TSSC3 promotes autophagy via inactivating the Src-mediated PI3K/Akt/mTOR pathway to suppress tumorigenesis and metastasis in osteosarcoma, and predicts a favorable prognosis

Specimens were obtained from 58 patients with histopathologically confirmed osteosarcoma with no preoperative anticancer treatment from Southwest Hospital and Xinqiao Hospital, Third Military Medical University (TMMU), Chongqing, China between February 2011 and November 2015. The last follow-up time was November 2017. The clinicopathological features of the patients are listed in Table 1. Two certified pathologists classified all the specimens as high-grade osteosarcoma. Lung metastasis and local recurrence were diagnosed by both imaging and pathology. The surgical margins and stage were classified according to the Enneking system. Patients with primary osteosarcoma were classified as with or without developed distant metastasis at diagnosis or after surgery. Written informed consent for the experimental studies was obtained from the patients or their guardians. All experiments were approved by the Institutional Ethics Committee of TMMU.

The details of the human benign bone and soft tissue tumor specimens are show in Additional file 1.

The human osteosarcoma cell line SaOS2 was obtained from Cellcook Biological Technology Co., Ltd. (Guangzhou China). The malignant transformed hFOB1.19 cell line (MTF cells) was produced in our laboratory, as previously reported [7]. All the cells were maintained in high-glucose Dulbecco’s modified Eagle’s medium (DMEM, Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS, BI, Kibbutz Beit Haemek, Israel) and 1% penicillin-streptomycin (Hyclone). All the cells were cultured at 37 °C in 5% CO2 and humidified atmosphere.

Cells were treated with pYEEI (100 mM, Src activator, Enzo Life Science, The Netherlands), BEZ235 (500 nM, PI3K inhibitor, MedChem Express, USA) or IGF-1 (100 ng/ml, Akt activator, R&D System, USA), chloroquine (CQ, 8 μM autophagy flux inhibitors, Sigma-Aldrich, USA) for 12 h, separately or in combination.

The pLVX-mCMV-ZsGreen expressing lentiviral vector for TSSC3 and the control was synthesized and obtained from Chongqing Maobai Technology Co., Ltd. (Maobai, Chongqing, China) and the coding sequence of TSSC3 was amplified using the primers 5′-CCGGAATTCGCCACCATGAAATCCCCCGACGAGGTGCTAC-3′ and 5′-CGCGGATCCTCACTTATCGTCGTCATCCTTGTA-3′. The short hairpin RNA (shRNA) in the pHBLV-U6-Puro lentiviral vector targeting ATG5 and its control (scrambled) were purchased from Hanbio Biotechnology Co., Ltd. (Hanbio, Shanghai, China). MTF and SaOS2 cells were infected with the lentivirus vectors according to the manufacturer’s instructions and as previously described [9, 10]. The shRNA target sequences are listed in Additional file 2: Table S1). The overexpression and knockdown function was verified using quantitative real-time PCR and western blotting analysis.

For the cell proliferation assay, a cell light 5-ethynyl-2′-deoxyuridine (EdU) imaging kit (C0075S, Beyotime, Shanghai, China) was used according to the manufacturer’s instructions (Edu; 10 μM, 2 h, for MTF cells and 20 μM, 2 h, for SaOS2 cells). Cells were observed under an inverted phase contrast fluorescence microscope (Olympus, Tokyo, Japan). Cell counting kit-8 (CCK8, C0038, Beyotime) and colony formation assays were carried as described previously [8, 30].

The transmission electron microscopy (TEM) assay was performed as previously described [8]. Briefly, cells (1 × 10⁶) were harvested after transfection with lentivirus overexpressing TSSC3 (overTSSC3) and its control (overCtrl), while fresh tissues were dissected, fixed with 4% glutaraldehyde in phosphate-buffered saline (PBS) at 4 °C overnight, post-fixed in 1% osmium tetroxide, dehydrated with ethanol, embedded in Epon, and then stained with aqueous uranyl acetate and lead citrate before being observed. Images were acquired using a HT7700 electron microscope (HITACHI, Tokyo, Japan). The number of autophagic vacuoles (AVs), including autophagosomes and autolysosomes, in each cell was quantified in 20 randomly selected cells of each group [31].

