Highlights
TIAM1 is part of the WNT-regulated destruction complex regulating TAZ/YAP stability
WNT induces TIAM1 nuclear translocation where it reduces TAZ/YAP-TEAD interaction
Nuclear TIAM1 blocks the TAZ/YAP genetic program inhibiting migration of CRC cells
Nuclear TIAM1 predicts good prognosis in CRC
Summary
Aberrant WNT signaling drives colorectal cancer (CRC). Here, we identify TIAM1 as a critical antagonist of CRC progression through inhibiting TAZ and YAP, effectors of WNT signaling. We demonstrate that TIAM1 shuttles between the cytoplasm and nucleus antagonizing TAZ/YAP by distinct mechanisms in the two compartments. In the cytoplasm, TIAM1 localizes to the destruction complex and promotes TAZ degradation by enhancing its interaction with βTrCP. Nuclear TIAM1 suppresses TAZ/YAP interaction with TEADs, inhibiting expression of TAZ/YAP target genes implicated in epithelial-mesenchymal transition, cell migration, and invasion, and consequently suppresses CRC cell migration and invasion. Importantly, high nuclear TIAM1 in clinical specimens associates with increased CRC patient survival. Together, our findings suggest that in CRC TIAM1 suppresses tumor progression by regulating YAP/TAZ activity.
Significance
Colorectal carcinogenesis typically commences with constitutive WNT signaling leading to nuclear accumulation of transcriptional co-activators including TAZ and YAP. Thereafter, mutational and epigenetic events ensue inducing a genetic program that propels invasion and metastasis. We herein identify TIAM1 as a component of the cytoplasmic destruction complex regulating TAZ/YAP stability. We further show that when, as in initiated intestinal epithelial cells, the destruction complex is inactivated, TIAM1 along with TAZ/YAP accumulate and translocate from the cytoplasm to the nucleus. In the nucleus, however, TIAM1 continues to antagonize nuclear TAZ/YAP function despite constitutive WNT signaling, and thereby suppresses cell migration and invasion. In keeping with these findings, nuclear TIAM1 is downregulated in advanced colorectal cancer, and low nuclear TIAM1 predicts poor prognosis.
Introduction
Intestinal epithelial cells transform into malignant cells following mutations in various oncogenes and tumor suppressors that drive changes in programs of gene expression. Typically, the initiating event is inactivation of the tumor suppressor adenomatous polyposis coli (APC) or, less frequently, activation of β-catenin, resulting in constitutive activation of the canonical WNT pathway. At the core of this pathway is the regulation of β-catenin by a cytoplasmic destruction complex. In the absence of a WNT signal, the central scaffold protein AXIN within the destruction complex binds APC which efficiently delivers cytosolic β-catenin to be phosphorylated by CK1 and then GSK3β. Phosphorylated β-catenin is subsequently ubiquitylated by the E3 ligase βTrCP and degraded by the proteasome. Binding of WNT to its receptor causes inactivation of the destruction complex, stabilization, and nuclear translocation of β-catenin, and subsequent transcription by a β-catenin/TCF complex (Clevers, 2006).
The transcriptional co-activators TAZ and YAP (TAZ/YAP; originally identified as targets of the HIPPO pathway) have recently been established as additional regulatory components but also as effectors of canonical WNT signaling. In the cytoplasm, TAZ/YAP interact with β-catenin directly and restrict β-catenin nuclear translocation (Imajo et al., 2012). Cytoplasmic TAZ/YAP also interact with DVL, a regulator of the destruction complex, inhibiting DVL phosphorylation by CK1 and consequently inhibiting β-catenin activation (Varelas et al., 2010). A further study showed that in the absence of WNT, TAZ/YAP are essential for βTrCP recruitment to the destruction complex and β-catenin degradation (Azzolin et al., 2014). Thus, cytoplasmic TAZ/YAP emerge as inhibitors of WNT signaling and, therefore, potential suppressors of colorectal cancer (CRC). Inactivation of the destruction complex, either through WNT stimulation or, as occurs in CRC cells, inactivation of APC, results in not only β-catenin but also TAZ/YAP stabilization and the opportunity for TAZ/YAP to translocate to the nucleus (Azzolin et al., 2012, Azzolin et al., 2014). Significantly, in contrast to the ability of TAZ/YAP to downregulate β-catenin in the cytoplasm, nuclear YAP seems to cooperate with β-catenin to transactivate WNT target genes (Azzolin et al., 2012, Rosenbluh et al., 2012). TAZ/YAP stimulate target gene expression involved in cell proliferation, stem cell self-renewal, and tumorigenesis through binding TEAD family transcription factors in the nucleus (Zhang et al., 2009, Zhao et al., 2008). In keeping, TAZ/YAP are required for the overgrowth of intestinal crypts and formation of adenomas following APC inactivation (Azzolin et al., 2014, Gregorieff et al., 2015) and TAZ/YAP expression is associated with poor prognosis of CRC patients (Wang et al., 2013, Yuen et al., 2013). Thus, the interaction between TAZ/YAP and the WNT pathway is complex, highlighting our incomplete understanding of how the pathway is regulated. In the present study, we investigated the role of TIAM1, a guanine nucleotide exchange factor specific for RAC1, in colon cancer progression.
