MicroRNA-200a induces immunosuppression by promoting PTEN-mediated PD-L1 upregulation in osteosarcoma

miRNA-200a promotes doxorubicin-induced PD-L1 upregulation in osteosarcoma cells

To screen for the miRNAs involved in the regulation of PD-L1 in osteosarcoma, we performed miRNA-sequencing (miR-seq) on U2OSR2 and U2OS cells (Figure 1A, 1B). We initially identified 36 differentially expressed miRNAs using a 4-fold change cutoff and P<0.05 (Supplementary Table 1). Then, 10 miRNAs were selected from the differentially expressed miRNAs based on P values for additional detection in doxorubicin-treated osteosarcoma cells (1.6 μM for 24h) (Table 1, Figure 1C). Among them, the change of miR-200a was the most obvious, and we carried out further experiments especially for miR-200a. We observed a dose-dependent increase in the expression of PD-L1 (Figure1D–1I) and miR-200a (Figure 1J, 1K) in doxorubicin-treated U2OS and K7 cells. These results suggest that miR-200a regulates PD-L1expression in doxorubicin-treated osteosarcoma cells.

Figure 1. Doxorubicin-induced the simultaneous upregulation of PD-L1 and miR-200a in osteosarcoma cell lines. (A) Heat map illustrating hierarchical clustering of 36 differential miRNAs in UOSR2 and paired U2OS. (B) Volcano plot of differential miRNAs. Up-regulated miRNAs (red) and down-regulated miRNAs (green) were determined under the conditions of fold change (U2OSR2 versus U2OS) and false discovery rate-adjusted p-value ≤ 0.05. P-value was transformed by −log10. The fold change was also log-transformed. (C) Expression of microRNAs in U2OS after doxorubicin stimulation. (D–E) Western blot analysis of PD-L1 in U2OS with stimulation of different concentration of doxorubicin. (F–G) Western blot analysis of PD-L1 in K7 with stimulation of different concentration of doxorubicin. (H) qRT-PCR analysis of PD-L1 in U2OS with stimulation of different concentration of doxorubicin. (I) qRT-PCR analysis of PD-L1 in K7 with stimulation of different concentration of doxorubicin. (J) qRT-PCR analysis of miR-200a in U2OS with stimulation of different concentration of doxorubicin. (K) qRT-PCR analysis of miR-200a in K7 with stimulation of different concentration of doxorubicin. *P<0.05, **P<0.01, ***P<0.001.

miRNA-200a overexpression upregulates PD-L1 in osteosarcoma cells

First, we investigated the role of miR-200a in osteosarcoma by analyzing the expression of miR-200a in human osteosarcoma cell lines, U2OS, HOS, MG63, 143B, and WELL5, and murine osteosarcoma cell lines, K7, K7M2, and DUNN. As shown in Supplementary Figure 1A, 1B, miR-200a levels were lowest in the U2OS and K7 cells, and highest in the 143B cells. Thus, we chose U2OS, K7, and 143B cells for further experiments. Next, to further study the relationship between miR-200a and PD-L1, we generated control and miR-200a-overexpressing osteosarcoma cells using lentiviral vectors called miR-200a control and miR-200a OE, respectively (Figure 2A, 2B). Western blot and quantitative RT-PCR (qRT-PCR) analysis showed significantly higher expression of PD-L1 in the miR-200a OE U2OS and K7 cells compared to the corresponding controls (Figure 2C–2H). Moreover, flow cytometry analysis also showed higher cell surface expression of PD-L1 in miR-200a OE U2OS and K7 cells compared to the corresponding controls (Figure 2I–2L). The miR-200a OE 143B cells also showed significantly higher expression of PD-L1 compared to the controls (Supplementary Figure 2A–2F). These data show that miR-200a overexpression does upregulate PD-L1 expression in osteosarcoma cells.

