Down-regulation of OGT promotes cisplatin resistance by inducing autophagy in ovarian cancer.

Cisplatin resistance significantly affects the survival rate of patients with ovarian cancer. However, the main mechanism underlying cisplatin resistance in ovarian cancer remains unclear.
Methods: Immunohistochemistry was used to determine the expression of OGT, OGA and O-GlcNAc in chemoresistant and chemosensitive ovarian cancer tissues. Functional analyses (in vitro and in vivo) were performed to confirm the role of OGT in cisplatin resistance. Autophagy-related proteins were tested by western blot. Transmission electron microscopy and mRFP-GFP-LC3 adenovirus reporter were used for autophagy flux analysis. Immunoprecipitation assay was utilized to detect protein-protein interactions.
Results: We found that O-GlcNAc and O-GlcNAc transferase (OGT) levels were significantly lower in chemoresistant ovarian cancer tissues than in chemosensitive tissues, whereas O-GlcNAcase (OGA) levels did not differ. The down-regulation of OGT increased cisplatin resistance in ovarian cancer cells but had no effect on the efficacy of paclitaxel. The down-regulation of OGT improved tumor resistance to cisplatin in a mouse xenograft tumor model. OGT knockdown enhanced cisplatin-induced autophagy, which reduced apoptotic cell death induced by cisplatin, and promoted autolysosome formation. A reduction in O-GlcNAcylated SNAP-29 levels caused by the down-regulation of OGT promoted the formation of the SNARE complex and autophagic flux.
Conclusion: Our findings suggest that down-regulation of OGT enhances cisplatin-induced autophagy via SNAP-29, resulting in cisplatin-resistant ovarian cancer. OGT may represent a novel target for overcoming cisplatin resistance in ovarian cancer.

Introduction

Ovarian cancer seriously threatens the health of women 1-3. Owing to a lack of reliable early detection methods, about 70% patients are initially diagnosed as stage III or IV 4-7. At present, the standard treatment plan for patients with ovarian cancer is aggressive surgical debulking followed by platinum-based chemotherapy. Progress in radical surgery and chemotherapy has improved the treatment of ovarian cancer 8. However, the intrinsic or acquired chemotherapy resistance of cancer cells is an important factor affecting the efficacy of chemotherapy, resulting in recurrence, metastasis, and mortality; this is a major issue for clinical treatment 9, 10. Therefore, determining the mechanism underlying chemotherapy resistance and developing effective preventive and therapeutic measures have become top priorities in studies of ovarian cancer.

Cisplatin is a first-line chemotherapeutic agent for ovarian cancer; it induces cell apoptosis primarily by inducing cellular DNA damage 11. Although cisplatin-based chemotherapy is very effective for ovarian cancer, drug resistance greatly reduces its efficacy. Recent studies have shown that autophagy plays a crucial role in chemotherapy resistance 12-14. Autophagy is a self-protection mechanism that occurs as an emergency response but can cause cell death 15-17. It is precisely this self-protection ability that increases the resistance of cancer cells to chemotherapeutic drugs. Autophagy consists of the formation of autophagosomes and later stages of autolysosomes, which mainly include the fusion of autophagosomes and lysosomes and the degradation of autophagosome contents 18. The SNARE complex, composed of Syntaxin 17 (Stx17), SNAP-29, and VAMP8, is involved the autophagosome-endosome/lysosome fusion process 19, 20. Previous studies have shown that Stx17, which is recruited by autophagosomes, interacts with SNAP-29 in the cytoplasm and further combines with VAMP7 or VAMP8 on the lysosomal membrane to promote the fusion of autophagosomes and lysosomes 21. However, the regulatory mechanism of SNARE proteins is not clear.

O-GlcNAcylation is a post-translational modification that transfers β-linked N-acetylglucosamine (GlcNAc) to the serine or threonine residues by O-GlcNAc transferase (OGT). In contrast, O-GlcNAcase (OGA) can remove O-GlcNAc from proteins, thus the whole process is reversible and dynamic 22, 23. O-GlcNAcylation plays an important role in maintaining the normal physiological functions of cells 24. Many tumor-associated proteins are modified by O-GlcNAcylation and accordingly it has a pivotal role in both tumor progression and prognosis 25-28. In this study, the role of OGT in cisplatin-induced drug resistance and its relationship with autophagy were examined.

