Sec62 promotes stemness and chemoresistance of human colorectal cancer through activating Wnt/β-catenin pathway

CRC cell lines were obtained from the Cell Resource Center (National Infrastructure of Cell Line Resource, NSTI) and cultured in DMEM or RPMI 1640 medium. All medium were supplemented with 10% fetal bovine serum. Cells were maintained in 5% CO₂ at 37 °C in incubators with 100% humidity. Cell line authentications were performed by the provider. The antibodies and reagents for this study are listed in supplementary Table S1.

Tissue microarrays consist of 102 formalin-fixed, paraffin-embedded CRC tissues and non-tumorous colorectal tissues obtained from the CRC patients who underwent curative surgical resection without prior neoadjuvant therapy from January 2004 to December 2008 in Peking University Cancer Hospital. The clinical pathologic characteristics of patients including age, gender, tumor location, carcinoembryonic antigen (CEA) level, tumor size, clinical stage, and distant metastasis are summarized in Supplemental Table S2.

Cells were transfected with plasmid DNA or siRNA RNA duplexes by Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. In transient transfection experiments, plasmid DNA was kept constant with empty vector. shRNAs were delivered by lentiviral infection with lentiviruses produced by transfection of HEK293T cells with the vector pLKO.1. Cells infected with lentiviruses delivering scrambled shRNA (shCtrl) were used as negative control cells. Short interfering RNA (siRNA) sequences were directly synthesized (GenePharma, Shanghai, China). The sequences of shRNAs and siRNAs are listed in the Table S3.

For sphere formation assay, a total of 800 cells were suspended in a serum-free medium and were plated into an ultralow attachment plate. Then, the cells were cultured in DMEM/F12 medium (Invitrogen) supplemented with insulin (Sigma), B27 (GIBCO), EGF (Sigma) and basic FGF (Sigma). For the serial passaging, the primary spheres were collected and resuspended in DMEM/F12 medium with the above supplements after trypsin dissociation. Finally, the number of spheres was counted under microscope and the size of spheres was estimated using Image J software.

For MTT assay, cells were seeded in a 96-well plate and cultured with indicated drugs for 72 h. Then, MTT assay was employed to assess cell viability according to the manufacturer’s protocol (Promega). For colony formation, cells were treated with DMSO or chemotherapeutic agents for 24 h, and subsequently seeded into 6-well plate (500 cell per well). After cultured for 12 days, the colonies were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet. The visible colonies were counted and summarized.

Apoptosis assay was performed using Annexin V-FITC and propidium iodide staining kit (Keygen, Nanjing, China) according to the manufacturer’s protocol. To detect CD133⁺/CD44⁺ cells, CD133-PE (#130–090-853) and CD44-APC (#130–098-110) antibodies (Miltenyi Biotec.) were utilized to label cells. Then, labelled cells were subjected to flow cytometric analyses.

DLD-1 cells were subcutaneously implanted into 4–6 weeks old female nude mice. When tumors reached a size of about 50 mm³, the nude mice were randomly divided into 6 groups. Group 1, 3 and 5 received an intratumoral injection of lentivirial-ctrl shRNA once per week for 5 weeks. Group 2, 4 and 6 received an intratumoral injection of lentivirial-Sec62 shRNA once per week for 5 weeks. At the same time, group 3 and 4 received an additional intraperitoneal injection of 5-Fu (0.15 mg/kg) twice per week for 4 weeks. Group 5 and 6 received an additional intraperitoneal injection of oxaliplatin (30 mg/kg) twice per week for 5 weeks. The tumor size was measured every 5 days and tumor volume was calculated using the formula V = 0.5 × W² × L (V, volume; L, Length; W, Width).

Cells were harvested and cell lysates were prepared in Buffer A (25 mM Tris-Cl pH 7.5, 150 mM KCl, 1 mM DTT, 2 mM EDTA, 0.5 mM PMSF, and 0.2% Nonidet P-40) and used for immunoprecipitation. The indicated antibodies were coupled with a 50% suspension of protein A-Sepharose beads (GE Healthcare) in Buffer IPP500 (500 mM NaCl, 10 mM Tris-Cl pH 8.0, 0.2% Nonidet P-40). Coupled beads were incubated with cell lysates for 2 h at 4 °C. After washing, the precipitates were examined by Western blot using the indicated antibodies.

