CHAC2, downregulated in gastric and colorectal cancers, acted as a tumor suppressor inducing apoptosis and autophagy through unfolded protein response

Tumor suppressor genes play a key role in cancer pathogenesis. Through massive expression profiling we identified CHAC2 as a frequently downregulated gene in gastric and colorectal cancers. Immunohistochemistry and western blot revealed that CHAC2 was downregulated in most tumor tissues, and 3-year survival rate of patients with high CHAC2 expression was significantly higher than that of patients with low CHAC2 expression (P<0.001 and P=0.001, respectively). The data of univariate analysis and multivariate analysis suggested that CHAC2 could serve as an independent prognostic marker. Our results showed for the first time that CHAC2 was degraded by the ubiquitin-proteasome pathway and CHAC2 expression inhibited tumor cell growth, proliferation, migration in vitro and in vivo. Mechanistic study showed that CHAC2 induced mitochondrial apoptosis and autophagy through unfolded protein response. So in gastric and colorectal cancer CHAC2 acted as a tumor suppressor and might have therapeutic implication for patients.

CHAC2, downregulated in gastric and colorectal cancers, acted as a tumor suppressor inducing apoptosis and autophagy through unfolded protein response Shuiping Liu 1 , Weiqiang Fei 1 , Qinglan Shi 1 , Qiang Li 1 , Yeye Kuang 1 , Chan Wang 1 , Chao He 1 and Xiaotong Hu* ,1 Tumor suppressor genes play a key role in cancer pathogenesis. Through massive expression profiling we identified CHAC2 as a frequently downregulated gene in gastric and colorectal cancers. Immunohistochemistry and western blot revealed that CHAC2 was downregulated in most tumor tissues, and 3-year survival rate of patients with high CHAC2 expression was significantly higher than that of patients with low CHAC2 expression (Po0.001 and P = 0.001, respectively). The data of univariate analysis and multivariate analysis suggested that CHAC2 could serve as an independent prognostic marker. Our results showed for the first time that CHAC2 was degraded by the ubiquitin-proteasome pathway and CHAC2 expression inhibited tumor cell growth, proliferation, migration in vitro and in vivo. Mechanistic study showed that CHAC2 induced mitochondrial apoptosis and autophagy through unfolded protein response. So in gastric and colorectal cancer CHAC2 acted as a tumor suppressor and might have therapeutic implication for patients. Gastric and colorectal cancers remain highly prevalent tumors, causing high morbidity and mortality in the world. Its molecular pathogenesis is believed to be a multistep process involving a number of genetic and epigenetic factors including the inactivation of tumor suppressor genes (TSGs). 1,2 Exploring and characterization of more novel TSGs will greatly facilitate to develop novel effective individualized therapeutic strategies.
We searched for candidate TSGs in digestive tumors through massive expression profiling, and found that CHAC2 was a frequently downregulated gene in gastric and colorectal cancers. CHAC2 is a member of cation transport regulator-like protein (CHAC) family which contain the BtrG/γ-glutamyl cyclotransferase (γ-GCT) fold and belong to the family of γ-GCTs. CHAC family members are reported to function as γ-GCT acting specifically to degrade glutathione but no other γ-glutamyl peptides using in vivo studies and novel in vitro assays. 3,4 The CHAC family is conserved across all phyla and represents the first cytosolic pathway for glutathione degradation in mammalian cells. Until now, two CHAC members, CHAC1 and CHAC2, have been identified in human genome. As we know, few literatures about CHAC2 have been reported, while there are several reports about CHAC1. CHAC1 overexpression enhanced apoptosis with an increase in TUNEL staining, cleaved PARP, and AIF nuclear translocation. 3 Later, Kumar et al. 4 reported that overexpression of mouse wild-type CHAC1 but not its catalytically inactive mutant (E4Q) led to glutathione depletion and enhanced apoptosis which could be reversed by addition of glutathione in yeasts. Overexpression of human CHAC1 also led to a robust depletion of glutathione in HEK293 cells. 5 Moreover, Goebel et al. 6 reported that elevated mRNA expression of CHAC1 splicing variants is associated with poor outcome for breast and ovarian cancer patients.
