Fatty acid oxidation inhibitor etomoxir suppresses tumor progression and induces cell cycle arrest via PPARγ-mediated pathway in bladder cancer

Bladder cancer (BCa), currently the ninth most common cause of cancer with a gradually increasing global incidence in the last 30 years, is relatively common in more developed areas and occurs more frequently among males than females (sex ratio worldwide of 3.5:1) [1]. As one of the most common malignances in the urinary system, 70% of newly diagnosed patients are non-muscle-invasive BCa (NMIBC) [2], which is treated by transurethral resection, followed by intravesical instillations of immunotherapy or chemotherapy [3]. At the present stage, the gold standard therapeutic method for muscle-invasive BCa (MIBC) is radical cystectomy, which still has many unfavorable outcomes [4,5]. However, even after receiving radical cystectomy, the BCa recurrence and distant metastasis rate remains approximately 50%, and its 5-year survival rate is only 50–66% [6–8]. Therefore, more effective therapeutic strategies need to be detected for BCa treatment.

Reprogramming of energy metabolism has been recognized as a hallmark of cancer [9,10]. Otto Warburg first described that cancer cells prefer aerobic glycolysis as the main energy resource even when oxygen is sufficient, which is known as the Warburg effect [11–13]. However, recent studies have stressed the crucial role of fatty acid oxidation (FAO) in providing adenosine triphosphate (ATP), nicotinamide adenine dinucleotide phosphate (NADPH), and other important anabolic materials in cancer cells [14–16]. Lin et al. [17] reported that FAO was active in glioma tissues and inhibition of FAO with etomoxir reduced energy production and cell proliferation. Furthermore, some kind of cancers rely on lipid fuel more than on aerobic glycolysis [18], which is greatly different from the Warburg effect. Recent studies showed that alterations in lipid metabolism were attached to bladder carcinogenesis [19]. As some metabolomics profiling analysis showed [20], there may exist an active fatty acid β-oxidation in promoting development and progression of BCa.

To be fully oxidized, fatty acids, derived from fatty acid synthesis (FAS) or extracellular environment, with the assistance of fatty acid binding proteins (FABP), are located at the outer membrane of mitochondria and then activated to fatty acyl-CoAs by fatty acyl-CoA synthetase. For the lack of permeability to the inner mitochondrial membrane, long-chain acyl-CoA is converted into acylcarnitine catalyzed by carnitine palmitoyltransferase I (CPT1), which is the first rate-limiting step of FAO [21]. Then, acylcarnitine is transported by carnitine acylcarnitine translocase (CACT) and converted back into acyl-CoA by carnitine palmitoyltransferase II (CPT2) in mitochondrial matrix for subsequent oxidation [22]. The most prominent regulators of FAO are peroxisome proliferator-activated receptors (PPARs), consisting of three isoforms PPARα, PPARβ/δ and PPARγ. Expressed in various tissues at different levels, they have different functions with partial overlapping. By regulating gene expression and biological processes associated with lipid metabolism [23,24], they are critical in modulating FAO.

Etomoxir is an irreversible inhibitor of CPT1 and has been clinically used for chronic heart failure and type 2 diabetes mellitus for a long time [25,26], but retired due to its adverse side effects in the liver [27]. Figure 1A illustrated part of the process by which fatty acids were transported into mitochondria for subsequent β-oxidation and the inhibitory site of etomoxir. Pharmacological inhibition of FAO with etomoxir was mostly used in research, and some recent experiments indicated promising anti-tumor results [28,29]. But studies reported that the effects of etomoxir on tumor growth are still not adequate, and there are no such studies published in BCa. In our study, we investigated the effects of FAO inhibition with etomoxir in BCa cell viability, fatty acid metabolism, motility,and molecular mechanisms. Our results indicated that FAO, inhibited by etomoxir in BCa cells, is important in sustaining tumor growth, decreasing cell migration and invasion, and inducing cell cycle arrest at G₀/G₁ phase through PPARγ-mediated pathway. Inhibition of FAO with etomoxir may provide a novel avenue to investigate new therapeutic approaches to human BCa.