RIP assay was carried out as previously described [19, 20]. Briefly, cells were treated with UV irradiation and lysed in high salt lysis buffer. Then, magnetic beads coated with 5 μg of antibodies were incubated with the prepared cell lysates overnight at 4 °C. The RNA-protein complexes were washed for 6 times and incubated with proteinase K digestion buffer. RNA was extracted by phenol-chloroform RNA extraction methods. The relative expression of RNA was determined by RT-qPCR and normalized to the input.

To investigate the effect of Sec62 on tumor self-renewal, an in vivo limiting dilution assay was performed. Female BABL/c nude mice (4 to 5 weeks old) were randomly divided into 4 groups (5 mice per group). Sec62-knockdown or control DLD1 cells were injected subcutaneously in the flank with a serial dilution of cells. At the end of 80 days, the tumor incidence of each group was observed and the stem cell frequency was estimated using an online tool at http://​bioinf.​wehi.​edu.​au/​software/​elda.

Me-RIP assay was performed following published protocols [20, 21]. Briefly, Poly (A) mRNAs were isolated from total RNA with poly-T oligo attached magnetic beads (Promega). The cleaved RNA fragments were incubated for 2 h at 4 °C with m⁶A-specific antibody (No. 202003, Synaptic Systems, Germany) in IP buffer (50 mM Tris-HCl, 750 mM NaCl and 0.5% Igepal CA-630) supplemented with BSA. The mixture was then incubated with protein-A beads and eluted with elution buffer (1 × IP buffer and 6.7 mM m⁶A). Eluted RNA was precipitated by 75% ethanol. RNA was finally extracted by phenol-chloroform RNA extraction methods. The relative expression of RNA was assessed by qPCR and normalized to the input.

Tissue sections were deparaffinized in xylene and rehydrated in ethanol. Then, the sections were treated with peroxidase solution and citrate buffer. After treatment with blocking buffer, sections were incubated with a primary antibody at 4 °C overnight. Next, the sections were visualized using an UltraVision Quanto Detection System HRP DAB Kit (ZSGB-Bio, China) according to the manufacturer’s protocols. The staining intensity were evaluated independently by two observers blinded to the clinical outcome. The percentages of positive tumor cells were scored as follows: 1%, 0 points; 1–25%, 1 point; 26–50%, 2 points; 51–75%, 3 points; and 75%, 4 points. The staining intensity was scored as follows: no staining, 0 points; weak staining, 1 point; moderate staining, 2 points; and strong staining, 3 points. Then, the two scores were multiplied to acquire a combined score ranging from 0 to 12.

Cells were fixed with 4% paraformaldehyde for 15 min and permeabilized using 0.2% Triton X-100 for 10 min at room temperature. After blocking with 10% goat serum, the cells were incubated with primary antibodies overnight at 4 °C. After washing with PBS, a FITC-conjugated anti-rabbit antibody and a TRITC-conjugated anti-mouse antibody were added, and the samples were incubated for 1 h at room temperature. Finally, the cells were stained with DAPI to visualize the nuclei.

The significance of the differences was determined via one-way ANOVA or Student’s t-test. Spearman’s correlation coefficient was used to calculate the correlations between the two groups. Kaplan-Meier analysis was employed for survival analysis and the differences in the survival probabilities were estimated using the log-rank test. The statistical analyses were performed using GraphPad Prism or SPSS version 17.0 (SPSS, Inc.). All data was presented as mean ± SEM. A two-sided P < 0.05 was considered to indicate statistical significance. *P < 0.05 and **P < 0.01 for all the analyses.

