Mechanism of RBM15 in Regulating PD-L1-Mediated Immune Escape in Ovarian Cancer Through the JAK2/STAT3/STAT5 Pathway

All animal experimental protocols were approved by the Ethics Committee of Affiliated Hospital of North Sichuan Medical College, and the experimental operations were carried out in accordance with the Guidelines for the Care and Use of Laboratory Animals.

The human ovarian epithelial cell line (IOSE80) and human OC cell lines (SKOV3, A2780, OVCAR8, OVCAR3) were sourced from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The mouse OC cell line (ID8) was purchased from Shanghai Jihe Biotechnology Co., Ltd. (Shanghai, China) CD8⁺ T cells were purchased from the American Type Culture Collection (Manassas, VA, USA). Human-derived cells (cultured in RPMI-1640 medium, Gibco, Grand Island, NY, USA) and ID8 and CD8⁺ T cells (cultured in Dulbecco's Modified Eagle's Medium) were incubated at 37°C in a 5% CO₂ incubator. All culture media contained10% fetal bovine serum (FBS) and 1% 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco).

Small interfering RNAs targeting RBM15 (si-RBM15) and EIF4A3 (si-EIF4A3) were designed and synthesized by RiboBio (Guangzhou, China). The pcDNA3.1 vector targeting circFGFR3 (overexpressed-circFGFR3, oe-circFGFR3) was synthesized by GenePharma (Shanghai, China). OC cells were transfected employing Lipofectamine 3000 transfection reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Cells were harvested 48 h post-transfection for downstream experiments.

RBM15 shRNA (sh-RBM15) and negative control shRNA (sh-NC) were synthesized by Genewiz (South Plainfield, NJ, USA) and cloned into lentiviral shRNA-overexpression plasmids (Hanbio Biotechnology Co., Ltd., Shanghai, China). ID8 cells were seeded in 12-well plates, infected with the supernatants of LV-sh-RBM15 or LV-sh-NC at 37°C for 48 h, and then used for subsequent experiments.

CCK-8 method was used to measure cell viability (Yeasen, Shanghai, China). OC cells and CD8⁺ T cells were seeded in 96-well plates (1000 cells per well). At the specified time points, 10 μL of CCK-8 reagent was added to each well for a 3-h incubation. The absorbance at 450 nm was measured using an Infinite M200 (Tecan, Switzerland).

Cells were seeded into 6-well plates (500 cells per well) and cultured for approximately 2 weeks. When the cell colonies reached the optimal confluence, they were fixed with 4% paraformaldehyde at room temperature for 30 min. The colonies were then stained with a 1% crystal violet solution (Sigma, St. Louis, MO, USA) at room temperature for 30 min. The cell colonies were counted using ImageJ software (version 1.46r; National Institutes of Health, Bethesda, MD, USA). The images are shown in supplementary figures.

Transwell inserts (with an 8 μm pore size, Corning, NY, USA) were used to measure the migration and invasion of OC cells. Subsequently, 600 μL of medium containing 20% FBS was added to the lower chamber of the Transwell, either without or with a reconstituted polycarbonate membrane (coated with Matrigel). Afterwards, 5 × 10⁵ OC cells were seeded into the upper chamber and cultured at 37°C with 5% CO₂ for 24 h. The migrated or invaded cells were fixed with 4% paraformaldehyde for 30 min, then stained with 0.1% crystal violet. Photographs were taken, and the cells were quantified under an inverted microscope (IX73, Olympus, Tokyo, Japan). The images are shown in supplementary figures.

OC cells were seeded into 6-well plates at a density of 2 × 10⁵ cells per well. CD8⁺ T cells were co-cultured with OC cells for 24 h. A single-cell suspension was generated and incubated with a CD8 fluorescent antibody (PE; ab28017, Abcam, Cambridge, MA, USA) at 4°C in the dark for 45 min. Apoptosis of CD8⁺ T cells was detected by Annexin-V/PI staining and analyzed using CytExpert software (Beckman Coulter, Miami, FL, USA). The images are shown in supplementary figures.