CHAC family members function as γ-GCTacting specifically to degrade glutathione, followed by disturbing the oxidation reduction potential of cell. 7 Some pathophysiological conditions including disturbing the oxidation reduction potential of cell, known as endoplasmic reticulum (ER) stress, disrupt ER function and cause the accumulation of unfolded and/or misfolded proteins in the ER. 8 ER stress leads to the activation of an adaptive signaling program, known as the ER stress response or unfolded protein response (UPR), which is mediated by three ER resident stress sensors, including PERK, IRE1, and ATF6. 9 In a preliminary study, we found that CHAC2, an important member of CHAC family, was downregulated and might be a candidate TSG in gastrointestinal cancer. The role of CHAC2 in cancer progression is poorly explored. So in this study, we investigated the regulation mechanism of CHAC2 expression and explored its functions and signaling pathway in the development of gastric and colorectal cancers.

Results
Expression of CHAC2 was downregulated in cancer cell lines. Firstly, we tested the CHAC2 expression in three gastric cancer cell lines and eight colorectal cancer cell lines.
CHAC2 mRNA was expressed in all the tested cell lines, but the protein was hardly detected in most cell lines except SW48 (Figure 1a). Then we quantified the relative mRNA to protein levels in cell lines by real-time PCR (Supplementary Figure S1A). The results revealed that the protein expression levels still correlated with the mRNA expression levels, suggesting that CHAC2 expression was regulated at both translational and post-translational level.

CHAC2 induced apoptosis and autophagy S Liu et al
To investigate the biofunction of CHAC2, CHAC2 expression plasmid was transfected into cancer cells to construct stable monoclonal CHAC2-expressed cell lines AGS-CHAC2 and SW620-CHAC2. Meanwhile, control vector plasmid was transfected to construct control cells AGS-vector and SW620vector. CHAC2 expression in these transfected cell lines was confirmed by RT-PCR, qRT-PCR, and western blot (Figure 1b and Supplementary Figure S1B). We further investigated the effects of 5-fluorouracil (5-FU) and Brefeldin A (BFA) on the expression of CHAC2 in these cells. Western blot results showed that the CHAC2 protein expression was significantly increased after 5-FU or BFA treatment (Figure 1c).
Ubiquitin-proteasome pathway mediated CHAC2 degradation. Since CHAC2 expression might be regulated at posttranslational level, we examined the stability of CHAC2 in AGS-CHAC2 and SW620-CHAC2. It revealed that CHAC2's half-life (t 1/2 ) was shorter than 1 h (Figure 1d) by treatment with protein synthesis inhibitor Cycloheximide (CHX). To identify which pathway might involve in CHAC2 degradation, proteasomal (MG-132) or lysosomal (leupeptin (Leu)/chloroquine diphosphate salt (CQ)) inhibitor was used together with CHX. Among them, only MG-132 blocked CHAC2 degradation (Figure 1d). We also observed that the degradation of CHAC2 protein was gradually enhanced with increasing concentration of ubiquitin added to the culture media ( Figure 1e). In addition, ERK that mediates serine phosphorylation, ubiquitination and degradation of cortactin 10 showed correlation with CHAC2 degradation (Figures 1d and e). To explore the specific ubiquitin pathway genes acting on CHAC2, we used the Human Ubiquitination Pathway RT 2 Profiler PCR Array to compare the expression of 84 key genes involved in the regulated degradation of cellular proteins by the ubiquitin-proteasome between CHAC2 expression silenced SW620 cells and SW48 cells with CHAC2 high expression. The results showed that most genes were upregulated in SW620 cells (Supplementary  Table S1). However, what impressed us was that among these genes RNF148 expression significantly increased nearly 1863 times in SW620 as compared to SW48 (Figure 1f). It probably was the specific ubiquitin ligase acting on CHAC2. Then we checked RNF148 expression in all used cell lines, it expressed highly in all CHAC2 nearly silenced cells but not expressed in SW48 (Figure 1g and Supplementary Figure S1C). So its expression was negatively correlated with CHAC2 protein expression. Moreover, we found that 5-FU and BFA treatment could remarkably decrease RNF148 expression (Figure 1h and Supplementary Figure S1D). This is further supported that RNF148 may be the specific ubiquitin ligase acting on CHAC2 since CHAC2 protein expression was significantly increased after 5-FU or BFA treatment. Anyway, more detailed molecular mechanisms of the ubiquitin-proteasome regulation of CHAC2 expression will be studied in the near future.