Quantitative real-time PCR (qPCR) and western blotting analyses were performed as previously described [32]. The primers used in this study are listed in Additional file 2: Table S2 and the details of western blotting are listed in Additional file 1.

The assays were performed as previously described [11, 32]. The details are shown in Additional file 1.

Wound healing and Transwell assays were performed as previously described [4, 11]. Briefly, after appropriate treatments, cells were seeded in 6-well plates and cultured until they reached 90% confluence. A 10-μl micropipette tip was used to make a wound. Cells were monitored at 0 h and 48 h after scratching and images of wound healing were captured (magnification of 100×) using a inverted phase contrast light microscope (Olympus, Tokyo, Japan) with DP Controller software (Olympus Life Science, Tokyo, Japan). Cell migration was quantified by measuring the wound healing index; i.e., the wound area healed by the cells at 48 h after scratching relative to the wound area at 0 h, using ImageJ software. For the Transwell migration or invasion assays, cells were resuspended in DMEM without serum and seeded into the upper chamber of 8-μm Transwell filters (Merck Millipore, Berlin, Germany). The invasion assay was performed using filters pre-coated in 1:3 diluted matrigel (BD Biosciences, Bedford, MA, USA) while the migration assay was not. DMEM containing 15% FBS was added to the lower chambers (24-well plate) and the cells were incubated 16 h for the migration assay and 24 h for the invasion assay. The invaded or migrated cells were quantified after 0.1% crystal violet staining in five randomly selected fields (magnification of 200×).

Xenograft models were generated in 6-week-old female nude mice (Laboratory Animal Center, Xinqiao Hospital, TMMU). Eighteen Mice were randomly divided into three groups (six for overCtrl and scrambled MTF cells; six for overTSSC3 and scrambled MTF cells; and six for overTSSC3 and shATG5 MTF cells) and weighed every four days. Cells suspensions of 4 × 10⁶ cells/ml in PBS were injected subcutaneously into a single side of the infra-axillary of each mouse in a volume of 0.1 ml. Xenografts were observed and measured every three days. The volume of xenografts were calculated as V (mm³) =1/2 × (length × width²) [33]. Mice were sacrificed at 20 days after injection and tumors were harvested and measured. A small part of tumors were excised and fixed with 4% glutaraldehyde quickly on ice for TEM. The remainder was fixed with 10% neutral buffered formalin, sliced, subjected to hematoxylin and eosin (H&E) staining, and further analyzed by immunohistochemistry (IHC).

Twenty-one nude mice were randomly divided into three groups as mentioned above. Cell suspensions of 5 × 10⁷ cells/ml were injected into the tail vein of 4-week-old female nude mice (Laboratory Animal Center, Xinqiao Hospital, TMMU) in a volume of 100 μl. The mice were evaluated every two days for weight and emaciation incapacitating tumor burden. Natural deaths of the nude mice were recorded to calculate the lifetime at 38 days after injection. The surviving nude mice were sacrificed at 38 days. All the lungs were resected and fixed in 10% neutral buffered formalin, and the fixed lungs were embedded in paraffin, sectioned, and stained with H&E. Microscopic lung metastases were counted under a light microscope after H&E staining. Five discontinuous and deep sections were used to define whether there was a metastasis nodule in the fixed lungs and the number of lung metastasis nodules was counted in one section with the most nodules.

All the animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee of Xinqiao Hospital, TMMU, and were performed according to the Guide for the Care Use of Laboratory Animals.

