P-Rex1 (PIP3-dependent Rac exchanger 1, PREX1) is a Dbl-type guanine-nucleotide exchange factor (GEF) that activates Rac-like small G proteins (small GTPases). P-Rex1 was discovered in 2002, during a search for targets of the lipid second messenger phosphatidyl inositol (3,4,5)-trisphosphate (PIP3) which is generated by phosphoinositide 3-kinase (PI3K) (Welch et al. 2002). At the time, it was known that PI3K can activate Rac and that this activation must occur indirectly, through unknown mediators. P-Rex1 was purified on the basis of its ability to activate recombinant Rac1 in the presence of PIP3 (Welch et al. 2002). The Gβγ subunits of heterotrimeric G proteins, which are released upon the activation of G protein-coupled receptors (GPCRs), were also known to be powerful but indirect activators of Rac in neutrophils. This led to the discovery that P-Rex1 can be also activated by Gβγ subunits and that PIP3 and Gβγs can activate P-Rex1 synergistically, making this Rac-GEF a coincidence detector for cell signaling through PI3K-coupled receptors and GPCRs (Welch et al. 2002).
Since these early discoveries, several other mechanisms of P-Rex1 regulation have been discovered. Knockdown, knockout, and mutational strategies have served to elucidate the varied physiological roles of P-Rex1 in leukocytes, platelets, and endothelial cells during the inflammatory response, in the migration of melanoblasts during development, in the thermogenic potential of adipose tissue, and in social recognition. The evaluation of cancer patients and of murine tumor models has shown that deregulation of P-Rex1 expression occurs in many types of cancer and promotes tumor growth and metastasis. Several detailed reviews have described these discoveries, including our own recent review (Welch 2015). Please refer to this review for all original papers and reviews of the P-Rex literature published up to mid-2014. This allows us to focus here on the most recent discoveries.
Gene and Protein
Structure and Function
The P-Rex1 gene is conserved throughout vertebrates. In the human genome, PREX1 (NM_020820) is located on chromosome 20q13.13, near a region associated with type 2 diabetes. PREX1 encodes the 185 kDa protein P-Rex1 (NP_065871). The possibility of splice variants has been proposed but requires further corroboration. The P-Rex1 gene has one homologue, P-Rex2, and both are found throughout the vertebrate kingdom.
Lucato et al. recently reported the first crystal structure of the P-Rex1 DHPH domain tandem in complex with constitutively active Rac1, at a resolution of 1.95 Å (Lucato et al. 2015). The P-Rex1 DH domain is composed of six main α-helices, arranged into an elongated bundle, as is typical of Dbl-family Rho-GEFs. The extended helix α6 bridges both the DH and PH domains and allows significant flexibility between these domains, which is also common in Dbl-family Rho-GEFs. This arrangement effectively orientates the PH domain away from the DH domain. Similar structures were recently reported by Cash et al., who also crystalized the P-Rex1 DHPH domain tandem in complex with nucleotide-free Rac1 or Cdc42, at a resolution of 3.3 Å and 3.2 Å, respectively (Cash et al. 2016) (Fig. 1b).
These crystal structures showed that the P-Rex1 DH domain alone confers Rac1 activation (Lucato et al. 2015; Cash et al. 2016). The DH domain makes multiple contacts with Rac1, whereas the PH domain makes none. Among three highly conserved regions of the P-Rex1 DH domain (CR1–3), CR1 and CR3 make extensive contacts with both the switch 1 and switch 2 regions of Rac1. Interactions with switch 1 are formed by Glu56 in CR1 and by Gln197 in CR3, both of which are highly conserved. Interactions with switch 2 are formed by Lys201, Asn238, and Arg242. No contacts occur between P-Rex1 and the Rac1 P-loop, which is consistent with the catalytic domain structures of other Rac-GEFs. Furthermore, Glu56 and Asn238 were confirmed to be essential for catalytic activity and Rac1 binding (Fig. 1b). These residues had previously been predicted to be important, and Glu56/Asn238 mutants are thus already in widespread use as “GEF-dead” forms of P-Rex1.
Through its catalytic GEF activity, P-Rex1 promotes the release of GDP from Rac, allowing excess free cellular GTP to bind and thus inducing the active conformation of Rac that is able to engage its downstream target proteins. In this manner, P-Rex1 controls a wide range of Rac-dependent cellular functions, including responses that depend on the structure of the actomyosin cytoskeleton, such as cell adhesion and migration, but also reactive oxygen species (ROS) production and gene expression (Welch 2015).