Correlation of CHAC2 expression with clinicopathological parameters of patients. We evaluated the CHAC2 protein expression in primary gastric and colorectal cancer tissues by immunohistochemistry (IHC) and western blot. Representative immunohistochemical staining of CHAC2 protein in normal and cancer tissues was shown in Figure 2a. CHAC2 protein was highly expressed mainly in the cytoplasm of normal tissues. In contrast, CHAC2 was silenced or just low expressed in gastric and colorectal cancer tissues. In addition, representative western blot staining of CHAC2 protein was shown in Supplementary Figure S2. CHAC2 expressed higher in paired non-tumor tissues compared with tumor tissues in 81.25% (13/16) gastric and 87.5% (14/16) colorectal cancer cases, respectively.
The correlations between CHAC2 protein expression and clinicopathological parameters of gastric and colorectal cancer patients were summarized in Supplementary Tables S2 and S3. CHAC2 expression was significantly correlated with histopathological grading (P = 0.002), depth of invasion (P = 0.039), lymph node metastasis (P = 0.017) and TNM stages (P = 0.011) in gastric cancer patients. In colorectal cancer patients, CHAC2 expression was significantly correlated with lymph node metastasis (Po0.001), distant metastasis (P = 0.017) and TNM stages (Po0.001). To further determine the role of CHAC2 expression in gastric and colorectal cancer development, all patients were followed up for overall survival after surgery. Kaplan-Meier survival indicated that the cumulative survival rate of patients with high CHAC2 expression was significantly higher than those with low CHAC2 expression in both gastric (Po0.001) and colorectal cancer (P = 0.001) cases ( Figure 2b).
Effects of CHAC2 on the growth and proliferation of cancer cells. Results above suggested that CHAC2 was a candidate TSG. We investigated the effects of ectopic CHAC2 on the growth of tumor cells. Colony formation assays showed that ectopic CHAC2 expression significantly suppressed the colony formation efficiency of gastric and colorectal cancer cells ( CHAC2 suppressed the migration of gastric and colorectal cancer cells. We investigated the effects of CHAC2 expression on migration of cancer cells by wound-healing assays and transwell migration assays. The results of woundhealing assays showed that the CHAC2-transfected cells took much longer time to close a scratch wound than that of controls ( Figure 3d). Similar results were found in transwell migration assays, CHAC2 expression led to significant suppression of gastric and colorectal cancer cell migration. Moreover, CHAC2 knockdown showed promotion effects on wound healing (Supplementary Figure S3) and migration (Figure 3e).

CHAC2 enhanced apoptosis of cancer cells.
To determine whether tumor cell growth and proliferation inhibition by CHAC2 was related to cell cycle arrest or apoptosis, cell cycle and apoptosis analysis was performed. There was no significant difference in cell cycle phase distribution between CHAC2-transfected cells and controls (Supplementary Figure S4). Regarding cell apoptosis, we checked the spontaneous and 5-FU induced apoptosis of CHAC2transfected cells and the controls. The percentage of spontaneous and 5-FU induced apoptotic cells was significantly higher in CHAC2-transfected cells relative to the controls ( Figure 4a). We further knocked down the CHAC2 expression in SW48 and then checked effects of knockdown on cell apoptosis. The percentage of spontaneous and 5-FUinduced apoptotic cells was significantly lower in CHAC2 knockdown cells as compared with the negative control cells (Figure 4a), which further confirmed that CHAC2 induced tumor cell apoptosis.