Quantitative data are presented as the mean ± SD. Quantitative data were analyzed using unpaired Student’s t-tests for two groups and analysis of variance (ANOVA) with Bonferroni’s multiple comparisons for three or more groups. Categorical data were analyzed using the Chi-squared test or Fisher’s exact test, and Spearman rank correlation coefficients. Survival analysis was carried out using the Kaplan–Meier method with the log-rank test. The independent prognostic factors were determined by Cox regression analysis. P < 0.05 was considered statistically significant. All analyses were performed using SPSS 20.0 software (version 20.0, SPSS Inc., Chicago, IL, USA) or GraphPad Prism (version 7.00, GraphPad Software, Inc., San Diego, CA, USA). In Vitro experiments were performed at least triplicate.

Previously, we proved that TSSC3 mRNA expression is decreased in human osteosarcoma tissues compared with that in non-tumor tissues [8]. To examine the correlation between TSSC3 and autophagy, we first applied immunohistochemical staining to examine the TSSC3, ATG5, and P62 expression in human benign bone and soft tissue tumors, and osteosarcoma. We observed higher rate of TSSC3 and ATG5 positivity in fibrous dysplasia and osteoblastoma compared with that in osteosarcoma (P < 0.05), while the positive rate of P62 expression was lower but not statistically significant (P > 0.05) in fibrous dysplasia and osteoblastoma compared with that in osteosarcoma (Additional file 3: Figure S1 a-c). The results showed that TSSC3 and autophagy-related proteins expression were in good agreement with various tendencies of benign bone and soft tissue tumors and osteosarcoma suggesting a potential correlation between them in osteosarcoma. To further confirm this, 58 human osteosarcoma tissues were used to investigate the relevance of TSSC3, ATG5, and P62 expression in vivo. The representative images of positive or negative IHC staining of TSSC3, ATG5, and P62 are shown in Fig. 1a. Interestingly, a higher IHC score for ATG5 was found in osteosarcoma patients with positive expression of TSSC3 (Fig. 1b). Further analysis demonstrated that the TSSC3 expression is positively correlated with that of ATG5. (Additional file 2: Table S3). However, we failed to detect a significant correlation between TSSC3 and P62 expression (Fig. 1c and Additional file 2: Table S3). Overall, these data suggested that ATG5 expression positively correlated with TSSC3 expression in osteosarcoma tissues.

We then used IHC staining of above tissues to investigate the correlations between TSSC3, ATG5, and P62 expression and clinicopathological features of osteosarcoma. Among the 58 patients, there were 34 males and 24 females. The median age was 22 years old (range: 8–59 years old). The follow-up time ranged from 4 (death) to 80 months. Twenty-two patients had local recurrence after chemotherapy and surgery of the primary osteosarcoma, 12 patients had lung metastasis at diagnosis, and eight more patients developed lung metastasis after the treatment (Table 1). Moreover, we found that expression of TSSC3 was significantly associated with a lower local recurrence rate (P < 0.05), but not with age, sex, stage, location, size, and pathological subtype. The rate of TSSC3 positivity was 36.8% (14/38) in osteosarcoma without metastasis and 15.0% (3/20) in osteosarcoma with metastasis; however, the decrease was not significant (P = 0.082) (Table 1). Surprisingly, neither the expression of ATG5 nor P62 seemed to be associated with the clinicopathological features of osteosarcoma (Table 1). As expected, Kaplan-Meier curves showed positive expression of TSSC3 was significantly related to an improved OS (Fig. 1d), while the differential expression of ATG5 and P62 had no notable impact on OS (Fig. 1e). Furthermore, TSSC3 protein expression, tumor size, and metastasis were validated to be independent prognostic markers for OS of osteosarcoma by univariate and multivariate Cox regression (Additional file 2: Table S4 and Table 2). We further observed that positive expression of TSSC3 and ATG5 was significantly associated with an earlier Enneking stage and there were markedly lower metastasis and recurrence rate (7.1 and 14.3%) for positive ATG5 expression compared to negative ATG5 expression (66.7 and 33.3%) in TSSC3(+) osteosarcoma; however, these differences were not statistically significant (Additional file 2: Table S5). Thus, we inferred that positive ATG5 expression in patients with TSSC3(+) osteosarcoma might suggest a favorable prognosis. To test this hypothesis, Kaplan–Meier analysis showed the patients with positive expression of both TSSC3 and ATG5 displayed a more favorable OS (Fig. 1f). Taken together, these results showed that TSSC3 is an independent prognostic marker for OS and positive expression of both TSSC3 and ATG5 is a potential predictor of favorable prognosis in osteosarcoma.