Human BCa cell lines UM-UC-3 (Cat. #TCHu217) and T24 (transitional cell carcinoma, Cat. #SCSP-536) were obtained from Chinese Academy of Sciences in Shanghai, China. The cell lines were authenticated by the China Center for Type Culture Collection in Wuhan. UM-UC-3 cells were maintained in DMEM (Gibco) and T24 cells were maintained in RPMI-1640 medium (Gibco) supplemented with 10% fetal bovine serum (FBS) in a humidified atmosphere composed of 5% CO₂ and 95% air at 37°C. All human cell lines authenticated using STR profiling within the last 3 years have been included. All experiments were performed with mycoplasma-free cells.

A total of 3000 BCa cells in 200 μl medium were pipetted into 96-well plates to grow for 24 h, and then treated by etomoxir (Sigma–Aldrich, Cat. #E1905) (0, 25, 50, 75, 100, 150, and 200 μM, diluted in pure water) or ST1326 (Avanti, Cat. 870853P) (0, 0.1, 0.5, 1, 5, 10, 25, and 50 μM, diluted in phosphate buffer saline) for 24, 48, and 72 h. After adding 20 μl MTT in each well, incubating 4 h at 37°C, discarding the medium and dissolving formazan precipitate in 150 μl DMSO, the absorbance at 490 nm was detected by a microplate reader (Molecular Devices, Cat. #SpectraMax M2). And the following experiments of etomoxir-treated BCa cells were generated as follows: cells were first cultured for 24 h and then treated by etomoxir at 0, 75, and 150 μM for 48 h.

BCa cells were first incubated for 24 h and subsequently treated with PPARγ antagonist GW9662 (Sigma–Aldrich, Cat. #M6191) at final concentrations of 0, 20, and 40 μM for 24 h. Next, cells were treated by etomoxir for another 48 h and then collected for relevant experiments. GW9662 was initially dissolved in DMSO as stock solution with a concentration of 50 mM, and the 0 μM group added DMSO to a concentration of 0.1% as control.

Total RNA was extracted from BCa cells and bladder tissues using the Qiagen RNeasy Mini Kit (Qiagen, Cat. #74101) and QIAshredder from Qiagen (Qiagen, Cat. #79654) according to the manufacturer’s instructions. Genomic DNA contamination was removed with DNase I digestion (Qiagen, Cat. #79254) in each RNA sample. The quantity control of isolated RNA was assessed by NanoDrop® ND-2000 UV-Vis spectrophotometer (Thermo Scientific).

Microarray analysis was proceeded as previously described [31]. Briefly, according to standard Affymetrix protocol, biotinylated cDNA was prepared from 250 ng extracted RNA with Ambion® WT Expression Kit. And then, 5.5 μg cDNA was hybridized, washed, stained, and scanned. Data were processed and analyzed with Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway. The microarray data of bladder tissues were submitted to the Gene Expression Omnibus (GEO) database and the accession number is GSE76211. All data were MIAME compliant.

The cDNA was synthesized from 1 μg of total RNA with RevertAid Ace qPCR RT kit (Toyobo). Then 1 μg cDNA was added in each reaction to a final volume of 20 μl with iQ™ SYBR®-Green Supermix (Bio-Rad). All primer sequences with its annealing temperatures were listed in Supplementary Table S1. And the cycle numbers of threshold (C_T) values were normalized to GAPDH.

Free fatty acid (FFA) level of tissues and cells was measured by Free Fatty Acid Quantitation Kit (Sigma–Aldrich, Cat. #MAK044). The ATP and NADPH level were determined using ATP Assay Kit (Beyotime, Cat. #S0026) and NADP/NADPH Quantification Colorimetric Kit (Biovision, Cat. #K347-100), respectively. All the procedures were performed according to the manufacturers’ instructions.