Results
Nuclear TIAM1 Expression in Human Tumors is a Prognostic Marker for CRC Progression
Previously, using recombinant mouse models, we have shown that TIAM1 cooperates with WNT signaling during initiation of intestinal epithelial neoplasia but then appeared to antagonize tumor progression (Malliri et al., 2006). However, how TIAM1 influences intestinal tumor initiation and progression remained elusive. To further address the impact of TIAM1 on intestinal neoplasia and increase our understanding of its clinical significance, we probed a tissue microarray (TMA) comprising 650 samples, including Dukes stages A to C, from a well-characterized patient cohort (Brown et al., 2014) (Table S1), with a pre-validated TIAM1 antibody (Marei et al., 2016, Vaughan et al., 2015, Whalley et al., 2015). Intriguingly, we detected TIAM1 not only in the cytoplasm but also in cell nuclei (Figure 1A). Furthermore, nuclear TIAM1 staining intensity decreased with advancing Dukes stage (χ2 = 54.165, p < 0.001) (Figures 1A, 1B, and S1A). TIAM1 cytoplasmic staining also decreased with advancing Dukes stage (χ2 = 19.057, p = 0.001) (Figures 1A and S1B). Thus, TIAM1 expression is negatively associated with colon cancer progression, consistent with our previous finding that TIAM1 antagonized progression of intestinal tumors in ApcMin/+ mice (Malliri et al., 2006). To determine the association between TIAM1 expression and overall survival, we grouped nuclear TIAM1 staining into low (scores 0–1) or high (scores 2–3) categories and found that patients having CRC with high nuclear TIAM1 had significantly better survival than patients having CRC with low nuclear TIAM1 (Figures 1C and S1C). However, no difference was found in overall survival between patients having CRC expressing low or high cytoplasmic TIAM1 (Figure S1D). Thus, high levels of nuclear TIAM1 could serve as a good prognostic factor for CRC patients.
TIAM1 Nucleocytoplasmic Shuttling is Regulated by a Functional Nuclear Localization Signal
To validate TIAM1 nuclear localization, we analyzed its expression in CaCo2, DLD1, and SW620 CRC cell lines. TIAM1 was detected predominantly in nuclei from all three cell lines (Figure 2A). Furthermore, depletion of TIAM1 using two different small interfering RNAs (siRNAs) reported previously (Vaughan et al., 2015, Whalley et al., 2015) dramatically reduced the nuclear signal in all three lines (Figure 2A), verifying the specificity of the staining. Specificity of the nuclear TIAM1 staining was also verified in two independent clones of SW480 cells with TIAM1 ablated using CRISPR (Figures S2A–S2C).
We next addressed the mechanism of TIAM1 recruitment to the nucleus. Nuclear entry of proteins as large as TIAM1 that are unable to diffuse passively through nuclear pores is selective, dependent on the presence of a nuclear localization signal (NLS) (Freitas and Cunha, 2009). We therefore examined whether TIAM1 has a functional NLS. In silico analysis revealed two putative bipartite NLSs, but no monopartite NLS (Figure S2D). Between the two putative NLSs, the second had a high confidence score of 9.4 (Figure S2D), and aligning peptide sequences from different species revealed that NLS2 was more highly conserved (Figure S2E), implying an important regulatory function.