To explore novel targets in chemosensitivity, we analyzed the DEGs (differentially-expressed genes) in two CRC cohorts consisting of chemoresistant and chemosensitive CRCs in dataset GSE28702. Eighteen genes were upregulated and 10 genes were downregulated in chemoresistant CRCs (Fig. 1a). AUC (area under curve) analysis were further performed to evaluate the involvement of candidate genes in CRC chemosensitivity using two independent datasets, GSE72968 and GSE81005 (Fig. 1b). Among these DEGs, the most significantly upregulated gene SEC62 is of particular interest since its upregulation has been found to be associated with tumorigenesis [16] and the roles of Sec62 in CRC have not yet been illustrated. We evaluated the Sec62 expression levels in human CRC tissues and show that Sec62 expression was higher in the chemoresistant CRC tissues than that in the chemosensitive ones (Fig. 1c), suggesting that Sec62 might act in the CRC chemoresistance.

Next, we depleted Sec62 with two independent shRNAs (shRNA1 and shRNA2) in CRC cell lines and evaluated response of these cells to 5-Fu (fluorouracil) or oxaliplatin. Depletion of Sec62 sensitized the CRC cells to 5-Fu or oxaliplatin treatment ((Fig. 1d and e). Additionally, Sec62 depletion results in the increase of apoptotic cells in response to chemotherapeutic drugs (Fig. 1f). Consistently, the cleavage of Caspase 3 increased in Sec62-depleted cells under the drug treatment (Fig. 1g), confirming that depletion of Sec62 sensitizes CRC cells to the drug treatment. These data suggest that Sec62 suppresses the chemosensitivity of CRC cells.

Since chemoresistance is a key feature of cancer stemness [22], we wanted to know if Sec62 modulates the stemness of CRC cells. Linear regression analyses using The Cancer Genome Atlas (TCGA) CRC dataset indicated that Sec62 mRNA level is positively correlated with the CSC markers/regulators including CD44, CD133, EPCAM, ITGB1 and BMI1 (Additional file: Figure S1A), indicating that Sec62 might regulate the stemness of CRC.

To determine the effect of Sec62 on the stemness of CRC, tumor sphere assay was employed. The number and average diameter of the spheres derived from the Sec62-depleted CRC cells were less than those derived from the control cells (Fig. 2a, b and Additional file: Figure S1B), confirming that Sec62 enhances the stemness of CSCs. Accordingly, Sec62 overexpression enhanced primary and secondary spheroid formation (Fig. 2c, d and Additional file: Figure S1C). Importantly, two CSC markers CD133 and CD44 [23, 24], were downregulated in Sec62-depleted cells (Fig. 1e). Flow cytometry analysis also showed that the percentage of cells coexpressing CD44 and CD133 was reduced when Sec62 was depleted (Fig. 2f). These data revealed that Sec62 maintains the stemness of CRC. The limited dilution assay in mice xenografts showed that loss of Sec62 caused a great reduction of CSC frequency (Fig. 2g). The shSec62 xenografts displayed lower expression levels of CD44 and CD133 than shCtrl xenografts (Additional file: Figure S1D). These results further confirmed that Sec62 promotes the stemness of CRC.

To uncover the mechanism by which Sec62 maintains the stemness in CRC cells, we purified the Flag-Sec62-specific immunocomplex from DLD1 cells and identified Flag-Sec62-interacting proteins by mass spectrometry (Fig. 3a). Among the Flag-Sec62-interacting proteins, β-catenin is of particular interest due to its pivotal role in cancer stemness [25].

The endogenous interaction between Sec62 and β-catenin was confirmed by co-immunoprecipitation (co-IP) (Fig. 3b). The co-localization of Sec62 and β-catenin was also observed in cells (Fig. 3c). The GST pull-down experiments performed with purified Sec62 and GST-β-catenin show that Sec62 specifically bound to GST-β-catenin other than GST-GSK3β (Fig. 3d), indicating that the Sec62 interacts with β-catenin in vitro. Further, the Sec62-binding region of β-catenin were narrowed down to the 8–9 arm repeats of armadillo (ARM) domain by GST pull-down experiments (Fig. 3e and f). The C-terminus of Sec62 was required for the interaction with β-catenin (Fig. 3g). Most of the β-catenin’s binding partners binds to β-catenin ARM repeats through the conserved DxqqxFx2-7E binding motifs (q and F denote hydrophobic and aromatic residues, respectively) [26]. We found a similar region in Sec62 C-terminus with high sequence consensus to the β-catenin binding motif and this β-catenin binding like (BCBL) motif is highly conserved in various species (Additional file: Figure S2A). We thus constructed a GST-Sec62 deletion mutant (GST-Sec62 DA) lacking the BCBL (Additional file: Figure S2B). Indeed, the GST-Sec62 DA muant lost the capability of binding to β-catenin (Fig. 3h). Together, these results indicate that Sec62 is a novel β-catenin binding protein.