P-Rex1 can activate all Rac-like small G proteins (the ubiquitous Rac1, hematopoietic Rac2, neuronal Rac3, and widely expressed RhoG). P-Rex1 can also activate Cdc42-like GTPases in vitro, but not other Rho-family GTPases. (Please see references in Welch (2015) for further reading on substrate specificity). The recent crystal structures of the P-Rex1 DHPH domains in complex with Rac1 or Cdc42 showed that P-Rex can accommodate both these GTPases in vitro. Thus, it remains unclear why P-Rex1-dependent activation of Cdc42 has not yet been seen in vivo. In contrast, modeling of P-Rex1 and RhoA suggested that steric hindrance would not allow RhoA binding, which explains why P-Rex1 cannot activate this GTPase (Cash et al. 2016).
Regulation of Expression
The original purification of P-Rex1 showed that the Rac-GEF is highly abundant in neutrophils, where it makes up 0.1% of the cytosolic protein (Welch et al. 2002). P-Rex1 is also expressed in other types of leukocytes as well as in platelets, endothelial cells, and neurons and at lower levels in many other cell types. Tissue-wise, P-Rex1 is expressed widely throughout the brain and is also present in bone marrow, thymus, spleen, lymph nodes, and lung.
The Tu lab showed in 2011 that P-Rex1 expression can be driven by the binding of transcription factor Sp1 to the PREX1 promoter, which is situated 190–198 base pairs 5′ of the coding region. In normal prostate epithelial cells, this transcription was repressed epigenetically by the binding of histone deacetylases (HDACs) to the promoter and the deacetylation of promoter-associated histone H4. In metastatic prostate cancer cells, this HDAC-mediated repression was lost, thus resulting in increased H4 acetylation and deregulated P-Rex1 expression. It was proposed that a locus-specific increase in H4 acetylation induces an open chromatin conformation which favors PREX1 transcription. It remains to be shown how the association of Sp1 and HDACs with the locus is controlled. A further mechanism of P-Rex1 expression was recently discovered that causally links P-Rex1 deregulation to drug resistance of prostate carcinoma cells. In drug-resistant cells, the transcription factor Myc was upregulated, and binding of Myc to a consensus site in the PREX1 promoter induced PREX1 transcription (see also section 5) (Goel et al. 2016).
A recent study by the Kazanietz lab identified hypermethylation of CpG islands in the promoter and transcriptional start regions of PREX1 as a major mechanism to suppress P-Rex1 expression in normal breast epithelial cells, together with regulation by HDACs (Barrio-Real et al. 2014). Treatment of mammary epithelial cells with a demethylating agent and HDAC inhibitors was sufficient to drive P-Rex1 expression. PREX1 promoter methylation was found to be reduced in luminal breast tumors and was associated with ER-positive status. Importantly, hypomethylation was also correlated with high P-Rex1 expression and with poor long-term survival of breast cancer patients (Barrio-Real et al. 2014).
Full-length P-Rex1 has a low basal catalytic Rac-GEF activity and is activated by through the release of intramolecular inhibition, which can be brought about by several mechanisms in response to cell stimulation. This mode of activation was first proposed in a P-Rex1 mutagenesis study by Hill et al. in 2005 and was supported by the recent crystal structures of the DHPH domain tandem (Lucato et al. 2015; Cash et al. 2016). Please also see reference (Welch 2015) for a detailed review and citations of the original literature up to 2014 on the mechanisms that regulate the catalytic activity and subcellular localization of P-Rex1.
PIP3 and Gβγ
P-Rex1 can be directly activated by PIP3 and Gβγ, either independently or synergistically. Synergistic activation allows for coincidence detection of signaling through PI3K-coupled receptors and GPCRs, a unique feature of the P-Rex family that enables the integration of highly complex signaling pathways. P-Rex1 Rac-GEF activity is stimulated more than 30-fold by PIP3 (EC50 1.5 μM) and with similar efficacy by Gβγ subunits, which bind full-length P-Rex1 at 1:1 stoichiometry (KD 0.3 μM).
Hill et al. first showed that PIP3 binds to the PH domain of P-Rex1. This was confirmed by a crystal structure of the P-Rex1 PH domain, which was solved in complex with the soluble PIP3 analogue inositol-1,3,4,5-tetraphosphate (IP4) at a resolution of 2 Å (Cash et al. 2016). The PH domain was shown to form a seven-stranded antiparallel β-sandwich, capped at one end by a C-terminal α-helix, whereas the other end is open and features loops that are typically involved in ligand binding or membrane interactions. IP4 bound within the open end of the β-sandwich, in a pocket lined with basic residues that tightly coordinate the 3- and 4-phosphates of the inositol ring. Deep within the binding pocket, Lys280 interacted with the 3- and 4-phosphates, Arg289 with the 3-phosphate, and Lys368 with the 4- and 5-phosphates of IP4. Accordingly, mutations of Lys280, Arg289, and Lys368 all reduced IP4 binding (Cash et al. 2016). The affinity of IP4 for the PH domain was KD 0.4 μM, whereas a soluble analog of phosphatidyl inositol (3,4)-bisphosphate (PI(3,4)P2) bound to a lesser extent (Cash et al. 2016). This confirmed the previous observation that although P-Rex1 can bind both PIP3 and PI(3,4)P2, only PIP3 can stimulate the GEF activity (Welch et al. 2002).