The cell mixtures and the digested tumor tissues were centrifuged at 1500 g for 5 min, and the supernatants were collected. The concentrations of interferon (IFN)-γ (ab46025/ab282874), interleukin (IL)-2 (ab100566/ab46096), and tumor necrosis factor (TNF)-α (ab181421/ab46105) were measured using ELISA kits (Abcam) according to the manufacturer's instructions.

Total RNA (1 μg) from OC cells was treated with 3 U of RNase R (Epicentre Technologies, Madison, WI, USA) for 15 min, and the products were analyzed by real-time quantitative polymerase chain reaction (RT-qPCR). OC cells were incubated with 2 μg/mL actinomycin D (Sigma), and then detected at different time points (0 h, 4 h, 8 h, 12 h, or 24 h). Data are shown in supplementary figures.

The m6A RNA methylation quantification kit (Abcam) was used to determine the RNA m6A levels in OC cells. First, the RNA was incubated in the experimental wells at 37°C for 90 min. Then, the capture antibody, detection antibody, and enhancer solution were added. After adding the colorimetric chromogenic solution, the absorbance at 450 nm wavelength in each well was measured using a microplate reader to determine the RNA m6A levels.

RIP was performed utilizing the Magna RIP RNA-binding protein immunoprecipitation (IP) kit (Millipore, Cambridge, MA, USA) following the manufacturer's protocol. OC cells were collected and lysed in IP lysis buffer, and then mechanically sheared using a homogenizer. Antibodies against the specified proteins (EIF4A3: ab180573, Abcam and m6A: ab208577; Abcam) and the negative control immunoglobulin G (ab170190; Abcam) were incubated with the cell lysates overnight at 4°C. After washing three times, streptavidin-coated magnetic beads were added and incubated for 2 h. The extracted RNA was detected and analyzed by RT-qPCR.

The expression of RNA in OC cells was detected using 50 nM biotin-labeled circFGFR3. The cells were harvested and washed with phosphate-buffered saline (PBS) and incubated in lysis buffer for 10 min. Then, streptavidin-coated agarose beads (Invitrogen) were added and further incubated at room temperature for 1 h. After washing with PBS, the proteins were collected and subjected to Western blot analysis.

Female C57BL/6 mice (4 weeks old, 20–22 g weight) were housed with free access to food and water, a constant humidity of 45%–50%, and a constant temperature of 25–27°C. ID8 cells (5 × 10⁶) were subcutaneously injected into the right side of C57BL/6 mice after being mixed with Matrigel and PBS at a ratio of 1:1. When the tumors grew to be palpable (approximately 50 mm³), the tumor volume was evaluated every 5 days. Fifty days later, the mice were euthanized by intraperitoneal injection of 150 mg/kg sodium pentobarbital. The tumors were harvested, weighed, and subjected to further examination.

Tissue sections were blocked with 5% bovine serum albumin (BSA) and incubated with primary antibodies against Ki67 (ab15580; Abcam), PD-L1 (ab213480; Abcam), and CD8 (ab217344; Abcam) at room temperature, followed by incubation with the secondary antibody (ab205718; Abcam) for 1 h at room temperature and staining with diaminobenzidine. The sections were counterstained with Mayer's hematoxylin, dehydrated, cleared, and mounted. Each section was observed under an optical microscope (Zeiss, Shanghai, China). The results were independently analyzed by two experienced pathologists (double-blind manner).

The m6A modification status of circFGFR3 was predicted according to the sequence-based RNA adenosine methylation site predictor (SRAMP) database (http://www.cuilab.cn/m6asiteapp/old) (Zhou et al. 2016). RNA binding proteins that bind to circFGFR3, JAK2, STAT3, and STAT5 were predicted according to the Starbase database (http://starbase.sysu.edu.cn/index.php) (Li et al. 2014) and the CircInteractome database (https://circinteractome.nia.nih.gov/mirna_target_sites.html) (Dudekula et al. 2016), and the intersection was taken. Results of the prediction are shown in supplementary figures.