To investigate the functionality of the putative NLSs, we generated GFP-tagged TIAM1 constructs with the predicted NLSs deleted. The TIAM1ΔNLS1 construct contained a deletion in amino acids 269–298 and the TIAM1ΔNLS2 construct in amino acids 684–710. The constructs were transiently overexpressed in the DLD1 cell line and the number of cells with nuclear or non-nuclear localization defined by confocal microscopy. As expected, exogenously expressed full-length TIAM1 (FL-TIAM1-GFP) was detected both in the nucleus and the cytoplasm (Figure S2F), as was TIAM1-ΔNLS1 (data not shown); however, we observed significantly less TIAM1-ΔNLS2 in nuclei (Figure S2F). To further interrogate the role of NLS2 in controlling nuclear localization of TIAM1, we generated three different GFP-tagged TIAM1 mutants in which the clusters of basic residues 688KRR690 and 704KKK706were mutated to AAA (Figure 2B). Cells expressing any of the three mutations in NLS2 displayed a significant decrease in nuclear TIAM1 (Figure 2C). Our results indicate, therefore, that TIAM1 shuttles between the cytoplasm and the nucleus mediated by at least one functional NLS.
TIAM1 Antagonizes Transcription of TAZ/YAP Target Genes
The nuclear localization of TIAM1 suggested a possible role in regulating gene expression, whose interrogation might allow us to elucidate the contribution of TIAM1 to CRC development. By performing RNA sequencing (RNA-seq) on SW620 cells transfected with either siTIAM1#1 or a negative control siRNA (siNT) (Figure S3A), we identified TIAM1 differentially expressed genes (Table S2). The expression profiles of siNT and siTIAM1#1 samples were clearly separated based on principal component analysis (Figure S3B), and 5,027 genes were found to be differentially expressed (false discovery rate <0.05) between the two populations (Figure 3A). To surmise molecular pathways associated with TIAM1 regulated genes, we performed gene set enrichment analysis (GSEA) using gene sets for the oncogenic pathways in the Molecular Signature Database (MSigDB) (Subramanian et al., 2005). GSEA revealed significant enrichment among TIAM1-regulated genes for apical polarity and epithelial-mesenchymal transition (EMT) signatures (Figure 3B), cellular processes well known to be regulated by TIAM1 (Mack et al., 2012, Vaughan et al., 2015, Woodcock et al., 2009). Interestingly, we also found that TIAM1-regulated genes overlapped significantly with the YAP-associated molecular signature (Cordenonsi et al., 2011) (Figure 3B). Indeed, numerous well-characterized TAZ/YAP target genes were found to be upregulated in the RNA-seq dataset upon knockdown of TIAM1 (Figure 3C). Furthermore, five target genes (AMOTL2, ANKRD1, AXL, CTGF, and CYR61) were validated to be upregulated by qRT-PCR in both SW620 and SW480 cells using two independent siRNA sequences targeting TIAM1 (Figures 3D, S3C, and S3D). These findings indicated that TIAM1 might be antagonizing induction of TAZ/YAP target genes.
Given the important role of TAZ/YAP in CRC, we decided to investigate further the interplay of TIAM1 with TAZ/YAP. First, we tested whether the effect of TIAM1 depletion on the expression of CTGF and CYR61 was also observed in vivo. As shown in Figure 3E, the levels of both CTGF and CYR61 were upregulated in the small intestines of Vil-Cre-ERT2 Apcfl/fl Tiam1−/− compared with Vil-Cre-ERT2 Apcfl/flmice treated with tamoxifen. Second, using RNA-seq we generated a list of TAZ/YAP differentially expressed genes from SW620 cells transfected with either specific siRNAs for TAZ/YAP or a negative control siRNA (Figure S3E and Table S3) and compared it with the siTIAM1#1 RNA-seq dataset. Interestingly, we found that 50% of TAZ/YAP-regulated genes were also TIAM1 dependent (Figure 3F).