CHAC2 led to glutathione depletion, increased intracellular Ca 2+ concentration and reactive oxygen species level. Since glutathione depletion is an important factor for apoptosis initiation and execution, the effect of CHAC2 expression on glutathione concentration was evaluated. Both total GSH (T-GSH) and GSH in CHAC2-transfected cells were significantly lower than those of controls ( Figure 4b). It was reported that increase of intracellular Ca 2+ played a central role in cell death, 11 we subsequently investigated the effects of CHAC2 expression on intracellular Ca 2+ concentration. It revealed that CHAC2-transfected cells showed higher intracellular Ca 2+ concentration than that of controls ( Figure 4c). Mitochondrial Ca 2+ handling regulates several important processes in cellular physiology including mitochondrial reactive oxygen species (ROS) generation. 12 Similar to the alteration of intracellular Ca 2+ concentration, CHAC2-transfected cells showed increased ROS level comparing with that of controls ( Figure 4d).
CHAC2 induces UPR. CHAC2 led to GSH depletion, which would disturb the oxidation reduction potential of cell and then resulted in triggering UPR. We subsequently investigated the correlation of CHAC2 expression with UPR by western blot. All the three ER resident stress sensors including PERK, IRE1 and ATF6 were significantly upregulated in CHAC2transfected cells exposed with or without 5-FU. After treatment with or without 5-FU, the important components of UPR, including ATF4, CHOP and XBP-1s, were upregulated in CHAC2-transfected cells than those of controls ( Figure 5a). On the contrary, in SW48 cells, these important components of UPR were downregulated in CHAC2 knockdown cells as compared with the negative control cells (Figure 5a).
CHAC2 promoted mitochondrial apoptosis. In response to ER stress, the intrinsic pathway is triggered, where mitochondria plays a central role in the control of apoptosis. 13 The collapse of the mitochondrial transmembrane potential led to releasing of mitochondrial apoptogenic factor cytochrome c (cyto c) and Smac, so the cytosolic and mitochondrial fractions were analyzed by immunoblotting. Moreover, most stimuli that induce mitochondria-dependent apoptosis is controlled by proteins of the Bcl-2 family, so we checked pro-apoptotic proteins PUMA, Bim, Bid, p-p53, Bax and pro-survival members Bcl-2 by western blot. Our results showed that lots of pro-apoptotic proteins were upregulated in CHAC2-transfected cells (Figure 5a). However, in SW48 cells, the pro-apoptotic proteins were downregulated in CHAC2 knockdown cells than that of the negative control cells (Figure 5a). Meanwhile, anti-apoptotic protein Bcl-2 was decreased in CHAC2-transfected cells than that of controls, and was upregulated in CHAC2 knockdown SW48 cells as compared with the negative control cells (Figure 5a). Further detection showed that mitochondrial cyto c of CHAC2transfected cells was decreased, while its cytoplasmic cyto c was increased as predicted. Similar results were observed regarding to Smac (Figure 5b).
Effect of CHAC2 on autophagy of gastric and colorectal cancer cells. Autophagy is widely involved in the pathogenesis of many diseases, especially cancers. 14 Current evidence supports the idea that autophagic cell death suppresses tumorigenesis. 15 We evaluated the effect of CHAC2 expression on cells which were transiently transfected with GFP-LC3 plasmids. More punctate fluorescence was observed in CHAC2-transfected cells than that of controls, along with increased autophagy LC3 II and decreased p62 (Figure 5c). These results suggested that CHAC2 expression contributed to autophagy induction in gastric and colorectal cancer cells.
CHAC2 expression was positively correlated with XBP-1s, active caspase-3, or Beclin 1 in clinical samples.  Table  S8). Representative IHC results are shown in Figure 2c. Therefore, in vivo study also showed that CHAC2 could induce mitochondrial apoptosis and autophagy through UPR.