Total RNA was extracted according to the instructions provided by the Trizol kit (Invitrogen). Then, the extracted RNA was reverse-transcribed into complementary DNA (cDNA) according to the manufacturer's instructions of the Prime Script RT kit (Takara, Dalian, China). The obtained cDNA was stored at −80°C for subsequent experimental use. Real-time PCR analysis was performed using SYBR Premix Ex Taq II (Takara) on an ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal reference. The relative expression levels were calculated by the 2^−ΔΔCt method (Livak and Schmittgen 2001). The primers are shown in Table 1. Some data are shown in supplementary figures.

Proteins were extracted using radioimmunoprecipitation assay lysis buffer (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) containing phenylmethylsulfonyl fluoride, and then the protein concentration was measured using a bicinchoninic acid kit (Wuhan Boster Biological Technology Co., Ltd., Wuhan, Hubei, China). Subsequently, the proteins were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto a polyvinylidene fluoride membrane (Sigma-Aldrich). Then, the membrane was blocked with 5% BSA and incubated at room temperature for 1 h. After that, the membrane was incubated with diluted primary antibodies overnight at 4°C as follows: RBM15 (1:1000, ab96544), PD-L1 (1:1000, ab228415/ab213480), EIF4A3 (1:1000, ab180573), JAK2 (1:5000, ab108596), STAT3 (1:1000, ab68153), STAT5 (1:1000, ab230670), p-JAK2 (1:1000, ab32101), p-STAT3 (1:2000, ab76315), p-STAT5 (1:1000, ab278764), and GAPDH (1:2500, ab9485). All the above antibodies were purchased from Abcam. The next day, the membrane was washed three times with Tris-buffered saline with Tween-20 (TBST) (5 min each time), and then incubated with the corresponding secondary antibody (1:2000, ab205718/ab205719; Abcam) at room temperature for 1 h. After rinsing three times in TBST (5 min each time), the immunocomplexes on the membrane were visualized using enhanced chemiluminescence reagents and a gel imager (Gel Doc EZ Imager; Bio-Rad Laboratories, Hercules, CA, USA). GAPDH was used as an internal reference, and the relative expression of the protein was presented as the ratio of the gray value of the target protein to that of the internal reference. The gray value intensity of the target protein bands was analyzed using ImageJ software. Some images are shown in supplementary figures.

All data were statistically analyzed and plotted using SPSS 21.0 statistical software (IBM SPSS Statistics, Chicago, IL, USA) and GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA). Initially, normality and homogeneity of variance tests were performed, and the data were verified to be normally distributed with homogeneous variances. One-way or two-way analysis of variance (ANOVA) was used for data comparison among multiple groups, and Tukey's multiple comparisons test was applied for post hoc testing. p < 0.05 was considered to indicate a statistically significant difference.

To elucidate the impact of RBM15 on the progression of OC cells, RBM15 expression in cultured cells was detected. RBM15 was highly expressed in OC cell lines (p < 0.05; Figures 1a,b). Subsequently, siRBM15 was transfected into A2780 and OVCAR8 cells with relatively high RBM15 expression (p < 0.01; Figures 1c,d), and two siRNAs with better transfection efficiency were selected for subsequent detection. After the downregulation of RBM15, the proliferation and colony formation of the two OC cell lines were decreased (p < 0.01; Figures 1e,f; Figure 1a in Supplementary Material). Compared with the small interfering RNA negative control (si-NC) group, migrated and invaded OC cells in the si-RBM15 group were reduced (p < 0.01; Figure 1g; Figure 1b in Supplementary Material). In conclusion, RBM15 is elevated in OC cell lines, while silencing RBM15 inhibits the proliferation, migration, and invasion of OC cells.

Fig 1.

