SLC25A21 Suppresses Cell Growth in Bladder Cancer Via The Reactive Oxygen Species-Mediated Mitochondrion-Dependent Apoptosis Pathway

Background: Bladder cancer (BCa) is a commonly diagnosed malignancy worldwide that has poor survival depending on its intrinsic biologic aggressiveness and a peculiar radio- and chemoresistance features. Gaining a better understanding of tumorigenesis and developing new diagnosis and treatment strategies for BCa is important for improving BCa clinical outcome. SLC25 family member 21 (SLC25A21), a carrier transporting C5-C7 oxodicarboxylates, has been reported to contribute to oxoadipate acidemia. However, the potential role of SLC25A21 in cancer remains absolutely unknown. Methods: The expression levels of SLC25A21 in BCa and normal tissues were examined by real-time PCR and immunohistochemistry. Gain-of- and loss-of-function experiments were performed to detect the biological functions of SLC25A21 in vitro and in vivo by CCK-8 assay, plate colony formation assay, cell migration, invasion assay and experimental animal models. The subcellular distribution of substrate mediated by SLC25A21, mitochondrial membrane potential and ROS production were assessed to explore the potential mechanism of SLC25A21 in BCa. Results: We found that the expression of SLC25A21 was downregulated in BCa tissues compared to normal tissues. A signicant positive correlation between decreased SLC25A21 expression and poor prognosis was observed in BCa patients. Overexpression of SLC25A21 signicantly inhibited cell proliferation, migration and invasion and induced apoptosis in vitro. Moreover, the enhanced SLC25A21 expression signicantly suppressed tumor growth in a xenograft mouse model. Furthermore, we revealed that SLC25A21 suppressed BCa growth by inducing the eux of mitochondrial α-KG to the cytosol, decreasing to against oxidative stress, and activating the ROS-mediated mitochondrion-dependent apoptosis pathway. Conclusions: Our ndings provide the rst link between SLC25A21 expression and BCa and demonstrate that SLC25A21 acts as a crucial suppressor in BCa progression, which may help to provide new targets for BCa intervention. and immunohistochemistry. We further demonstrated the effects of SLC25A21 on cellular biological behaviors both in vitro and in vivo by gain- and loss-of-function experiments. We further presented evidence that SLC25A21 induces the eux of mitochondrial α-KG to the cytosol and activates the ROS-mediated apoptosis pathway. Our ndings provide novel insights into the roles of SLC25A21 in the tumorigenesis and progression of BCa, which may help to provide new therapeutic targets for BCa. in accordance with the results from our real-time RT-PCR and and GSEA database analyses, and we determined that low expression of SCL25A21 is signicantly associated with invasion and poor survival in BCa patients. These data indicate that the downregulation of SLCA25A21 expression could serve as a predictor of a higher risk of BCa progeression. Moreover, to highlight the function of SLC25A21 in BCa, we further explored the critical effects of SLC25A21 on cell behaviors in BCa by gain- and loss-of-function experiments. Our results identied that SLC25A21 acts as a tumor suppressor results from the overexpression of SLC25A21 inhibits cell growth and facilitates cell apoptosis in BCa in vitro and in vivo. To the best of our knowledge, our study provides the rst evidence that SLC25A21 plays a critical role in tumors, especially BCa. Our results also indicate that the elevated expression of SLC25A21 can inhibit the malignant phenotype of BCa, suggesting SLC25A21 as a potential predictor and therapeutic target for BCa progression. the inner mitochondrial family mitochondrial

grade and stage assessments were made according to the World Health Organization (WHO) 2004 standard and Union for International Control Cancer (UICC) TNM system.