Methods

Clinical materials and immunohistochemistry

Paraffin-embedded tissues were obtained from patients diagnosed with ovarian cancer from 2011 to 2013. All patients underwent primary surgery followed by completed cisplatin-based chemotherapy. Patient characteristics are listed in Table . Patients were divided into two groups, chemoresistant and chemosensitive, defined as follows. Chemoresistant: in the course of postoperative chemotherapy, the disease continued to progress; CA125 did not fall to the normal range after 6 courses of chemotherapy; or patients suffered recurrence within 6 months after the primary chemotherapy. Chemosensitive: patients suffered recurrence beyond 6 months or no recurrence occurred. Before the investigation, the participants were informed about the objectives of this study and signed informed consent forms. Ethics approval was obtained from the Institutional Review Board at Xijing Hospital, Fourth Military Medical University.

The tissue microarray was incubated with antibodies after deparaffinization, rehydration, antigen retrieval, and quenching of endogenous peroxidase. The slides were incubated with anti-OGT (1:100, Abcam, ab96718), anti-O-GlcNAc (1:100, Abcam, ab2739) or anti-OGA (1:100, SIGMA, SAB4200267) overnight at 4 °C. After washing with PBS thrice, slides were incubated with biotinylated secondary antibodies for 1 h at room temperature, and then, antibody labeling was visualized with diaminobenzidine. Finally, the tissue microarray was scanned by tissue microscopy (Pannoramic MIDI, 3D HISTECH, Budapest), and the histochemistry score (H-SCORE) was obtained from Quant center analysis. Statistical analysis was performed according to the H-SCORE.

Cell culture

A2780 and SKOV3 were obtained from the Shanghai Cell Collection (Shanghai, China) and were maintained in RPMI 1640 (Hyclone, SH30809.01) and McCoy 5A medium (Gibco, 16600082), supplemented with 10% fetal bovine serum (Tianhang, 11012-8611) at 37 °C and 5% CO2.

Cell viability assays

Equal numbers of the two cell lines (5,000 cells per well for A2780 and 8,000 cells per well for SKOV3) were added to 96-well plates. Cells were treated with cisplatin at a concentration gradient (6 replicates for each concentration) for 48 h before harvesting. Cell counting kit-8 (CCK-8, Bestbio, BB-4202) reagent was added to the cells according to the manufacturer's instructions and incubated for 2 h. The spectrophotometric absorbance of the samples was measured using Infinite M200PRO (Tecan, USA) at 450 nm.

Western blotting analysis

Western blotting was performed as previously described 29. Briefly, after protein extraction and denaturation, an appropriate SDS-PAGE protocol (15% SDS-PAGE for VAMP8; 12% SDS-PAGE for LC3, SNAP-29 and Stx17; 10% SDS-PAGE for O-GlcNAc and p62; 8% SDS-PAGE for OGT) was selected according to the molecular weight of the protein. The primary antibody was incubated at the appropriate concentration overnight at 4 °C. The membranes were then incubated with secondary antibodies conjugated with horseradish peroxidase for 2 h at room temperature and visualized using ECLTM Western Blotting Detection Reagent (Bio-Rad, USA). The antibodies used in this study were as follows: OGT (1:1,000, Abcam, ab96718), LC3 (1:1,000, CST, 2775s), p62 (1:1,000, Abcam, ab56416), O-GlcNAcylation (1:1,000, Abcam, ab2739), SNAP-29 (1:1,000, Abcam, ab138500), Stx17 (1:1,000, MBL, PM045), VAMP8 (1:5,000, Abcam, ab76021), Actin (1:2,000, CWBIO, CW0096A).

Generation of stable cell lines and transient transfection

A2780 (50,000 cells/well) and SKOV3 (80,000 cells/ well) cells were cultured in 6-well plates for 18 h before being treated with lentiviral vectors for 8 h. The cells were incubated further in normal culture medium for 48 h, and then, puromycin was added for 3-4 weeks to establish stable cell lines. The cDNA encoding SNAP-29, SNAP-29 (Mut) and LC3 were cloned in pcDNA3.1, pEGFP-N1 or pmCherry-C1 vectors. The cells were transfected with Lip2000 (Invitrogen, 11668019). The siRNA sequences were: SNAP-29, 5'-AGACAGAAAUUGAGGAGCA-3'; Stx17, 5'- CCGAAAGGAUGACCUAGUA-3'; VAMP8, 5'-GCAACAAGACAGAGGAUCU- 3'. The siRNA was delivered into cells with Lipo2000 according to the instructions of the manufacturer.

