SKOV3 and CAVO3, two ovarian cancer cell lines, were purchased from the Cell Bank of Chinese Academy of Sciences (Shanghai, China) and maintained in RPMI 1640 medium (Invitrogen, USA) supplemented with 10% fetal bovine serum (BI). Normal ovarian cells, IOSE-80 (a gift from China Medical University), were cultured in Dulbecco's Modified Eagle's Medium (DMEM) (Invitrogen, USA) added with 10% BI. All cells were cultured in a humidified environment at 37°C, 5% CO2.

GPR12 overexpression vector was constructed by inserting GPR12 DNA fragment into pSin-EF2-Nanog-Pur plasmid to replace Nanog as previously described (14). Briefly, pSin-EF2-Nanog-Pur was cut by SpeI and EcoRI New England Biolabs (NEB) and become a linearized plasmid, followed by purification with an agarose gel DNA extraction kit (TaKaRa MiniBEST). Human GPR12 cDNA was obtained from Vigene Biosciences (Shandong, China). Then, clone PCR was performed to amplify and form double-strand DNA (Takara R045Q PrimeSTAR Max DNA Polymerase), which was purified by using an agarose gel DNA extraction kit (TaKaRa MiniBEST). The primer sequences used in the reactions for clone PCR were listed in Supplementary Table 1 . The 5′ terminals of this set of primers contained 16-nt sequence overlapping with the two terminals of the linearized pSin-EF2-Pur plasmid, respectively, and 5-nt sequence of corresponding restriction enzyme. Then, the DNA fragment was inserted into the linearized plasmid by Gibson Assembly (ClonExpress Ultra One Step Cloning Kit Vazyme C115-02). The empty vector pSin-EF2-Pur (a gift from Zhongshan medical university) was used as a control. The plasmids were verified by sequencing. The short hairpin RNA (shRNA) for human GPR12 was cloned into a hU6-MCS-CBh-gcGFP-IRES-puromycin lentiviral vector (GV493, Genechem, China). The primers used for clone reactions were listed in Supplementary Table 1 . Cell line stable expression of GPR12, shGPR12#1, and shGPR12#2 was constructed by filtered lentivirus infection using HEK293T cells (a gift from China Medical University) and selected and maintained with puromycin (0.5 mg/ml) for 14 days. The overexpression/knockdown efficiency of GPR12 was evaluated by Western blot.

The individual 369 EOC tissues were collected from patients during surgery from the Shengjing Hospital of the China Medical University between January 2011 and December 2019. All patients have been diagnosed on the basis of clinical and pathological evidence, and the tissues were collected and stored at −80°C. Patient consent and approval from the Institute Research Ethics Committee were obtained prior to the use of these clinical materials for research purposes.

To determine the effect of GPR12 on cell viability of SKOV3 and CAVO3, CCK-8 assay (APExBIO, USA) was carried out. Cells were seeded into 96-well plates at the density of 5 × 103 cells per well, and the performance was followed to the protocol. Optical density values were read at 450 nm by a microplate reader (Bio-Rad, USA).

Protein extraction was performed by using the radioimmunoprecipitation assay buffer supplemented with phosphatase inhibitor cocktails. Equal volumes of protein lysates were separated using 10% (vol/vol) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). For immunoblotting, antibodies against G protein-coupled receptors 12 (GPR12) (Abcam, USA), poly (ADP-ribose) polymerase (PARP), caspase-3, Cleaved caspase-3, B cell lymphoma-2 (BCL2), BCL-2-Associated X Protein (BAX), phosphorylation of extracellular-regulated protein kinases 1/2 (p-ERK1/2), total extracellular-regulated protein kinases 1/2 (t-ERK1/2) (Cell Signaling Technology Cambridge, MA, USA), α-tubulin, and β-actin (ProteinTech, China) were used.