RNA isolation and real-time PCR
Total RNA from tissues and cell lines were extracted using Trizol reagent (Takara, Dalian, China) according to the manufacturer's protocol. The cDNA was synthesized with 500 ng total RNA by using PrimeScript RT Reagent Kit (Takara). Real-time PCR was carried out using SYBR® Rremix Ex TaqTM (Takara) as described previously [16]. GAPDH was used as an internal control. The assay was performed in triplicate. All primer sequences are listed in Supplement Information S1. IHC staining and scoring IHC was performed on 3-μm sections of para n-embedded tissue samples. The sections were depara nized with xylene and rehydrated with descending ethanol concentrations, and then were treated with Peroxidase Blocking Reagent (Dako, Glostrup, Denmark) for 5 min, followed by incubation with SLCA5A21 primary antibody (1:80, A nity Biosciences, OH, USA, Cat# DF4172) at 4 °C overnight. After incubation with the secondary antibody at room temperature for 1 h. Immunodetection was performed with the diaminobenzidine (DAB) reagent (Dako) according the manufacturer's protocol, and the reaction time of each section was consistent, followed by counterstaining with hematoxylin.
The IHC-stained tissue sections were reviewed and scored separately in a double-blinded manner. Scores were determined based on both the intensity and proportion of SLC25A21-positive cells as described previously [17,18]. The staining score ranged from 0 to 4, corresponding to the percentage of immunoreactive tumor cells (0%, 1-25%, 26-50%, 51-75%, or 76-100%). The staining intensity scores were as follows: negative (0), weak (score = 1), medium (score = 2) or strong (score = 3). A total score of 0 to 12 was calculated by multiplying the staining degree score by the intensity score. The nal staining score of ≥ 3 was considered to indicate high level of SLC25A21.

Construction of cell lines with stably overexpressed SLC25A21
The full-length open reading frame of human SLC25A21 was ampli ed and cloned into the pcDNA3.1 vector. BCa cell lines EJ and T24 were transfected with pcDNA3.1-SLC25A21 using Lipofectamine 3000 reagent (Thermo Fisher Scienti c, IL, USA). The empty pcDNA3.1 vector was used as the control. G418 was used to select stable SLC25A21-overexpressing cells. The expression levels of SLC25A21 were determined by real-time PCR and western blotting analysis.

Oligonucleotide transfection
The siRNAs targeting SLC25A21 and negative control siRNA (silencer negative control siRNA) were synthesized (GenePharma, Shanghai, China). Oligonucleotides transfection was performed with Lipofectamine 3000 following the manufacturer's protocol. Target sequences for siRNAs were shown in Supplement Information S2.

Cell proliferation
Cell proliferation was assessed using the Cell Counting Kit-8 kit (Dojindo Laboratories, Kyushu Island, Japan) according to the manufacturer's instructions. In brief, BCa cells were seeded in 96-well plates at 8 × 10 2 cells per well. The results are re ected in the form of the absorbance optical density at 450 nm using a microplate reader (BioTek, VT, USA). The experiments were conducted in triplicate.

Colony formation assay
Colony formation assays in vitro were performed according to standard protocols as previously described [19,20]. In brief, BCa cells were seeded in 6-well plates at 2 × 10 2 cells per well and cultured in medium containing 10% FBS. After 2 weeks of culture, cell colonies were xed with methanol for 30 min and stained with Giemsa for 15 min. The surviving colonies (≥50 cells per colony) were then counted under a microscope. The experiments were performed at least in triplicate.

Cell cycle analysis and apoptosis analysis
Apoptosis analysis was performed by ow cytometry by a BD FACSCanto II Flow Cytometer (BD Bioscience, San Jose, CA, USA). The cells were collected and washed twice with ice-cold PBS, resuspended in 200 μl binding buffer and stained with the Annexin V-FITC Detection Kit (Genechem, Shanghai, China) according to the manufacturer's instructions. Early apoptosis was determined based on the percentage of cells with annexin V + /PIstaining, while late apoptosis was determine based on the percentage of cells with annexin V + /PI + staining. The experiments were performed at least in triplicate.
For cell cycle, cells were harvested after 48 h transfection and xed in ice-cold 70% ehanol at 4°C overnight. Cells were then resuspended in propidium iodide solution (Genechem) according to the manufacturer's protocol and subjected to analysis using ow cytometry (BD).

Wound healing assays
A wound healing assay was conducted to measure the migration of BCa cells. Cell migration was assessed by measuring the movement of cells into a scraped, acellular area by a 200 μL pipette tube in 6well plates. The spread of wound closure was observed every 24 h. Migration was quanti ed by measuring the distances of cells that migrated toward the original wound eld.