A study by the Garrison lab in 2006 showed that a range of different Gβ and Gγ protein dimers can stimulate P-Rex1 Rac-GEF activity, with the exception of Gβ5γ2. The potency with which the various Gβγ combinations could activate P-Rex1 matched their effects on another Gβγ effector, PI3Kγ. We showed that Gβγs activate P-Rex1 primarily through the DH domain in vitro, and the Itoh lab found that other P-Rex1 domains contribute to Gβγ binding in vivo (Welch 2015). The Gβγ binding site was recently modeled on the crystal structures of P-Rex1:Rac1 and GRK2:Gβγ complexes, which suggested that a negatively charged surface patch of P-Rex1 could accommodate Gβγ with few steric clashes (Lucato et al. 2015). In this model, Gβγ would contact both the DH and PH domains, sit on the opposite side of P-Rex1 to Rac1, and not make any contacts with the GTPase (Fig. 1b).
The Garrison lab first showed in 2006 that cAMP-dependent kinase (PKA) inhibits P-Rex1 by blocking Gβγ binding and the activation by PIP3. A recent study identified the PKA binding site and one of the PKA target sites (Fig. 2). The regulatory subunit of PKA was shown to interact with the PDZ1 domain of P-Rex1 through its C-terminal cAMP-binding domain, and the catalytic domain of PKA was found to phosphorylate Ser436 in the DEP1 domain (Chavez-Vargas et al. 2016). In endothelial cells, PKA inhibited P-Rex1-dependent Rac1 activity and cell migration, whereas an S436A mutant of P-Rex1 was insensitive to PKA inhibition (Chavez-Vargas et al. 2016). Furthermore, the DEP1 domain of P-Rex1 was shown to interact with the isolated DHPH domain tandem, whereas a DEP1-S436A mutant could not, suggesting that PKA-dependent phosphorylation of Ser436 confers an inhibitory intramolecular interaction. Similarly, PKA also strengthened the inhibition of the DHPH core by the C-terminal half of P-Rex1, indicating that additional PKA target residues may lie in the C-terminus (Fig. 2). Interestingly, phosphorylation of Ser436 did not affect the interaction of P-Rex1 with Gβ1 (or mTOR) (Chavez-Vargas et al. 2016), suggesting that PKA regulates Gβγ binding through a separate mechanism. Finally, it is unlikely that PKA is the kinase which counteracts the activation of P-Rex1 by PP1α, because the main PP1α target residue, Ser1165, does not lie within a PKA consensus sequence.
Recent work identified phosphorylation by p21-activated kinase (PAK) as a further mechanism to inhibit P-Rex1 (Barrows et al. 2016). PAKs were shown to phosphorylate P-Rex1 in response to stimulation of receptor tyrosine kinases (RTKs), PAK2 could phosphorylate P-Rex1 directly in vitro, and phosphorylation by PAK1 reduced the PIP3-stimulated Rac-GEF activity of P-Rex1 in a broken-cell assay (Barrows et al. 2016). As PAKs are activated by Rac1-GTP, the existence of a negative feedback loop to limit P-Rex1 activity was proposed (Barrows et al. 2016). The PAK target residue in P-Rex1 remains to be identified, but the Parsons lab recently found Ser1107 in the P-Rex1 homologue P-Rex2 to be phosphorylated by PAK. Interestingly, the Pandiella lab had previously demonstrated that phosphorylation of the equivalent site in P-Rex1, Ser1169, by unknown kinases in breast cancer cells is associated with activation rather than inhibition.
The Pandiella lab had also identified that phosphorylation of Ser313 limits P-Rex1 Rac-GEF activity. Interestingly, the crystal structure of the P-Rex1 DHPH domain revealed that Ser313 lies within the β3/β4 loop of the PH domain, which is thought to be highly mobile. On this basis, Lucato et al suggested that phosphorylation of Ser313 might enable the β3/β4 loop to interact with the DH domain to sterically inhibit Rac1 binding (Lucato et al. 2015). Apart from PAK, further kinases and phosphatases that modulate P-Rex1 phosphorylation in response to RTK stimulation remain to be identified. mTOR was investigated by the Pandiella lab as a possible candidate in breast cancer cells and was largely ruled out. Furthermore, PKCδ and Akt were shown by the Ye and Bae labs to activate P-Rex1 in various cell systems, but it is unknown if the effects of these kinases are direct. In addition, various reports suggested that Akt can be a downstream target as well as an upstream regulator of P-Rex1 (e.g., in breast cancer).