Apoptosis analysis

The percentage of cells that underwent apoptosis or necrosis was determined using the ANXA5-FITC and propidium iodide (PI) analysis kit (Beyotime, C1062). Twenty-four hours after treatment with cisplatin (5 µg/mL), cells were harvested, washed once in PBS, and resuspended at a final concentration of 1×106 cells/mL. ANXA5 and PI staining was performed according to the manufacturer's instructions. Samples were analyzed by a flow cytometer (Cytomics FC 500; Beckman-Coulter, Danvers, MA) equipped with CXP 2.1 software.

Immunocytochemistry

A2780 (5,000 cells/well) and SKOV3 (8,000 cells/well) were added on a Millicell EZ slide (Millipore, PEZGS0816). After culturing for 12 h, cells were treated with cisplatin for 24 h. Cells were washed with PBS three times and fixed with 4% paraformaldehyde for 20 min at room temperature and then treated with 0.1% Triton X-100 for 20 min. Cells were blocked with 0.1% BSA for 1 h at room temperature and subsequently incubated with anti-LC3 (1:100, CST, 2775s) and Lamp1 (1:200, Abcam, ab25630) at 4 °C overnight. After washing three times with PBS, the cells were incubated with the secondary antibody for 1 h and DAPI for 30 min at room temperature. The slides were washed 3 times in PBS and then examined under a fluorescence microscope (Olympus BX51TF, Japan).

Autophagic flux analysis

As previously described 30, cells in the confocal dishes were transfected with mRFP-GFP-LC3 adenovirus (Hanbio, China) at a multiplicity of infection of 80. Twenty-four hours after transfection, cells were cultured for 24 h with or without cisplatin. Then, cells were scanned with a confocal microscope (FV10i, Olympus, Japan). The number of RFP and GFP puncta were calculated in 5 fields, and no fewer than 50 cells in each group were counted.

Transmission electron microscopy

As previously described 30, cells were fixed with phosphate buffer containing 2.5% glutaraldehyde and 2% paraformaldehyde solution for 2 h. After washing with PBS three times, cells were post-fixed with 1% osmium tetroxide for 2 h. The specimens were then dehydrated and embedded in Araldite 812 and cut with a Leica EM UC7 (Leica, Solms, Germany) ultra-microtome. Ultra-thin sections were stained with 1% uranyl acetate and 1% lead citrate. The specimens were observed with a Tecnai G2 Spirit (FEI, Eindhoven, Netherland) transmission electron microscope operating at 120 kV.

Coimmunoprecipitation

Cells were lysed on ice with lysis buffer (20 mM Tris pH 7.5, 150 mM NaCl, 1% Triton X-100, and protein inhibitor) for 20 min. The extracts were centrifuged at 12,000×g for 20 min at 4 °C. After preclearing with protein A/G (7seabiotech, P001-3), the supernatants were incubated with primary antibody and rocked at 4 °C overnight. Then, protein A/G was added and cells were incubated at 4 °C on a rocker platform for 1 h. The beads were washed with lysis buffer 4 times and then subjected to immunoblotting analysis.

Tumor xenograft assay

Five-week-old female BALB/c nude mice (Vital River Laboratories, China) were raised in pathogen-free conditions. SKOV3 cells stably transfected LV-OGT-RNAi or LV-Control-RNAi were resuspended in sterilized PBS. A total 1×106 cells in 200 µL PBS were injected subcutaneously into the flank of nude mice. Cisplatin treatment was initiated on the 23rd day after subcutaneous cell injection; mice in each group were injected with cisplatin (3 mg/kg) or PBS in the abdominal cavity 3 times a week for 2 weeks. Every other day, the tumor volume was measured with a caliper, and the weight of the nude mice was recorded. The formula volume = width × length2 × 0.5 was used to calculate the tumor volumes. All study protocols were approved by the Animal Care and Use Committee of The Forth Military Medical University.