RBM15 is upregulated in OC cells and promotes cell proliferation, migration, and invasion. (A,B): The expression of RBM15 in each cell was detected by RT-qPCR and Western blot. si-RBM15 was transfected into A2780 and OVCAR8 cells, with si-NC transfection used as a negative control. (C,D): The expression of RBM15 in cells was detected by RT-qPCR and Western blot. (E,F): Cell proliferation was detected by CCK-8 and colony formation assays. (G) Cell migration and invasion were detected by Transwell. Three independent repeated tests were carried out, and the data were expressed as mean ± standard deviation. One-way ANOVA was used for data comparison among multiple groups in panels (A,B), and two-way ANOVA was used for data comparison among multiple groups in panels (C–G). Tukey's multiple comparisons test was used for post hoc test. *p < 0.05, and **p < 0.01. ANOVA, analysis of variance; CCK-8, cell counting kit-8; OC, ovarian cancer; RBM15, RNA binding motif protein 15; RT-qPCR, real-time quantitative polymerase chain reaction; si-RBM15, small interfering RNAs targeting RBM15.

The impact of RBM15 on the immune escape of OC cells remains unclear. Compared with the si-NC group, PD-L1 expression in the si-RBM15 group was remarkably decreased (p < 0.01; Figure 2a). Subsequently, we co-cultured the si-RBM15-transfected OC cells with CD8⁺ T cells. After RBM15 downregulation, CD8⁺ T cell activity was increased, apoptosis was decreased (p < 0.05; Figure 2b,c; Figure 1c in Supplementary Material), and the concentrations of IFN-γ, IL-2, and TNF-α in the supernatant were upregulated (p < 0.01; Figure 2d). These results suggest that RBM15 downregulation inhibits PD-L1-mediated immune escape of OC cells.

Fig 2.

Downregulation of RBM15 inhibits PD-L1-mediated immune escape of OC cells. (A) The expression of PD-L1 in cells was detected by Western blot. A2780 and OVCAR8 cells in each group were co-cultured with CD8⁺ T cells. (B) CD8⁺ T cell viability was determined by the CCK-8 method. (C) Apoptosis of CD8⁺ T cells was detected by flow cytometry. (D) The secretion of cytokines was detected by the ELISA method. Three independent repeated tests were carried out, and the data were expressed as mean ± standard deviation. Two-way ANOVA was used for data comparison among multiple groups in panels (A–D), and Tukey's multiple comparisons test was used for post hoc test. *p < 0.05, and **p < 0.01. ANOVA, analysis of variance; CCK-8, cell counting kit-8; ELISA, enzyme-linked immunosorbent assay; OC, ovarian cancer; PD-L1, programmed death-ligand 1; RBM15, RNA binding motif protein 15.

The database prediction indicated that circFGFR3 has m6A modification (Figure 2a in Supplementary Material). CircFGFR3 is upregulated in OC (Zhou et al. 2020). However, the relationship between circFGFR3 and RBM15 remains unclear. Firstly, circFGFR3 was upregulated in OC cell lines (p < 0.01; Figure 3a), and further experiments verified the stability of circFGFR3 in OC cell lines (p < 0.01; Figure 2b,c in Supplementary Material). Additionally, the m6A level was decreased with the reduction of RBM15 expression (p < 0.01; Figure 3b). MeRIP showed that after the downregulation of RBM15, the m6A enrichment on circFGFR3 was decreased (p < 0.01; Figure 3c). Moreover, compared with the si-NC group, circFGFR3 expression in the si-RBM15 group was decreased (p < 0.01; Figure 3d). In conclusion, RBM15 upregulates circFGFR3 expression through m6A modification.

Fig 3.

RBM15 upregulates circFGFR3 expression through m6A modification. (A) The expression of circFGFR3 in each cell was detected by RT-qPCR. (B) The m6A level in cells was analyzed by m6A quantification. (C) The m6A enrichment of circFGFR3 in cells was analyzed by MeRIP. (D) The expression of circFGFR3 in cells was detected by RT-qPCR. Three independent repeated tests were carried out, and the data were expressed as mean ± standard deviation. One-way ANOVA was used for data comparison among multiple groups in panel (A), and two-way ANOVA was used for data comparison among multiple groups in panels (B–D). Tukey's multiple comparisons test was used for post hoc test. ^**p < 0.01. ANOVA, analysis of variance; circFGFR3, circRNA fibroblast growth factor receptor 3; MeRIP, methylated RNA immunoprecipitation; RBM15, RNA binding motif protein 15; RT-qPCR, real-time quantitative polymerase chain reaction.