Cell samples were homogenized individually with 5% concentrated sulfuric acid methanol and 0.2% butylated hydroxytoluene methanol in ice water bath for ultrasonic for 15 min. After mixing and extracting in a constant temperature water bath at 90–95°C for 1.5 h, cooling to room temperature, saturated NaCl and n-hexane were added to vortex for 1 min. Then the following steps were performed: centrifuging at 5 min, 4°C, 3500 rpm, transferring supernatants to a new centrifuge tube, and vortexing with nonadecanoic acid methyl ester for 10 s. Then fatty acids were identified and quantified by 7890A GC-5975C FID (Agilent, U.S.A.) with an Agilent DB-225 capillary column (10 m × 0.1 mm i.d., 0.1 µm film thickness). The carrier gas was helium. The initial temperature of column was set at 50°C, then increased to 205°C at a rate of 30°C/min, held for 1 min, increased again to 230°C, and then held isothermally. The inlet and FID temperature was 250°C, and the split ratio was 1:15. All the data were analyzed and processed with FID Chem Station (G1701EA.02.00.493) and ACD/Spectrus Processor 2015 (S30S41) software.

Etomoxir-treated BCa cells were fixed with paraformaldehyde (PFA) for 15 min, rinsed in distilled water and 60% isopropanol, and stained with freshly prepared Oil Red O (Sigma–Aldrich, Cat. #SLBP5248V) working solution for 10 min at room temperature. Then cell slides were rinsed with 60% isopropanol and distilled water, counterstained with Hematoxylin (Sigma–Aldrich, Cat. #H9627), mounted in glycerol gelatin medium and observed with a phase contrast microscope (Leica, Cat. #DMI 1).

A total of 800 UM-UC-3 cells/well and 1000 T24 cells/well were added in six-well plates to grow into colonies for 7–10 days. The medium was removed, the cells were fixed with 4% PFA, then stained with Crystal Violet for 30 min, followed by photographing and counting.

After collecting and washing with cold PBS, BCa cells were resuspended with 1× DNA Staining Solution containing permeabilization solution and propidium iodide (Multi sciences), then incubated in the dark for 30 min at 37°C. Flow cytometry (Beckman, Cat. #FC500) was used to analyze the cell cycle. According to its manufacturer’s instructions, cell apoptosis analysis was conducted using annexin V-fluorescence isothiocyanate (FITC)/PI apoptosis detection kit (BD biosciences) by flow cytometry.

Cells were cultured in six-well plates and scratched with a tip, washed with PBS. Medium containing different concentrations of etomoxir was added and results were photographed by a phase contrast microscope in several pre-marked spots at 0 and 12 h. The rate of migration was analyzed statistically.

A total of 4 × 10⁴ BCa cells in 200 μl serum-free medium were seeded in the upper transwell chamber (Corning), with 600 μl medium containing 10% FBS in the lower chamber to induce cell migration. After incubation for 24 h at 37°C, removing the cells in the upper chamber, fixing and staining with 0.1% Crystal Violet, then photographing and counting migrated cells by phase contrast microscope. To conduct invasion assay, the transwell chambers were percolated with ECM Matrix gel solution (Sigma–Aldrich) for 1 h at 37°C, and 1 × 10⁵ cells were seeded as described previously. Subsequent procedures were the same as those of the migration assays.

BCa cells were lysed and sonicated on ice in RIPA buffer including protease inhibitor (Sigma–Aldrich) and phosphatase inhibitor for approximately 30 min. Then after centrifuging at 12000×g for 10–15 min, the supernatant was collected. Protein was separated using 7.5–12.5% SDS/PAGE gels and transferred to the PVDF membrane (Millipore), which was then blocked in 5% fat-free milk (BD Biosciences) and incubated with primary antibodies (Supplementary Table S2) overnight and secondary antibodies (Supplementary Table S3) for 2 h. Bands were detected by enhanced chemiluminescence kit (Bio-Rad) and visualized by Molecular Imager Chemi Doc XRS⁺ Imaging system (Bio-Rad).

PPARγ and PPARα DNA-binding activity was detected using PPARγ and PPARα transcription factor assay kit (Abcam) independently according to the manufacturer’s instructions. Briefly, the nuclear fraction was prepared from BCa cells using Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific). Then, conjunction of primary and secondary antibodies was conducted according to the protocols. The absorbance at 450 nm was detected by a microplate reader (Molecular Devices, Cat. #SpectraMax M2).