Statistical analysis

Results are presented as the mean ± SD. Statistical significance was assessed using a 2-tailed Student's t-test or chi-squared test. P < 0.05 was considered statistical significance.

Results

Reductions in OGT and O-GlcNAc are associated with chemoresistance in ovarian tumors

Previous studies have shown that aberrant levels of OGT are associated with tumor development and poor prognosis 23, 31, 32. We investigated the expression levels of OGT, OGA, and O-GlcNAc in chemoresistant and chemosensitive ovarian cancer tissues by immunohistochemistry (Table and Figure ). The levels of OGT and O-GlcNAc were significantly lower in tissues from chemoresistant patients than in chemosensitive ovarian cancer tissues (Figure ). The histochemistry scores (H-SCORE) for OGT and O-GlcNAc were significantly lower in chemoresistant ovarian cancer samples than in chemosensitive samples (P < 0.05; Figure ). However, the level of OGA did not differ appreciably between the two groups. (P = 0.3231; Figure ).

Down-regulation of OGT reduces the efficacy of cisplatin in ovarian cancer cells

Cisplatin and paclitaxel are first-line chemotherapy drugs for the treatment of ovarian cancer. We examined the effects of two different chemotherapy drugs on OGT-knockdown ovarian cancer cell lines. We generated stable OGT-deficient A2780 and SKOV3 cell lines using OGT-specific small hairpin RNA (shRNA). The protein level of OGT was reduced in both LV-OGT-RNAi-infected ovarian cancer cell lines compared to control cells (Figure ). We found that OGT knockdown did not affect cell proliferation or apoptosis (Figure ). Then, we treated A2780 and SKOV3 cells with cisplatin in OGT-knockdown and control cells to evaluate cell proliferation. As shown in Figure , compared with the control cells, OGT knockdown significantly improved the viability of cisplatin-treated ovarian cancer cells. Furthermore, flow cytometry was performed by ANXA5 and PI staining to evaluate cisplatin-induced apoptosis in OGT-knockdown and control cells. OGT knockdown significantly decreased cisplatin-induced apoptosis in the two ovarian cancer cells (Figure ). Then, we examined the effects of altered OGT expression on paclitaxel sensitivity in A2780 and SKOV3 cells. The down-regulation of OGT did not lead to a decrease in paclitaxel sensitivity in either cell line (Figure ). Based on flow cytometry, the down-regulation of OGT did not reduce apoptosis induced by paclitaxel (Figure ). These results indicate that the down-regulation of OGT reduces the sensitivity of ovarian cancer cells to cisplatin but has no effect on paclitaxel sensitivity.

OGT knockdown improves ovarian tumor cisplatin resistance in vivo

Next, we investigated the effect of OGT on tumor formation in mice using a xenograft model. Either control (Con-sh) or OGT-deficient (OGT-sh) SKOV3 cells were implanted subcutaneously in the flanks of female nude mice. After 5 days, mice injected with SKOV3-control cells and SKOV3-deficient cells began to form tumors and the tumor volume did not differ significantly between the two groups (Figure ). As shown by IHC staining, the expression of OGT in the OGT-sh group was less than that of the Con-sh group (Figure ). Then, we investigated whether the down-regulation of OGT improves cisplatin resistance in vivo. After 23 days, we treated tumor-bearing mice with cisplatin or PBS and found that the tumor size in cisplatin-treated groups was significantly reduced compared with that in the PBS-treated groups. Moreover, the tumors size in the OGT cisplatin-treated (OGT sh+cis) group was significantly larger than that of the control cisplatin-treated (Con sh+cis) group (Figure ).

Cisplatin resistance due to down-regulation of OGT is associated with autophagy

Previous studies have shown that cisplatin can induce autophagy activation 13, 14, which has a protective role in cells. Therefore, autophagy may be related to cisplatin resistance. We next investigated whether cisplatin induces autophagy in ovarian cancer cells. We found that cisplatin significantly increased the ratio of LC3-II to LC3-I (Figure ). We also examined the expression of p62, a well-known substrate of autophagy, and found that the cisplatin-treated group exhibited significantly reduced p62 expression (Figure ). We next explored whether the down-regulation of OGT enhances the autophagy induced by cisplatin. The ratio of LC3-II to LC3-I was further increased in OGT-deficient cells compared to that in control cells with or without cisplatin treatment (Figure and Figure ). Moreover, we examined the number of LC3 puncta in A2780 and SKOV3 cells by immunofluorescence and found that OGT shRNA markedly increased the number of LC3 puncta (Figure and Figure ). We also observed that the addition of cisplatin after OGT knockdown could further increase the autophagy level of ovarian cancer cells (Figure ).

