaction also regulates clathrin-mediated EGFR endocytosis and degradation [21, 64]. However, it is not clear whether MIG-6 and ACK1 cooperate in regulating EGFR turnover or whether they bind to EGFR in a mutually exclusive way to accomplish individual inhibitory roles under different circumstances. The internalized EGFR is guided to late endosome through the binding of MIG-6 to the endosomal SNARE complex component STX8, en route to degradation in the lysosome [61, 65].

The magnitude of MIG-6-mediated inhibition of EGFR signaling is likely maximized and integrated by the second level of molecular regulation, that is, direct inhibition of downstream signaling molecules (Figure 2). MIG-6 binds to the SH3 domain-containing protein GRB2 [14, 17], the key molecule in linking activated RTK to the intracellular signaling cascade; GRB2 brings RAS together with SOS to the phosphorylated receptor for activation [66]. Binding of MIG-6 may block GRB2’s interaction with activated EGFR and other RTKs or disrupt GRB2-SOS-RAS complex formation, thereby preventing activation of the RAS-MEK-MAPK pathway. MIG-6 also interacts with Src, the PI3K p85 subunit, PLCγ, and Fyn [17], although those interactions were demonstrated in an artificial system and their biological meaning remains to be determined. More complexity is likely in the dynamics of MIG-6-mediated RTK signaling regulation due to the interaction of MIG-6 with 14-3-3 proteins [12, 67], an adaptor family that may interact with diverse signaling molecules and regulate many biological activities [68]. Nonetheless, most of these direct downstream signaling regulations by MIG-6 remain largely speculation and require further investigation.

Assessing the effect of MIG-6 on RTK pathways other than those of the EGFR family, however, may provide insightful answers to such speculation, because direct regulation of the RTK itself by MIG-6 appears to be unique to the EGFR family. For instance, MIG-6 can be induced by HGF and function as negative feedback regulator of the MET pathway by inhibiting HGF-induced cell migration and proliferation [14], yet no physical interaction between MIG-6 and MET has been observed. Through its CRIB domain, MIG-6 can bind to and inhibit the activity of the Cdc42 small GTPase [12, 14], and this inhibitory activity is required for blocking of HGF-induced cell migration [14]. The CRIB domain also interacts with IκBα, thereby activating NFκB for transcriptional regulation of its target gene expression [15, 16]. Whether the inhibition of HGF-induced cell proliferation by MIG-6 is mediated by its ability to bind GRB2 or to bind other downstream molecules is still unknown. Negative feedback inhibition of other RTKs (such as FGFR and IGFR) by MIG-6 is also likely to be mediated by its inhibitory activities on the downstream signaling molecules rather than on the receptor itself.

The inhibitory activity of MIG-6 on RTK signaling seems to be modulated by phosphorylation; Liu et al. recently reported that MIG-6 can be phosphorylated by Chk1 upon EGF stimulation [67]. EGF activates Chk1 via the PI3K-AKT-S6K pathway, which in turn phosphorylates Ser251 of MIG-6 and results in inhibition of MIG-6 [67]. Thus, Chk1 counter-balances the EGFR inhibition of MIG-6, positively regulating EGFR signaling. Interestingly, Ser251 resides in the 14-3-3 binding motif of MIG-6 and its phosphorylation is likely involved in the MIG-6 and 14-3-3ζ interaction, because that interaction is abolished by Chk1 depletion [67]. In addition, two tyrosine residues (Tyr394 and Tyr458) in MIG-6 are phosphorylated by EGFR activation [69-71], but the underlying mechanism and the biological significance remain unclear.

3.2. Regulation of RTK pathways by SPRY2

SPRY2 renders another layer of modulation on RTK signaling via a negative feedback loop [10, 36, 37, 72]; its expression is induced by many activated RTKs including EGFR, FGFR, MET, and VEGFR [10, 72]. As with MIG-6, SPRY2-mediated regulation can occur on two levels: on the RTK itself and on the downstream signaling molecules (Figure 3). However unlike MIG-6, the most prominent inhibitory activity of SPRY2 is derived from its abilities to interact with downstream signaling molecules centering on the RAS-RAF-MAPK pathway, while its effect on the RTK itself seems to be indirect and may provide some signaling specificity for different RTKs [10, 36, 37].