Figure 2. miR-200a up-regulated PD-L1 expression in U2OS and K7. (A) qRT-PCR analysis of miR-200a in U2OS miR-200a control and miR-200a OE. (B) qRT-PCR analysis of miR-200a in K7 miR-200a control and miR-200a OE. (C–D) Western blot analysis of PD-L1 in U2OS miR-200a control and miR-200a OE. (E–F) Western blot analysis of PD-L1 in K7 miR-200a control and miR-200a OE. (G) qRT-PCR analysis of PD-L1 in U2OS miR-200a control and miR-200a OE. (H) qRT-PCR analysis of PD-L1 in K7 miR-200a control and miR-200a OE. (I–J) Flow cytometry analysis of PD-L1 in U2OS miR-200a control and miR-200a OE. (K–L) Flow cytometry analysis of PD-L1 in K7 miR-200a control and miR-200a OE. *P<0.05, **P<0.01, ***P<0.001.

miRNA-200a inhibits the activation and function of CD8⁺ T cells by upregulating PD-L1 in vitro

To explore the effects of miR-200a on anti-tumor immunity, we examined the functional status of CD8⁺ T cells after co-culturing with control and miR-200a OE osteosarcoma cells. CD8⁺ T cells co-cultured with miR-200a OE U2OS cells showed significantly higher apoptosis compared to the controls (Figure 3A–3B). The same phenomenon was also noted in 143B and K7 cells (Supplementary Figure 3A–3D).

Figure 3. miR-200a inhibited the function of CD8⁺ T cells through PD-L1/PD-1 pathway in vitro. (A–B) Examine the apoptosis of CD8⁺ T cells after co-cultured with U2OS miR-200a control or miR-200a OE separately. (C–D) Examine the proliferation rate of CD8⁺ T cells after co-cultured with U2OS miR-200a control or miR-200a OE separately. (E–F) Enzyme-Linked ImmunoSorbent Assay (ELISA) of the secretion of granzyme B and perforin. *P<0.05, **P<0.01, ***P<0.001.

We then analyzed the proliferation status of CD8⁺ T cells when co-cultured with control and miR-200a OE osteosarcoma cells. There was an increased proliferation rate of CD8⁺ T cells after the stimulation of CD3 and CD28 (Figure 3C, 3D). However, the proliferation rate of CD8⁺ T cells was significantly lower when co-cultured with miR-200a control U2OS, 143B, K7 cells, and lowest when co-cultured with miR-200a OE U2OS, 143B, K7 cells (Figure 3C, 3D; Supplementary Figure 4A–4D).

Furthermore, we analyzed the secretory function of CD8⁺ T cells by estimating the concentrations of secreted granzyme B and perforin. We observed that the concentrations of granzyme B and perforin were markedly decreased after co-cultured with miR-200a U2OS OE cells compared to the controls (Figure 3E, 3F).

We then evaluated whether immunosuppression induced by miR-200a was related to the upregulation of PD-L1 via adding anti-PD-L1 antibody during co-culture in vitro. We observed that the functions of CD8⁺ T cells were partially restored when co-cultured with U2OS, 143B and K7 cells in the presence of anti-PD-L1 antibody, but was still lower than CD8⁺ T cells cultured alone (Figure 3A–3F; Supplementary Figures 3A–3D, 4A–4D). These data suggest that miR-200a suppresses the function of CD8⁺ T cells and promotes the survival of osteosarcoma cells in vitro.

miRNA-200a promotes the growth of osteosarcoma by impairing anti-tumor immunity via PD-L1 in vivo

We used the K7-derived syngeneic mouse model to analyze the role of miR-200a in osteosarcoma in vivo. As shown in Figure 4A, 4B, the tumor volume of miR-200a-overexpressing K7 cells (miR-200a OE group) was significantly larger than that of the control group. Furthermore, the positive rate of ki-67 was significantly higher in miR-200a OE group compared to the controls (Figure 4C, 4D). This suggests that miR-200a promotes tumorigenesis partly by increasing the viability of osteosarcoma cells. Moreover, the overexpression of miR-200a could upregulate the expression of PD-L1 continuously in vivo as the expression of PD-L1 in the miR-200a OE group was significantly higher compared to the controls (Figure 4E–4G).