Immunofluorescence staining of BCa cells was previously described [31]. For immunohistochemistry (IHC) staining, paraffin sections from xenograft tumors were deparaffinized. Then antigen retrieval was conducted in citrate buffer. Inactivating endogenous peroxidase activity in 0.3% H₂O₂ and blocking in goat serum for another 30 min, the section was subsequently incubated in the indicated primary and secondary antibodies. After counterstaining with Hematoxylin, the slices were eventually visualized by Olympus BX53 biological microscope.

All the tissue paraffin sections from xenograft tumors were stained with Hematoxylin and Eosin (H&E). The slides were serially deparaffinized, rehydrated by xylene, 100% ethanol, 96% ethanol, 80% ethanol, 70% ethanol, and H₂O. Sections were subsequently stained with 10% Hematoxylin (Sigma–Aldrich) for 7 min. And 1% Eosin (Sigma–Aldrich) with 0.2% glacial acetic acid was taken to stain the cytoplasm. Images of the slices were eventually taken by an inverted phase contrast microscope (Leica, Cat. #DMI 1).

Male BALB/c nude mice (3 weeks old) were obtained from Beijing HFK Bioscience Co., Ltd. in Beijing, China. After adaptation of 1 week in the laboratory of animal facility of Zhongnan Hospital of Wuhan University, mice were injected with T24 cells of 3 × 10⁶ dispersed in 0.2 ml PBS into the right flank. Twelve days later, mice with approximately 3 mm × 3 mm tumors were assigned to two groups (n=5) randomly. Etomoxir (40 mg/kg) and vehicle were injected intraperitoneally every other day for 20 days. The tumor size was measured every 2 days and calculated as: tumor volume (mm³) = length × width² × 0.5. After killiing, tumor tissues were dissected for tumor weight and fixed in 4% PFA preparing for subsequent staining. Animal experiment complied with the ARRIVE guidelines and carried out according to the U.K. Animals (Scientific Procedures) Act, 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments. The animal experiment was approved by the Institutional Animal Care and Use Committee at Center for Animal Experiment, Wuhan University (approval no. 2018152).

All experiments were performed at least three times, and representative data were from three iterations. One-way analysis of variance (ANOVA) and two-tailed Student’s t test were used to evaluate statistical significance of differences between subgroups. All the statistical analyses were conducted with SPSS 16.0, and the cut-off level was set at probability values of P<0.05.

The differentially expressed genes (DEGs) were detected in three MIBC tissues compared with three normal bladder tissues by microarray (Figure 1B). GO analysis revealed that DEGs were associated with a variety of metabolic process especially the fatty acid metabolic process (Figure 1C). Path-net analysis based on KEGG pathway in GCBI platform showed that fatty acid degradation, cell cycle, PPAR signaling pathway, and p53 signaling pathway were significantly related to BCa (Figure 1D). We detected the FFA levels in nine pairs of BCa tissues and paracancerous tissues, and found that there were significantly increased FFA levels in BCa tissues (Figure 1E). To verify if the genes involved in metabolism were misregulated, quantitative real-time PCR (qRT-PCR) was performed in paired BCa and paracancerous tissues. Although the alteration of CPT1A gene expression in BCa tissues was not significant, our results showed that ACSL1, SCD1, CD36, and PLA2 were up-regulated, while ACAD7 and PTGIS were down-regulated in BCa compared with paracancerous tissues (Supplementary Figure S1A–G). Taken together, we hypothesized that fatty acids catabolism was altered in BCa, and inhibition of fatty acids oxidation might have a great impact on BCa (Figure 1A).