Subsequently, circFGFR3 expression was upregulated in A2780 cells (p < 0.01; Figure 4a) and combined with si-RBM15-2. After overexpressing circFGFR3, the cell proliferation and colony formation were increased (p < 0.01; Figure 4b,c; Figure 2d in Supplementary Material), and migrated and invaded cells were elevated (p < 0.01; Figure 4d; Figure 2e in Supplementary Material). In addition, circFGFR3 overexpression upregulated PD-L1 expression (p < 0.01; Figure 4e). The A2780 cells with combined treatment were co-cultured with CD8⁺ T cells. The results showed that the activity of CD8⁺ T cells was reduced and apoptosis was enhanced (p < 0.05; Figure 4f,g; Figure 2f in Supplementary Material). ELISA demonstrated that overexpression of circFGFR3 decreased the concentrations of IFN-γ, IL-2, and TNF-α (p < 0.01; Figure 4h). The above results indicate that overexpression of circFGFR3 promotes immune escape and attenuates the inhibitory effect of RBM15 knockdown on the progression of OC cells.

Fig 4.

CircFGFR3 overexpression promotes immune escape and alleviates the inhibitory effect of RBM15 downregulation on OC cell progression. oe-circFGFR3 was transfected into A2780 cells, with over-expression negative control (oe-NC) transfection used as a negative control. (A) The transfection efficiency of circFGFR3 in cells was detected by RT-qPCR. (B,C) Cell proliferation was detected by CCK-8 and colony formation assays. (D) Cell migration and invasion were detected by Transwell. (E) The expression of PD-L1 in cells was detected by Western blot. A2780 cells in each group were co-cultured with CD8⁺ T cells. (F) The viability of CD8⁺ T cells was determined by the CCK-8 method. (G) Apoptosis of CD8⁺ T cells was detected by flow cytometry. (H) The secretion of cytokines was detected by the ELISA method. Three independent repeated tests were carried out, and the data were expressed as mean ± standard deviation. The t-test was used for data comparison between two groups in panel (A). One-way ANOVA was used for data comparison among multiple groups in panels (C–G), and two-way ANOVA was used for data comparison among multiple groups in panels (B) and (H). Tukey's multiple comparisons test was used for post hoc test. *p < 0.05, and **p < 0.01. ANOVA, analysis of variance; CCK-8, cell counting kit-8; circFGFR3; circRNA fibroblast growth factor receptor 3; ELISA, enzyme-linked immunosorbent assay; OC, ovarian cancer; PD-L1, programmed death-ligand 1; RBM15, RNA binding motif protein 15; RT-qPCR, real-time quantitative polymerase chain reaction.

RNA-binding proteins that could bind to circFGFR3, JAK2, STAT3, and STAT5 (proteins related to the JAK/STAT pathway) were predicted through the Starbase and CircInteractome databases. After taking the intersection, we obtained EIF4A3 (Figure 3a in Supplementary Material). RIP and RNA pull-down showed that circFGFR3 could bind to EIF4A3 (p < 0.01; Figures 5a,b). The RIP results demonstrated that circFGFR3 could bind to JAK2, STAT3, and STAT5 (p < 0.01; Figure 5c). In addition, circFGFR3 did not affect the expression of EIF4A3 (p > 0.05; Figure 5d; Figure 3b in Supplementary Material). Overexpression of circFGFR3 stabilized the mRNAs of JAK2, STAT3, and STAT5 (p < 0.01; Figure 5e), but after downregulating EIF4A3 expression (p < 0.01; Figure 5d; Figure 3b in Supplementary Material), their messenger RNA (mRNA) stabilities were decreased (p < 0.05; Figure 5e). The expression of JAK2, STAT3, and STAT5 was increased after circFGFR3 upregulation, but was decreased after EIF4A3 downregulation (p < 0.01; Figures 5d,f). Moreover, after downregulating RBM15, the levels of p-JAK2/JAK2, p-STAT3/STAT3, and p-STAT5/STAT5 were decreased, which were reversed upon circFGFR3 overexpression (p < 0.05; Figure 3c in Supplementary Material). In conclusion, RBM15 upregulates circFGFR3 to activate the JAK2/STAT3/STAT5 signaling pathway.