Upon growth factor stimulation, SPRY2 translocates from the cytosol to the plasma membrane where it binds to phosphatidylinositol-4,5-bisphosphate [PtdIns(4,5)P2] via the conserved SPRY domain [73-75]. Membrane translocation appears to be triggered by the activated Rac small GTPase and is essential for SPRY2 function [76]. The phosphorylation of the Tyr55 residue in SPRY2 is likely regulated by a SRC family kinase, which upon FGF stimulation is recruited to activated FGFR by FRS2 [77]. While Tyr55 phosphorylation enables the binding of SPRY2 to GRB2 via the SH2-binding motif in the N-terminus [78], the SH3-binding motif in its C-terminus may also play a role in the SPRY2-GRB2 interaction [39-41] (Figure 1B). The latter is likely regulated by the protein phosphatase 2A (PP2A), a serine/threonine phosphatase that interacts with Tyr55 and dephosphorylates certain serine residues in SPRY2 for permitting access of GRB2 to the SH3-binding motif on SPRY2 [41, 45]. Consequently, the binding of SPRY2 to GRB2 prevents the recruitment of the GRB2-SOS complex to the FRS2 adaptor or SHP2 phosphatase proximal to the activated RTK, thereby inhibiting the activation of RAS and its downstream molecules [10, 36, 39, 77, 78]. The RTK-RAS-RAF-MAPK/ERK pathway can also be inhibited by direct binding of SPRY2 to RAF [75, 79, 80]. On the other hand, TESK1 negatively regulates SPRY2 inhibitory activity by interfering with the SPRY2 interaction with GRB2 and PP2A [81], while DYRK1A binding results in Thr75 phosphorylation on SPRY2 and suppression of SPRY2 inhibitory activity, thereby enhancing FGF-induced ERK activation [42].

The regulation of EGFR signaling by SPRY2 appears to be more complicated than that of FGFR signaling, because SPRY2 can indirectly influence the turnover of EGFR through its interaction with CBL (Figure 3) [10, 36, 37]. This regulation is mediated by the SH2-binding motif, which includes Tyr55 in the N-terminus of SPRY2. When Tyr55 is phosphorylated, SPRY2 interacts with the SH2-domain on CBL, and prevents CBL from binding to the activated EGFR, thereby interfering with CBL-mediated EGFR endocytosis and degradation [43, 82-84]. This action can prolong the activation of EGFR and the downstream RAS-RAF-MAPK pathway, contrary to the direct inhibitory activity of SPRY2 on the downstream signaling molecules. These two opposite activities render SPRY2 a delicate role in fine-tuning EGFR signaling, negative in some situations and positive in others, depending on the threshold and balance of these two activities.

It is conceivable that RTKs such as MET and PDGFR that are also substrates of CBL might also be affected by SPRY2 like that of EGFR [8, 85], while non-CBL-substrate RTKs like FGFR appears unaffected by SPRY2-CBL interaction [36]. However, it is known that MET protein level is not affected by SPRY2 overexpression that inhibits HGF-induced ERK and AKT activation, indicating that the effect of SPRY2-CBL interaction on EGFR and on MET might not be the same [72, 86]. The sequestration of CBL by SPRY2 can also affect downstream signaling molecules, as it may free proteins like GRB2 from CBL-mediated ubiquitination and degradation [85]. Furthermore, SPRY2 itself can be ubiquitinylated and degraded by the CBL binding, thereby influencing the RTK signaling [83].

4.1. MIG-6 as a tumor suppressor gene

Human chromosome 1p36, a locus frequently associated with many human cancers [87-92], harbors the MIG-6gene. In fact, loss or reduction of MIG-6expression has been observed in non-small cell lung cancer (NSCLC) [35, 93-95], breast carcinoma [30, 96], melanoma and skin cancer [30, 94], ovarian carcinoma [30], pancreatic cancer [30], endometrial cancer [34], thyroid cancer [97, 98], hepatocellular carcinoma [28], and glioblastoma [91]. Prognostically, low MIG-6expression is often associated with poor prognosis or poor patient survival [93, 96, 97].