We treated BCa cells with etomoxir and ST1326 at different concentrations (0, 25, 50, 75, 100, 150, 200 μM and 0, 0.1, 0.5, 1, 5, 10, 25, 50 μM, respectively) for 24, 48, and 72 h (Figure 2A,B and Supplementary Figure S1H–K), and found that etomoxir and ST1326 could inhibit UM-UC-3 and T24 cell viability in a dose-dependent manner. Based on the MTT assay, in the following experiments, cells were treated with etomoxir at 75 and 150 μM for 48 h as the low- and high-dose groups and vehicle as the control group for subsequent tests. Oil Red O staining revealed that lipid accumulation was increased in etomoxir-treated UM-UC-3 cell (Figure 2C). Consistently, etomoxir treatment increased the FFA level in UM-UC-3 cell (Figure 2D). Compared with fatty acid standard by gas chromatography-flame ionization detection (GC-FID), the absolute qualitative and quantitative analysis of fatty acid composition was conducted in etomoxir-treated UM-UC-3 cells (Supplementary Figure S1L–M). Fatty acids, such as C16:0, C18:0, C18:2n6t, C20:2, C20:3n3, C22:1n9, and C24:0 were significantly elevated, while C12:0 and C20:4n6 were decreased in UM-UC-3 cells caused by the FAO inhibition of etomoxir (Figure 2I).

To further confirm the FAO inhibition of etomoxir, we detected the ATP and NADPH level in BCa UM-UC-3 cells, and found that ATP and NADPH level were significantly decreased in etomoxir treated groups (Figure 2E,F) and NADP⁺/NADPH ratio was increased (Figure 2G). By inhibition of FAO, the proliferation inhibitory effect of etomoxir and ST1326 was validated with clonogenic survival assay (Figure 2H,K), and confirmed by statistical analysis (Figure 2J,L), respectively.

We established a mouse model by transplanting T24 cells subcutaneously to further evaluate the effect of etomoxir on BCa cell growth in vivo (Figure 3A). The results indicated that etomoxir could significantly suppress tumor growth compared with the vehicle group (Figure 3B). After the indicated time, all tumors were dissected from the mice, and the tumor weight of etomoxir treatment group was lower in comparison with the vehicle group (Figure 3C). The tumor tissues dissected from mice were stained with H&E, and the IHC staining analysis results showed that Ki-67, an important marker of cell proliferation [32], was considerably decreased in the etomoxir-injected group (Figure 3D). Moreover, the positive staining of PPARγ was up-regulated in etomoxir group compared with the vehicle group, but the alteration of CPT1A was not significant between the two groups (Figure 3D).

The flow cytometry analysis indicated that the cell cycle was arrested at G₀/G₁ phase in etomoxir-treated BCa UM-UC-3 (Figure 4A) and T24 cells (Supplementary Figure S2A) in a dose-dependent manner. However, we did not observe significant apoptosis even in the high concentration group in UM-UC-3 cells (Figure 4B), as well as in T24 cells (Supplementary Figure S2B). Consistent with the flow cytometry analysis, the Western blot (WB) results showed that G₀/G₁-related proteins such as cyclin D1, CDK2, CDK4, and CDK6 were strongly down-regulated in the low- and high-concentration groups (Figure 4C and Supplementary Figure S2C). Moreover, we detected the upstream indicators of the cell cycle pathway, and found that etomoxir treatment up-regulated p21 and acetyl-p53 proteins in BCa UM-UC-3 cells. Furthermore, the PPARγ-SIRT1 pathway related proteins were analyzed that PPARγ was up-regulated and total and phosphorylated SIRT1 were strongly down-regulated in UM-UC-3 cells (Figure 4C).

Cell migration was detected by wound healing and transwell migration assays. After 12-h incubation, the migration rate of wound healing was calculated in UM-UC-3 cells that the migration rate of etomoxir-treated groups was significantly reduced (Figure 4D). And the transwell migration and invasion assays indicated that etomoxir could decrease the cell invasion and migrations rate of BCa UM-UC-3 (Figure 4E) and T24 cells (Supplementary Figure S2D). In accordance with the transwell assay, WB analysis showed a down-regulation of N-cadherin, Vimentin, β-catenin, Slug, MMP2, and MMP9, and an up-regulation of E-cadherin proteins associated with the epithelial–mesenchymal transition (EMT) process. Furthermore, phosphorylated/total AKT, the upstream regulator of EMT, was down-regulated both in UM-UC-3 (Figure 4F) and T24 cells (Supplementary Figure S2E).