Down-regulation of OGT enhances cisplatin-induced autophagy by inducing autolysosome formation

To further examine whether autophagy plays a role in cisplatin resistance caused by OGT knockdown, we inhibited autophagy using 3-methyladenine (3-MA) and bafilomycin A1 (Baf), which prevent the formation and fusion of autophagosomes, respectively (Figure ). Neither 3-MA or Baf affected cell proliferation and apoptosis (Figure ). After the addition of Baf, the viability of cisplatin-treated cells decreased significantly, but only slightly decreased after the addition of 3-MA (Figure ). Furthermore, flow cytometry showed that Baf led to a stronger increase in the sensitivity of OGT-deficient cells to cisplatin and induced more cell apoptosis than did 3-MA (Figure ). Thus, we consider that OGT mediate cell autophagy by affecting the fusion of autophagosomes. Subsequently, we monitored autophagy flux using an mRFP-GFP-LC3 adenovirus reporter. Because GFP is sensitive to acidic conditions, when autophagosomes fuse with lysosomes, GFP fluorescence is quenched and only red fluorescence can be detected 15. When GFP and RFP fluorescence appear simultaneously, it indicates that the autophagosomes have not combined with lysosomes, e.g., the phagophore or autophagosome. The red puncta, which exhibited no or weak GFP signals, were obviously increased in OGT-deficient ovarian cancer cells compared to control cells (Figure and Figure ). The total autophagic structures (GFP-RFP+ and GFP+RFP+ LC3 puncta) also increased in the two cell lines. However, the down-regulation of OGT did not lead to a change in the yellow puncta (RFP+GFP+ LC3) in the two cell lines (Figure ). Morphological analysis by transmission electron microscopy revealed that the number of autophagic vacuoles per cell increased significantly in OGT-deficient cells after cisplatin treatment (Figure ). In addition, we compared the number of RFP-LC3 signals that co-localized with Lamp1, a late endosome-lysosome marker, and found that there were more mRFP signals co-localized with Lamp1 in OGT-deficient cells (Figure ). Thus, the down-regulation of OGT promotes the fusion of autophagosomes and lysosomes.

Aberrant autophagy induced by OGT deregulation is associated with O-GlcNAcylation of SNAP-29

Many autophagy-related proteins could be modified by O-GlcNAcylation, including SNAP-29, which is involved in the fusion of autophagosomes and lysosomes 33, 34. When si-SNAP-29 was added to SKOV3 cells, LC3 accumulated and the expression of p62 increased after cisplatin treatment. This phenomenon was not reversed by OGT knockdown (Figure ). The addition of si-SNAP-29 reduced the number of red puncta induced by the down-regulation of OGT but increased the yellow puncta (Figure ). Compared with control cells, the down-regulation of OGT had no effect on the expression of SNAP-29 (Figure ). However, O-GlcNAcylation levels of SNAP-29 were lower in OGT-deficient cells than in control cells (Figure ). In addition, the co-localization of SNAP-29 and LC3 in OGT-deficient cells was significantly increased based on the result of immunofluorescence (Figure ). Therefore, the enhanced autophagy caused by the down-regulation of OGT depends on the O-GlcNAcylation levels of SNAP-29.

Inhibition of O-GlcNAcylation of SNAP-29 promotes autophagy flux in ovarian cancer cells

