TGF-β Suppressed Succinat Dehydrogenase to Promote Osteosarcoma Chemo-Resistance Through An HIF1α Dependent Manner

Background (cid:0) Despite the widespread adoption of chemotherapy, drug resistance has been the major obstacle in tumor elimination of cancer patients. Our aim was to explore the role of TGF-β in osteosarcoma chemo-resistance. Methods (cid:0) We performed cytotoxicity analysis of methotrexate (MTX) and cisplatin (CIS) in TGF-β treated osteosarcoma cells. Then a metabolite prole of the core energetic routes was analyzed by 1 H-NMR in Saos-2 and MG-63 cell extracts. We detected the expression of succinate dehydrogenase (SDH), STAT1, and Hypoxia-inducible factor 1α (HIF1α) in TGF-β treated osteosarcoma cells, and further tested the effects of these molecules on the cytotoxicity of chemotherapeutic agents. In vivo, we examined the tumor growth and survival time of Saos-2 bearing mice given the combination therapy of chemotherapeutic agents and a HIF1α inhibitor. Results (cid:0) Metabolic analysis revealed an enhanced succinate production of osteosarcoma cells after TGFβ treatment. We further found the decrease in SDH expression and the increase in HIF1α expression of TGF-β treated osteosarcoma cells. Consistently, blockade of SDH aggravated the resistance of Saos-2/MG-63 cells to MTX and CIS. Also, a HIF1α inhibitor signicantly strengthened the anti-cancer ecacy of chemotherapeutic drugs in mice with osteosarcoma cancer. Conclusion (cid:0) Our study demonstrated that TGF-β attenuated the expression of SDH through reducing the transcription factor STAT1. The reduction of SDH then caused the up-regulation of HIF1α, thereby rerouting the glucose metabolism and aggravating chemo-resistance in osteosarcoma cells. Linking tumor cell metabolism to the formation of chemotherapy resistance, our study may guide the development of more effective treatments of osteosarcoma.

Results Metabolic analysis revealed an enhanced succinate production of osteosarcoma cells after TGFβ treatment. We further found the decrease in SDH expression and the increase in HIF1α expression of TGF-β treated osteosarcoma cells. Consistently, blockade of SDH aggravated the resistance of Saos-2/MG-63 cells to MTX and CIS. Also, a HIF1α inhibitor signi cantly strengthened the anti-cancer e cacy of chemotherapeutic drugs in mice with osteosarcoma cancer.
Conclusion Our study demonstrated that TGF-β attenuated the expression of SDH through reducing the transcription factor STAT1. The reduction of SDH then caused the up-regulation of HIF1α, thereby rerouting the glucose metabolism and aggravating chemo-resistance in osteosarcoma cells. Linking tumor cell metabolism to the formation of chemotherapy resistance, our study may guide the development of more effective treatments of osteosarcoma.

Background
Osteosarcoma is the most prevalent malignant bone tumor with a high occurrence in children and adolescents. Surgery combined with chemotherapeutic agents is considered the major strategy for treatment. However, patients with osteosarcoma still undergo pulmonary metastasis and relapse, resulting in low survival rate [1,2]. Over the past decades, resistance of tumor cells to drugs such as methotrexate (MTX), cisplatin (CIS) and doxorubicin has emerged as a major impediment to eradicating osteosarcoma [3]. Therefore, further research is warranted to better understand osteosarcoma chemoresistance and formulate e cacious therapeutic strategies.
Drug resistance, a multifaceted process, is attributed to a combination of factors comprising apoptosis induction, autophagy induction, cancer stem cell regulation, DNA damage and repair, and epigenetic regulation [4]. Emerging evidence has supported the role of tumor metabolism in promoting drug resistance [5,6]. Cancer cells rewire their metabolism to satisfy the high demand for both energy and biosynthesis. In this context, glycolysis is given a high-priority in cancer cells even under physiological oxygen conditions, which is named "Warburg effect". Components of the glycolytic pathway such as glucose transporters and hexokinase-2, the rst rate-limiting enzyme in the glycolytic pathway, are intimately linked to chemo-resistance [7,8]. These all raise awareness of intervening tumor metabolism to combat drug resistance.
It has been substantiated that transforming growth factor-β (TGF-β) may function as a signi cant contributor to drug resistance in several types of tumor [9][10][11][12][13]. TGF-β, a pleiotropic cytokine, plays a key role in regulating multiple biological processes, including cell proliferation, immune response and in ammation [14]. Reportedly, up-regulation of TGF-β signaling has been found in erlotinib-resistant lung cancer cells [15]. Upon attaching to its receptors, TGF-β spurs a series of events, among which epithelialmesenchymal transition (EMT) imparts cancer cells with metastatic and invasive properties. Compelling reports have demonstrated that EMT confers chemo-resistance on cancer cells through increasing drug e ux pumps and anti-apoptotic effects [16]. TGF-β has thus become a promising target in cancer therapy, and treatments targeting TGF-β pathway such as neutralizing antibodies and soluble TGF-β receptors have been evaluated in pre-clinical tests and even clinical trials [17].
In the current study, we observed a stronger TGF-β expression in chemo-resistant osteosarcoma patients, and in vitro TGF-β treatment obviously strengthened the multi-drug resistance of osteosarcoma cell lines.
Mechanistically, we demonstrated that TGF-β caused the decrease of STAT1 to inhibit the metabolic enzyme succinate dehydrogenase (SDH), giving rise to succinate accumulation in osteosarcoma cells.
We further elucidated that the increased succinate promoted the expression of Hypoxia-inducible factor 1α (HIF-1α), and blockade of HIF1α in turn augmented the antitumor effects of MTX and CIS. These ndings revealed the molecular basis of TGF-β implicated in the regulation of metabolic pathways and subsequent chemo-resistance of osteosarcoma.