Transwell invasion assays
Cell invasion assays were performed in Ttranswell chambers (BD Biosciences, San Jose, CA, USA) containing 8-μm pores in 24-well plates. A total of 1× 10 5 cells were starved overnight and seeded into the upper chamber of the transwell, which had been pretreated with 60 μl Matrigel for 4 h, in 200 μl of serumfree medium. The lower chamber was supplemented with 500 μl of medium with 10% FBS. After incubation at 37 °C for 24 to 72 h, noninvading cells in the upper chamber were scraped off with a cotton swab, and invading cells stuck to the lower surface of the membrane were xed in 100% methanol for 30 min and stained with Giemsa solution for 15 min. The cells were counted randomly for ve elds of each membrane under a light microscope. The experiments were performed at least in triplicate.
In vivo functional assays in nude mouse models Male, 4-to 6-week-old Balb/C-nu/nu nude mice were obtained from the Laboratory Animal Centre of Southern Medical University. Stably overexpressing SLC25A21 EJ cells and control cells were used to produce subcutaneous xenograft models. The nude mice were dIvided randomly into two groups (5 mice per group). A total of 2 × 10 6 SLC25A21-overexpressing EJ cells or control cells were subcutaneously injected into the right ank of mice, respectively. Tumor growth was measured every 3 days with a digital caliper. Tumor volume was calculated according to the formula: V (volume) = (D × d 2 )/2 (D indicates the longest diameter and d indicates the shortest diameter). After 30 days of monitoring, the mice were sacri ced by cervical dislocation, and the xenografts were removed and subjected to histological examination. This investigation was carried out in strict accordance with ethical standards in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committee on the Ethics of Animal Experiments of Southern Medical University.

Mitochondria isolation and determination of α-KG
Mitochondria extraction was conducted according to the instructions of the Mitochondria Isolation Kit (Beyotime Biotechnology, Jiangsu, China). Brie y, at least 2 × 10 7 cells were harvested and then centrifuged at 600 × g at 4 °C for 5 min. Mitochondrial isolation buffer was used to resuspend the pellet. The mixture was placed on ice for 15 min and then ground and centrifuged at 1000 × g at 4 °C for 10 min.
The supernatant was collected and centrifuged at 3500 × g at 4 °C for 10 min. The pellet contained the mitochondria. The supernatant was collected and centrifuged at 12,000 × g at 4 °C for 10 min. The supernatant contained the cytosolic fraction. Mitochondria were resuspended in mitochondria isolation buffer and centrifuged again at 3500 × g at 4 °C for 10 min to obtain puri ed mitochondria. The levels of α-KG in the mitochondria and cytosolic fraction were determined using α-ketoglutarate (α-KG) enzymelinked immunosorbent assay (ELISA) kits according to the documents of the manufacturer (CUSABIO Technology, Wuhan, China).

Mitochondrial membrane potential (Δψm)
Changes in the Δψm during the early stages of apoptosis were assayed using the Mitochondrial Membrane Potential Assay kit with JC-1 (Beyotime, Shanghai, China) as described previously [21]. Brie y, 5 × 10 4 cells were harvested and incubated with JC-1 at 37 °C for 20 min in the dark. The stained cells were washed with ice-cold working solution twice and then analyzed by ow cytometry (BD). JC-1 aggregates in the polarized mitochondrial matrix and forms J-aggregates, which emit red uorescence at 595 nm when excited at 525 nm. However, JC-1 cannot aggregate in the depolarized mitochondrial matrix and exists as JC-1 monomers, which emit green uorescence at 525 nm when excited at 485 nm. Mitochondrial depolarization is indicated by a decrease in the red/green uorescence intensity ratio.

ROS production
The ROS levels were detected with the Human ROS ELISA Kit and analyzed with a microplate reader according to the documents of the manufacturer (JianglaiLab, Shanghai, China).

Statistical analysis
All statistical analyses were performed using SPSS version 16.0 software (SPSS, Chicago, Illinois, USA). Differences between groups were identi ed using a two-tailed Student's t-test. Associations between SLC25A21 expression and clinicopathological characteristics were determined by the χ 2 test. Survival curves were plotted by the Kaplan-Meier method and compared by the log-rank test. The signi cance of various variables for survival was analyzed by the Cox proportional hazards model for multivariate analyses. A probability value of 0.05 or less was considered to be signi cant.