CHAC2 inhibited tumor growth and peritoneal metastasis in nude mice. At last, we evaluated the effects of CHAC2 on tumor cell growth in vivo by injection of SW620-CHAC2 and SW620-vector into BALB/c nude mice cells (n = 6, respectively). The growth curve of CHAC2-transfected and vectortransfected SW620 cells in nude mice is shown in Figure 3f. No obvious tumor was observed from the BALB/c nude mice injected with CHAC2-transfected cells. We further investigated whether CHAC2 expression had a suppressive effect on peritoneal metastasis by in vivo model. Peritoneal dissemination visualized as innumerable whitish nodules was observed in the abdominal cavity, mesenterium and liver of all mice inoculated with SW620-vector cells. While no obvious peritoneal dissemination in mice inoculated with SW620-CHAC2 cells (Figure 3g).

Discussion
In this study, we checked the expression of CHAC2 in gastric and colorectal cancer cell lines and primary tissues. All the tested cell lines exhibited CHAC2 mRNA expression, but only SW48 showed obvious CHAC2 protein expression, suggesting that the CHAC2 expression might be regulated at the level of translation or post translation. Though the quantified results of the relative mRNA to protein levels suggested that CHAC2 expression was regulated at both translational and posttranslational level. We found CHAC2 half-life was shorter than 1 h. The ubiquitin-proteasome pathway but not the lysosome pathway was mainly involved in CHAC2 degradation. Similar result was once reported for mouse CHAC1 where CHAC1 protein expression was only detected in the presence of proteasome inhibitor (MG-132) though its mechanism is still unknown. 16 In addition, ERK, which mediates serine phosphorylation, ubiquitination and degradation of cortactin, was also involved in CHAC2. 10 Our study results have pointed out that RNF148 may be the specific ubiquitin ligase acting on CHAC2. RNF148 was just identified as a novel gene recently. 17 It is abundantly expressed in testis and slightly expressed in pancreas but not in other tissues including normal gastric and colorectal mucosa. RNF148 encodes a polypeptide of 305 amino acids (AA), including a C3H2C3type RING-H2 finger domain at the C-terminus from AA residues 258 to 299. This domain has a high identity with that of other human E3 ubiquitin ligases such as RNF128, 18 RNF130, 19 and mouse Rnf133. 20 In vitro ubiquitination reaction has shown that the recombinant RNF148 protein has E3 ubiquitin ligase activity. 21 Now we found for the first time that RNF148 may have an important role in human gastric and colorectal cancer development through its E3 ubiquitin ligase function though more detailed molecular mechanisms of its ubiquitin-proteasome regulation of CHAC2 expression will be studied further.
In both gastric and colorectal cancer cases, CHAC2 was found to be frequently downregulated. CHAC2 expression was significantly correlated with many clinicopathological parameters of patients. These findings, together with the results that restoration CHAC2 expression inhibited tumor cell growth, proliferation and migration in vitro and in vivo, suggested that CHAC2 acted as a tumor suppressor in gastric and colorectal cancers. Sequence alignment of CHAC2 homologues revealed that it shares 50% AA identity with CHAC1. Similar to CHAC1, CHAC2 also contains signature motif and catalytic glutamate residue and belongs to the family of γ-GCTs (Supplementary Figure S5). We showed that CHAC2 led to depletion of T-GSH and GSH. GSH acts as a major antioxidant within cells by maintaining a tight control of the redox status. 22 GSH depletion disturbing the oxidation reduction potential of cell will result in ER stress and triggering UPR. In this study, all the three major ER stress sensors: IRE1, ATF6 and PERK, together with XBP-1 s and ATF4 were upregulated in CHAC2transfected cells. Each sensor uses a unique mechanism to activate and regulate ER stress target genes. 23 Among them, IRE1α excises the mRNA of XBP-1s to generate spliced XBP-1 (XBP-1s) which is translocated to the nucleus to induce the upregulation of its target genes. [24][25][26][27][28] Under ER stress conditions, ATF6 translocates to the Golgi, following releasing a cytosolic fragment that directly controls genes encoding ERAD components and XBP-1s. Finally, 29,30 PERK phosphorylates eIF2α. eIF2α selective translate of the transcription factor ATF4. 