TIAM1 Regulates TAZ Stability and Nuclear Translocation
The inverse association between TAZ/YAP and TIAM1-regulated genes prompted us to investigate the molecular mechanism(s) by which TIAM1 might antagonize TAZ/YAP. Depletion of TIAM1 in HEK293 cells (Figure S4A) using siTIAM1#1 resulted in a statistically significant increase of TAZ protein both in the cytoplasm and in the nuclei of HEK293 cells comparable with that induced by APC depletion (Figure 4A). A similar effect on TAZ was observed after depletion of TIAM1 by a second siRNA sequence (siTIAM1#2) (Figure S4B). Because TAZ mRNA levels were not significantly upregulated by TIAM1 depletion (Figure S4C), this suggested that TIAM1 depletion promoted the stabilization and nuclear translocation of TAZ. YAP stability, and nuclear translocation also appeared to increase upon TIAM1 and APC depletion (Figures 4A and S4B).
To further investigate the relationship between TIAM1 and TAZ/YAP, we tested the effect of TIAM1 depletion on the activity of 8xGTIIC-Lux, a synthetic luciferase reporter containing multimerized response elements of TEAD, the main DNA-binding cofactor of TAZ or YAP (Dupont et al., 2011). As shown in Figure 4B, TIAM1 depletion strongly induced the activity of 8xGTIIC-Lux, indicating that TIAM1 regulates TAZ/YAP activity. Depletion of endogenous TAZ alone partially rescued the effect of TIAM1 depletion (Figures S4D and S4E), while depletion of YAP or a combination of YAP and TAZ was even more effective at reversing the effect (Figures 4B, S4D, and S4E). Moreover, the expression of the endogenous TAZ/YAP target genes CTGF and CYR61 was also shown to be upregulated by TIAM1 depletion in both HEK293 and HaCaT cells (Figures 4C and S4E–S4G). Again this effect could be reversed by depleting TAZ or YAP alone, and particularly through combined depletion of both (Figures 4C and S4E–S4G). Analysis of target gene expression and activity of a synthetic reporter thus confirmed the ability of TIAM1 to antagonize the function of both TAZ and YAP.
Finally, we decided to investigate the mechanism by which TIAM1 depletion induced the stability of TAZ. As previous studies have shown that the E3 ubiquitin ligase βTrCP interacts directly with both TAZ (Azzolin et al., 2012, Huang et al., 2012, Liu et al., 2010) and TIAM1 (Magliozzi et al., 2014, Zhu et al., 2014), we hypothesized that TIAM1 might regulate the interaction of TAZ with βTrCP. By carrying out immunoprecipitations for endogenous βTrCP from confluent control or TIAM1-depleted HEK293 cells, we found that the interaction of TAZ with βTrCP was decreased in cells in which TIAM1 expression was downregulated (Figure 4D). Taken together, these results indicate that TIAM1 is important for the interaction of TAZ with βTrCP through which it regulates the stability and cellular distribution of TAZ.
TIAM1 is Regulated by the Canonical WNT Pathway
As stated above, previous studies have shown that a major factor regulating the stability and nuclear translocation of TAZ is the destruction complex (Azzolin et al., 2012, Azzolin et al., 2014), and our aforementioned findings suggested that TIAM1 is also implicated in this mechanism of TAZ regulation. This prompted us to investigate whether TIAM1 was a component of the WNT-regulated destruction complex. By immunoprecipitating endogenous TIAM1 from cytosolic and nuclear fractions of confluent HEK293 cells, which carry a functional destruction complex, we found that TIAM1 interacted with AXIN, β-catenin, and (as shown previously by Magliozzi et al., 2014, Zhu et al., 2014], and above) βTrCP, which are established elements of the destruction complex as well as TAZ and YAP (Figure 5A). Even though low levels of TIAM1 were detected in the nucleus, TIAM1 interaction with AXIN, β-catenin, βTrCP, TAZ, and YAP was observed only in the cytoplasm of HEK293 cells. The interaction of TIAM1 with TAZ was confirmed in vitro using recombinant proteins and was inferred to be direct (Figure 5B).