Immunohistochemistry (IHC) staining was carried out with anti-GPR12 antibody according to the standard protocol. Two independent investigators evaluated the scores of the proportion of tumor cells according to the following criteria: 0 (no positive cells), 1 (positive cells, <10%), 2 (positive cells, 10%–35%), 3 (positive cells, 35%–70%), and 4 (positive cells, >70%). The SI represents staining intensity. Revised version: The staining intensity (SI) was graded as follows: 0 (no staining), 1 (light yellow, faint staining), 2 (yellow brown, moderate staining), and 3 (brown, strong staining). The final immunohistochemical score staining index was the product of the score of SI and the score of the proportion of positive cells (range from 0 to 12).

Cell apoptosis was determined with a commercial Annexin V–Fluorescein isothiocyanate (FITC) Apoptosis Detection Kit (Beyorime, China). Briefly, cells were harvested with trypsin, and 1 × 106 cells were resuspended in 100 μl of the binding buffer with 5 μl of Annexin V–FITC and 5 µl PI. Subsequently the cells were incubated at room temperature for 15 min in the dark and then were assayed by a FACScan flow cytometer with CellQuest Pro software.

Female 6-week-old BALB/c-nu mice were randomly assigned to two groups (six mice per group). SKOV3 cells (1 × 106) were subcutaneously implanted into the inguinal folds of each nude mouse. Tumor height and width were measured with a caliper every 2 days to calculate tumor volume (= width2 × height × π/6). For subcutaneous tumor growth, the maximum single tumor cannot exceed 1.2 cm in diameter in mice, and no experiments in this study generated tumor burden over this limit. Tumor tissues were collected from the nude mice, fixed in 4% paraformaldehyde, and embedded in paraffin.

All the data were analyzed using GraphPad 5.0 software (USA). The results are presented as means ± standard deviation (SD). Student’s t-test was used to was used to make comparisons between two groups. One-way ANOVA was performed to make comparisons among multiple testing. Chi-square test was applied to detect the relationship between GPR12 expression and clinicopathological characteristics. Kaplan–Meier survival curves were analyzed by log-rank test. Multivariable cox regression analysis was used to determine independent risk factors of EOC. A value of P < 0.05 was considered statistically significant.

To investigate whether the prognosis of EOC is associated with GPR12 expression, we first performed Kaplan–Meier survival analysis through Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia.cancer-pku.cn/) by using transcriptome and survival data from The Cancer Genome Atlas (TCGA) dataset. Results showed a significant reduction in overall survival (OS) rate in patients with EOC with high GPR12 mRNA expression than those with low GPR12 mRNA expression ( Figure 1A ). In addition, we also performed a Cox regression analysis by using EOC data from TCGA database. GPR12 mRNA expression and clinical characteristics including age, stage, and grade were selected into the multivariable Cox regression model. Results showed that GPR12 expression (p = 0.017, hazard ratio = 1.1), and age (p < 0.01, HR = 1) were significantly associated with EOC prognosis ( Figure 1B ). Then, a total of 369 patients with EOC were recruited in our study. The GPR12 protein expressions in low-grade serous ovarian cancer (LGSOC), high-grade serous ovarian cancer (HGSOC), mucinous adenocarcinoma (MC), endometrioid adenocarcinoma (EC), clear cell adenocarcinoma (CCA), and normal ovarian epithelial tissues were examined by immunohistochemical staining, respectively. Significantly increased GPR12 expression was observed in all histological subtypes of EOC tissues compared with the normal ovarian epithelial tissues ( Figures 1C, D ). All patients with EOC composed of HGSOC patients and other patients with EOC, were further divided into GPR12 low expression group and GPR12 high expression group based on the immunohistochemical results, and clinicopathological characteristics of patients were assessed, respectively. As shown in Table 1 , the patients with high GPR12 expression have higher proportion of the positive in progression status than the patients with low GPR12 expression. In addition, the patients with high GPR12 expression have lower sensitivity to first chemotherapy than the patients with low GPR12 expression. Furthermore, progression-free survival (PFS) of all patients with EOC, patients with HGSOC, and the patients with other types of EOC was assessed, separately. Our results showed that PFS was significantly decreased in all patients with EOC, HGSOC, or the other types of EOC with high GPR12 levels than that in the patients with low GPR12 levels, respectively ( Figures 1E–G ). These results suggest that GPR12 may play a critical role in the progression of EOC and serves as a biomarker for prognosis in EOC.