Etomoxir, an irreversible inhibitor of CPT1 as reported [25,26], altered the mRNA expression of FAO-related genes in UM-UC-3 cells (Figure 5A). However, we did not observe a significant change of CPT1A both at transcriptional and protein levels (Figure 5A–C). PPARα expression was slightly up-regulated in high-dose group but not in low-dose group at mRNA level. However, at the protein level, PPARα did not change significantly in both groups (Figure 5B). The results indicated a strong up-regulation of PPARγ expression both at mRNA and protein levels in etomoxir-treated UM-UC-3 cells (Figure 5A,B). The Enzyme-linked immunosorbent assay (ELISA) showed that PPARγ DNA-binding activity was increased in high-dose group compared with the control group in UM-UC-3 cells, but there was no change of PPARα DNA-binding activity (Figure 5D).

Next, we blocked the PPARγ activity with distinct concentrations of PPARγ antagonist GW9662. The flow cytometry analysis showed that the cell cycle G₀/G₁ arrest caused by etomoxir was significantly recovered by GW9662 in a dose-dependent manner in UM-UC-3 (Figure 5E) and T24 cells (Supplementary Figure S2F), confirmed by statistical analysis (Figure 5F and Supplementary Figure S2G). The clonogenic survival assay revealed that GW9662 could partially restore the suppressed proliferation caused by etomoxir (Supplementary Figure S2H–I). Consistently, WB results revealed that the reduced proteins related to G₀/G₁ phase such as CDK2, CDK4, CDK6, and cyclin D1 were also inverted by PPARγ antagonist in BCa UM-UC-3 (Figure 5G) and T24 cells (Supplementary Figure S2J).

Reprogramming energy metabolism is considered to be a hallmark of cancer [33]. Otto Warburg first made the point that cancer cells turned to aerobic glycolysis rather than oxidative phosphorylation to sustain the energy supply [34]. However, recent studies have indicated that other substrates such as glutamine and fatty acids are also critical in cancer cell survival, and prostate cancer displays a low rate of glucose utilization and a high rate of fatty acids uptake [18,35]. Moreover, B-cell lymphoma and malignant glioma are highly dependent on FAO for growth and survival [17,36]. Whether the aerobic glycolysis or FAO is predominant in BCa, is still unknown. Related metabolomics analysis and studies revealed that fatty acid metabolism was dysregulated and was promising in treatment and diagnosis of BCa [19,37]. In this study, our microarray and bioinformatics analysis revealed that DEGs between normal bladder and BCa tissues are enriched to lipid metabolism, especially fatty acid metabolism. Genes involved in metabolism were also misregulated in BCa tissues (Supplementary Figure S1A–G). Moreover, KEGG pathway analysis showed that fatty acid degradation was significantly linked to BCa via PPAR signaling pathway. The FFA, which could be transported into mitochondria for oxidation, was also increased in BCa tissues compared with paracancerous tissues (n=9, Figure 1). Thus, on this basis, we hypothesized that FAO was closely related to BCa.

Etomoxir was an inhibitor of FAO, and was developed for treating heart disease and diabetes mellitus for a long time [38,39]. A few studies have reported that etomoxir, which could irreversibly inhibit the function of CPT1, had an influence on CPT1A expression [18]. However, in our study, we did not see the significant change of CPT1A expression at mRNA and protein levels (Figure 5A–C) in BCa cells and in vivo experiment (Figure 3D), which might be because of the enzyme activity inhibition rather than the transcriptional regulation of etomoxir. Although recently some studies identified an off-target effect of etomoxir [40,41], still we could see the FAO inhibition of etomoxir in research. Consistently, the FAO inhibition of etomoxir could be validated by increased fatty acid and lipid accumulation and decreased ATP and NADPH level in the present study (Figure 2C–G). Etomoxir could suppress the transportation of long-chain fatty acid into mitochondrial for β-oxidation, thus increasing the level of long-chain fatty acids (Figure 2I). We also observed a significant decrease in C12:0, which might be the compensatory results of long-chain FAO inhibition.