SLC25A21 is downregulated in BCa and is negatively associated with poor prognosis in BCa patients
To evaluate the expression levels of endogenous SLC25A21 in BCa, we rst examined SLC25A21 mRNA expression in 20 BCa tissues and their pair-matched noncancerous tissue samples using real-time PCR. As shown in Figure 1A, SLC25A21 mRNA was signi cantly downregulated in BCa tissues compared to nontumor tissues (P<0.001). To further con rm the levels of SLC25A21 expression in BCa, we also analyzed the SLC25A21 expression in public datasets from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) database. Consistent with our results, SLC25A21 was expressed at lower transcript levels in BCa tissues than in normal tissues in the TCGA cohort, GSE3167 and GSE68020 ( Figure 1B-1D).
To evaluate whether SLC25A21 expression was associated with the clinicopathologic features of BCa patients, we further performed immunostaining with SLC25A21 antibody in tissue microarrays using IHC and observed that SLC25A21 was highly expressed in 60.00% (12 of 20) of normal tissues and only in 33.93% (19 of 56) of BCa samples. The levels of SLC25A21 were signi cantly downregulated in BCa tissues (P =0.042, Figure 1E). The SLC25A21 immunoreactivity scores showed that the lower levels of SLC25A21 expression were associated with tumor invasion and T stage (Table 1). In addition, we also analyzed the prognostic signi cance of SLC25A21 expression in patients with BCa. As shown in Figure   1F, it was revealed that low expression of SLC25A21 was signi cantly correlated with poor survival in BCa patients (log rank P = 0.044). Taken together, these ndings suggest that SLC25A21 expression is downregulated in BCa and that low SLC25A21 expression is a poor prognostic biomarker for patients with BCa, which indicates that SLC25A21 might behave as a suppressor in BCa.

Overexpression of SLC25A21 inhibits the growth of BCa cells in vitro
To investigate whether SLC25A21 plays an important role in the biological behavior of BCa carcinogenesis, we examined the SLC25A21 expression in BCa cell lines to establish stable SLC25A21overexpressing BCa cells and SLC25A21-depleted BCa cells. EJ and T24 cells which have relatively low endogenous mRNA expression (Supplementary Information S3) were used to establish two BCa cell lines with stable SLC25A21 overexpression. Real-time PCR and western blotting analyses revealed that the relative level of SLC25A21 in BCa cells was signi cantly increased compared to that in control cell lines (Figure 2A and B). CCK8 assays revealed a signi cantly slower proliferation rate in SLC25A21overexpressing EJ and T24 cells than in control cells (P<0.001, Figure 2C). In the colony formation assay, both SLC25A21 overexpression obviously suppressed colony formation in both BCa cell lines (EJ, P = 0.0028 and T24, P = 0.0103, Figure 2D).

Overexpression of SLC25A21 induces cell apoptosis in BCa cells
Based on the results of CCK-8 and colony formation assays, we next investigated the potential mechanisms underlying the growth-inhibitory effects of SLC25A21 overexpression. The results of the ow cytometry cell cycle assays in two SLC25A21-overexpressing cell lines were inconsistent (data not shown). However, the upregulation of SLC25A21 expression resulted in an increased percentage of early apoptosis (EJ, P = 0.0092 and T24, P = 0.0138, Figure 2E) in both BCa cell lines compared with controls. In short, these data indicate that the SLC25A21-mediated decrease in cell proliferation is modulated by apoptosis, rather than by the G 1 -S checkpoint.

Overexpression of SLC25A21 inhibits BCa cell migration and invasion in vitro
We also examined the effects of SLC25A21 overexpression on BCa cell migration and invasion. Wound healing assays were performed to explore the effects on the migration ability of BCa cells after enforcing the expression of SLC25A21. The results suggested that SLC25A21 upregulation inhibited the migration of BCa cells (EJ cells, P=0.0211 and T24 cells, P=0.0181, Figure 2F). Cell invasion analysis demonstrated that the exogenous SLC25A21 overexpression in two different BCa cell lines signi cantly inhibited cancer cell invasion through Matrigel, a basement-membrane-like extracellular matrix (P<0.001, Figure 2G).

SLC25A21 overexpression suppressed tumor growth of BCa cells in vivo
In light of our in vitro ndings, we also assessed the effects of SLC25A21 overexpression on the growth of xenograft tumors in vivo. EJ cells with SLC25A21-overexpressing and control cells were subcutaneously inoculated into nude mice. As shown in Figure 3A-3B, compared with control cells, SLC25A21 overexpression signi cantly inhibited BCa cell growth in vivo (P=0.0392). The tumor weight was lower in the stable SLC25A21overexpression group than in the control groups (P=0.0293, Figure 3C). These results suggest that SLC25A21 exerts a signi cant inhibitory effect on BCa tumorigenesis in vivo.