P-Rex1 is largely cytosolic in basal cells and must translocate to the plasma membrane in order to activate Rac upon cell stimulation. PIP3 and Gβγ are both membrane-bound signals. Early studies by us and by the Bokoch lab showed that these signals synergistically mediate the membrane translocation of P-Rex1, as well as its activation, although PIP3 alone appears to be sufficient in some situations (Fig. 2). Barber et al. showed that membrane-derived purified P-Rex1 is more active than the cytosolic protein. Furthermore, membrane-derived P-Rex1 exhibits different mobility on gels compared to the cytosolic protein. As similar gel mobility shifts are seen upon phosphatase treatment, this suggested that P-Rex1 phosphorylation may contribute to subcellular localization. Finally, we recently identified the GPCR adaptor protein Norbin as a further regulator of P-Rex1 subcellular localization. Norbin could bind and stimulate the Rac-GEF activity of P-Rex1 directly, but the biggest effect of co-expressing the two proteins was a robust membrane translocation of both. We proposed that Norbin-dependent membrane translocation brings P-Rex1 into closer proximity with its activators PIP3 and Gβγ, thus stimulating the GEF activity (Pan et al. 2016).
Apart from Rac and the regulators described above, few binding proteins of P-Rex1 have been identified. Only proteins known to bind P-Rex1 directly are considered here. For further reading on P-Rex1 interactions, please see reference (Welch 2015).
The Vazquez-Prado lab identified in 2007 that the serine kinase mTOR binds P-Rex1 directly, through the DEP domains. P-Rex1 interacts with both mTOR-containing protein complexes, TORC1, which is central in cell growth, and TORC2 (also known as PDK2), which controls cell morphology and migration (Fig. 2). More research is required to elucidate if P-Rex1 can regulate both TORC1- and TORC2-dependent cellular processes and to build on evidence by the Vazquez-Prado and Bae labs, which suggests that P-Rex1 can signal both upstream and downstream of mTOR.
The Malliri lab recently performed a SILAC screen to identify proteins that interact with Rac1 specifically in the presence of P-Rex1. This identified the actin remodeling protein FLII (flightless-1 homolog), a member of the gelsolin superfamily, as a P-Rex1 binding protein (Marei et al. 2016a). P-Rex1 bound directly to FLII, independently of its Rac-GEF activity. P-Rex1 interacted with the C-terminal GEL domain of FLII, whereas Rac1 bound to the N-terminal LRR domain of FLII. In addition, the actin capping protein TMOD3 was shown to bind Rac1 and FLII more strongly in the presence of P-Rex1 (Marei et al. 2016b). Importantly, P-Rex1 enabled the interaction between Rac1 and FLII in fibroblasts preferentially when Rac1 was GTP-loaded, and the ability of P-Rex1 to induce cell migration and myosin contractility required FLII. Based on these findings, a model was proposed where P-Rex1 activates Rac1 and acts as a scaffold that enables FLII to interact with this active Rac1, thus inducing cell responses that require both Rac1-GTP and FLII (Marei et al. 2016a).
Leukocytes, Platelets, and Inflammation
Knockdown of P-Rex1 in neutrophil-like NB4 cells provided the first evidence that P-Rex1 is important for myeloid cell function (Welch et al. 2002), and this was corroborated later in Prex1 −/− mouse strains generated by us and by the Wu lab (please see also review Welch 2015). Prex1 −/− neutrophils have defects in toll-like receptor 4 (TLR4)-dependent priming and in GPCR-dependent Rac2 and RhoG activity, ROS production, F-actin polymerization, polarity, and migration. The migration defect comes from a reduction in cell speed (chemokinesis), whereas the directionality of migration (chemotaxis) is normal. The Zarbock lab showed that P-Rex1 mediates the E-selectin-dependent activation of the neutrophil integrin LFA1 under flow conditions, thus controlling the slow rolling of neutrophils along the endothelial vessel wall, as well as the activation of the integrin Mac1, to regulate neutrophil crawling along the vessel wall. In vivo, Prex1 −/− mice show impaired neutrophil and macrophage recruitment during sterile and septic inflammation (please see references in Welch (2015) for more details).