Knockdown of Sec62 significantly decreased the β-catenin protein level without affecting its mRNA level in DLD1, HT29 and LOVO cells (Fig. 4a and b). This effect was reversed by the proteasome inhibitor MG132 (Fig. 4c), suggesting that Sec62 depletion promotes proteasome-mediated β-catenin degradation. Since APC is truncated in these cells, we analyzed β-catenin level after depleting Sec62 in HEK293 and RKO cells which express wild-type (WT) APC with little autocrine Wnt signaling. Depletion of Sec62 reduced β-catenin levels only when these cells were treated by Wnt3a (Fig. 4d). These findings suggest that Sec62 inhibits β-catenin degradation in APC-truncated cells or in WT APC cells when Wnt signal is present. Further, we determined if Sec62 regulates β-catenin level dependent on the BCBL motif. Flag-Sec62 and Flag-Sec62 C-terminal increased β-catenin level, while the Flag-Sec62 NM or Flag-Sec62 DA mutants lost this capability (Fig. 4e and f), indicating that Sec62 controls β-catenin level through binding to β-catenin.

The ubiquitin-proteasome system is critical for determination of protein stability, including β-catenin stability [27]. Thus, we evaluated the effect of Sec62 intervention on ubiquitination of β-catenin. Sec62 depletion enhanced poly-ubiquitination of β-catenin in DLD1 cells (Fig. 4g). Knockdown of Sec62 increased poly-ubiquitination of β-catenin in RKO cells treated with Wn3a (Fig. 4h). Accordingly, Flag-Sec62, but not the Flag-Sec62 NM or DA mutant reduced the poly-ubiquitination of β-catenin in DLD1 or RKO cells (Fig. 4i and j). These results demonstrated that Sec62 inhibits poly-ubiquitination and degradation of β-catenin.

The activity of β-catenin destruction complex plays essential role in controlling β-catenin stability [8, 28]. To determine if Sec62 impairs the activity of β-catenin destruction complex, the ratio of phosphorylated β-catenin/total β-catenin was evaluated as previously described [29]. The ratio of phospho-β-catenin/total β-catenin was increased upon knockdown of Sec62 in DLD1 and HT29 cells (Fig. 5a). Moreover, Sec62 depletion increased the ratio of phospho-β-catenin/total β-catenin in Wnt3a-treated HEK293 or RKO cells (Fig. 5b). In contrast, Flag-Sec62 reduced the ratio of phospho-β-catenin/total β-catenin, while the Sec62 DA mutant lost this function (Fig. 5c). Furthermore, Sec62 depletion decreased the expression of AXIN2, one of β-catenin downstream genes, and had no effect on AXIN1 level, which participates in the assembly of the β-catenin destruction complex (Additional file: Figure S3A). Taken together, our results demonstrated that Sec62 inhibits the activity of the β-catenin destruction complex to enhance β-catenin activation.

Since Sec62 binds to the 8–9 ARM repeat domain of β-catenin which also interacts with APC, Sec62 might compete with APC for binding to β-catenin. Flag-Sec62, but neither the Flag-Sec62 DA nor Flag-Sec62 NM mutant, inhibited the interaction between APC and β-catenin (Fig. 5d and Additional file: Figure S3B, C), indicating that Sec62 disrupts the β-catenin-APC interaction through binding to β-catenin. Since the 15aa and 20aa repeats of APC binds to the ARM repeats 6–10 of β-catenin [30], we thus examined the interaction between GFP-APC 15aaR1 or GFP-APC 20aaR3 and β-catenin when HA-Sec62 was overexpressed. HA-Sec62, but not the HA-Sec62 NM mutant, blocked the interaction between β-catenin and GFP-APC 15aaR1 or GFP-APC 20aaR3 (Fig. 5e), confirming that Sec62 competes with APC for the binding to β-catenin. However, the interaction between β-catenin and AXIN1 was not affected by Flag-Sec62 (Fig. 5f). Moreover, the interaction between APC and β-catenin was enhanced by knockdown of Sec62 (Fig. 5g), indicating that endogenous Sec62 attenuates the interaction between APC and β-catenin.