Figure 4. miR-200a promoted tumor growth by up-regulating the expression of PD-L1 in vivo. (A) miR-200a promoted the growth of osteosarcoma through up-regulating PD-L1. (B) Quantification of tumor volume. (C–D) Immunohistochemical staining analysis of ki-67 expression in tumor tissues. Scale bar represents 50μm. (E) Immunohistochemical staining analysis of PD-L1 expression in tumor tissues. Scale bar represents 100μm. (F–G) Flow cytometry analysis of PD-L1 expression in tumor tissues. *P<0.05, **P<0.01, ***P<0.001.

We further examined the proportions of CD4⁺ T cells, CD8⁺ T cells, Foxp3⁺ regulatory T lymphocytes (T-regs), and IFN-γ⁺ cytotoxic T lymphocytes (CTLs) in the different groups of tumor tissues to determine the effects of miR-200a on anti-tumor immunity in vivo. The decreased proportions of intra-tumoral CD4⁺ and CD8⁺ T cells were observed in miR-200a OE group compared to the controls, for which the decrease in the proportion of CD8⁺ T cells was more obvious (Figure 5A, 5B). Moreover, the percentage of Foxp3⁺ T-regs (Figure 5C–5D) was significantly higher and the percentage of IFN-γ⁺ CTLs (Figure 5E–5F) was significantly lower in the miR-200a OE group compared to the control group. However, treatment with the anti-PD-L1 antibody significantly restored the anti-tumor immunity in the miR-200a OE group (Figures 4A–4D, 5A–5F). These data demonstrate that miR-200a promotes the growth of osteosarcoma by impairing anti-tumor immunity in vivo, but at the same time, miR-200a OE group is also more sensitive to the treatment of anti-PD-L1 antibodies.

Figure 5. miR-200a impaired anti-tumor immunity in vivo. (A–B) Flow cytometry analysis of intratumoral proportion of CD4⁺ and CD8⁺ T cells. (C–D) Flow cytometry analysis of intratumoral proportion of Foxp3⁺ Treg cells. (E–F) Flow cytometry analysis of intratumoral proportion of IFN-γ⁺ CTLs. *P<0.05, **P<0.01, ***P<0.001.

High expression of miR-200a predicts the poor response to chemotherapy in osteosarcoma patients

We performed qRT-PCR and immunohistochemical analysis of the tumor specimens from 32 osteosarcoma patients to verify the relationship between miR-200a and PD-L1 (Table 2). The patients were divided into high- and low- miR-200a expression groups. As expected, patients with high miR-200a expression showed higher PD-L1 positivity than patients with low miR-200a expression (Figure 6A–6C). Our previous research showed that the therapeutic efficacy of doxorubicin was reduced due to the doxorubicin-induced-upregulation of PD-L1 expression in osteosarcoma [3]. Therefore, we tested if the expression of miR-200a was related to the efficacy of chemotherapy in osteosarcoma by analyzing biopsy tissues from a cohort of 62 osteosarcoma patients from the GEO database (GSE39058-GPL15762). The tumor necrosis rate (TNR) after conventional chemotherapy was significantly lower in osteosarcoma patients with high miR-200a expression compared to patients with lower miR-200a expression (Figure 6D; 60.10% vs. 75.59%; P=0.0407). Similar results were not observed for miR-200b and miR-200c (data not shown). Subsequently, we validated these results using fresh resection samples from 32 osteosarcoma patients (Figure 6E; 29.38% vs 67.81%; P < 0.001). Furthermore, the correlation analysis proposed to a negative correlation between miR-200a expression and TNR (Figure 6F; R=-0.789; P<0.001). Our results indicate that elevated expression of miR-200a is associated with increased expression of PD-L1 and predicts a poor response to chemotherapy.