To investigate the role of GPR12 in EOC, SKOV3, and CAOV3 cell lines was selected for the study. First, we examined the protein level of GPR12 in normal ovarian epithelial cell line (IOSE-80) and two EOC cell lines (SKOV3 and CAOV3) by using Western blot. Results showed that the protein expression of GPR12 was significantly increased in SKOV3 and CAOV3 cells compared with IOSE-80 cells ( Figure 2A ). Two different small hairpin RNA (shRNA) (sh-GPR12-1 and sh-GPR12-2) were used to knock down the expression of GPR12 in SKOV3 and CAOV3 cells ( Figure 2B ). CCK-8 assay was used to determine the cell viabilities of SKOV3 and CAOV3 cells transfected with GPR12 shRNAs or empty vector (scramble). After 24 h, there was no significant change in cell viability of SKOV3 and CAOV3 cells transfected with GPR12 shRNAs than that in the cells transfected with scramble sequences. However, 48 and 72 h after the transfection, GPR12 downregulation significantly decreased cell viability in both SKOV3 and CAOV3 cells compared with the scramble group, respectively ( Figures 2C, D ). Furthermore, we overexpressed GPR12 in SKOV3 and CAOV3 cell lines ( Supplementary Figure 1A ) and performed CCK8 assays to evaluate cell viability. Results showed that GPR12 overexpression promoted the viability and proliferation of SKOV3 and CAOV3 cell lines compared with control group ( Supplementary Figures 1B, C ).

Then, we further assessed the effect of GRP12 on apoptosis of SKOV3 and CAOV3 cells using flow cytometric analysis. Results showed that GPR12 downregulation significantly increased Annexin V–positive cell population in both SKOV3 and CAOV3 cells compared with scramble group, respectively ( Figures 2E, F ), whereas GPR12 overexpression exerted opposite effects in both SKOV3 and CAOV3 cells ( Supplementary Figures 1D, E ). Western blot results showed that GPR12 knockdown significantly increased the protein levels of BAX, cleaved caspase-3, and cleaved PARP and decreased the expression of BCL-2 in both SKOV3 and CAOV3 cells compared with scramble group, respectively. Meanwhile, GPR12 knockdown significantly decreased the protein level of PARP in CAOV3 cells. However, no significant change in caspase3 protein level was observed in both SKOV3 and CAOV3 cells compared with scramble group ( Figures 2G, H ). These results suggested that GPR12 knockdown inhibits EOC cell proliferation by inducing cellular apoptosis. In GPR12-overexpressed SKOV3 and CAOV3 cells, significantly decreased protein levels of cleaved PARP and cleaved caspase-3 and significantly increased BCL-2 protein expression were observed ( Supplementary Figure 1F ). Although there was no change in BAX protein level in SKOV3 and CAOV3 cells compared with control group, the protein expression ratio of BCL-2 to BAX, a critical index for evaluating endogenous cell apoptosis (15), was significantly increased in SKOV3 and CAOV3 cells. In addition, no significant change in the protein levels of PARP and caspase3 was observed in both SKOV3 and CAOV3 cells compared with control group ( Supplementary Figure 1F ). These results suggest that GPR12 overexpression could inhibit the apoptosis of SKOV3 and CAOV3 cells.

To investigate the potential underlying mechanisms by which GPR12 knockdown enhanced EOC cell apoptosis, bioinformatic analysis was carried out to explore and reveal the possible molecular mechanism. First, differential expressed genes (DEGs) between the high GPR12 expression and the low GPR12 expression were screened using the R software with the “DEGSeq2” package. There were 548 DEGs, including 428 downregulated and 120 upregulated genes (adjusted P < 0.05 and |logFC| ≥ 1), as shown in Figure 3A . Then, gene ontology (GO) and pathway enrichment analysis was conducted using Metascape online tool to explore the functional characteristics of the DEGs. The analysis results showed that upregulated DEGs were significantly enriched in “fatty acids”, “antimicrobial humoral immune response mediated by antimicrobial peptide”, “NABA MATRISOME ASSOCIATED”, “positive regulation of monocyte chemotaxis”, and “ERK1 and ERK2 cascade” ( Figure 3B ). For downregulated DEGs, significant enrichments were identified in “pattern specification process”, “embryonic morphogenesis”, “appendage morphogenesis”, “endocrine system development”, and “NABA MATRISOME ASSOCIATED” ( Figure 3B ).