Given that β-catenin bridges the interaction between the truncated APC and AXIN1 in the assembly of the β-catenin destruction complex [14], we wondered if Sec62 affects the APC-AXIN1 interaction. Flag-Sec62 reduced the APC-AXIN1 interaction in DLD1 cells (Fig. 5h). Thus, Sec62 inhibits the assembly of β-catenin destruction complex by binding to β-catenin in the APC-truncated CRC cells. Then, we evaluated the effect of Sec62 on the APC-AXIN1 interaction in the cells expressing WT APC. Wnt3a treatment resulted in the dissociation of APC with AXIN1 as previously reported [14] and Flag-Sec62 enhanced this effect, indicating that Sec62 reinforces the dissociation of APC-AXIN1 upon Wnt signal (Fig. 5i). Together, Sec62 protects β-catenin from degradation through inhibiting APC-β-catenin interaction.

Based on the observed effect of Sec62 on β-catenin levels, we further investigated whether Sec62 regulated the nuclear translocation of β-catenin and activation of β-catenin signaling. Overexpression of Sec62 increased the nuclear accumulation of β-catenin (Fig. 6a). Moreover, Sec62 knockdown decreased β-catenin recruiment to the promoter of Wnt target genes, such as MYC and LGR5 (Fig. 6b). Accordingly, depletion of Sec62 decreased the expression of Wnt target genes including AXIN2, LGR5 and MYC (Fig. 6c), all of which play critical roles in maintaining cancer cell stemness and CRC progression [4, 31, 32]. To determine if Sec62 regulates the downstream gene of β-catenin in CRC cells expressing WT APC, we evaluated AXIN2 mRNA levels in RKO and HEK293 cells. Knockdown of Sec62 repressed the transcription of AXIN2 in the presence of Wnt3a in these cells (Fig. 6d). Moreover, Flag-Sec62, but not Flag-Sec62 DA mutant, upregulated AXIN2 mRNA levels and this effect was blocked by a specific β-catenin/TCF transcription inhibitor iCRT14 (Fig. 6e). Consistently, depletion of Sec62 downregulated the AXIN2 mRNA level and iCRT14 blocked this effect (Fig. 6e). Together, Sec62 promotes the β-catenin/TCF mediated transcription. We further examined if Sec62 activates β-catenin signaling dependent on BCBL motif in vivo. Overexpression of Sec62 promoted tumour growth in the mouse xenografts while Sec62 DA mutant failed to do so (Fig. 6f and g). Moreover, the mRNA levels of AXIN2 and LGR5 in the xenografts were upregulated by Sec62 but not by Sec62 DA mutant (Fig. 6h). Thus, Sec62 activates the Wnt/β-catenin signaling by binding to β-catenin.

Next, we assessed if Sec62 controls the stemness of CRC through activating β-catenin signalling. Overexpression of Sec62 increased the number of oncosphere while Sec62 DA mutant lost this capability (Fig. 6i and j). Further, overexpression of Sec62 failed to potentiate the stemness when β-catenin was depleted, demonstrating that Sec62 maintains the stemness of CRC dependent on β-catenin. Clony formation was decreased by 5-Fu or oxaliplatin and this effect was reversed by overexpression of Sec62 but not the Sec62 DA mutant, indicating that Sec62 attenuates the chemosensitivity of CRC cells through activating β-catenin (Fig. 6k). Collectively, we demonstrated that Sec62 maintains the stemness by activating the Wnt signalling pathway in CRC.