Figure 6. High expression of miR-200a predicts the poor response to chemotherapy. (A–C) Immunohistochemical staining analysis of PD-L1 expression in samples of osteosarcoma patients. The quantification of immunohistochemical staining was based on its positive area. Scale bar represents 50μm and 100μm separately. (D) Comparison of TNR in biopsy tissue of 62 osteosarcoma patients acquired from the GEO database (GSE39058-GPL15762). Divided them into two groups according to miR-200a expression. (E) Comparison of TNR in resection tissues of 32 osteosarcoma patients. Divided them into two groups according to miR-200a expression. (F) Correlation analysis between miR-200a and TNR in resection tissues of 32 osteosarcoma patients. The expression of miR-200a was log-transformed. *P<0.05, **P<0.01, ***P<0.001.

miRNA-200a induces PD-L1 expression by direct targeting PTEN

To investigate the molecular mechanism by which miR-200a induced upregulation of PD-L1, we identified potential targets of miR-200a using Targetscan (http://www.targetscan.org). According to the prediction of Targetscan, PTEN was found as the target gene of miR-200a. The interaction between miR-200a and PTEN 3’-UTR in U2OS cells was analyzed using the dual-luciferase reporter assay. As shown in Figure 7A, 7B, there was a direct binding between PTEN 3’-UTR and miR-200a as the luciferase activity of the wild type-PTEN (wt-PTEN), but not the mutant-PTEN (mut-PTEN) was suppressed by the overexpression of miR-200a. We also observed that the expression of PTEN in miR-200a OE U2OS and K7 cells was significantly reduced compared to the controls (Figure 7C–7H).

Figure 7. miR-200a up-regulated PD-L1 expression by targeting PTEN. (A) miR-200a target sequence binding to PTEN 3-UTR was predicted with TargetScan. mut-PTEN, mutated from the seed matches, was indicated. (B) Transfected miR-200a overexpressing plasmid with wt-PTEN or mut- PTEN separately in U2OS, measured with luciferase assays. A Renilla luciferase plasmid was co-transfected as a transfection control. (C) qRT-PCR analysis of PTEN in U2OS miR-200a control and miR-200a OE. (D) qRT-PCR analysis of PTEN in K7 miR-200a control and miR-200a OE. (E and F) Western blot analysis of PD-L1 in U2OS miR-200a control and miR-200a OE. (G and H) Western blot analysis of PD-L1 in K7 miR-200a control and miR-200a OE. (I and J) Immunohistochemical staining analysis of PTEN expression in tumor tissues. Scale bar represents 50μm. (K and L) Flow cytometry analysis of PD-L1 expression of U2OS miR-200a control and miR-200a OE after the addition of VO-Ohpic. (M and N) Flow cytometry analysis of PD-L1 expression of K7 miR-200a control and miR-200a OE after the addition of VO-Ohpic. (O) qRT-PCR analysis of PTEN in U2OS miR-200a control and miR-200a OE after PTEN overexpressing. (P and Q) Western blot analysis of PD-L1 in U2OS miR-200a control and miR-200a OE after PTEN overexpressing. (R and S) Flow cytometry analysis of PD-L1 expression of U2OS miR-200a control and miR-200a OE after PTEN overexpressing. *P<0.05, **P<0.01, ***P<0.001.

Furthermore, the expression of PTEN was significantly lower in tumor tissues from the miR-200a OE group compared to the miR-200a control group in vivo (Figure 7I, 7J). Moreover, miR-200a control U2OS and K7 cells treated with VO-Ohpic, a PTEN selective inhibitor, showed increased expression of PD-L1 as analyzed by flow cytometry (Figure 7K–7N). However, miR-200 OE U2OS and K7 cells did not further increase the already elevated expression of PD-L1 when treated with VO-Ohpic (Figure 7K–7N).

To further confirm the role of PTEN in miR-200a-induced upregulation of PD-L1, we constructed a stable PTEN overexpression cell line by lentiviral transfection in U2OS cells (Figure 7O). As shown in Figure 7P–7S, the overexpression of PTEN significantly attenuated the upregulation of PD-L1 induced by miR-200a overexpression. And we also showed that PTEN acted as the target gene of miR-200a in 143B cells (Supplementary Figure 5A–5E).

Taken together, our results demonstrate that miR-200a upregulates the expression of PD-L1 by binding to PTEN 3’-UTR directly. Furthermore, miR-200a overexpression promotes the growth of osteosarcoma by inducing immunosuppression via PD-L1. And the high expression of miR-200a predicts a better response to PD-L1-targeted immunotherapy and a poorer response to chemotherapy in osteosarcoma.