A variety of studies have indicated that reduced ATP by FAO inhibition could suppress tumor cell proliferation. Ricciardi et al. [16] reported that pharmacological inhibition of CPT1A could impair cell proliferation of acute myeloid leukemia. Consistently, our results showed that FAO inhibition by etomoxir could suppress BCa cell growth both in vitro (Figure 2A,B,H,K) and in vivo (Figure 3). Moreover, we observed an inhibitory effect of etomoxir on invasion and migration, with down-regulated N-cadherin and up-regulated E-cadherin proteins related to EMT, which have been certified to have a critical role in cancer cell invasion and migration [42]. The phosphorylation of AKT at Ser⁹ could regulate the EMT pathway [43], which was consistent with our WB results of down-regulated phosphorylated/total AKT (Figure 4F and Supplementary Figure S2E). Furthermore, accumulating studies reported that FAO activation was an important mechanism for tumor cells to develop drug resistance [44]. Galicia-Vazquez et al. [45] reported that ibrutinib-resistant cells of chronic lymphocytic leukemia showed metabolic rewiring to FAO, and FAO inhibition with etomoxir could re-sensitize resistant cells. Intravesical instillations of chemotherapy drugs is an important adjuvant therapeutic method for NMIBC. However, the 5-year recurrence rate and the risk of progression to MIBC is not optimistic [46]. The effect of etomoxir combination with chemotherapy drugs on BCa cells has not been detected. That would be an interesting research subject.

Microarray analysis revealed the connection of BCa with cell cycle and PPAR signaling pathway (Figure 1D). And cell cycle could be blocked at G₀/G₁ phase by etomoxir in BCa UM-UC-3 (Figure 4A) and T24 cells (Supplementary Figure S2A), with changes in G₀/G₁-related proteins (Figure 4C and Supplementary Figure S2C). However, we did not notice a significant change in apoptosis. It might because of the unchanged expression of CPT1A, which could protect cells from apoptosis [47]. CDK inhibitor p21 has been reported as an important tumor suppressor and plays a crucial role in inhibiting cell growth [48]. p53 could induce variable expression of p21 and cyclin D1, which almost exclusively determines the cell cycle status [49]. We observed a strong increase in p21 and acetyl-p53 protein in UM-UC-3 cells when treated with etomoxir. Our previous studies demonstrated the cell cycle regulation of PPARγ-SIRT1 feedback loop in BCa [50]. In this research, we noted an up-regulation of PPARγ and down-regulation of phosphorylated and total SIRT1 in etomoxir-treated UM-UC-3 cell (Figure 4C). We also observed the change of FAO-related gene expression (Figure 5A). Fatty acid synthesis (FAS) was thought to be conflicting with FAO at the same time [51]. But some studies have proved that different isoforms of acetyl-CoA carboxylase, that was ACACA and ACACB, had different subcellular locations and yielded different outcomes of FAO and FAS [21]. It means that FAO and FAS may be active simultaneously and independently in cell and support each other. Simultaneous suppression of FAO and FAS was quite effective in anti-tumor therapy [18,52]. Our results showed a down-regulated expression of ACACA, which could synthesize malonyl-CoA as substrate for FAS with acetyl-CoA produced by FAO, and an up-regulated expression of ACACB, which could control FAO by producing malonyl-CoA and was reported to be regulated by SIRT1 [53]. Furthermore, after treatment of UM-UC-3 cells with etomoxir, the gene expression and protein level of PPARγ was up-regulated, and the PPARγ DNA-binding activity was also significantly increased. Consequently, we pretreated cells with PPARγ antagonist GW9662 and found that cell cycle arrest induced by etomoxir could be rescued by GW9662. Even though our previous study indicated the anti-tumor effect of GW9662 [54], when PPARγ was pre-inhibited by GW9662, BCa cell proliferation was inhibited to a less degree compared with etomoxir treatment alone (Supplementary Figure S2H-I). In conclusion, our results suggested that FAO inhibition by etomoxir could induce cell cycle arrest at G₀/G₁ phase via PPARγ-mediated pathway.