We next determined the relationship between the Sec62 expression and prognosis of the CRC patients. Sec62 expression was evaluated by IHC in 102 cases of CRC samples. Sec62 was upregulated in the cancer tissues compared with non-cancerous tissues and significantly increased at clinical stage IV (Fig. 7a and b). Importantly, the high Sec62 expression was correlated with shorter overall survival and disease-free survival in the patients (Fig. 7c and d). The GSE17538 data also showed that high Sec62 expression predicts shorter survival in 271 cases of CRC patients (Fig. 7e). These findings indicate that Sec62 is upregulated in CRC and predicts poor prognosis of CRC patients. The expression of Sec62 and β-catenin on 16 paired cancerous and matched non-cancerous sections of human CRC tissues was evaluated by immunoblotting. Sec62 expression was significantly correlated with β-catenin in CRC tissues (Fig. 7f). Moreover, Sec62 expression was positively correlated with the expression of β-catenin downstream genes, including CD44 and MYC, in CRC tissues (Additional file: Figure S4A), indicating that Sec62 activates β-catenin in CRC tissues.

We further explored if Sec62 affects the chemosensitivity of CRC cells in vivo. The nude mice bearing tumor xenograft were injected with control lentivirus (shCtrl) or Sec62 shRNA (shSec62)-lentivirus. Meanwhile, 5-Fu or oxaliplatin was injected. The results showed that Sec62 knockdown effectively attenuated tumor growth as 5-Fu or oxaliplatin did (Fig. 7g and h). Tumors treated with Sec62-shRNA lentivirus were more sensitive to 5-Fu or oxaliplatin when compared with those treated with control-shRNA lentivirus, indicating that Sec62 suppresses the chemosensitivity of CRC cells. IHC confirmed that the expression of Sec62 was inhibited by lentiviral-shSec62 in xenografts (Fig. 7i). Accordingly, Sec62 depletion caused more apoptotic cells upon 5-Fu or oxaliplatin treatment (Fig. 7i). The shSec62 xenografts displayed lower expression levels of CD44 and c-Myc than shCtrl xenografts, indicating that Sec62 promoted the activation of β-catenin (Additional file: Figure S4B). These data demonstrated that depletion of Sec62 sensitized CRC cells to chemotherapeutics in vivo.

Finally, we explored how Sec62 is overexpressed in cancers. Since both the mRNA and protein levels of Sec62 are upregulated in CRC tissues, Sec62 expression might be regulated at mRNA level. Numerous m⁶A sites were found in Sec62 mRNA by RMBase prediction, suggesting that Sec62 mRNA might be controlled by m⁶A modification. We show that m⁶A was significantly more enriched in Sec62 mRNA in DLD1 or SW480 cells than that in normal colorectal cell NCM460 (Fig. 8a). We found that METTL3 depletion reduced Sec62 mRNA and protein levels (Fig. 8b and c). Importantly, depletion of METTL3 decreased m⁶A modification of Sec62 mRNA (Fig. 8d). Accordingly, Flag-METTL3 increased Sec62 protein level (Fig. 8e). These data revealed that Sec62 is upregulated by METTL3-induced m⁶A modification of Sec62 mRNA.

Since m⁶A modification controls mRNAs stability [33], we evaluated the Sec62 mRNA stability upon METTL3 knockdown. The Sec62 mRNA stability was decreased by METTL3 depletion (Fig. 8f), while it was increased by Flag-METTL3, indicating that METTL3 enhances Sec62 mRNA stability by controlling its m⁶A modification. It’s reported that IGF2BP1 (Insulin-like growth factor 2 mRNA-binding protein 1), a“reader” of m⁶A, plays a specific role in reading m⁶A and controlling the stability of m⁶A-modified mRNA [34]. We showed that IGF2BP1 interacted with Sec62 mRNA in CRC cells (Fig. 8g) and silencing IGF2BP1 decreased Sec62 expression levels (Fig. 8h). The half-life of Sec62 mRNA was reduced upon IGF2BP1 depletion (Fig. 8i). Finally, the interaction between IGF2BP1 and Sec62 mRNA was impaired after METTL3 depletion (Fig. 8j), indicating that IGF2BP1 binds to the m⁶A-modified Sec62 mRNA to maintain the Sec62 mRNA stability. Additionally, METTL3 or IGF2BP1 expression was positively correlated with Sec62 expression in CRC tissues (Fig. 8k). Together, we demonstrated that Sec62 is upregulated by m⁶A-mediated stabilization of Sec62 mRNA in CRC.