circMRPS35 promotes malignant progression and cisplatin resistance in hepatocellular carcinoma

Abstract

Hepatocellular carcinoma (HCC) is one of the major causes of cancer-related death worldwide. Circular RNAs (circRNAs), a novel class of non-coding RNA, have been reported to be involved in the etiology of various malignancies. However, the underlying cellular mechanisms of circRNAs implicated in the pathogenesis of HCC remain unknown. In this study, we identified a functional RNA, hsa_circ_0000384 (circMRPS35), from public tumor databases using a set of computational analyses, and we further identified that circMRPS35 was highly expressed in 35 pairs of HCC from patients. Moreover, knockdown of the expression of circMRPS35 in Huh-7 and HCC-LM3 cells suppressed their proliferation, migration, invasion, clone formation, and cell cycle in vitro, and it suppressed tumor growth in vivo as well. Mechanically, circMRPS35 sponged microRNA-148a-3p (miR-148a), regulating the expression of Syntaxin 3 (STX3), which modulated the ubiquitination and degradation of phosphatase and tensin homolog (PTEN). Unexpectedly, we detected a peptide encoded by circMRPS35 (circMRPS35-168aa), which was significantly induced by chemotherapeutic drugs and promoted cisplatin resistance in HCC. These results demonstrated that circMRPS35 might be a novel mediator in HCC progress, and they raise the potential of a new biomarker for HCC diagnosis and prognosis, as well as a novel therapeutic target for HCC patients.

Keywords: HCC; circMRPS35; circRNA; cisplatin resistance; proliferation; protein coding.

Graphical abstract

Figure 1

Expression and characteristics of circMRPS35 in HCC tissues and cells (A) Schematic illustration showing the significantly different expressions of circRNAs predicted by overlapping GEO: GSE77509, GSE114564, and GSE159220 data (left) and an expression heatmap of those overlapping circRNAs (right). (B) Quantitative real-time PCR analysis of circMRPS35 in 35 pairs of HCC and adjacent tissues. (C) Quantitative real-time PCR analysis of circMRPS35 in HCC cell lines compared to L02 cells. (D) ROC curve of the diagnostic value of circMRPS35. (E) RT-PCR analysis of circMRPS35 in HCC cell lines and L02 cells. (F) RT-PCR analysis of circMRPS35 and MRPS35 with divergent and convergent primers after RNase R treatment. (G) Quantitative real-time PCR analysis of circMRPS35 and MRPS35 after ACTD treatment. (H) Quantitative real-time PCR analysis of circMRPS35 after RNA nucleocytoplasmic separation, with U6 and GAPDH as markers of the nucleus and cytoplasm in Huh-7 and HCC-LM3 cells, respectively. Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 2

circMRPS35 acts as an oncogene in HCC cells (A) Schematic representation of target sequences of shRNAs of circMRPS35. (B) Quantitative real-time PCR analysis of circMRPS35 and MRPS35 of circMRPS35 KD Huh-7 and HCC-LM3 cells. (C) Cell viability assays were used to test proliferation of circMRPS35 KD or control Huh-7 and HCC-LM3 cells. (D) Colony formation assays were performed to test cell growth of circMRPS35 KD or control Huh-7 and HCC-LM3 cells (scale bars, 1 cm). (E) Wound-healing experiments were used to detect cell migration of circMRPS35 KD or control Huh-7 and HCC-LM3 cells (scale bars, 100 μm). (F) Transwell assays of invasion and migration in circMRPS35 KD or control Huh-7 and HCC-LM3 cells (scale bars, 100 μm). (G and H) Cell cycle assays were used to detect cell cycle arrest levels of circMRPS35 KD or control Huh-7 and HCC-LM3 cells: G₀/G₁, green only; S, yellow; G₂/M, blue only. (I) BALB/c nude mice (n = 6 each group) were injected sh-circMRPS35-1 or sh-scramble Huh-7 and HCC-LM3 cells. Sizes of xenograft tumors were measured every 5 days, and weights of xenograft tumors were summarized after animals were sacrificed (scale bars, 1 cm). (J) IHC analysis of Ki67 for sh-circMRPS35-1 or sh-scramble Huh-7 and HCC-LM3 xenograft tumors tissues (scale bars, 100 μm). Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 3

circMRPS35 serves as a sponge for miR-148a in HCC cells (A) Schematic illustration of the target miRNAs of circMRPS35 predicted by overlapping miRanda, ENCORI, and circBank databases. (B) Kaplan-Meier analysis of miR-148a in HCC (n = 364). (C) Quantitative real-time PCR analysis of circMRPS35 and miR-148a with AGO2-RIP in Huh-7 and HCC-LM3 cells. (D) Western blot analysis of AGO2 protein level in Huh-7 and HCC-LM3 cells. (E) Quantitative real-time PCR analysis of miR-148a/148b/152 in 35 pairs of HCC and adjacent tissues normalized to U6 or U48. (F) Quantitative real-time PCR analysis of miR-148a/148b/152 in HCC cell lines compared to L02 cells. (G) Correlation analysis of circMRPS35 and miR-148a/148b/152 expression (n = 35). (H) Predicted complementary binding sites between circMRPS35 and miR-148a (up) and a luciferase reporter assay were used to test the binding of miR-148a and circMRPS35 in Huh-7 and HCC-LM3 cells (down). (I–K) Co-transfection with miR-148a mimics and circMRPS35 to test the proliferation assays (I), colony formation assays (scale bars, 1 cm) (J), and migration and invasion assays (scale bars, 100 μm) (K) in Huh-7 and HCC-LM3 cells. Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 4