Cell lines and reagents
Human osteosarcoma cell lines, Saos-2 and MG-63, were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured in Dulbecco's Modi ed Eagle's medium (DMEM) with 10% fetal bovine serum (FBS) (Gibco). All cells were grown at 37℃ in a 5% CO2 incubator. MTX, CIS, succinate, and KC7F2 were from Sigma-Aldrich (ST, USA). Recombinant protein human and mouse TGF-β1 were bought from Pepro Tech (Rocky Hill, NJ).

Patients and specimens
This study was approved by the Ethics Committee of the A liated Hospital of Southwest Medical University. Written informed consent was attained from all patients, and all methods were performed according to the Declaration of Helsinki. A total of 20 osteosarcoma patients were recruited and provided informed consent between January 2017 and January 2020. These patients were divided into the chemosensitive and chemo-resistant groups according to the Response Evaluation Criteria in Solid Tumors (RECIST). Then formalin-xed, para n-embedded osteosarcoma specimens were collected for further detection.
For 3D matrix gel culture, tumor tissues were minced and digested with collagenase (Sigma-Aldrich, MA, USA) followed by ltration (BIOFIL). After centrifugation and removal of the red blood cells, osteosarcoma cells were seeded into 3D matrix gels in DMEM medium with 10% FBS.

Metabolic assessment of cells
Metabolic assessment of Saos-2 and MG-63 cells (1×10 7 cells per sample) was performed by NMR as mentioned before [18,19]. Brie y, after using a methanol-chloroform-water extraction method, the upper aqueous phase was lyophilized and then redissolved in 550 μl of phosphate buffer solution (60 mM K 2 HPO 4 /NaH 2 PO 4 , pH 7.4, 99.9% D 2 O) [20]. A Bruker 600-MHz spectrometer was used for the 1 H-NMR experiments at 277 K temperature. Quantitative analysis of metabolites was performed using TopSpin (version 3.5) software. Metabolites were assigned according to published data. Metabolite concentrations were quanti ed per million cells and mean cell metabolite concentrations (fold change) were then calculated.

Metabolite quanti cation
Quantitative analysis of succinate and fumarate was conducted using succinate (succinic acid) and fumarate colorimetric assay kit (BioVison, SF, USA), respectively, under the supplied instructions.
Immunohistochemical staining was performed using the DAB Horseradish Peroxidase Color Development Kit (Beyotime, Shanghai, China) according to the supplied protocol. In brief, tissue slides were stained with HRP-conjugated secondary antibodies for 1 hour at room temprature and conterstained with hematoxylin (Solarbio, Beijing, China). For immuno uorescent staining, tissue slides were incubated with secondary antibody followed by incubation with DAPI (Solarbio, Beijing, China). The intensity of immunostaining was analyzed by Image J 9.0 software.