SLC25A21 knockdown promotes BCa cell proliferation, migration and invasion and inhibits cell apoptosis in vitro
To assess the effect of reducing SLC25A21 expression in BCa, we next knocked down SLC25A21 levels by siRNA transfection in BIU-87 and UMUC3 cells, which have relatively high endogenous SLC25A21 expression. A considerable reduction in SLC25A21 was observed in both SLC25A21-depleted BCa cell lines ( Figure 4A and 4B). The results of CCK8 and colony formation assays showed that both BCa cell lines with SLC25A21 knocked down had a signi cant growth advantage over the respective cells transfected with negative control siRNA (P<0.01, Figure 4C and 4D). Moreover, SLC25A21 knockdown inhibited cell apoptosis in BIU-87 (P=0.0014) and in UMUC3 cells (P<0.0001, Figure 4E). In addition, markedlly higher migration rates and stronger invasiveness were observed in SLC25A21-depleted BIU-87 and SLC25A21-depleted UMUC3 cells, as revealed by the wound healing assay (BIU-87 cells, P=0.0254 and UMUC3 cells, P=0.0214, Figure 4F) and Transwell migration assay (P<0.0001, Figure 4G).

SLC25A21 overexpression induces cell apoptosis by the α-KG-mediated ROS pathway in BCa
We next studied the molecular mechanisms by which SLC25A21 affects proliferation. Because the function of SLC25A21 is linked to the transport of C5-C7 oxodicarboxylates, including α-KG, across the IMM, we speculated that aberrant SLC25A21 expression directly changes α-KG levels in the mitochondria. Therefore, we asessed the subcellular distribution of α-KG mediated by SLC25A21. SLC25A21 overexpression signi cantly increased α-KG levels in the cytosol and decreased α-KG levels in the mitochondria in T24 cells (cytosol, P=0.0150 and mitochondria, P=0.0284, Figure 5A, left). Moreover, SLC25A21 knockdown mediated by siRNA signi cantly elevated α-KG levels in the mitochondria of UMUC3 cells compared those after transfection with a negative control siRNA (P<0.001, Figure 5A, right). The oxidative metabolism of α-KG produces succinate in the mitochondria. Indeed, decreased levels of the succinate were observed in the mitochondria of BCa cells with SLC25A21 overexpressing (P=0.0032, Figure 5B, left). In contrast, the levels of succinate in the mitochondria were reduced in SLC25A21depleted UMUC3 cells, further supporting the view that upregulation of SLC25A21 promoted the e ux of α-KG from mitochondria (P=0.0028, Figure 5B, right).
Because ROS generation is augmented by mitochondrial dysfunction, we next investigated whether SLC25A21 induces ROS production. The results showed that with the reduction of α-KG levels in the mitochondria, the upregulation of SLC25A21 in the two cell lines resulted in signi cantly higher ROS accumulation compared with the corresponding control group, which indicated a weaker ability to survive oxidative stress (T24 cells, P=0.0341 and EJ cells, P=0.0348, Figure 5C). We also evaluated the effects of SLC25A21 on Δψm in both SLC25A21-overexpressing cells by ow cytometry. As shown in Figure 5D, a decreased Δψm was observed in both SLC25A21-overexpressing BCa cell lines. T24 cells with SLC25A21overexpressing exhibited a Δψm decline of more than 36.3% (P<0.0001), and EJ cells showed a 13.3% decline (P=0.0012) compared to the control. It has been reported to participate in triggering caspasedependent apoptosis by regulating Δψm. The results of western blotting showed that SLC25A21 induced cytochrome C (cyto C) transfer from mitochondria to the cytosol in BCa cells. In addition, we also found that SLC25A21 overexpression increased the activation of caspase-3 and caspase-9 in BCa cell lines and xenografts tissues ( Figure 5E and 5F), suggesting that SLC25A21 triggers mitochondrial apoptotic pathways in BCa. Hence, these data indicate that SLC25A21 activates the mitochondrial apoptosis pathway by inducing α-KG e ux from the mitochondria to the cytosol, and then increasing ROS levels, thereby activating cell apoptosis ( Figure 5G).