We found ATG5 expression is positively correlated with TSSC3 expression in osteosarcoma samples, and the relationship between TSSC3 and autophagy had never been reported before. Therefore, we examine the effect of TSSC3 in regulating autophagy in osteosarcoma cells. First, we stably overexpressed TSSC3 using lenti-overTSSC3 in MTF and SaOS2 cells and examined the formation of AVs by using TEM. The results demonstrated that there was an increased number of large diameter autophagosomes and autolysosomes in MTF or SaOS2 cells transfected with lenti-overTSSC3 transfected (Fig. 2a). Consistent with this, western blotting analysis revealed an accumulation of ATG5, BECN1, and lipid-bound LC3-II in TSSC3-overexpressing cells (Fig. 2b), which indicated an increase in the synthesis of autophagosomes [23]. In complementary experiments, immunofluorescence assays demonstrated that BECN1, ATG5, and LC3B levels were significantly increase in TSSC3-overexpressing cells (Additional file 3: Figure S2 a). In addition, overexpression of TSSC3 in osteosarcoma cell lines increased BECN1 and ATG5 mRNA levels (Additional file 3: Figure S2b). Furthermore, to investigate the effect of TSSC3 on autophagic flux, the expression of P62, also known as sequestosome 1 (SQSTM1), a substrate of autophagy [34], was examined. As expected, TSSC3 overexpression decreased P62 protein levels (Fig. 2b), suggesting enhanced autophagic flux. To verify this result, the lysosomal autophagy inhibitor CQ was used, which resulted in increased LC3-II and P62 accumulation in TSSC3-overexpressing cells (Fig. 2c), supporting the notion that TSSC3 overexpression did not block autophagic flux. Taken together, these data suggested that TSSC3 overexpression induced autophagy and enhanced autophagic flux in osteosarcoma cells in vitro.

Autophagy has been reported as a tumorigenesis suppressor in some situations [35, 36] and we have found TSSC3 overexpression enhanced autophagy flux and inhibited cell proliferation (Additional file 3: Figure S3 a) in osteosarcoma cells [8]. Thus, to determine whether TSSC3-induced inhibition of osteosarcoma malignant proliferation was related to autophagy, we applied the CCK-8 assay to detect the cell viability with or without TSSC3-overexpression, with or without autophagy suppression by CQ. We found that overexpression of TSSC3 significantly decreased the cell viability of MTF and SaOS2 cells, and no obvious cytotoxicity was found compared with the control group when the cells were treated with CQ alone. By contrast, CQ significantly blocked the autophagic flux (Additional file 3: Figure S2 c), and cell viability was partly rescued in TSSC3-overexpressed cells when treated with CQ (Fig. 3a). Subsequently, we found that blocking autophagy with CQ attenuated the TSSC3-induced suppression of proliferation and colony-formation ability of MTF and SaOS2 cells, as assessed using EdU and colony formation assays (Fig. 3b and c). These data indicated that TSSC3-induced inhibition of malignant proliferation of osteosarcoma cell lines might be related to autophagy. Moreover, to further confirm this notion, enhanced autophagy was eliminated via ATG5 knockdown using shRNA lentiviruses (two effective sequences were chosen), which were stably transfected into TSSC3-overexpressing cells; western blotting analysis confirmed this result (Fig. 3d). We also found that blocking autophagy (ATG5 knockdown) reduced TSSC3-induced inhibition of cell proliferation and colony-formation ability in MTF and SaOS2 cells (Fig. 3e and f). Accumulating evidence shows that autophagy can promote apoptosis to suppress tumor progression [37]. Therefore, to determine whether TSSC3-mediated autophagy could suppress malignant proliferation depending on apoptosis in osteosarcoma cell lines, Hoechst and Annexin V-PE/7-AAD staining were used. As shown in Additional file 3: Figure S3 b and c, inhibition of autophagy by CQ partly abrogated TSSC3-induced apoptosis in SaOS2 cells, which was consistent with the markedly upregulation of cleaved caspase-3 and Bax, and reduced Bcl2 (Additional file 3: Figure S3 d). Intriguingly, autophagy inhibition had no visible effect of TSSC3-induced apoptosis in the MTF cell line (Additional file 3: Figure S3 b-d). These differences may be derived from the genetic background of the cell lines. Collectively, these results suggested that TSSC3-mediated autophagy contributes to the anti-osteosarcoma effect of TSSC3 by suppressing cell malignant proliferation, which may not completely depend on autophagy-triggered apoptosis.

