RASAL2 promotes tumor progression through LATS2/YAP1 axis of hippo signaling pathway in colorectal cancer

Primary and metastatic tumor and normal frozen tissues from 8 patients who underwent colectomy and/or metastasectomy at the Prince of Wales Hospital, Hong Kong were used for genomic studies (Additional file 1: Table S1). Archival formalin-fixed paraffin-embedded (FFPE) tissue specimens from 208 CRC patients who underwent colectomy at the same hospital between 1995 and 2014 were retrieved. All specimens were reviewed by an expert gastrointestinal pathologist (KFT) to confirm histological diagnosis and tumor cell content. Clinicopathological information was retrieved from the hospital database last updated in December 2015. The study was approved by Committees for Clinical Research Ethics of Joint Chinese University of Hong Kong-New Territories East Cluster.

Microarray-based comparative genomic hybridization (array-CGH) analysis was performed using SurePrint G3 Human CGH Microarray Kit, 1 × 1 M (Agilent Technologies, Santa Clara, CA, USA). Gene expression array analysis was carried out by Macrogen (Seoul, South Korea).

Total RNA was isolated according to the protocol of TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). Reverse transcription (RT) to synthesize complementary DNA (cDNA) was performed using High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). The aliquots of cDNA were amplified using SYBR Green Master Mix (Thermo Fisher Scientific). β-actin was used as endogenous control. The RASAL2 and β-actin primers (Additional file 1: Table S2) were designed using Primer3web [20, 21].

Immunohistochemistry (IHC) was performed on 5-μm sections from tissue microarray (TMA) blocks stained using anti-RASAL2 antibodies (ab121578, 1:400, Abcam, Cambridge, MA, USA) and anti-YAP1 antibodies (ab52771, 1:500, Abcam). The protein expression on the TMA slides was assessed using the histoscore (H-score) method.

Protein was extracted in ice-cold RIPA lysis buffer containing complete protease inhibitor cocktail tablets (Roche, Rotkreuz, Switzerland). RASAL2 was detected with a polyclonal anti- RASAL2 antibody (1:250, ab121578, Abcam). Anti-YAP1 antibody (1:20000, ab52771) was provided by Abcam. Other primary antibodies were from Cell Signaling (Danvers, MA, USA), including antibodies to LATS2 (1:1000, #5888S), phospho-YAP1 (S127) (1:1000, #4911) and Cyclin D1 (1:1000, #2926). β-actin (1:100000, A5441, Sigma-Aldrich, St. Louis, MO, USA) expression was used as an equal loading control. The secondary antibodies were anti-Mouse IgG-HRP (1:15000, 00049039, Dako, Agilent Technologies, Santa Clara, CA, USA) and anti-Rabbit IgG-HRP (1:5000, 00028856, Dako).

Immortalized human normal colon epithelial cells NCM460 [22] and human CRC cell lines Caco2, CL-14, DLD-1, HCT 116, HT-29, LoVo, LS 180, SW480 and SW620 were used in this study. The NCM460 cell line was obtained from Dr. Jun YU [23, 24] (Department of Medicine & Therapeutics, The Chinese University of Hong Kong), who purchased from INCELL Corporation LLC (San Antonio, TX, USA). The cell line was cultured using M3BaseTM medium (INCELL) with 10% FBS. Caco2, DLD-1, HCT 116, HT-29, LoVo, LS 180, SW480 and SW620 were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). CL-14 cell line was obtained from Deutsche Sam lung von Mikroorganismen und Zellkulturen (Braunschweig, Germany). Constructs of ectopic RASAL2 expression, siRNAs, and shRNAs for RASAL2 knockdown can be found in the Additional file 1: Tables S2 and S3.

1 × 10⁶ transfected CRC cells suspended in 100 μl PBS were injected subcutaneously into the dorsal region of anaesthetized nude mice (5 mice/construct, control in left and treatment in right). When tumor was formed, tumor diameter was recorded every three days for three consecutive weeks. At the end of investigation, mice were sacrificed and xenografts were then collected for diameter and weight measurements. All animal handling and experimental procedures were approved by the Department of Health of Hong Kong and the CUHK.