Since the main function of the destruction complex is regulating protein stability through degradation, we wondered whether its inactivation would affect TIAM1 stability. The first approach we used for inactivation was downregulation of APC expression by a siRNA (Figure S5A). APC depletion resulted in a statistically significant increase in TIAM1 levels in the cytoplasm of HEK293 cells (Figure 5C). Since TIAM1 mRNA levels were not increased by APC depletion (Figure S5B), this suggested that APC depletion promoted the stabilization of TIAM1 protein. Interestingly, we also observed a significant increase in the nuclear levels of TIAM1 (Figure 5C). The stabilization and nuclear translocation of TIAM1 following APC depletion was also confirmed by immunofluorescence microscopy (Figure S5C) and is consistent with the nuclear accumulation of TIAM1 we observe in colorectal tumors (Figure 1A) and CRC cell lines with defective APC (Figure 2A). As a second approach to inactivate the destruction complex, we used conditioned medium from WNT3A expressing cells. Similar to APC depletion, treatment of confluent RKO cells with WNT3A-conditioned medium induced the stabilization and translocation of TIAM1 from the cytoplasm into the nucleus (Figure 5D). Finally, activation of the WNT pathway by BIO, a small molecule that inhibits GSK3β activity (Meijer et al., 2003), also induced the stabilization and nuclear accumulation of TIAM1 in RKO cells (Figure 5E). In summary, these results indicate that the WNT pathway regulates the stability and nuclear translocation of TIAM1 through the destruction complex, as described previously for TAZ (Azzolin et al., 2012, Azzolin et al., 2014).
We next addressed how the destruction complex regulates TIAM1 stability. It is known that βTrCP is responsible for the degradation of β-catenin (Hart et al., 1999, Kitagawa et al., 1999, Marikawa and Elinson, 1998) and TAZ (Azzolin et al., 2012, Huang et al., 2012, Liu et al., 2010). Previous studies have shown that TIAM1 also interacts with βTrCP, resulting in its degradation (Magliozzi et al., 2014, Zhu et al., 2014). This prompted us to investigate whether the increased stabilization of TIAM1 observed upon WNT3A treatment or APC depletion is due to its decreased interaction with βTrCP. By carrying out immunoprecipitations for endogenous βTrCP from confluent control or APC-depleted HEK293 cells, we found that the interaction of TIAM1 with βTrCP was almost abolished in cells in which APC was depleted (Figure 5F). A dramatic decrease in the interaction between TAZ and βTrCP was also observed under these conditions (Figure 5F). Furthermore, following activation of the WNT pathway by WNT3A treatment, TIAM1 interaction with proteins of the destruction complex including AXIN and βTrCP, as well as TAZ, was abolished (detected from 0.5 hr; Figure 5G). Taken together, these data imply that activation of the canonical WNT pathway (inactivation of the destruction complex) induces TIAM1 stability and nuclear translocation by regulating the interaction between TIAM1 and βTrCP.
Nuclear TIAM1 Antagonizes TAZ Transcriptional Activity
The role of TIAM1 in promoting TAZ degradation by the destruction complex could offer one explanation for the stimulation of TAZ transcriptional activity following TIAM1 depletion (Figures 4B and 4C). However, our findings also indicated that TIAM1 antagonizes TAZ transcriptional activity in CRC cells with a constitutively defective destruction complex (Figure 3), pointing to an additional mechanism by which TIAM1 antagonizes TAZ. Since inactivation of the destruction complex induces nuclear translocation of both TIAM1 and TAZ, we hypothesized that the two proteins might also interact in the nucleus, potentially interfering with TAZ transcriptional activity. Immunoprecipitation of endogenous TIAM1 from control or APC-depleted HEK293 cells revealed that inactivation of the destruction complex indeed resulted in the appearance of a nuclear TIAM1/TAZ complex accompanied by reduction of this complex in the cytoplasm (Figure 6A). An interaction between TIAM1 and TAZ could also be detected in nuclear extracts from three different CRC cell lines (Figures 6B, S6A, and S6B), and was confirmed by Duolink proximity ligation assay (Figure 6C). Next, we tested the effect of TIAM1 knockdown on TAZ transcriptional activity in these CRC cells. Similar to HEK293 and HaCaT cells (Figures 4C and S4E–S4G), TIAM1 depletion in SW620 and DLD1 cells strongly induced the expression of CTGF and CYR61, which could be suppressed by depletion of endogenous YAP and TAZ (Figures 6D and S6C), indicating that TIAM1 inhibits the transcriptional activity of YAP and TAZ in CRC cells.