31 Unmitigated ER stress induces apoptosis to eliminate irreversibly damaged cells. Cell death under ER stress depends on the core mitochondrial apoptotic pathway. 32 It is reported that sustained PERK signaling makes ATF4 upregulating the pro-apoptotic transcription factor CHOP. CHOP upregulates the transcription of Bim and downregulates the anti-apoptotic protein Bcl-2 which is crucial for the control of ER stress-induced apoptosis. 33,34 In CHAC2-transfected cells, we observed that CHOP, Bax, and Bak were upregulated, while Bcl-2 was dowregulated. CHOP can also upregulate BH3-only proteins which regulate the activation of Bax or Bak to trigger apoptosis. In addition, Bax or Bak can be also upregulated by IRE1α-JNK and ATF6-Bid, which were upregulated in CHAC2-transfected cells. Meanwhile, CHOP also upregulates growth arrest and DNA damage-inducible34 (GADD34) which may induce the generation of ROS. This was further confirmed by the fact that the level of ROS was increased in CHAC2-transfected cells than that of controls. In addition to ROS, altered calcium homeostasis, which was observed in CHAC2-transfected cells, may also contribute to the opening of the mitochondrial permeability transition pore (PTP) and promote apoptosis. Further detection showed that both cyto c and Smac were decreased in mitochondria and increased in cytoplasm, suggesting that opening of the mitochondrial PTP makes the mitochondrial cyto c and Smac releasing to cytoplasm. Release of cyto c leads to caspase-3 activation, which was upregulated in CHAC2-transfected cells, and subsequent apoptosis. Our findings showed that CHAC2 induced UPR and stimulated the mitochondrial apoptotic pathway by releasing cyto c that further activates caspasedependent signaling ( Figure 6). Our results showed that CHAC2 induced mitochondrial apoptosis and autophagy through UPR. Both apoptosis and autophagy play an important role in the cell death, normal physiology, and cellular homeostasis, and disturbance of these regulatory pathway may lead to different human diseases including cancers. 14,35 Autophagy plays dual roles during cell life; on the one hand, it facilitates cell survival under stressful conditions, on the other hand, excessive or persistent autophagy promotes cell death. Enhanced UPR and autophagy were observed in CHAC2-transfected cells. It is reported that UPR attenuate protein synthesis at the ER by inhibiting translation, activating mRNA decay, and activating autophagy through the IRE1α-JNK or PERK-ATF4 pathway. 36,37 Our study showed that all these proteins, IRE1α, JNK, PERK and ATF4, were upregulated in CHAC2-transfected cells. Accumulating evidences suggest that the effects of anticancer therapies are not confined to apoptosis but also involve autophagy. The crosstalk between autophagy and apoptosis is quite complicated; on the one hand, some common upstream signals result in the activation of combined autophagy and apoptosis, on the other hand, they may be also mutually exclusive under certain conditions. 38 Further studies are needed to explore the mechanisms of crosstalk between apoptosis and autophagy mediated by CHAC2.
In conclusion, this study provides the first evidence that CHAC2 was downregulated and degraded by ubiquitinproteasome pathway in gastric and colorectal cancers. And CHAC2 acted as a tumor suppressor inducing mitochondrial apoptosis and autophagy simultaneously through UPR. A better understanding of the molecular mechanisms of CHAC2 in gastrointestinal cancer development may lead to a more effective management of gastric and colorectal cancer patients with the low expression of CHAC2. Protein extraction and western blot. Cells were collected from cultured dishes and were lysed in a RIPA lysis buffer (Beyotime, Hangzhou, Zhejiang, China) supplemented with inhibitors of proteases. Protein concentration was determined using a BCA Protein Assay Kit (Beyotime). Cell lysates (40 μg protein/line) were separated on a 6-15% SDS-PAGE and transferred to nitrocellulose membrane (Bio-Rad, Hercules, CA, USA). The blotted membranes were blocked in 5% skim milk for 1 h and incubated over-night at 4°C with the following primary antibodies (1:1000). Afterward, the blotted membranes were incubated with HRP-labeled secondary antibody (1:2000) in TBST for 1 h. Primary and secondary antibodies used were purchased from Cell Signaling Technologies (Danvers, MA, USA). Detection was carried out using ECL Kit (Pierce Chemical Co., Rockford, IL, USA) and the blots were developed using a Fujifilm Las-4000 Imaging System.