Cells grown on coverslips were fixed with 4% paraformaldehyde, permeated with 0.3% Triton X-100 and blocked with 1% BSA. Then cells were in turn incubated with a mixture of two primary antibodies (rabbit anti-YAP1, 1:500 and mouse anti-β-actin, 1:1000) at 4 °C overnight, as well as a mixture of secondary antibodies (Alexa Fluor 488 and 594, 1:2000, Thermo Fisher Scientific) in the dark for 1 h. Nuclei were counterstained by 4′,6-diamidino-2-phenylindole (DAPI, Thermo Fisher Scientific). Images were captured using an Axio Imager2 microscope (Carl Zeiss, Germany).

Cells were transiently co-transfected with siRASAL2 and HA-Ubiquitin plasmids [25]. Then the cells were kept in 20 μM MG132 (Calbiochem, Millipore, Billerica, MA, USA) for 4 h before protein collection. 500 μg cell lysate was incubated with a mixture of 50 μl Protein G bead and 5 μg conjugated antibody to YAP1 at 4 °C overnight with rotation. Finally, samples were loaded onto a SDS-PAGE gel for western blot analysis. Ubiquitination was detected with anti-HA-tag antibody (1:7000, ab9110, Abcam).

The transfected cells were seeded into 96 well plates and starved for 24 h. Fresh medium with 2-fold diluted drug verteporfin concentration was added to each well. MTT assay was performed after incubation for 24 h.

By high-resolution microarray-based comparative genomic hybridization (array-CGH), we found recurrent gains of chromosome 1q in 5 out of 8 metastatic CRC patients (Additional file 1: Table S4). We further employed gene expression microarrays to identify genes differentially expressed among primary tumor, metastasis and normal colon mucosa samples (Additional file 2: Figure S1), one of which was RASAL2. RASAL2 was overexpressed in primary and metastatic tumor samples in both array-CGH (Fig. 1a) and expression microarray analyses (Additional file 1: Table S5).

We examined the mRNA expression level of RASAL2 in a panel of CRC cell lines including 5 lines established from early stage CRC (LS 180, CL-14, HT-29, SW480 and Caco-2) and 4 lines from advanced stage CRC (LoVo, SW620, HCT 116 and DLD-1). RASAL2 mRNA expression was higher in CRC cell lines than in normal colons, and significantly higher in advanced stage cell lines [American joint committee on cancer (AJCC) stage III and IV] than in early stage ones (AJCC stage I and II) (P < 0.05, Fig. 1b).

We evaluated RASAL2 mRNA expression in multiple groups of primary tumors and normal colon mucosa. Upregulation of RASAL2 mRNA in CRC was validated independently in the Hong Kong local cohort (P < 0.01), Gene Expression across Normal and Tumor tissue (GENT) [26] (GSE20916 [27], P < 0.01 and GSE21510 [28], P < 0.01) and The Cancer Genome Atlas (TCGA) [29, 30] (P < 0.05) (Fig. 1c). We further assessed RASAL2 mRNA expression in 18 pairs of primary tumor and matched adjacent normal tissues. The mean RASAL2 expression was significantly higher in primary tumors compared with their matched adjacent normal colon mucosa (P < 0.01, Fig. 1d). A similar trend was observed in 32 paired primary tumor and normal samples in GENT data [26] (GSE8671 [31], Fig. 1d, P < 0.01). We also examined RASAL2 mRNA expression in 27 metastatic CRC specimens from patients with liver metastasis underwent hepatic resection. Metastatic tumors showed the highest RASAL2 mRNA expression (P < 0.05 vs. primary tumors), followed by primary tumors (P < 0.05 vs. normal), and normal colon mucosa (P < 0.01 vs. metastasis, Fig. 1e). Out of 15 cases with both primary and metastatic tumor samples, 12 (80%) had a higher RASAL2 mRNA expression level in metastatic tumors than in their primary counterparts (P < 0.05, Fig. 1f).