The antagonistic effect of nuclear TIAM1 on TAZ/YAP activity prompted us to examine whether nuclear TIAM1 might prevent the interaction of TAZ and YAP with TEADs. Immunoprecipitating endogenous TEADs from nuclear extracts of parental SW480 cells or the two CRISPR-mediated TIAM1 knockout SW480 clones revealed that TIAM1 ablation increased the interaction of TEADs with TAZ and YAP (Figures 6E and S6D). Duolink proximity ligation assays confirmed that the interaction between TAZ or YAP with TEADs was increased in the absence of TIAM1 (Figures 6F and S6E). The inhibitory effect of nuclear TIAM1 on TAZ activity was further demonstrated by chromatin immunoprecipitation sequencing (ChIP-seq) analysis. Parental SW480 cells or the TIAM1 knockout SW480 clone #1 cells were used to determine the recruitment of TAZ to chromatin in the presence or absence of TIAM1. We found that the number of TAZ peaks identified in the TIAM1 knockout cells (35,450 peaks) was increased compared with the parental cells (24,913 peaks) (GEO: GSE90492). As shown in Figure S6F, of the total peaks discovered by MACS peak caller, a subset of the peaks were exclusively found in either the TIAM1 knockout SW480 clone #1 cells (23,849 peaks) or parental cells (13,101 peaks). We then examined locations with overlapping TAZ peaks. In total there were 5,740 overlapping peaks, of which 2,082 had higher (>1.5 fold) TAZ recruitment to the chromatin in the TIAM1 knockout cells compared with parental cells (Figure 6G). In contrast, only 84 peaks were found to have a higher TAZ recruitment in the parental cells. Furthermore, consistent with the above findings that TIAM1 knockdown induces the interaction of TAZ with TEAD and induces the expression of CYR61, we found that deletion of TIAM1 increased the binding of TAZ to the promoter region of CYR61 (Figure 6H). Taken together, our results indicate that nuclear TIAM1 functions as a negative regulator of the transcriptional activity of TAZ and YAP.
The Effect of Nuclear TIAM1 on CRC Cell Proliferation, Migration, and In Vivo Invasion
Having demonstrated that TIAM1 antagonizes TAZ/YAP transcriptional activity, we analyzed the functional significance of their interaction. Previous studies have shown that both TIAM1 and TAZ/YAP regulate cell proliferation and migration (Minard et al., 2004, Malliri et al., 2006, Zanconato et al., 2016). First, we evaluated the potential effect of TIAM1 depletion on the proliferation of CRC cell lines and their dependence on TAZ/YAP. TIAM1, TAZ/YAP, or a combination of all three was depleted using siRNAs, and cell number was determined 6 days post transfection. As shown in Figure S7A, downregulation of either TAZ/YAP or TIAM1 decreased the proliferation of DLD1, SW480, and SW620 cells. However, the combined depletion of TIAM1 and TAZ/YAP on the proliferation of CRC cells was not additive, indicating that the effect of TIAM1 depletion on cell proliferation is TAZ/YAP independent.
To investigate specifically the role of nuclear TIAM1 on cell proliferation, we generated an NLS-TIAM1 construct by fusing full-length TIAM1 with the well-characterized NLS of SV40 large T antigen. We verified that the NLS-TIAM1 construct was exclusively nuclear compared with full-length TIAM1 (FL-TIAM1) (Figure S7B). Expression of NLS-TIAM1 (resistant to siTIAM1#1) at levels comparable with endogenous TIAM1 in DLD1 or SW620 cells had no effect on the proliferation of TIAM1-depleted cells (Figure S7C). However, FL-TIAM1 partially restored proliferation in TIAM1-depleted cells. Similarly, expression of FL-TIAM1, but not NLS-TIAM1, in SW480 CRISPR-mediated TIAM1 knockout cells restored proliferation to levels observed in parental cells (Figure S7D). Taken together, these results indicate that the effect of TIAM1 depletion on cell proliferation is TAZ/YAP independent and not related to its nuclear localization.