Animal experiments
NSG mice (4-6 weeks old) were purchased from HFK Bioscience Company (Beijing, China) and maintained under pathogen-free conditions. For tumor growth analysis, 2×10 6 Saos-2 or MG-63 cells were subcutaneously injected into NSG mice. Then these mice were randomized into different groups 10 days later after inoculation and treated with or without TGF-β (20 μg/kg), MTX (5 mg/kg), CIS (1 mg/kg), KC7F2 (10 mg/kg) twice a week for 14 days. The mice of control groups received an equal volume of saline. Tumor growth was examined every other day and the survival of mice was recorded. Tumor volume was calculated using the formula: tumor volume=length×width 2 /2.
For tumorigenesis analysis, NSG mice received injections of 2×10 5 Saos-2 or MG-63 cells subcutaneously. 10 days later after inoculation, these mice were treated with or without TGF-β (20 μg/kg), KC7F2 (10 mg/kg) twice a week for 14 days. The tumorigenesis was calculated 20 days after injection. The above experimental procedures were approved by the Ethics Committee of the A liated Hospital of Southwest Medical University, according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All animal studies were conducted in accordance with the Public Health Service Policy and complied with the ARRIVE guidelines for the humane use and care of animals.

Statistical analysis
All experiments were independently performed in triplicate. Results are presented as mean ± SEM, and analyzed by Student's t-test or one-way ANOVA. The survival rates were determined by Kaplan-Meier survival analysis. P value < 0.05 was considered statistically signi cant. The analysis was performed using Graphpad 8.0 software.

TGF-β promoted osteosarcoma chemo-resistance
Compelling reports have implicated that tumor tissues with increasing TGF-β expression revealed enhanced migratory features and poor prognosis. Our study further explored the role of TGF-β in osteosarcoma progression, including drug resistance and tumorigenicity. To do this, we isolated tumor tissues from chemo-resistant/sensitive osteosarcoma patients and examined the expression of TGF-β in tumor tissues. Notably, those tumor tissues isolated from chemo-resistant patients exhibited elevated TGF-β expression compared to the chemo-sensitive group (Fig. 1A), suggesting the potential role of TGF-β in chemo-resistance development of osteosarcoma. Given the limited samples size, we further performed cytotoxicity analysis in TGF-β treated osteosarcoma cell lines Saos-2 and MG-63. As a result, TGF-β treatment signi cantly strengthened the resistance of Saos-2/MG-63 to chemotherapeutic MTX (Fig. 1B) and CIS (Fig. 1C). Similar results were observed in TGF-β treated Saos-2/MG-63 bearing mice (Fig. 1D). Those results suggested that TGF-β could facilitate the chemo-resistance in osteosarcoma cells. Subsequently, we further assessed the in uence of TGF-β in tumor growth. Though no signi cant difference was observed in tumor volumes of Saos-2/MG-63 bearing mice (Fig. 1E), Saos-2 and MG-63 cells treated with TGF-β revealed strengthened capability of tumorigenesis (Fig. 1F), reminding us of the potential relationship between TGF-β and stem associated transcription factors in osteosarcoma. Together, those results suggested TGF-β could promote osteosarcoma chemo-resistance, resulting in poor outcome in patients.
TGF-β suppressed succinat dehydrogenase to promote chemo-resistance Alteration of energy metabolism is a biological ngerprint of tumor cells, and enhanced lactate production (caused by glycolysis) correlated with drug resistance in several tumor types. Here, a metabolite pro le of the core energetic routes was analyzed by 1 H-NMR in Saos-2/MG-63 cell extracts. After TGF-β treatment, Saos-2/MG-63 cells revealed obviously higher consumption of glucose and increasing production of lactate ( Fig. 2A), suggesting that tumor cells might enhance glycolysis by preferential conversion of pyruvate to lactate instead of oxidation. Meanwhile, enhanced succinate production and an increased succinate/fumarate ratio were observed in Saos-2/MG-63 cells treated with TGF-β, as shown in Fig. 2A. These results prompted us to speculate that TGF-β might affect the SDH metabolism, resulting in the accumulation of succinate and suppression of oxidation in the tricarboxylic acid cycle. In that case, the expression of SDHD, an submit of SDH, was examined. Intriguingly, suppression of SDHD expression was observed in mRNA (Fig. 2B) and protein (Fig. 2C) levels of TGF-β treated Saos-2/MG-63 cells. To further con rm the role of SDH in osteosarcoma progression, siRNA inference was conducted to suppress the expression of SDHD (Fig. 2D). Blockade of SDHD obviously strengthened the resistance of Saos-2/MG-63 cells to MTX and CIS (Fig. 2E and F). Meanwhile, addition of succinate to suppress the oxidation also promoted MTX/CIS resistance of osteosarcoma cells (Fig. 2G  and H). Consistently, suppression of SDHD was observed in chemo-resistant tumor tissues from patients ( Fig. 2I). Together, those results suggested that TGF-β could suppress SDH to promote chemo-resistance in osteosarcoma.