Discussion
In the present study, we illustrated that the downregulation of SLC25A21 plays a key role in promoting human BCa development. Our data showed that SLC25A21 is strongly downregulated in BCa tissues and is especially positively correlated with a invasion and poor outcome in BCa patients. We validated SLC25A21 as an important tumor suppressor gene in BCa by gain-and loss-of-function experiments in vitro and in vivo. Interestingly, we also found that SLC25A21 promoted the e ux of α-KG from the mitochondria and increased ROS levels, resulting in the activation of the mitochondrial apoptosis pathway. Thus, our study provides strong evidence supporting the tumor suppressor roles of SLC25A21 in BCa.
SLC25 is a large family of nuclear-encoded transporters embedded in the IMM and, in a few cases, other organelle membranes. The members of this superfamily are widespread in eukaryotes and are involved in numerous metabolic pathways and cell functions [9]. SLC25 members vary greatly in the nature and size of their transported substrates, modes of transport and driving forces, although the molecular mechanism of substrate translocation may be basically the same. In recent years mutations in the SLC25 genes have been shown to be responsible for some diseases [9], highlighting the important role of SLC25 in physiological and pathological implications.
Human SLCA25A21 is ubiquitously expressed at both the mRNA and protein levels, with very little variation between tissues [14]. The best substrates for human SLC25A21 are 2-oxoadipate and α-KG [14]. The physiological role of SLC25A21 suggests its possible involvement in 2-oxoadipate acidemia, which is accompanied by the accumulation and excretion of large amounts of 2-oxoadipate, 2-aminoadipate and 2-hydroxyadipate in urine [14]. Moreover, Harris and colleagues rst reported that the deletion of SLC25A21 appears to be associated with lung cancer [15]. However, the effects and mechanisms responsible for these diseases have not been characterized. In the present study, we revealed that the expression levels of SLC25A21 are signi cantly downregulated in BCa tissues compared with normal bladder epithelial tissues, in accordance with the results from our real-time RT-PCR IHC and TCGA and GSEA database analyses, and we determined that low expression of SCL25A21 is signi cantly associated with invasion and poor survival in BCa patients. These data indicate that the downregulation of SLCA25A21 expression could serve as a predictor of a higher risk of BCa progeression.
Moreover, to highlight the function of SLC25A21 in BCa, we further explored the critical effects of SLC25A21 on cell behaviors in BCa by gain-and loss-of-function experiments. Our results identi ed that SLC25A21 acts as a tumor suppressor results from the overexpression of SLC25A21 inhibits cell growth and facilitates cell apoptosis in BCa in vitro and in vivo. To the best of our knowledge, our study provides the rst evidence that SLC25A21 plays a critical role in tumors, especially BCa. Our results also indicate that the elevated expression of SLC25A21 can inhibit the malignant phenotype of BCa, suggesting SLC25A21 as a potential predictor and therapeutic target for BCa progression.
Many new functions have been ascribed to SLC25 in tumors from data generated over the last year [10,12]. For example, SLC25A22, a glutamate carrier that facilitates the transport of glutamate across the IMM into the mitochondria matrix, promotes cell proliferation and tumor progression of colorectal cancer with KRAS mutations via tntrascellular synthesis of aspartate [10]. The knockdown of SLC25A10 changes the growth properties to a less malignant phenotype and casues increased glutamine dependency and sensitivity to oxidative stress in human lung adenocarcinoma cells, suggesting SLC25A10 as a novel target for anticancer strategies [13]. However, the mechanisms by which SLC25A21 mediates the progression of BCa remain unknown. The main physiological function of human SLC25A21 is probably to catalyze the uptake of 2-oxoadipate into the mitochondria and the e ux of α-KG from the mitochondria to the cytosol. Here, we observed that SLC25A21 downregulation, directly restricted the e ux of α-KG, resulting in elevated levels of α-KG in the mitochondria. α-KG is an intermediate of the tricarboxylic acid cycle (TAC) that plays important roles in cell metabolism and physiology. As a biological substance, α-KG is essential for the oxidation of fatty acids, amino acids and glucose, and serves as the carbon skeleton for the synthesis of glutamate and glutamine [22]. Additionally, emerging evidence shows that α-KG can effectively protect different animal models from oxidative stress [23]. It has been established that α-KG can act as a true antioxidant, because it directly reacts with hydrogen peroxide (H 2 O 2 ) to form succinate, water, and carbon dioxide [23,24]. The supplementation of α-KG can directly or indirectly stimulate endogenous antioxidant defense [25,26]. Therefore, we speculated that SLC25A21 depletion-mediated α-KG arrest in mitochondria may reduce ROS generation in BCa. Indeed, when higher levels of mitochondrial α-KG are present in SLC25A21depleted BCa cells, ROS production is signi cantly decreased, accompanied by an increase in succinate in the mitochondria and vice versa. These results suggest that SLC25A21 can indirectly protect against oxidative stress by restricting the e ux of α-KG in the mitochondria.