We next investigated the potential effect of TIAM1 depletion on the migration of SW480 and SW620 cells and its dependence on TAZ/YAP. TIAM1, TAZ/YAP, or a combination of all three was depleted using siRNAs, and cells were assayed for their migratory potential over 24 hr using the Boyden chamber assay. As shown in Figure 7A, TIAM1 depletion in both SW480 and SW620 cells induced a strong migratory response. Interestingly, this increase in migration was dependent on TAZ/YAP, as depletion of endogenous YAP and TAZ dramatically reduced the effect of TIAM1 depletion (Figure 7A). To verify that specifically nuclear TIAM1 suppresses cell migration, we analyzed the migratory capacity of SW480 cells overexpressing NLS-TIAM1. Notably, compared with parental SW480 cells, cells expressing NLS-TIAM1 migrated significantly less (Figure 7B). Furthermore, overexpression of NLS-TIAM1 abolished the increased migration observed in the two CRISPR-mediated TIAM1 knockout SW480 clones (Figure 7B).
To investigate the role of nuclear TIAM1 on migration in vivo, we turned to a zebrafish xenograft model. Fluorescently labeled parental SW480 cells or the aforementioned SW480 CRISPR-mediated TIAM1 knockout cells inducibly expressing either FL-TIAM1 or NLS-TIAM1 (Figure S7D) were injected into the pericardial cavity of 2-day-old zebrafish embryos. Ninety-six hours after injection, tumor cells were observed to have largely filled the approximately conical-shaped cavity (excluding the volume occupied by the heart) in all cases. Compared with parental SW480 cells, significantly more non-induced (minus doxycycline [Dox]) SW480 TIAM1 knockout cells disseminated away from the pericardial region (Figure 7C), whereas cells expressing (plus Dox) either FL-TIAM1 or equally NLS-TIAM1 showed a decreased ability to disseminate (Figure 7C), indicating the inhibitory role of nuclear TIAM1 on invasion. Taken together, these results suggest that nuclear TIAM1 suppresses TAZ/YAP-induced cell migration and in vivo invasion of CRC cells.
Discussion
Herein, we present evidence indicating that TIAM1 suppresses CRC cell migration and invasion by regulating TAZ/YAP transcriptional activity. Furthermore, we propose that this function of TIAM1 underpins our earlier observation that TIAM1 antagonizes malignant progression of intestinal neoplasia in the ApcMin/+ mouse model (Malliri et al., 2006). Interestingly, we found that the role of TIAM1 in regulating TAZ/YAP entails not only functions performed in the cytoplasm but also events in the nucleus, and propose the following model (Figure 8). (1) In the absence of WNT signaling, TIAM1 and TAZ are collectively recruited to the destruction complex for targeted ubiquitylation and degradation. (2) Upon WNT stimulation (or following initiating oncogenic events that inactivate the destruction complex), both TIAM1 and TAZ are released from the destruction complex and stabilized, and translocate to the nucleus. However, TAZ transcriptional activity is still greatly suppressed by the nuclear pool of TIAM1, impairing the interaction of TAZ with TEADs. (3) Upon WNT stimulation/inactivation of the destruction complex and in the absence of TIAM1, TAZ present in the nucleus transactivates the TEAD transcriptional program, which in the context of neoplasia would promote a more malignant phenotype.
One of the main findings of this study is that TIAM1 associates with the destruction complex. Inactivation of the destruction complex by various means was shown to promote TIAM1 accumulation and, concomitantly, its nuclear translocation. We further demonstrated that TIAM1 modulates the interaction between TAZ and βTrCP, influencing TAZ stability. As TIAM1 and βTrCP interact (Magliozzi et al., 2014, Zhu et al., 2014) as do TIAM1 and TAZ, it is probable that TIAM1 functions as a physical scaffold bringing TAZ and βTrCP into proximity. Whether TIAM1 directly interacts with AXIN, the main structural component of the destruction complex, or APC is presently unknown. However, the ability to immunoprecipitate a complex containing all these components suggests that the interactions are relatively stable.