Two independent investigators, who were blinded to patient-related clinical information, evaluated slides on staining intensity, subcellular localization, and the percentage of positive cells three times. To evaluate the association of CHAC2 expression with clinical and pathological parameters, the patients were then grouped into two categories, low-expression and high-expression, based on the percentage of positive cells and the intensity of stained cells.
Construction of stable CHAC2-expressing cell lines. To construct stable CHAC2-expressing cell lines, MegaTran 1.0 transfection reagent (Origene, Rockville, MD, USA) was used according to its instruction. Gastric cancer cell line AGS and colorectal cancer cell line SW620 were transfected with pCMV6 entry-CHAC2 plasmid (pCMV6 entry-CHAC2; Origene) or control vector (pCMV6 entrymock; Origene). Stable CHAC2-expressing and control clones were selected for further study.
Colony formation assays. For colony formation assays, 1000 stable transfected cells were plated in a 10 cm dish and allowed to grow for 2 weeks at 37°C in 5% CO 2 . Surviving colonies (⩾ 50 cells/colony) were counted under a microscope after Giemsa staining. The experiments were performed in triplicate.
Cell viability assay. Cell viability assay was performed using the CCK8 detection kit (Dojindo, Kumamoto, Japan) as described by the manufacturer. Briefly, cells were seeded in 96-well flat-bottomed microtitre plates at a density of 5000 cells/well. After 24 h, the medium was replaced by medium without or with indicated 5-fluorouracil (5-FU; 120 μg/ml for AGS; 200 μg/ml for SW620) and incubated for 24 h. Absorbance was measured on a microplate reader (Synergy H4 Hybrid Reader, Bio-Tek, Winooski, VT, USA) at a wavelength of 450 nm. The experiments were performed in triplicate.
Wound-healing assay. Cell motility was assessed using a scratch wound assay. The transfected cells and the controls were cultured in 6-well dishes with 10 mg/l mitomycin C (MCE, Princeton, NJ, USA) until confluent. The cell layers were carefully wounded using sterile tips and washed twice with fresh medium. Cells were incubated with fresh medium and photographed under a phase contrast microscope at different hours after wounding. The experiments were performed in triplicate.
Cell migration assay. For the transwell assay, cells were trypsinized and resuspended in corresponding medium containing 1% FBS at a density of 1 ×10 6 cells/ml with 10 mg/l mitomycin C. One hundred microliters of the cell suspension was added into the upper chamber of a transwell (Corning, Corning, NY, USA) consisted of inserts containing 8-mm pore-size PET membrane. Six hundred microliters medium containing 10% FBS was placed in the lower chamber. After indicated hours incubation at 37°C in 5% CO 2 , cells remained in the upper chamber was removed carefully by cotton swab, and cells on the bottom side of the chamber membrane were fixed, stained with 0.25% crystal violet, photographed and counted under a light microscope. The experiments were performed in triplicate.
Cell cycle and apoptosis analysis. Cell cycle distribution and the percentage of apoptosis were determined by flow cytometry according to the manufacturer's instruction (BD, San Jose, CA, USA).
Measurement of total GSH, GSH and GSSG concentration. The protein concentration of treated cells was determined using a BCA Protein Assay Kit (Beyotime). The concentrations of reduced glutathione (GSH) and total GSH (T-GSH) were measured using GSSG Assay Kit (Beyotime) as described by the manufacturer. 39 Briefly, T-GSH was assayed using the 5,5-dithio-bis (2-nitrobenzoic) acid (DTNB)-GSSG reductase recycling. GSSG was measured by measuring 5-thio-2-nitrobenzoic acid (TNB) which was produced from the reaction of reduced GSH with DTNB. The concentration of reduced GSH in the sample was obtained by subtracting GSSG from T-GSH.