We evaluated RASAL2 protein expression in primary and metastatic colorectal tumors as well as normal tissues by IHC on tissue microarrays. RASAL2 proteins were predominantly cytoplasmic in the tumor cells. Figure 2a show representative IHC positive (H-score > 100) and negative (H-score ≤ 100) images. Upregulation of RASAL2 protein expression was observed in 40.4% (84/208) of primary CRCs. Moreover, there was no significant difference in RASAL2 protein expression level between RAS mutant and RAS wild-type CRC cases (Fig. 2b). RASAL2 overexpression was also independent of KRAS/NRAS mutation status as well as gene expression subtypes in the TCGA data (Additional file 2: Figure S2). As shown in Additional file 1: Table S6, overexpression of RASAL2 was positively associated with advanced TNM stages (P = 0.018), including lymphatic metastasis (P = 0.013) and distant metastasis (P = 0.007). Kaplan-Meier survival analysis showed that CRC patients with RASAL2 positive expression had significantly shorter disease-free survival and overall survival compared with those without (P = 0.043 and 0.004, Fig. 2c). Multivariate analysis by Cox’s proportional hazards regression model revealed that RASAL2 positive expression is an independent prognostic factor for shorter overall survival in CRC (relative risk (RR): 1.561; 95% CI: 1.074 ~ 2.269; P = 0.02, Additional file 1: Table S7). We validated the prognostic significance of RASAL2 in TCGA (n = 520), Kaplan-Meier survival analysis also showed that copy number gain of RASAL2 and higher RASAL2 mRNA expression was correlated with aggressive disease-free survival (Fig. 2d) and overall survival in CRC patients (Fig. 2e).

Figure 3a shows RASAL2 protein expression in normal colon and CRC cell lines by western blot. RASAL2 was frequently upregulated in CRC, suggesting its oncogenic role. Its expression could be silenced with siRNA in the CRC cell lines. Successful knockdown of RASAL2 in KRAS/NRAS wild-type (DLD-1, HCT 116 and SW620) and mutant (Caco2) cell lines were confirmed by western blot (Fig. 3b). Using MTT cell proliferation and colony formation assays, a significant decrease in cell proliferation rate and anchorage dependent growth were observed in both KRAS/NRAS wild-type and mutant cell lines (Fig. 3c). Using soft agar assay, reduced anchorage-independent growth ability was seen in RASAL-knockdown cells (Fig. 3d). Flow cytometry showed that siRASAL2 inhibited cell growth through inducing cell cycle arrest at G1 phase in the CRC cell lines (Fig. 3e). Western blot analysis revealed that siRASAL2 suppressed the G1-S transition promoter cyclin D1, confirming that siRASAL2 blocked the cell cycle at the G1/S checkpoint (Additional file 2: Figure S3). Next, we examined the effect of RASAL2 knockdown on cell motility as well as invasiveness. We observed that knockdown of RASAL2 by two different siRNAs suppressed cell invasion and migration in DLD-1 and HCT 116 cell lines using Matrigel transwell invasion and cell migration assays, respectively (Fig. 3f). SW620 and Caco2 cells were not studied for the two assays due to endogenous weakness for invasion and migration abilities. To evaluate the effect of siRASAL2 in vivo, early passage of SW620 cells transfected with siControl or siRASAL2 were subcutaneously inoculated on the left or right flank of the anaesthetized mice. The siRASAL2 group formed either no (2 of 5) or smaller tumor (3 of 5) within 24 days (Fig. 3g) when compared to siControl group. In addition, mice injected with siRASAL2 had significantly reduced mean tumor size (P < 0.01) and mean tumor weight (P < 0.01). shRNA stable knockdown system was also applied to xenograft mouse model and the tumorigenic ability of the gene was confirmed (Additional file 2: Figure S4).

Gain-of-function study was performed by stably transfecting RASAL2 expression vectors (empty vector pLPCX and pLPCX-RASAL2 plasmid) into SW480 cells. Ectopic expression of RASAL2 in these cells, as shown by western blot (Fig. 4a), caused a significant increase in cell viability in SW480 compared to empty vector transfected cells (P < 0.01, Fig. 4b). RASAL2-overexpressing cells also showed increased colony formation (1617.0 ± 72.7) compared to cells receiving empty vector (1217.0 ± 44.1) (P < 0.01, Fig. 4c). Moreover, the invasion and migration abilities enhanced significantly in SW480 cells stably expressing RASAL2 (Fig. 4d) compared to those without. In nude mice, SW480-pLPCX-RASAL2 had significantly shorter latency of tumor formation, larger tumor size (P < 0.01) and larger mean tumor weight (P < 0.05) than SW480-pLPCX (Fig. 4e). SW480 xenografts overexpressing RASAL2 showed increased cell proliferation by Ki-67 staining (P < 0.05, Fig. 4f).