On the basis of the computing method of enrichment analysis, changes in expression level of key genes contributing to enrichment in the gene set of ERK1 and ERK2 cascade may be the protein kinase cascade or downstream genes. We further performed correlation analysis of GPR12 mRNA level with ERK1/2 cascade expression including MAP3K1, MAP2K1, MAP2K2, and MAPK1 by using ovarian cancer data from TCGA database. Results showed that no significant correlation was observed between GPR12 level and ERK1/2 cascade expression ( Supplementary Figure 3A ). Accumulating research studies have shown that the phosphorylation of ERK1/2 is a key effector to exert an influence on cell proliferation and death by translocating into nucleus (16). More importantly, GPR12 has been reported to promote HEK 293T proliferation and induce neurite outgrowth in PC12 cells via the activation of ERK1/2 (8, 12). Therefore, to investigate whether the EOC cell apoptosis induced by GPR12 knockdown was mediated by ERK1/2 pathway, we measured the level of pERK1/2 by using Western blot. Results showed that GPR12 knockdown significantly reduced the phosphorylation levels of ERK1/2 in both SKOV3 and CAOV3 cells compared with scramble group ( Figure 3C ), whereas ERK1/2 phosphorylation was significantly increased in the GPR12-overexpressed SKOV3 and CAOV3 cells ( Supplementary Figure 1G ). To further investigate the role of ERK1/2 pathway in EOC cell apoptosis induced by GPR12 knockdown, both SKOV3 and CAOV3 cells were pretreated with LM22B-10, the selective activator of ERK1/2. As shown in Figure 3D , the decreased cellular viability induced by GPR12 knockdown was significantly attenuated by pretreatment with LM22B-10 in both SKOV3 and CAOV3 cells. In addition, the enhanced cell apoptosis induced by GPR12 knockdown was significantly inhibited by pre-incubation with LM22B-10 in both SKOV3 and CAOV3 cell lines ( Figure 3E ). Western blot results also proved that LM22B-10 can rescue the upregulated expressions of BAX and the cleaved caspase 3 in the cells of GPR12 knockdown ( Figure 3F ).

All these results proved that ERK1/2 pathway is involved in EOC cell apoptosis regulated by GPR12.

To further clarify the effects of GPR12 on tumor progression in vivo, a subcutaneous tumor xenograft model in nude mice was established and TUNEL assay was performed to evaluate cancer cell apoptosis and growth in vivo. About 5 × 105 SKOV3 cells transfected with GPR12-shRNAs or scramble vector were subcutaneously implanted into the inguinal folds of nude mice. Around 14 days later, tumor volumes and weight were measured and recorded when the average tumor volume reached around 100 mm3. As shown in Figures 4B, C knockdown of GRP12 significantly decreased tumor volumes and tumor weight in nude mice compared with those from scramble group. The representative images of individual tumors are shown Figure 4A . In addition, Hematoxylin-Eosin (HE) staining and TdT-mediated dUTP-biotin nick end labeling (TUNEL) staining showed that GPR12 knockdown significantly increased the tumor necrotic area ( Figure 4D ) and the number of apoptotic cells in the tumor tissues of mice compared with scramble group, respectively ( Figure 4E ). Western bolt results showed that cleaved caspase3 was significantly increased in tumor lysates of GPR12 knockdown group compared with scramble ( Figure 4F ), which indicated that GPR12 knockdown may lead to the tumor regression of EOC by inducing apoptosis. Analysis of tumor lysates from scramble and GPR12 knockdown mice showed a significant decrease in phosphorylated ERK1/2 protein levels in GPR12 knockdown tumors ( Figure 4G ).

Results of the in vivo and in vitro studies demonstrate that GPR12 knockdown contributes to the cell apoptosis in EOC cells via ERK1/2 pathway.