The destruction complex is a key checkpoint for regulating canonical WNT signaling. Inactivation of the destruction complex liberates two different effectors: the well-documented β-catenin (Clevers, 2006) and recently implicated TAZ/YAP (Azzolin et al., 2012, Azzolin et al., 2014). Indeed, both transcriptional co-activators appear to be required for expression of WNT target genes in intestinal epithelial progenitor cells and CRC cells (Azzolin et al., 2012, Fevr et al., 2007). However, the identities of the transcription factor complexes that regulate subsets of WNT-induced target genes responsible for distinct WNT-mediated biological responses (such as stem cell maintenance, transit amplification, EMT, or terminal differentiation) have not been fully delineated. Previous studies have shown that TAZ is not required for β-catenin/TCF-dependent transcription (Azzolin et al., 2012, Imajo et al., 2015) and conversely that TAZ/YAP transactivation is independent of β-catenin/TCF (Imajo et al., 2015). Nonetheless, a complex between β-catenin and TAZ/YAP that includes the transcription factor TBX5 and is critical for β-catenin-driven transformation has been described (Rosenbluh et al., 2012). These studies indicate the complexity of the system and suggest that additional studies are required to fully uncover the transcriptional molecular mechanisms that regulate CRC initiation and progression.
Our study sheds light on TAZ/YAP function in CRC progression. Our results showed that nuclear TIAM1 plays a critical role in regulating the TAZ/YAP transcriptional program that is responsible for cell migration, with TIAM1 suppressing this program even when the WNT pathway is activated. Interestingly, the inhibitory nuclear function of TIAM1 on TAZ/YAP transcriptional activity is not associated with cell proliferation. Since one of the main molecular signatures that were enriched in the TIAM1-depleted cells was associated with EMT, we speculate that TIAM1 may be associated with suppressing cancer stem cell identity, since cancer stem cells are characterized by a mesenchymal cell gene expression program (Oskarsson et al., 2014). Furthermore, TAZ and YAP have been implicated in the expansion of stem cells in the regenerating gut but also in adenomatous polyps (Cai et al., 2010, Gregorieff et al., 2015). Our previous work revealed a gradient of TIAM1 expression highest toward the bottom of intestinal crypts, overlapping with the stem cell and transit amplification (TA) compartments (Malliri et al., 2006). In accordance with TIAM1 playing a role in suppressing stemness in the TA compartment through inhibiting TAZ/YAP, in this study we found that the TIAM1 knockout-mediated induction of the TAZ target genes CTGF and CYR61 was predominantly observed in the expanded crypts of APC-deficient intestines.
Different studies have shown that APC loss alone is insufficient to promote intestinal tumor progression (Anderson et al., 2002, Blaker et al., 2003); additional mutations in other oncogenes such as KRAS and tumor-suppressor genes such as TP53 are required (Fujii et al., 2016). Our results suggest that TIAM1 is an additional protein that needs to be inactivated during CRC progression. In support of this statement, our clinical data revealed that TIAM1 expression levels decline with disease progression and that patients having CRC with high levels of nuclear TIAM1 survive longer. Based on the outcome of cancer genome resequencing, TIAM1 downregulation is not typically achieved through mutation. Downregulation could be achieved at the transcriptional level or post-translationally. Previous work has shown that oncogenic RAS expression suppresses TIAM1 transcription (Zondag et al., 2000). Furthermore, we and others have identified E3 ubiquitin ligases targeting TIAM1 for degradation (Genau et al., 2015, Magliozzi et al., 2014, Vaughan et al., 2015, Zhu et al., 2014), which may be deregulated in intestinal carcinoma. Potentially, factors resulting in exclusion of TIAM1 from the nucleus might also result in downregulation of this critical suppressor function of TIAM1, and we are currently investigating the mechanisms regulating TIAM1 nuclear import and export. Previous studies have indicated that YAP and TAZ expression is increased in CRC (especially high-grade CRC), and their expression may be used as a prognostic marker (Cai et al., 2010, Wang et al., 2013, Yuen et al., 2013). However, TAZ/YAP expression levels might not be a sufficient prognostic marker for CRC patients, as differences in nuclear levels of TIAM1 might differentially affect the transcriptional activity of TAZ/YAP. An index based on both TAZ/YAP expression and nuclear TIAM1 intensity could therefore provide clearer prognostication