To elucidate the molecular mechanism by which RASAL2 promotes development and progression in CRC, expression of RASAL2-modulated target genes in CRC cells transfected with siRASAL2 was compared to that in CRC cells transfected with control siRNA using Illumina HumanHT-12 Expression BeadChip arrays. Successful knockdown of RASAL2 by siRNA was confirmed using qRT-PCR and the efficiency ranged from 13.3 to 42.8% (Additional file 2: Figure S5). KEGG pathway enrichment analysis by STRING [32, 33] showed that the hippo signaling pathway was significantly enriched with differentially expressed genes (Fig. 5a). We confirmed the induction of LATS2 and YAP1 phosphorylation upon siRASAL2 treatment in a panel of CRC cell lines with or without KRAS/NRAS mutation (Fig. 5b). We next investigated whether the oncogenic effect of RASAL2 is dependent on its ability to downregulate LATS2. Given that RASAL2 abrogated the tumor suppressor function of LATS2, we conducted rescue experiment in which we co-transfected siLATS2, siRASAL2 or both into SW620 (KRAS G12 V) and Caco2 (KRAS/NRAS wild-type) cells and evaluated the effect on phospho-YAP1(S127) expression. LAST2 depletion reduced phospho-YAP1 (Ser127) but the level of total YAP1 remained unchanged (Fig. 5c). Double knockdown of LATS2 and RASAL2 increased YAP1 phosphorylation when compared to siLATS2 treated cells but could not fully restored the normal effect of LATS2, suggesting that siRASAL2 could partially rescue the effect of YAP1 dephosphorylation caused by LATS2 depletion (siRASAL2/LATS2 vs. siLATS2, Fig. 5c). Also, cell proliferation rate (Fig. 5d) and anchorage dependent growth (Fig. 5e) in siRASAL2/siLATS2 double knockdown cells were significantly reduced when compared with siLATS2 cells. These findings showed that LATS2 downregulation underlies oncogenesis induced RASAL2 in CRC. From another perspective, we observed a significantly lower level of phosphor-YAP1 in siRASAL2/LATS2 double knockdown cells when compared to siRASAL2-treated cells. This suggested that kinases other than LATS2 might be involved in siRASAL2-mediated YAP1 phosphorylation.

We then examined the subcellular localization of YAP1 upon siRASAL2 treatment in CRC cells by immunofluorescence. Knockdown of RASAL2 significantly increased translocation of YAP1 from nucleus to cytoplasm in SW620 (KRAS G12 V) and Caco2 (KRAS/NRAS wild type) cells (Fig. 6a). Since YAP1 phosphorylation promotes its binding to 14–3-3, cytoplasmic retention as well as subsequent poly-ubiquitination and degradation, we tested whether YAP1 ubiquitination could be stimulated by siRASAL2. As shown in Fig. 6b, YAP1 ubiquitination was elevated when SW620 and Caco2 cells were treated with RASAL2 siRNA, and this is consistent with the situation under which the hippo signaling pathway was turned on. To further delineate the effect of the pathway on RASAL2-mediated tumorigenesis, the pharmacological YAP1 inhibitor verteporfin, which is a small molecule inhibiting TEAD-YAP1 interaction [34], was introduced into SW620 and Caco2. The IC50 of verteporfin was 222.8 nM in SW620-siControl cells but decreased to 134.2 nM when treated with siRASAL2 at 24 h after verteporfin incubation (Fig. 6c). The results in Caco2 cells were similar. To evaluate the correlation of RASAL2 and YAP1 expression in clinical specimens, IHC of YAP1 was performed in 208 CRC samples and 42.3% samples had YAP1 protein detected in the nucleus. Moreover, YAP1 expression was positively correlated with RASAL2 expression in these samples (R = 0.403, P < 0.01, Fig. 6d), and the same results were observed in GENT [26] (GSE14333 [35], GSE17536 [36]) and TCGA [29, 30] (Fig. 6e).

In summary, RASAL2 showed high expression and targeted LATS2/YAP1 axis in CRC. YAP1 was therefore dephosphorylated and translocated to the cell nucleus. This prevented YAP1 from ubiquitination in the cell cytoplasm and functioning as a transcriptional co-activator to stimulate expression of pro-proliferation genes like CCND1 in CRC (Fig. 7). Collectively, we demonstrated the oncogenic role of RASAL2 in promoting tumorigenesis as well as metastasis and revealed that RASAL2 exerted its oncogenic property through LATS2/YAP1 axis of hippo signaling pathway in CRC.