ROS may exhibit a dual role in promoting or suppressing cancer formation. Although upregulation of ROS is known to promote tumorigenesis, growing evidence suggests that excessive accumulation of ROS leads to cell apoptosis [27,28]. Apoptosis, a type of programmed cell death, is an important pathway for regulating homeostasis and morphogenesis by eliminating redundant and abnormal cells and is associated with various diseases, especially cancer. The typical characteristic of cancer is deregulation of apoptosis. Apoptosis is controlled by two principal pathways, including the death receptor-mediated extrinsic pathway and the mitochondrion-dependent intrinsic pathway [27]. The intrinsic pathway is much more protean but mitochondrial membrane permeabilization (MMP) is a crucial step in the signaling cascade. MMP is triggered by a variety of stimuli, such as hypoxia, oxidative stress, DNA damage, toxins, viral infections or the absence of growth factors or cytokines. All of them can induce Δψm loss in the mitochondria. The mitochondrial apoptosis pathway starts with mitochondrial depolarization, triggering the release of cyto C into the cytosol, thereby activating caspase-9 and caspase-3, and ultimately leading to apoptosis [29]. Here, we also found that SLC25A21 reduced MMP and induced the subsequent following e ux of mitochondrial cyto C to the cytosol and activation of caspase 9 and caspase 3. The elevaton of intracellular ROS levels mediated SLC25A21 overexpression-induced apoptosis in BCa cell lines, which may explain why SLC25A21 overexpression decreased the cell proliferation and colony formation ability of BCa cells in vitro and in vivo. In addition, our results also revealed that low expression of SCL25A21 is signi cantly associated with invasion state of BCa patients and promoted tumor cell migration and invasion. Recently, Wu et al. founded that cancer-dervied succinate promotes macrophage polarization and cancer metastasis in murine and human lung cancer cells [30]. Moreover, previous studies have reported that the accumulation of fumarate in mouse and human cells elicits an epithelialto-mesenchymal-transition (EMT), a phenotypic switch associated with cancer initiation, invasion, and metastasis [31]. Frezza et al. reported that the fumarate inhibits Tet-mediated demethylation of a regulatory region of the antimetastastic miRNA cluster, leading to the expression of EMT-related transcription factors and enhanced migratory properties in renal cancer [32]. In TAC reaction, α-KG is oxidized and decarboxylated to succinyl CoA mediated by the α-KG dehydrogenase complex, which in turn producing succinate. Then, succinate is oxidized and hydrated to fumarate. Here, we observed that higher levels of mitochondrial α-KG promoted the accumulation of succinate in the mitochondria in SLC25A21-depleted BCa cells. Therefore, we speculated that SLC25A21 depletion-mediated α-KG arrest in mitochondria may result in an increase of succinate and fumarate, thereby promoting cell initiation, invasion, and metastasis in BCa. These data indicate that the downregulation of SLCA25A21 expression could serve as a predictor of a higher risk of BCa progeression. The mechanisms by which SLC25A21 mediates invasion inhibition present an interesting issue for further investigation.

Conclusions
Our research demonstrated that SLC25A21 is downregulated in BCa tissues and that its downregulation might play a promoting role in BCa tumorigenesis and progression through the ROS-mediated mitochondrion-dependent apoptosis pathway. These results indicate that the overexpression of SLC25A21, which speci cally induces ROS-mediated apoptosis in cancer cells, could be used to develop a highly effective treatment strategy for BCa patients.

Declarations
Ethics approval and consent to participate The use of tissues for this study has been approved by the ethics committee of Nanfang Hospital, Southern Medical University (Guangzhou, China). All of the patients signed an informed consent before the use of these clinical materials for research purposes. Moreover, the animal experiences were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committee on the Ethics of Animal Experiments of Southern Medical University.

Consent for pubilication
All authors have read the paper and approved of publication. Written informed consent for publication was obtained from all participants.

Con ict of interest statement
The authors have no con icts of interest.