Neutrophils from mice deficient both in P-Rex1 and the Vav-family Rac-GEF Vav1 have more profound defects in GPCR-dependent Rac activity, ROS formation, adhesion, and migration than cells which lack either the whole P-Rex or the whole Vav family. This suggested that P-Rex1 and Vav1 can cooperate to generate robust levels of Rac activity in neutrophils. We showed recently that Prex1 −/− Vav1 −/− and Prex1 −/− Vav3 −/− mice also have more profound impairments in neutrophil recruitment than mice lacking either GEF family (Pan et al. 2015). Intravital imaging revealed that this recruitment defect is caused by the loss of L-selectin- and E-selectin-dependent neutrophil adhesion to the postcapillary endothelial microvasculature prior to extravasation. Surprisingly, this adhesion defect was largely caused by platelets, as P-Rex1/Vav deficiency in platelets was sufficient to block neutrophil adhesion to the vessel wall. Prex1 −/− Vav1 −/− and Prex1 −/− Vav3 −/− platelets had low surface levels of the selectin-ligand PSGL-1, and the mice showed a reduced occurrence of platelet-neutrophil adhesion in the circulation, which is prerequisite for leukocyte extravasation (Pan et al. 2015). Furthermore, platelet P-Rex1 and Vav were also important for the recruitment of other types of inflammatory cells. During allergic inflammation, the pulmonary recruitment of eosinophils, monocytes and lymphocytes was compromised by Prex1 −/− Vav1 −/− or Prex1 −/− Vav3 −/− platelets, and airway inflammation was essentially abolished in Prex1 −/− Vav1 −/− and Prex1 −/− Vav3 −/− mice, resulting in improved airway responsiveness (Pan et al. 2015). In contrast work by the Ye lab in 2012 showed that Prex1 −/− mice only have a mild defect in blood clotting, and Prex1 −/− platelets show partial impairments in GPCR-dependent aggregation and dense granule secretion. Together, these studies suggest therefore that P-Rex1 plays a preferential role in inflammatory rather than hemostatic platelet functions.
Recently, the Ye lab showed P-Rex1 to be required for pulmonary fibrosis, a late phase of pulmonary inflammation which can result in loss of lung function. In a model of pulmonary fibrosis, both early leukocyte infiltration and the development of fibrosis were reduced in Prex1 −/− mice, which drastically increased survival (Liang et al. 2016). The lungs of Prex1 −/− mice showed reduced accumulation of extracellular matrix proteins and other markers of fibrosis, as well as impaired production of the cytokine TGFβ1, which is required for disease development. Finally, primary fibroblasts isolated from the lungs of Prex1 −/− mice showed reduced TGFβ1-induced signaling and cell migration (Liang et al. 2016).
In summary, P-Rex1 plays many different roles in inflammation. An important question for future inflammation research is whether deregulation of P-Rex1 expression occurs in human immune-deficiencies or inflammatory conditions.
Endothelial Cells and Vascular Biology
Control of endothelial cell function by P-Rex1 was first demonstrated in 2010 by the Vázquez-Prado lab who showed that knockdown in human microvascular endothelial cells (HMEC) inhibits SDF1-stimulated Rac1 activity, chemotaxis and in vitro angiogenesis. Since then, the Ye lab found that P-Rex1 also contributes to the temporary decrease in barrier function that the vascular endothelium undergoes under inflammatory conditions. This was demonstrated by P-Rex1 knockdown in human lung microvascular endothelial cells (HLMVEC), which increased endothelial barrier function through effects on Rac1 activity, VE-cadherin phosphorylation and ROS formation. TNFα-stimulated “transmigration” of neutrophils across HLMVEC monolayers with P-Rex1 depletion suggested that endothelial P-Rex1 may regulate leukocyte recruitment. In addition, the Voorberg lab showed in primary human umbilical vein endothelial cells (HUVECs) that endothelial P-Rex1 can also mediate Weibel-Palade body secretion, a process required for the upregulation of P-selectin on the endothelium and for the capture of leukocytes from the blood stream during inflammation. Thus, endothelial P-Rex1 appears to have roles in inflammation and angiogenesis that merit further study in vivo.
Neurons and Behavior
P-Rex1 is widely expressed throughout the nervous system. Its functional roles in neurons were first investigated in 2005 by the Hoshino lab, who used knockdown and a dominant-negative ΔDH mutant of P-Rex1 in PC12 cells to inhibit NGF-stimulated Rac1 activity, lamellipodia formation, membrane ruffling, cell spreading and migration. More recently, the Vanderhaeghen lab used ectopic expression of a ΔDH mutant in mouse embryos to suggest that P-Rex1 controls the ephrin-B1 dependent migration of pyramidal neurons within the cortex during development. However, it should be emphasized that Prex1 −/− mice show no overt defects in cerebral development. We also reported in 2008 that Prex1 −/− mice have normal cerebellar morphology and synaptic plasticity, as well as normal motor behavior. However, our studies did show that P-Rex1 contributes to the plasticity of Purkinje neurons, as Prex1 −/− Prex2 −/− mice have an exacerbated impairment in cerebellar long-term potentiation and motor coordination compared to Prex2 −/− mice. Please see (Welch 2015) for further reading.