circMRPS35 sponges miR-148a to regulate the expression of STX3 in HCC cells (A) Schematic illustration showing the target mRNAs of miR-148a predicted by overlapping TargetScan, miRWalk and miRDB databases (left) and the HCC TCGA database (right). (B and C) TCGA analysis of expression of STX3 in HCC tissues and correlation analysis of miR-148a and STX3 expression. (D) Quantitative real-time PCR assays of STX3 expression in HCC cell lines compared to L02 cells. (E) Quantitative real-time PCR assays of STX3 in miR-148a overexpression HCC-LM3 and Huh-7 cells. (F) Kaplan-Meier analysis of the expression of STX3 in HCC (n = 364). (G) Quantitative real-time PCR analysis of STX3 in 35 pairs of HCC and adjacent tissues (n = 35). (H) Western blot analysis of STX3 in five pairs of HCC and adjacent tissues. (I) Correlation analysis of circMRPS35 and STX3 expression (n = 35). (J) Predicted complementary binding sites between STX3 and miR-148a (upper); a luciferase reporter assay was used to test the binding of STX3 and miR-148a in Huh-7 and HCC-LM3 cells (lower). (K) Quantitative real-time PCR assays of STX3 in circMRPS35 KD or control Huh-7 and HCC-LM3 cells. (L) Western blot analysis of STX3 and PTEN in circMRPS35 KD and miR-148a overexpression and control Huh-7 and HCC-LM3 cells. Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 5

The stability of PTEN is regulated by circMRPS35 and STX3 in HCC cells v(A) Quantitative real-time PCR assays of STX3 in STX3 silenced or control Huh-7 and HCC-LM3 cells. (B) Western blot analysis of STX3 and PTEN in STX3 silenced or control Huh-7 and HCC-LM3 cells. (C) IP by anti-IgG, anti-PTEN, or anti-STX3 antibodies and western blot analysis of PTEN and STX3 in Huh-7 and HCC-LM3 cells. (D) 293T cells were transfected and treated with/without MG132, and then cells were immunoprecipitated with anti-FLAG antibodies and western blot by indicated antibodies. (E) STX3 silenced or control Huh-7 and HCC-LM3 cells were treated with MG132; cells were immunoprecipitated with anti-PTEN antibodies and western blot by indicated antibodies. (F) circMRPS35 KD or control Huh-7 and HCC-LM3 cells were treated with MG132; cells were immunoprecipitated with anti-PTEN antibodies and western blot by indicated antibodies. (G and H) circMRPS35 KD or control Huh-7 and HCC-LM3 cells were treated with CHX for indicated times; western blot and quantification of relative PTEN protein levels are shown. Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 6

Chemotherapy induces the expression of circMRPS35 and translation of circMRPS35-168aa (A) RNA-seq (GEO: GSE140202) analyses of circMRPS35 in sorafenib-treated cells compared to non-treated cells. (B) Quantitative real-time PCR analysis of circMRPS35 after DOX, etoposide, ACTD, and cisplatin treatment or non-treated Huh-7 and HCC-LM3 cells. (C) Schematic illustration showing that IRESs in circMRPS35 were cloned between R-Luc and F-Luc reporter genes with independent start and stop codons (upper). The relative luciferase activity of F-Luc/R-Luc in the above vectors was tested in Huh-7 cells (lower). The encephalomyocarditis virus (EMCV) IRES was used as a positive control. (D) Polysome fractionation and RT-PCR analysis of Huh-7 and HCC-LM3 cell lysate, with GAPDH as the positive control. (E and F) IP by FLAG antibody and SDS-PAGE separation of protein bands stained by Coomassie brilliant blue (CBB) and the band (red frame) (left) analyzed by LC-MS (right). (G) Western blot analysis of the expression of circMRPS35-168aa after treatment with five chemotherapy drugs in Huh-7 and HCC-LM3 cells, with GAPDH as the control. Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 7

circMRPS35-168aa resists cisplatin treatment in HCC cells (A) Cell viability assay of different circMRPS35-168aa-expressed Huh-7 and HCC-LM3 cells with different concentrations of cisplatin treatment. (B) IC₅₀ analysis of different circMRPS35-168aa-expressed Huh-7 and HCC-LM3 cells with different concentrations of cisplatin treatment. (C and D) Fluorescence-activated cell sorting (FACS) analysis of apoptosis for different circMRPS35-168aa-expressed Huh-7 and HCC-LM3 cells with cisplatin treatment (left), and statistical analysis of apoptosis rate (right). (E) Western blot analysis of cleaved caspase-3 in different circMRPS35-168aa-expressed Huh-7 and HCC-LM3 cells with cisplatin treatment, with GAPDH as the control. (F) In this model, circMRPS35 elicited its oncogenic role in HCC via sponging miR-148a to regulate the expression of STX3, which modulated the ubiquitination and degradation of PTEN, and circMRPS35 was further upregulated in chemotherapeutic drug treatment that stimulated the coding of circMRPS35-168aa peptide. circMRPS35-168aa suppressed the cisplatin-induced apoptosis via inhibiting the cleavage of caspase-3, which led to cisplatin resistance. Error bars represent the means ± SEM of three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.