Osteosarcoma patient surgical specimens

Samples from 32 osteosarcoma patients (11 females, 21 males) of ages 6 to 62 years (mean = 21.13 years) were selected for RNA extraction. 24 of the 32 samples were chosen for immunohistochemical staining of PD-L1. Tumor necrosis rate (TNR) was judged by 3 independent pathologists from the pathology department of Ruijin Hospital, Shanghai, China. In the correlation analysis between miR-200a and TNR, we performed a log transformation of miR-200a expression. General information on all 32 patients was presented in Table 2. This study was approved by the Ethics Committee of Ruijin Hospital, affiliated with Shanghai Jiaotong University School of Medicine.

Quantitative RT-PCR analysis

mRNA expression of miRNAs, PD-L1, and PTEN was analyzed by qRT-PCR as previously described [36]. All primer sequences were presented in Table 3. U6 or GAPDH served as the reference gene, and the expression of target genes was calculated using 2^—ΔΔCT.

Immunohistochemistry

Immunohistochemical staining of PD-L1, PTEN, and ki-67 was performed using the Super Sensitive IHC Detection System Kit (BD5001, Bioworld) according to the manufacturer’s instruction. Slides were incubated with antibodies against PD-L1 (10084-MB55, Sino biological, China), PTEN (9188, CST, USA) and ki-67 (ab16667, Abcam, USA) overnight at 4°C. Slides were evaluated by 3 independent investigators who were blinded to the identity of each slide. The positive rate of PD-L1 and ki-67 was quantified by Image J.

Cell culture

Human osteosarcoma cell lines 143B, MG63, HOS, and U2OS cells were obtained from ATCC (Manassas, VA, USA), human osteosarcoma multidrug resistance (MDR) cell line U2OSR2 cells [37] and mouse osteosarcoma cell line K7, K7M2, and DUNN cells were kindly gifted by Dr. Zhengdong Cai of Shanghai General Hospital. All cells were cultured in high glucose DMEM supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37°C in 5% CO₂. Different concentrations of doxorubicin (Sigma, USA) were added to the medium. VO-Ohpic (S8651, Selleck), a PTEN inhibitor, was added to the medium at 10 μM.

Lentiviral vectors and transfection

Lentiviral vectors for miR-200a or PTEN overexpression were constructed by GeneChem (Shanghai, China). Both a recombinant lentivirus and negative control lentivirus (GFP-lentivirus) were prepared. 72 h after the viral infection, cells were treated with puromycin (GeneChem, China) at 5 μg/ml to construct cell lines stably overexpressing miR-200a (miR-200a OE) or PTEN (PTEN OE).

microRNA sequencing (miRNA-seq)

RNA was isolated from U2OSR2 and U2OS cells using RNeasy kit (217004, Qiagen). miRNA-Seq libraries were performed using the Small RNA Library Prep (NEB E7300L/E7580L). miRNA-seq libraries were run on the HiSeq 2500 (Illumina) using Single-end 50-bp sequencing. miR-Seq toolbox with miRNA workflow was used. Briefly, reads were checked by FastQC v0.11.5 and aligned with Bowte v2.28. Differentially expressed genes (DEGs) analysis was performed with EdgeR. Data are presented as gene sequence fold-change in U2OSR2 cells versus U2OS cells.

Cell viability measurement

K7 miR-200a control and miR-200a OE were seeded at a density of 5x10³ cells/well in 96-well plates. Cell counting kit-8 (CCK-8) (Dojindo, Japan) was used to measure the viability of osteosarcoma cells in vitro following the manufacturer’s instructions.