To further establish the relationship between autophagy and O-GlcNAc-modification of SNAP-29, we determined the level of intracellular autophagy after SNAP-29 glycosylation site mutation. Compared with cells transfected with the wild-type SNAP-29 plasmid, the O-GlcNAcylation level of SNAP-29 was significantly reduced in cells transfected with the SNAP-29 (Mut) mutant plasmid (Figure ). We further examined the co-localization of SNAP-29 and LC3 after cisplatin treatment and found that the co-localization of SNAP-29 and LC3 increased in cells expressing SNAP-29 (Mut) (Figure ). Western blots also showed that the ratio of LC3-II to LC3-I increased in SNAP-29 (Mut) cells, while the expression of p62 decreased after cisplatin treatment. Furthermore, we found that the level of autophagy increased in the cisplatin-treated cells transfected with the SNAP-29 wild-type plasmid, but the effect of cisplatin was significantly weaker than that in SNAP-29 (Mut) cells (Figure 7C). The SNARE complex, including SNAP-29, VAMP8, and Stx17, mediates the fusion of autophagosomes and lysosomes. When si-Stx17 and si-VAMP8 were added to OGT-deficient cells, LC3 accumulated and the expression of p62 increased after cisplatin treatment, indicating that the autophagic flux was inhibited (Figure ). Coimmunoprecipitation showed that the down-regulation of OGT increased the levels of Stx17 and VAMP8 co-precipitated by SNAP-29 after treatment with cisplatin (Figure ). These results suggest that following treatment with cisplatin, the down-regulation of OGT induces the formation of the SNAP29-Stx17-VAMP8 complex, resulting in promotion of the autophagic flux.

Discussion

Ovarian cancer is still one of the most fatal gynecological malignancies. After treatment with cisplatin-based chemotherapeutic drugs, the immediate remission rate of advanced ovarian cancer can reach more than 80%, but most patients relapse within 2-3 years, and almost all recurrent ovarian cancers are resistant to chemotherapy 35, 36. Chemotherapy resistance has become a major challenge for sustainable treatment and new strategies are needed. The formation of drug resistance involves a variety of factors, particularly the acquisition of mutations that result in increased cell resistance to apoptosis induced by chemotherapy drugs 37, 38. An understanding of the molecular mechanisms underlying chemotherapy resistance may provide a basis for the development of treatments for ovarian cancer. Our results provide the first evidence that OGT expression is decreased in chemoresistant tissues, and this reduction in OGT increases the resistance of ovarian cancer cells to cisplatin and promotes cisplatin-induced autophagy to form drug resistance.

OGT exists extensively in the nucleus and cytoplasm and can participate in the regulation of gene expression, metabolic reactions, cellular stress responses, and other reactions related to the cell nutritional status 39. The deregulation of OGT is associated with a variety of human diseases, including cancer, cardiovascular disease, diabetes, and Alzheimer's disease 22, 31, 40. Previous studies have focused on differences in OGT levels between normal and tumor tissues, and relatively few studies have examined changes of OGT in chemotherapy resistance tissues. In our study, we found that OGT and O-GlcNAc levels were significantly lower in chemoresistant ovarian cancer tissues than in chemosensitive tissues. Previous studies have shown that OGT levels are elevated in tumors and are related to malignancy 41, 42, but our results indicated that OGT expression in drug-resistant tissues was reduced. An explanation for this difference is that OGT promotes the occurrence and development of tumors and a reduction in OGT helps cancer cells respond to the harsh environment. OGT is involved in the post-translational modification of various proteins, we evaluated whether changes in OGT are a cause of chemoresistance in ovarian cancer. We found that the down-regulation of OGT reduced cisplatin-induced apoptosis and increased cell viability. However, no changes were observed with down-regulation of OGT in paclitaxel-treated cells. Cisplatin mainly inhibits cell division by damaging DNA whereas paclitaxel, as a cell cycle inhibitor, mainly combines with intracellular microtubule protein to prevent the proliferation of cancer cells. The different effects of OGT down-regulation may be due to the different anti-tumor mechanisms of the two drugs. OGT knockdown had no effect on tumor growth in the mouse xenograft tumor model but promoted tumor resistance to cisplatin. Therefore, the down-regulation of OGT may be associated with cisplatin resistance in ovarian cancer.