An extraordinary recent study found SNPs, copy number deletions and reduced mRNA levels of P-Rex1 in children with autism spectrum disorders (Li et al. 2015). Evaluation of Prex1 −/− mice in behavioral models of autism revealed deficits in social recognition, reversal learning and fear extinction. Acute knockdown of P-Rex1 in the hippocampal CA1 region of young wild-type mice resulted in similar behavioral defects as seen in Prex1−/− mice, suggesting that P-Rex1 expression in the hippocampus is required for social interaction and flexible behaviors (Li et al. 2015). Furthermore, NMDAR-induced long term depression (LTD) was shown to be impaired in in the hippocampal CA1 region of Prex1 −/− mice. Remarkably, the impairments in NMDAR-dependent hippocampal LTD, social recognition and behavioral flexibility seen in Prex1 −/− mice could be rescued by the overexpression of P-Rex1 in pyramidal neurons of the CA1 region of the hippocampus (Li et al. 2015).
This study also provided important mechanistic insight: PPI phosphatases are important mediators of NMDA-stimulated hippocampal LTP, and as PP1α can activate P-Rex1 directly (see above), the interaction of PP1α with P-Rex1 in the hippocampus was investigated. Indeed, NMDA stimulation was found to increase the interaction of PP1α with P-Rex1 in hippocampal slices from the CA1 region (Li et al. 2015). Furthermore, hippocampal plasticity requires the internalization of the glutamate receptor GluR2 following NMDA stimulation. Hence, the subcellular localization of GluR2 was assessed in NMDA-treated hippocampal neurons, and P-Rex1 deficiency was found to impair the internalization of GluR2. This internalization could be rescued by the re-expression of wild-type P-Rex1 or Rac1, but neither by a P-Rex1 mutant that cannot interact with PP1α (VAFA mutant), nor by catalytically inactive P-Rex1. Thus, the following model was proposed: NMDAR stimulation induces Ca2+ influx into neurons, which activates PP1α to dephosphorylate and activate P-Rex1. P-Rex1 then activates Rac1 to mediate both AMPAR endocytosis and the large-scale actin cytoskeletal remodeling required for LTD. Together, the results of this important study have suggested that approaches to restore P-Rex1 function in hippocampal neurons may prove beneficial in the future in the treatment of autism spectrum disorders.
Melanocytes and Pigmentation
The Sansom lab showed in 2011 that Prex1 −/− mice on C57Bl6 genetic background have the expected black fur but white bellies, feet and tail tips, due to impaired melanoblast migration during development. This migration defect becomes relevant in melanoma, where melanocytes are transformed into melanoma cells. P-Rex1 expression is deregulated in melanoma cells and is critical for melanoma metastasis. A recent follow-on paper from the same lab described Prex1 −/− mice with an additional melanocyte-specific deletion of Rac1. These mice showed impaired melanoblast proliferation as well as the melanoblast migration defect (the mice had mostly white fur). This showed that P-Rex1 controls melanocyte development not only through Rac1 but possibly also through other Rac-GTPases, such as RhoG. Another possibility is that unknown GEF-independent adaptor functions of P-Rex1 are involved.
Adipocytes and Metabolism
The Mitchell lab showed in 2011 that P-Rex1 is expressed in 3T3-L1 adipocytes, where it controls insulin-stimulated trafficking of the glucose transporter Glut4 and glucose uptake. This is likely to be mediated by the Rac1-GEF activity of P-Rex1, as a dominant-negative Rac1 mutant could block the P-Rex1 dependent membrane localization of Glut4. The Bowden lab also found potential links of P-Rex1 to metabolism, through the identification of SNPs in the 3′ perigenic region of PREX1 associated with an increased risk of obesity developing into type-2 diabetes. However, it remains to be seen if these SNPs affect P-Rex1 expression or function.
P-Rex1 has also recently been identified as a biomarker of the thermogenic potential of human brown adipocytes (Xue et al. 2015). Immortalized preadipocytes were established from human brown and white adipose tissue, and were screened by microarray profiling for molecular determinants of thermogenic potential. This identified that P-Rex1 expression in preadipocytes correlates with the presence of the brown fat marker UCP1 in differentiated cells. CRISPR-mediated knockout of P-Rex1 in brown adipose precursors showed that the GEF has no effect on the ability of preadipocytes to differentiate into adipocytes. However, P-Rex1 deficiency significantly decreased the expression of UCP1 and other brown fat markers, as well as reducing the basal respiration, proton leak and maximal respiration capacity of differentiated adipocytes. Therefore, P-Rex1 is required for the development of the thermogenic potential of brown adipose cells (Xue et al. 2015). The mechanism through which P-Rex1 regulates gene expression in adipocytes, and its significance for the function of brown adipose tissue in vivo, remain to be investigated.