TGF-β suppressed STAT1 to decrease SDH expression
Current studies provided evidence to suggest that TGF-β could regulate JAK/STAT associated signaling pathways to promote cancer progression. To further clarify the mechanism in SDH associated osteosarcoma progression, the expression of JAK1, JAK2, STAT1, STAT2, and STAT3 was examined by western blotting. Notably, reduced expression of phosphorylated JAK2 and STAT1 were observed in TGFβ treated Saos-2 cells (Fig. 3A and B). Meanwhile, reduced nucleus entry of STAT1 in TGF-β treated Saos-2 and MG-63 cells were found in our immuno uorescent staining (Fig. 3C), suggesting TGF-β suppressed STAT1 activation in osteosarcoma cells. Next, siRNA interference was performed to suppress the STAT1 expression in Saos-2 and MG-63 cells (Fig. 3D). Consequently, silence of STAT1 down-regulated the expression of SDHD in Saos-2 and MG-63 cells (Fig. 3E). And enhanced succinate production and increased succinate/ fumarate ratio were found in STAT1 silenced Saos-2 and MG-63 cells (Fig. 3F). These results implicated that TGF-β suppressed STAT1 to down-regulate SDH activity. Subsequently, we observed similar chemo-resistance to MTX (Fig. 3G) and CIS (Fig. 3H) in STAT1 silenced Saos-2 and MG-63 cells, suggesting that STAT1 signals were involved in the TGF-β associated chemo-resistance. Consistently, STAT1 level was reduced in chemo-resistant tumor tissues from osteosarcoma patients (Fig.  3I). Those results suggested that TGF-β suppressed STAT1 signals to down-regulate SDH metabolism.
Metabolic succinate facilitated osteosarcoma chemo-resistance through an HIF1α dependent manner Compelling ndings implicated that succinate served as a crucial oncometabolite by suppressing PHDs, which target HIF1α for proteasomal degradation or glycolysis. Hence, we sought to investigate whether TGF-β could mediate the HIF1α up-regulation by controlling SDH metabolism. TGF-β or succinate treatment promoted the expression of HIF1α expression in Saos-2/MG-63 cells (Fig. 4A). And silence of STAT1 or SDHD also contributed to the elevated expression of HIF1α (Fig. 4B), indicating that TGF-β could up-regulate HIF1α through an SDH dependent manner. Next, we used HIF1α inhibitor KC7F2 to treat Saos-2/MG-63 cells for cytotoxicity analysis and no obvious cytotoxicity was observed in bulk Saos-2/MG-63 cells (Fig. 4C). However, blockade of HIF1α e ciently reversed the drugs resistance caused by TGF-β ( Fig. 4D and E), indicating that TGF-β promoted chemo-resistance through HIF1α. More importantly, our previous results implicated that TGF-β strengthened the capability of tumorigenesis in osteosarcoma cells (Fig. 1F), and HIF1α has been proved to be associated with tumor stemness regulation. Herein, we further suppressed HIF1α, then treated Saos-2/MG-63 cells with TGF-β. Intriguingly, blockade of HIF1α e ciently weakened the tumorigenesis of Saos-2 and MG-63 cells in vivo (Fig. 4F), demonstrating that TGF-β could further promote osteosarcoma stemness through HIF1α. Together, those results suggested that TGF-β suppressed SDH activity to up-regulate HIF1α, resulting in chemo-resistance in osteosarcoma.

Blockade of HIF1α improved outcome of chemotherapy in osteosarcoma
Given the crucial role of HIF1α in tumorigenesis and chemo-resistance, it might be feasible to target HIF1α to improve the outcome of chemotherapy in osteosarcoma. To assess our hypothesis, Saos-2 cells were subcutaneous injected into the NOD-SCID mice, following with HIF1α inhibitor KC7F2, and MTX/CIS treatment. Intriguingly, addition of KC7F2 signi cantly strengthened the anticancer effects of MTX (Fig.  5A) and CIS (Fig. 5B), though no TGF-β was used for tumor-bearing mice. We conjectured that osteosarcoma cells could produce TGF-β through an autocrine manner, thereby resulting in the activation of HIF1α pathway. Consistently, KC7F2 treatment also signi cantly prolonged the survival time of Saos-2 bearing mice ( Fig. 5C and D), indicating that suppression of HIF1α could e ciently improve anticancer effects of chemotherapy. To further evaluate the anticancer effects of KC7F2, we isolated tumor cells from chemo-resistant patients and seeded those tumor cells into matrix gels. Tumor cells isolated from patient #2 succeed to generate spherical colonies and revealed proliferative phenotypes in two months (Fig. 5E). Here, we treated those colonies with MTX, CIS and KC7F2 in matrix gels. As anticipated, slight cytotoxicity was observed in MTX or CIS treatment group, which might be due to the tumor cells isolated from chemo-resistant tumor tissue. However, addition of KC7F2 signi cantly strengthened the cytotoxicity of MTX and CIS ( Fig. 5F and G), indicating the potential anticancer effects of HIF1α inhibitors in clinic. Together, those results suggested that blockade of HIF1α could improve the outcome of chemotherapy, which described a novel strategy in clinical osteosarcoma treatment.