Having shown that TSSC3 mediates autophagy and induces inhibition of malignant proliferation of osteosarcoma cells in vitro, it was important to determine whether these effects could be detected in vivo. We established subcutaneous xenograft models by stably transfecting MTF cells into athymic nude mice. Then, we found that TSSC3 overexpression significantly reduced tumorigenesis, as reflected by the tumor size, volume, body weight, and tumor weight of the xenograft tumors. However, knockdown of ATG5 restored tumorigenesis in TSSC3-overexpressing cells (Additional file 3: Figure S4 a-b and Fig. 4a-b). Moreover, that overexpression of TSSC3 enhanced autophagy and knockdown of ATG5 blocked TSSC3-induced autophagy were further confirmed in vivo using TEM (Fig. 4c) and IHC staining of TSSC3 and autophagy related proteins of human xenograft tumor tissues (Fig. 4d and Additional file 3: Figure S4c). As expected, the rate of positive Ki67 IHC staining was high in the control groups, decreased in the TSSC3-overexpressing group, and restored in the TSSC3-overexpressing and ATG5-knockdown groups, suggesting that autophagy contributes to TSSC3-induced inhibition of malignant proliferation in vivo (Additional file 3: Figure S4d and Fig. 4e). Thus, all of the vitro and in vivo results indicated that TSSC3-induced autophagy plays a crucial role in TSSC3-mediated tumorigenesis in osteosarcoma cells.

Metastasis is a major factor predicting poor prognosis in patients with osteosarcoma [3] and autophagy depends on an environment that may limit the invasion and dissemination of tumor cells from a primary site [21, 36]. We found that positive expression of both TSSC3 and ATG5 is significantly associated with an improved prognosis and correlated negatively with metastasis in human osteosarcoma; therefore, we speculated that TSSC3-induced autophagy might have an effect on the regulation of osteosarcoma cell metastasis. To confirm this speculation, we used wound healing and Transwell assays. The results showed that TSSC3 overexpression significantly decreased the motility and invasive ability of MTF and SaOS2 cells, but had no significant effect when the cells were treated with CQ alone, compared with that in the control group. Although the inhibitory effects of TSSC3 overexpression on cell motility and invasive ability were markedly attenuated by CQ treatment (Fig. 5a and b), knockdown of I in TSSC3-overexpressing MTF and SaOS2 cells significantly improved cell motility and invasion (Fig. 5c and d). Thus, these results confirmed the findings that TSSC3 overexpression-induced autophagy suppresses osteosarcoma cell metastasis in vitro. To further investigate the mechanism, we observed that the epithelial cell marker E-cadherin was upregulated in TSSC3-overexpressing SaOS2 cells (but not in MTF cells), whereas the expression levels of mesenchymal markers N-cadherin, Vimentin, and MMP2 were inhibited. These protein expression levels did not increase significantly when the cells were treated with CQ alone to block autophagy. However, interestingly, the N-cadherin, Vimentin, and MMP2 levels increased markedly when TSSC3-overexpressing MTF and SaOS2 cells were treated with CQ (a decrease in the E-cadherin level was also detected in TSSC-overexpressing SaOS2 cells) (Fig. 5e and f). The epithelial-mesenchymal transition (EMT) is a metastatic behavior of cancer cells [38]. Thus, the impairment of MTF and SaOS2 cell migration and invasion might be caused by inhibition of the EMT, dependent on TSSC3 overexpression-induced autophagy.