During zebrafish development, endoderm and mesoderm formation is regulated by the TGFβ-like cytokine Nodal. In 2012, the Stainier lab used morpholino knockdown to show that P-Rex1 controls the persistence and speed of endodermal cell migration. Overexpression of P-Rex1 could partially rescue the migration defect caused by Nodal inhibition. The recent study by the Ye lab on the role of P-Rex1 in pulmonary fibrosis in mice suggests furthermore that this role of P-Rex1 in TGFβ-dependent zebrafish embryonic development be transformed into into a functional role in inflammatory TGFβ signaling in mammals (Liang et al. 2016).
Overexpression and Mutation
Deregulated P-Rex1 expression is seen in many types of cancer, including melanoma, breast, prostate, kidney, thyroid and colon cancer. P-Rex1 deregulation is sufficient to drive the growth of breast tumors and the metastasis of prostate cancer and melanoma. Many excellent reviews on the roles of P-Rex1 in cancer have recently been published, please see reference (Welch 2015) for citations of these reviews and for an overview. We will focus here on the most recent reports.
Overexpression of P-Rex1 can occur through amplification of the PREX1 gene, as was first discovered in breast cancer by the Kazanietz lab, or through loss of epigenetic repression in breast and prostate cancer, as described above. In contrast, somatic mutations of P-Rex1 in cancer are rare, although the Tsuchihara lab reported in 2013 mutations in the coding region to occur in lung carcinomas with a frequency of 5%. It remains to be shown if such mutations affect the function of the P-Rex1 in a manner that could promote the growth or spread of lung cancer. Finally, a SNP in the third intron of PREX1 (rs6066835) was recently identified in a GWAS study to be associated with multiple myeloma and with increased P-Rex1 expression in multiple myeloma plasma cells (Mitchell et al. 2016).
The Pandiella and Kazanietz labs first showed in 2010 that, although P-Rex1 is not detected in normal breast tissue, it is expressed in 58% of human breast tumors, particularly the luminal B subtype. Highest levels were reported in advanced tumors and metastases, correlated with estrogen receptor and ErbB2 expression. Disease-free survival of breast cancer patients with high levels of P-Rex1 was significantly reduced. P-Rex1 promoted the viability, proliferation and motility of breast cancer cells, in a Rac-GEF activity dependent manner, within a range of RTK-dependent pathways. Downregulation of P-Rex1 reduced tumor growth in xenograft mouse models of breast cancer.
Sosa et al. reported in 2010 that stimulation of breast cancer cells through the RTK ErbB3 leads to P-Rex1 activation via transactivation of the GPCR CXCR4 (SDF1 receptor), resulting in increased cell growth and migration. Recent follow-on studies by the same lab showed that Hypoxia-Inducible Factor 1a (HIF-1α) upregulates CXCR4 upon ErbB3 activation, and this likely contributes to the aberrant P-Rex1 signaling in breast cancer cells (Lopez-Haber et al. 2016). Furthermore, P-Rex1 expression led to the deregulation of 89 genes in these cells, including the prometastatic metalloproteinase MMP10 which is associated with poor prognosis.
The Engelman lab proposed in 2013 that activating PI3K mutations and ErbB2 amplification in breast cancer cells involve P-Rex1/Rac1 signaling through the Raf/Mek/Erk pathway, and two recent studies found similar pathways (Dillon et al. 2015; Liu et al. 2016). These studies showed furthermore that P-Rex1 affects the expression of the cell cycle regulators Cyclin D1 and p21Cip1 (Liu et al. 2016), and that PI3K pathway activity regulates P-Rex1 protein levels in breast cancer cells (Dillon et al. 2015).
The Sansom lab reported in 2011 that, although P-Rex1 is not detectable in adult human skin, it is expressed in 80% of melanomas. Highest expression was found in advanced melanoma and metastases, correlating with invasiveness rather than with N-Ras or B-Raf status. Knockdown of P-Rex1 in human melanoma cells reduced invasiveness, and Prex1 −/− mice exhibited a drastic reduction in melanoma metastasis and a significant improvement in survival. Furthermore, the metastasis-promoting role of P-Rex1 was dependent on its Rac1-GEF activity. In addition, a recent study showed that P-Rex1 expression in melanoma cells correlates with high ERK activity, and pharmacological inhibition of MEK or ERK reduced P-Rex1 expression in these cells (Ryan et al. 2016).