Discussion
TGF-β is a pleiotropic cytokine which has a dual action in tumor development. Activation of TGF-β pathway results in a variety of gene responses, which modulate cell-cycle arrest and apoptosis in earlystage cancer, as well as metastasis and angiogenesis in advanced cancer [21]. A growing body of research has indicated that TGF-β is also linked to chemo-resistance and the union application of drugs and TGF-β inhibitors achieve remarkable effects in several tumor types [9,22]. However, our current understanding of TGF-β induced drug resistance is incomplete. Given its role in gene expression and cell differentiation, cellular metabolism is emerging as a key player in tumor initiation and progression [23]. Here, our study provided evidence of the contribution of TGF-β to glucose metabolism of tumor cells, thus exacerbating drug resistance in osteosarcoma.
Variation in energetic routes is a hallmark of tumor cells, which is characterized by enhancement of glucose uptake, glycolysis hyperactivation, reduction of oxidative phosphorylation, and lactate accumulation [24]. The altered cellular metabolism not only brings an abundance of energy, but also gives rise to metabolic intermediates that exert important effects in supporting tumor proliferation, metastasis, and even chemo-resistance [25]. In a study of breast cancer, elevated expression of lactate dehydrogenase which converted pyruvate into lactate was proved to mediate trastuzumab resistance [26]. Shi et al. reported that knockdown of pyruvate kinase M2 (PKM2), the nal rate-limiting enzyme of the glycolytic pathway, caused an accumulation of docetaxel in lung cancer cells and synergistically strengthened the e ciency of chemotherapy in mice [27]. Apart from enzymes in control of glycolysis, glucose transporters also contributed to drug resistance [28]. Here, our study provided a quantitative analysis of metabolite levels and observed increasing glucose consumption as well as lactate generation in TGF-β treated osteosarcoma cells. Importantly, we demonstrated that the TGF-β down-regulated SDH expression in osteosarcoma cells, thus leading to the collection of succinate and conversion to glycolysis pathway. Previous research has revealed that TGF-β could control SDH expression through transcriptional and posttranslational regulation of STAT1 [29,30]. Consistently, we further indicated that TGF-β attenuated STAT1 phosphorylation so as to suppress SDH in osteosarcoma cells. Decrease of SDH expression in osteosarcoma cells brings about an accumulation of succinate. Succinate has been regarded as a crucial oncometabolite which functions through enhancing angiogenesis and suppressing histone and DNA demethylases. [31,32]. Our results indicated that succinate treatment caused a rise in HIF1α in osteosarcoma cells. Hypoxia is probably the most pervasive condition within the tumor tissue.
Tumor cells have to struggle with the low oxygen tension by activating several pathways, among which HIF1α signaling plays a vital role in cancer progression. There is increasing evidence that HIF1α mediates tumor metastasis, angiogenesis and development of resistance to various therapeutic modalities [33]. Sowa et al. reported that HIF-1 facilitated drug resistance of lung adenocarcinoma in part due to the induction of carbonic anhydrase IX (CAIX) [34]. In another study of bladder cancer, cisplatin-resistant cells exhibited higher levels of HIF1α, which was correlated with increased expression of MDR1 encoding multidrug e ux pump P-glycoprotein (P-gp) [35]. Here, we demonstrated that HIF1α up-regulation resulting from elevated succinate exerted regulative effects on the glycolytic state and chemotherapy resistance of osteosarcoma cells.

Conclusions
Our study indicated that TGF-β exhibited an inhibitory effect on SDH expression in osteosarcoma cells, therefore having a central function in tumor metabolism rerouting and subsequent drug resistance. Chemotherapeutic agents in combination with a HIF-1α inhibitor signi cantly abrogated TGF-β mediated chemo-resistance and enhanced the curative effects, which revealed a potentially promising method for combating osteosarcoma.

Declarations
Ethics approval and consent to participate

Consent for publication
Not applicable.

Availability of data and materials
The anonymized data used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.