Dual-luciferase reporter assay

Partial DNA sequences of PTEN 3’-UTR containing wild-type (wt) or mutant (mut) miR-200a binding sites were amplified by PCR and cloned into the GV272 reporter vector (GeneChem, China) to produce wt-PTEN and mut-PTEN reporter plasmids. Then, the constructed reporter plasmids and renilla luciferase reporter were co-transfected with miR-200a mimics or negative control into U2OS cells using lipofectamine 3000. 48h after transfection, luciferase activity was detected using a Dual-Luciferase Reporter Assay System (Promega, USA) and firefly luciferase units normalized against renilla luciferase units.

Western blot analysis

Western blot was performed as previously described [38]. The blots were incubated with primary antibodies against PD-L1 (ab213524, Abcam, USA), PTEN (9188, CST, USA) and GAPDH (5174, CST, USA) overnight at 4°C. For proteins with similar molecular weights, we used Western Blot Fast Stripping Buffer (PS107, Epizyme, China) for 15 minutes and then incubated with another antibody. Protein amounts were determined by densitometric analysis and normalized to GAPDH levels.

Flow cytometry

Osteosarcoma cells were digested by 0.25% trypsin with EDTA and incubated with anti-human PD-L1-APC (563741, BD bioscience, USA) or anti-mouse PD-L1-BV421 (564716, BD bioscience, USA) or isotype control antibodies for 30 min at room temperature in the dark. After incubation, cells were washed with phosphate buffered saline (PBS) then subjected to flow cytometry. When we analyzed the results, only osteosarcoma cells with GFP fluorescence were selected.

CD8⁺ T cell proliferation assay

CD8⁺ T cell proliferation assay was performed as described previously [3]. The human anti-PD-L1 antibody (16-5983-82, eBioscience, USA) or mouse anti-PD-L1 antibody (16-5982-85, eBioscience, USA) was added to the medium at 5 μg/ml. After co-culture, we selected CD8⁺ T cell (CD3 and CD8 positive lymphocytes) to detect the proliferation of it by CytoFlex S (Beckman, USA).

CD8⁺ T cell apoptosis assay

An Annexin V Apoptosis Detection Kit APC (88-8007-72, eBioscience, USA) was used to detect the apoptosis of CD8⁺ T cells. The detailed procedure was the same as described previously [3]. The binding of PD-L1 to PD-1 was blocked using the anti-PD-L1 antibody mentioned above. CD8⁺ T cell apoptosis was calculated as the percentage of annexin V⁺ cells in a gated CD8⁺ population.

Cytokine assay

After co-culture, the supernatant from each well was collected and centrifugated to remove cell debris. Concentrations of granzyme B and perforin were then determined using an ELISA kit (BMS6029, Invitrogen, USA; BMS2306, Invitrogen, USA) according to the manufacturer’s introductions.

Subcutaneous tumor models

The subcutaneous osteosarcoma models were performed as described previously using mouse osteosarcoma cell line K7 [3]. Mice were randomly divided into four groups: miR-200a control (with PBS only), miR-200a control (with anti-PD-L1 antibody), miR-200a OE (with PBS only) and miR-200a OE (with anti-PD-L1 antibody) (N = 5 per group). anti-PD-L1 antibody (BE0101, Bio X Cell, USA) was administered every 3 days intravenously at 10 mg/kg. Tumor size was measured every four days and calculated using the equation (length×width²) /2. This protocol was approved by the Ethics Committee of Ruijin Hospital, affiliated with Shanghai Jiaotong University School of Medicine.

In vivo lymphocyte detection

16 days after implantation, tumors were dissected and lysed to a single cell suspension using type IV collagenase (1 mg/ml) and DNase I (0.2 mg/ml). The detailed procedure was the same as described previously [3]. The antibodies used were listed as below: PD-L1-BV421 (564716, BD bioscience, USA), CD3-APC (17-0032-82, eBioscience, USA), CD4-FITC (11-0041-82, eBioscience, USA), CD8-PE (12-0081-82, eBioscience, USA), Foxp3-PE (12-5773-82, eBioscience, USA), IFN-γ FITC (11-7311-82, eBioscience, USA).

Statistical analysis

All data from three independent experiments are expressed as mean ± standard deviation. Statistical differences between groups were estimated using a Student's t-test or one-way ANOVA. Statistical analyses were performed using GraphPad Prism 5.0.