The Tu lab first reported in 2009 that P-Rex1 expression is low in normal human prostate and in primary prostate tumors, but high in advanced tumors and metastases. Knockdown of P-Rex1 in human metastatic prostate cancer cells reduced invasiveness, whereas overexpression induced metastasis of xenografts in immune-deficient mice in a GEF-activity dependent manner. A recent study also implicated P-Rex1 in drug resistance of prostate cancer. Drugs that target vascular endothelial growth factor (VEGF) or its receptor are used to treat prostate cancer, but drug resistance is common. VEGF signaling through the RTK NRP2 and Rac1 was shown to be important for the development of drug resistance by a subpopulation of stem-cell like cancer cells. Expression analysis revealed upregulation of P-Rex1 and the transcription factor Myc in this subpopulation (Goel et al. 2016). Myc was shown to regulate the transcription of P-Rex1 in the drug resistant prostate cancer cells by binding to a consensus site in the PREX1 promoter 246 bp 5′ of the transcriptional start. Importantly, downregulation of P-Rex1 restored drug sensitivity, and it increased the survival of mice in xenograft models and in a prostate cancer model driven by transgenic expression of Myc (Goel et al. 2016).
The biggest progress in P-Rex1 research over the past couple of year arguably came from the availability of structural information of P-Rex1 (Lucato et al. 2015; Cash et al. 2016). Other important contributions were the identification of new functional roles in thermogenic capacity and in social interactions, as well as new roles of P-Rex1 in inflammation, new P-Rex1 binding partners and new mechanisms of regulation. Future directions should include investigations into potential adaptor functions of P-Rex1, and into the possibility of P-Rex1 deregulation in human inflammatory disorders, immune-deficiencies and metabolic syndrome.
The roles of P-Rex1 in tumor growth and metastasis make this Rac-GEF an attractive target for cancer therapy. However, targeting Rac-GEFs is not trivial, because it entails (b)locking the interaction with Rac, which is difficult to achieve with high specificity and efficacy. The development of small-molecule inhibitors for Rac-GEFs is currently widely pursued, but compounds with particular efficacy or specificity for P-Rex1 have not been reported to date. One promising avenue for targeting P-Rex1 could be to inhibit its activation by phosphorylation. Among these phosphorylation events, effective and specific blockade of the kinase that phosphorylates Ser1169 may be a promising path to pursue.
This review was funded by Institute Strategic Programme Grant BB/J004456/1 from the Biotechnology and Biological Sciences Research Council (BBSRC) to the Babraham Institute Signalling Programme. CP is funded by a Targeted PhD studentship from the BBSRC Doctoral Training Programme. ET is funded by a CASE PhD studentship from the BBSRC in collaboration with the Cardiovascular and Metabolic Disease unit of MedImmune, Cambridge.
- Chavez-Vargas L, Adame-Garcia SR, Cervantes-Villagrana RD, Castillo-Kauil A, Bruystens JG, Fukuhara S, et al. Protein Kinase A (PKA) Type I interacts with P-Rex1, a Rac Guanine Nucleotide Exchange Factor: effect on PKA localization and P-Rex1 signaling. J Biol Chem. 2016;291:6182–99.PubMedPubMedCentralCrossRefGoogle Scholar
- Liu HJ, Ooms LM, Srijakotre N, Man J, Vieusseux J, Waters JE, et al. PtdIns(3,4,5)P3-dependent Rac Exchanger 1 (PREX1) Rac-Guanine Nucleotide Exchange Factor (GEF) activity promotes breast cancer cell proliferation and tumor growth via activation of Extracellular Signal-Regulated Kinase 1/2 (ERK1/2) Signaling. J Biol Chem. 2016;291:17258–70.PubMedPubMedCentralCrossRefGoogle Scholar
- Lucato CM, Halls ML, Ooms LM, Liu HJ, Mitchell CA, Whisstock JC, et al. The Phosphatidylinositol (3,4,5)-trisphosphate-dependent Rac Exchanger 1:Ras-related C3 Botulinum Toxin Substrate 1 (P-Rex1:Rac1) complex reveals the basis of Rac1 activation in breast cancer cells. J Biol Chem. 2015;290:20827–40.PubMedPubMedCentralCrossRefGoogle Scholar
- Ryan MB, Finn AJ, Pedone KH, Thomas NE, Der CJ, Cox AD. ERK/MAPK signaling drives overexpression of the Rac-GEF, PREX1, in BRAF- and NRAS-mutant melanoma. Mol Cancer Res. 2016;14:1009–18.Google Scholar