Unlike many other tumor suppressor genes whose expression is directly regulated by epigenetic modification of their promoter regulatory elements [3], the silencing of MIG-6expression seems otherwise [94]. Although high in CpG contents, the MIG-6promoter appears hypomethylated and is not directly affected by either the DNA methyltransferase inhibitor 5-aza-2’-deoxycytidine [5-aza-dC) or the histone deacetylase inhibitor trichostatin A (TSA), indicating indirect regulation [94]. Interestingly, MIG-6expression seems to be differently induced by 5-aza-dC and TSA in different cancer types: it is induced by 5-aza-dC in melanoma but not in NSCLC and neuroblastoma, and by TSA in NSCLC and neuroblastoma but not in melanoma, suggesting a possible tissue-specific transcriptional regulation for this gene [90, 94, 96]. However in some papillary thyroid cancer, it is reported that the MIG-6promoter is hypermethylated as determined by methylation-specific PCR [98]. It is unclear at this point what cause such differences.

Besides loss or reduction of expression, MIG-6can also be inactivated by genetic mutation, even though this occurs rarely [35, 90, 96]. To date, two homozygous mutations (Asp109 to Asn, and Glu83 to a stop codon) in MIG-6were identified in human lung cancer cell lines, while heterozygous germline mutations were found in a primary lung cancer (Ala373 to Val) and a neuroblastoma patient (Asn343 to Ser) [35, 90]. Evidence that MIG-6is a bona fide tumor suppressor gene also arises from mouse studies. Mice with targeted disruption of Mig-6are prone to neoplastic development ranging from epithelial hyperplasia to carcinoma at multiple sites including the lung, gallbladder, bile duct, uterus, gastrointestinal tract and skin [30, 34, 35]. The carcinogen-induced skin cancer seen in Mig-6-deficient mice is likely mediated by EGFR-ERK/MAPK pathway, because inhibiting EGFR kinase activity with gefitinib or replacing it with a kinase-defective EGFR rescues the phenotype [30].

4.2. SPRY2 as a tumor suppressor gene

There is compelling evidence supporting SPRY2 as a tumor suppressor [99]. The SPRY2gene is located on human chromosome 13q31, where loss of heterozygosity (LOH) is observed in prostate cancer and hepatocellular carcinoma [86, 100]. Down-regulation of SPRY2expression has been reported in breast cancer [55, 101], hepatocellular carcinoma [86, 102, 103], NSCLC [104], prostate cancer [100, 105, 106], endometrial carcinoma [107], gliomas [108], and B-cell lymphomas [109, 110]. However in colon cancer, both downregulation and upregulation of SPRY2have been reported [111, 112]. Low (or no)SPRY2expression is associated with advanced tumor stages and poor survival, and it may be a significant prognostic factor in breast cancer [101], hepatocellular carcinoma [86, 103], prostate cancer [100, 105], gliomas [108], and colon cancer [111]. Further, SPRY2expression is inversely correlated with the level of miR-21 microRNA expression in gliomas and colon cancer, indicating that SPRY2is a target of miR-21[108, 111].

Downregulation of SPRY2 expression may be attributed to DNA methylation; hypermethylation of its promoter has been observed in prostate cancer [100], hepatocellular carcinoma [86], endometrial carcinoma [107] and B-cell lymphomas [109, 110]. However, controversial results have also been reported, since no methylation in SPRY2promoter was found in other studies involving different cancer types [55, 102, 106, 111], indicating that other epigenetic mechanisms might as well be responsible for SPRY2down-regulation. Thus far, no mutation has been identified in the SPRY2gene in any human cancers, and no neoplastic phenotypes have been observed in any Spry2-deficient mice. This may be due to compensatory roles played by other family members such as SPRY1 or SPRY4.

Acknowledgments