Previous studies have shown that OGT is involved in the regulation of autophagic activity. Park et al. found that the overexpression of OGT reduces autophagy via Akt/dFOXO signaling 43. Guo et al. reported that OGT regulates autophagy by mediating the O-GlcNAcylation level of the SNARE protein SNAP-29 in Caenorhabditis elegans and HeLa cells 21. To investigate whether autophagy is involved in cisplatin resistance induced by OGT deregulation, we evaluated the expression of LC3 and p62. The autophagy marker protein LC3 exhibited elevated expression and p62 exhibited reduced expression in OGT-deficient cells in the presence of cisplatin. Autophagy is a self-degradation process, phagocytosing cytoplasmic proteins or organelles and fusing with lysosomes, therefore meeting the basic metabolic needs of cells. The strong protective effect mediated by autophagy enables cells to respond to various intrinsic or extrinsic stimuli 18. We found that chemotherapy can activate autophagy, and the activation of autophagy may help cells resist drug toxicity and develop resistance 13, 14. We found that cisplatin induced autophagy in ovarian cancer cells, and the inhibition of autophagy increased cisplatin-induced apoptosis (data not shown). These results suggest that the down-regulation of OGT increases cisplatin-induced autophagy, and this enhanced autophagy promotes cell resistance to cisplatin toxicity.

Autophagic flux includes the following steps: induction and autophagosome formation, autolysosome formation, and degradation. Our findings showed that OGT knockdown induces autolysosome formation, which promotes autophagic flux. Previous studies have indicated that autophagosomes fuse with lysosomes in a manner dependent on the Rab protein family and the SNARE protein family. Morelli et al. found that a lack of SNAP-29, one of the SNARE proteins, induces the accumulation of autophagosomes, highlighting a major role of SNAP-29 in autolysosome formation 44. Guo et al. found that SNAP-29 is O-GlcNAcylated and the O-GlcNAcylation of SNAP-29 regulates the formation of a complex comprising SNAP-29, autophagosome-localized Stx17, and lysosomal VAMP8 21. In this study, we found that LC3 accumulated and the expression of p62 increased when SNAP-29 was downregulated. P62, a ubiquitin and LC3 binding protein, is a selective autophagy substrate and cargo receptor for degradation of ubiquitinated substrates by autophagy. Since p62 is degraded during autophagy, after we down-regulated SNAP-29 to prevent the fusion of autophagosomes, p62 could not be degraded and then aggregated in cells. Moreover, the effect of OGT knockdown on cisplatin-induced autophagy was abolished after addition of si-SNAP-29. As the sole transferase for O-GlcNAc modification, changes in the expression of OGT directly affect the level of O-GlcNAcylation of intracellular proteins. We downregulated OGT and found that the O-GlcNAcylation of SNAP-29 was significantly reduced, consistent with previous results 21. Moreover, the co-localization of intracellular SNAP-29 and LC3 increased after OGT knockdown, indicating that the inhibition of the O-GlcNAc-modification of SNAP29 promotes the fusion of autophagosomes and lysosomes.

We previously found that increasing O-GlcNAc-modified β-catenin prevented it from combining with the destruction complex 29. Other studies have also shown that the O-GlcNAc modification competitively inhibits phosphorylation 39. Therefore, we evaluated whether the O-GlcNAc modification of SNAP29 prevents its interaction with Stx17 and VAMP8, thereby preventing fusion of autophagosomes and lysosomes, resulting in obstruction of autophagy. When the SNAP-29 glycosylation site was mutated, not only the level of autophagy was increased, but also the combination of SNAP-29 with Stx17 and VAMP8 was increased. Moreover, the effect of downregulation of OGT on cisplatin-induced autophagy was attenuated by si-Stx17 and si-VAMP8. Therefore, down-regulation of OGT causes a reduction in SNAP-29 O-GlcNAcylation, which promotes SNAP-29 interactions with Stx17 and VAMP8 and autophagic flux.

In conclusion, our results indicate that OGT plays an essential role in cisplatin resistance in ovarian cancer. Down-regulation of OGT reduced O-GlcNAc modification of SNAP-29 and subsequently promoted SNAP29-Stx17-VAMP8 complex formation and autophagy, thereby contributing to cisplatin resistance. Our findings suggest that OGT is a new molecular marker for chemotherapy resistance and provide new clinical targets for chemotherapy.

Supplementary figures.

Click here for additional data file.

Table 1

Patient characteristics.

Characteristic	Chemotherapy		P-value
	Sensitive (n=27)	Resistant (n=18)
Age (years)	53.41±1.355	50.61±3.316	0.3819
Clinical Stage
Ⅰ	4	1
Ⅱ	3	1	0.2067
Ⅲ	19	12
Ⅳ	1	4
Histology
Serous	25	15	0.3329
Mucinous	2	3