Fig 5.

RBM15 upregulates circFGFR3 to activate the JAK2/STAT3/STAT5 signaling pathway. (A,B) The binding of circFGFR3 and EIF4A3 was analyzed by RIP or RNA pull-down assay. (C) The binding of EIF4A3 and proteins related to the JAK/STAT pathway was analyzed by RIP. (D) The expression of EIF4A3 and proteins related to the JAK/STAT pathway in cells was detected by Western blot. (E) After actinomycin D treatment, the mRNA stability of proteins related to the JAK/STAT pathway was detected by RT-qPCR. (F) The expression of proteins related to the JAK/STAT pathway was detected by RT-qPCR. Three independent repeated tests were carried out, and the data were expressed as mean ± standard deviation. Two-way ANOVA was used for data comparison among multiple groups in panels (A,C,D–F). Tukey's multiple comparisons test was used for post hoc test. *p < 0.05, and **p < 0.01. ANOVA, analysis of variance; circFGFR3, circRNA fibroblast growth factor receptor 3; JAK/STAT, Janus kinase-signal transducer and activator of transcription; RBM15, RNA binding motif protein 15; RIP, RNA immunoprecipitation; RT-qPCR, real-time quantitative polymerase chain reaction.

Finally, we established a mouse xenograft tumor model using ID8 cells. RBM15 downregulation reduced the volume and weight of the tumors (p < 0.01; Figures 6a,b). Compared with the sh-NC group, the positive rates of Ki67 and PD-L1 in the sh-RBM15 group were decreased, while the positive rate of CD8 was increased (p < 0.01; Figures 6c,d). ELISA revealed that after downregulating RBM15, the concentrations of IFN-γ, IL-2, and TNF-α were elevated (p < 0.01; Figure 6e). Silencing RBM15 led to the reduction of circFGFR3, p-JAK2/JAK2, p-STAT3/STAT3, and p-STAT5/STAT5 (p < 0.05; Figures 6f,g). These results indicate that RBM15 downregulation suppresses the JAK/STAT signaling pathway by downregulating circFGFR3, thereby inhibiting PD-L1-mediated immune escape and reducing the proliferation of OC cells.

Fig 6.

RBM15 activates the JAK/STAT signaling pathway via circFGFR3 to promote PD-L1-mediated immune escape in OC. (A) The volume of the tumor was recorded every 5 days. (B) After euthanizing the nude mice on the 50th day, the tumor tissues were collected, weighed, and representative photos were taken. (C,D) The positive rates of Ki67, PD-L1 and CD8 in the tissues were detected by immunohistochemistry. (E) The secretion of cytokines was detected by ELISA. (F) The expression of RBM15 and circFGFR3 was detected by RT-qPCR. (G) The expression of RBM15 and proteins related to the JAK/STAT pathway was detected by Western blot. N = 6, and the data were expressed as mean ± standard deviation. The t-test was used for data comparison between two groups in panel (B). Two-way ANOVA was used for data comparison among multiple groups in panels (A,D–G). Tukey's multiple comparisons test was used for post hoc test. *p < 0.05, and **p < 0.01. ANOVA, analysis of variance; circFGFR3, circRNA fibroblast growth factor receptor 3; ELISA, enzyme-linked immunosorbent assay; JAK/STAT, Janus kinase-signal transducer and activator of transcription; OC, ovarian cancer; PD-L1, programmed death-ligand 1; RBM15, RNA binding motif protein 15; RT-qPCR, real-time quantitative polymerase chain reaction.