Measurement of Ca 2+ concentration. Measurement of Ca 2+ concentration was performed using the Fluo-3 AM (Beyotime) according to the protocol with little modification. 40 Calibrate and express the signals as [Ca 2+ ] i using a pseudo calibration method, as described by Trafford et al. 41 where F is the fluorescence signal, F bg is the fluorescence in the absence of Ca 2+ , F max is the fluorescence in the presence of saturating Ca 2+ , and K d (400 nM/l) the dissociation constant of Fluo-3AM.
Measurement of reactive oxygen species production. C-H 2 DCFDA (Invitrogen) was used to measure ROS production as described by the manufacturer. Briefly, 5 × 10 5 cells were seeded into six-well plate, indicated H 2 O 2 was added into the culture medium when cells were 70% confluent. After 24 h incubation, treated cells were stained with 8 μM C-H 2 DCFDA for 20 min, and washed thrice in PBS before flow cytometry analysis.
Autophagy detection. Autophagosome formation is one of the features of autophagy and can be detected by endogenous LC3 or GFP-LC3 puncta incorporating into autophagic vacuoles. Autophagy detection was determined as described previously. 42 Briefly, cells were seeded on 12 mm cover slips into six-well at a density of 5 × 10 5 cel|ls/well. After attachment overnight, the cells were transiently transfected with a GFP-LC3-expressing vector (Origene). After 32 h of treatment, the cells were fixed with 4% formaldehyde, then observed and photographed under fluorescent microscopy. Cells representing several intense punctate GFP-LC3B aggregates with no nuclear GFP-LC3B were classified as autophagic.
The human ubiquitination pathway RT 2 profiler PCR array. The Human Ubiquitination Pathway RT 2 Profiler PCR Array (Cat. no. 330231, Qiagen, Hilden, German) was used to profile the expression of 84 key genes involved in the regulated degradation of cellular proteins by the ubiquitin-proteasome. Total RNA was extracted from SW620 and SW48 with TRIzol. Single stranded cDNA was synthesized from 2 μg of total RNA by using the SuperArray reaction ready first strand cDNA synthesis kit. The cDNAs were mixed with SuperArray RT 2 Real time SYBR Green/ROX PCR master mix and real-time PCR performed in accordance with the manufacturer's instructions. Thermal cycling and fluorescence detection were performed using an ABI ViiA TM 7 Dx System (Applied Biosystems, Foster City, CA, USA), and expression of these 84 E3 ubiquiting ligases gene transcripts was compared between SW620 and SW48.
In vivo subcutaneous tumor model. All of the in vivo experimental protocols were approved by the animal care committee of Sir Run Run Shaw Hospital, Zhejiang University. Viable CHAC2-transfected cells and their controls (5 × 10 6 cells in 0.1 ml PBS) were injected subcutaneously into left forelimb flank of 5-week-old female BALB/c nude mice (six mice per group). Tumor volume was assessed every 2 days for 4 weeks. Tumor volume was calculated by the following formula: (short diameter) 2 × (long diameter)/2.
In vivo peritoneal metastasis model. Viable cells (1 × 10 7 cells in 0.5 ml PBS) were injected into the peritoneal cavity of 5-week-old female BALB/c nude mice (six mice per group). All mice were killed after 15 days. We determined the number of tumor nodules and the presence or absence of any bloody ascites as described before. 43 Statistical analysis. Results are expressed as values of mean ± standard deviation (s.d.). Statistical analysis was carried out using the SPSS 22.0 software package (SPSS, IBM, Chicago, IL, USA). Pearson's Chi-squared tests and Fisher's exact test were used to analyze the association of CHAC2 expression with clinicopathological parameters. The Kaplan-Meier method was used to analyze survival curve, and the differences were conducted by the log-rank test. Cox's proportional hazards regression model was utilized for univariate and multivariate analyses. We performed paired t-test (two-tailed) statistical analysis. Po0.05 was considered statistically significance.