To further investigate whether autophagy contributes to TSSC3-induced suppression of metastasis in osteosarcoma, we established a lung metastasis model in vivo by injecting stably transfected MTF cells into the tail vein of athymic nude mice. We found that overexpression of TSSC3 notably reduced the ability of MTF cells to induce lung metastases, as reflected by the decreased incidence and number of metastatic nodules and the increased survival time of the mice. However, knockdown of ATG5 restored the ability of TSSC3-overexpressing cells to establish lung metastases (Fig. 6a-c and Additional file 3: Figure S6b). In addition, using IHC staining, the metastatic nodules were negative for E-cadherin and CK18, but positive for Vimentin, indicative of a mesenchymal origin (Additional file 3: Figure S6a). Consistent with these findings, the expression change of E-cadherin and Vimentin in subcutaneous xenograft tumors induced using transfected cells were measured using IHC staining (Fig. 6d and e), which indicated that TSSC3-induced autophagy blocked the EMT of osteosarcoma cells in vivo. These results, together with our in vitro findings, highlighted a functional role for autophagy in the regulation of TSSC3-induced lung metastasis suppression in osteosarcoma.

Our previous studies suggested TSSC3 expression might regulate the phosphorylation of Src and Akt. In view of the important role of the PI3K/Akt/mTOR pathway in autophagy regulation [27] and the role of Src in activating the PI3K/Akt pathway [11, 28], to determine how TSSC3 mediated autophagy in osteosarcoma, we first measured Src, Akt, mTOR, and their phosphorylation levels in TSSC3-overexpressing MTF and SaOS2 cells. The results showed that overexpression of TSSC3 did not affect the total Src, Akt and mTOR levels, but significantly decreased the levels of their phosphorylated forms (Fig. 7a), which suggested that the Src and PI3K/Akt/mTOR pathways might be involved in TSSC3-induced autophagy. Then, we applied IGF-1 and p-YEEI for Src and PI3K/Akt pathway activation, separately or in combination, to further identify whether the Src and PI3K/Akt/mTOR signaling pathways underlie TSSC3-mediated autophagy. TSSC3 overexpression inhibited Src and PI3K/Akt/mTOR activity, increased the levels of ATG5 and LC3-II, and decreased P62 expression. Following p-YEEI or IGF-1 treatment, the Src or PI3K/Akt/mTOR pathway was significantly activated, remarkably reversing the effect of TSSC3 overexpression on autophagy related proteins in MTF and SaOS2 cells. This implied that PI3K/Akt/mTOR or Src inhibition triggered autophagy in TSSC3-overexpressing cells (Fig. 7b). We also found that the PI3K/Akt/mTOR pathway was activated when the cells were treated with p-YEEI, but no obvious increase of Src phosphorylation was found after applying IGF-1. This result suggested that TSSC3-induced PI3K/Akt/mTOR pathway inhibition is Src-dependent. Furthermore, we applied p-YEEI and BEZ235 together to activate Src signaling and inhibit PI3K/Akt/mTOR pathway. The result indicated that inhibition of Src phosphorylation without PI3K/Akt/mTOR inactivation could not decrease the ATG5 and LC3-II levels or increase the P62 levels in TSSC3-overexpressing MTF and SaOS2 cells (Fig. 7b). Taken together, these results allowed us to conclude that TSSC3 overexpression induces autophagy in osteosarcoma cells by inactivating the Src-mediated PI3K/Akt/mTOR pathway.