Multiple types of cancer have been reported to be of the ‘lipolytic phenotype’, which disobeys the Warburg effect [22,55]. However, the exact metabolic phenotype of BCa still needs to be further explored. Etomoxir was the most commonly used FAO inhibitor and well tolerated in a variety of scientific research, but its clinical trials showed a liver toxicity [27]. The high drug dose might contribute to the undesired side effects. But combined with glutaminolysis or glycolysis and chemotherapy drugs, etomoxir might have a great application prospect. Some other FAO inhibitors like ST1326, a CPT1A-specific inhibitor, also showed significant tumor growth suppression [16,56]. Our findings revealed a novel target and offered a new therapeutic avenue for the treatment of BCa.

• Cancer cells prefer aerobic glycolysis as the main energy source. Recent research highlighted the importance of lipid metabolism in tumor progression and alterations in fatty acid metabolism in BCa. However, there are few studies on fatty acid metabolism especially in BCa.

• Our main findings were that (1) fatty acid metabolism is activated in BCa compared with normal bladder; (2) etomoxir reduced FAO and motility of BCa cells and inhibited cell proliferation in vitro and in vivo; (3) the cell cycle arrest induced by etomoxir via PPAR-mediated pathway could be recovered by PPAR antagonist.

• Our results revealed that by inhibition of FAO with etomoxir could inhibit tumor cell proliferation and progression in BCa, which may provide a novel avenue to investigate new therapeutic approaches to human BCa.

We thank Ms. Yayun Fang, Ms. Yuan Zhu and Ms. Danni Shan for providing technical assistance in culturing cells. We also would like to acknowledge the KEGG database for free use.

As described by Cao et al. [30] in our group, the BCa and paracancerous tissues were obtained from patients after surgery at Zhongnan Hospital of Wuhan University, and normal bladder tissues were obtained from donors with accidental death. Tissue samples were collected and stored immediately in liquid nitrogen for the abovementioned experiments. Histological diagnosis of the bladder tissues was confirmed by two pathologists independently. The informed consent was obtained from all subjects and the study was conducted in accordance with the Declaration of Helsinki. The collection and use of human bladder tissues was approved by the Ethics Committee at Zhongnan Hospital of Wuhan University (approval number 2015029).

This work was supported in parts by the Science and Technology Department of Hubei Province Project [grant number 2018ACA159]; and the Wuhan Science and Technology Bureau Project (grant number 2018061005132294). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

The authors declare that there are no competing interests associated with the manuscript.

S.C., G.W. and X.W. were responsible for conceptualization. S.C. and L.C. were responsible for methodology. Y.W. was responsible for software. Y.X. and X.L. were responsible for validation. L.J. was responsible for formal analysis. K.Q. and L.J. were responsible for resources. G.W. and Y.W. were responsible for data curation. S.C. was responsible for writing-original draft preparation. G.W. and X.L. were responsible for writing-review and editing. X.W. was responsible for supervision. X.W. was responsible for project administration and funding acquisition.

ATP

adenosine triphosphate

BCa

bladder cancer

CPT1

carnitine palmitoyltransferase I

CPT2

carnitine palmitoyltransferase II

DEG

differentially expressed gene

EMT

epithelial–mesenchymal transition

FAO

fatty acid oxidation

FAS

fatty acid synthesis

FBS

fetal bovine serum

FFA

free fatty acid

GCBI

Gene-Cloud of Biotechnology Information

GO

gene ontology

H&E

Hematoxylin and Eosin

IHC

immunohistochemistry

KEGG

Kyoto Encyclopedia of Genes and Genomes

MIBC

muscle-invasive BCa

NADPH

nicotinamide adenine dinucleotide phosphate

NMIBC

non-muscle-invasive BCa

PFA

paraformaldehyde

PPAR

peroxisome proliferator-activated receptor

WB

Western blot