PTBP1 modulates osteosarcoma chemoresistance to cisplatin by regulating the expression of the copper transporter SLC31A1

Abstract Chemoresistance is the main obstacle of treatment in patients with osteosarcoma. RNA‐binding protein PTBP1 has been identified as an oncogene in various cancers. However, the role of PTBP1 in osteosarcoma, especially in chemoresistant osteosarcoma, and the underlying mechanism remain unclear. In this study, we aimed to explore the functions of PTBP1 in chemoresistance of osteosarcoma. We found that PTBP1 was significantly increased in chemotherapeutically insensitive osteosarcoma tissues and cisplatin‐resistant osteosarcoma cell lines (MG‐63CISR and U‐2OSCISR) as compared to chemotherapy‐sensitive osteosarcoma tissues and cell lines. Knock‐down of PTBP1 can enhance the anti‐proliferation and apoptosis‐induced effects of cisplatin in MG‐63CISR and U‐2OSCISR cells. Moreover, PTBP1 knock‐down significantly up‐regulated the expression of the copper transporter SLC31A1, as indicated by transcriptome sequencing. Through RNA immunoprecipitation, dual‐luciferase reporter assay and RNA stability detection, we confirmed that PTBP1 binds to SLC31A1 mRNA and regulates the expression level of SLC31A1 by affecting mRNA stability. Additionally, SLC31A1 silencing abrogated the chemosensitizing effect of PTBP1 knock‐down in MG‐63CISR and U‐2OSCISR cells. Using a nude mouse xenograft model, we further confirmed that PTBP1 knock‐down enhanced chemoresistant osteosarcoma responsiveness to cisplatin treatment in vivo. Collectively, the present study suggests that PTBP1 is a crucial determinant of chemoresistance in osteosarcoma.

purine bases on DNA and interfering with DNA repair mechanisms, subsequently causing DNA damage and eventually inducing apoptosis of cancer cells. 4,5 Although this treatment strategy is effective, the development of cisplatin resistance has emerged as a major cause of chemotherapy failure, resulting in overall survival stagnation over the past three decades. [6][7][8] Therefore, overcoming the chemoresistance of osteosarcoma and exploring potential molecular mechanisms have become key issues for improving the survival rate of osteosarcoma patients.
Cisplatin resistance attributes to a variety of factors, including abnormal genetic alterations, tumour microenvironment changes, epigenetic abnormalities and cancer stem cells. [9][10][11][12] It is noteworthy that drug accumulation disorder has emerged as a crucial contributor in the cisplatin resistance. 13 Traditionally, platinum drugs penetrate cell membranes mainly through passive diffusion, but recent studies have found that ATP-consuming active transport weighs more in platinum drug transport, and copper transporters (CTRs) are responsible for this transport pattern. 14,15 Copper transporters are mainly distributed in cell membranes, endoplasmic reticulum, lysosomes and other organelles, and the main function of CTRs is to transport copper and maintain intracellular copper homeostasis. 16 CTRs involved in platinum drug transport mainly depend on copper transporter 1 (SLC31A1), copper transporter 2 (SLC31A2), ATPase copper transporting alpha (ATP7A) and ATPase copper transporting beta (ATP7B). Among them, SLC31A1 is the major member that affects the intracellular concentration of platinum drugs and mediates the uptake of 60%~70% cisplatin and 30%~40% carboplatin in cells. 17 Previous studies have demonstrated that the expression level of SLC31A1 is associated with the chemoresistance of cancers. Yong et al found that oleandrin increases the chemosensitivity of osteosarcoma by inhibiting the degradation of copper transporter 1. 18 Lv et al 19  Post-transcriptional regulation, as a key regulation mode that affects various biological behaviours of tumours, has been extensively studied by researchers. 21,22 Post-transcriptional regulation depends on the RNA-binding proteins (RBPs), which directly bind to target RNA and affect the transport, stability, splicing and degradation of RNA, thereby regulating the expression of target genes and playing a role in pro-cancer or anti-cancer process. 23,24 PTBP1 is a newly discovered RNA-binding protein which specifically binds to polypyrimidine sequences on RNA and regulates the expression of downstream target genes with different or even opposing functions. 25 Increasing evidence has demonstrated that PTBP1 plays a critical role in the malignant characteristics of tumours and may be a target for cancer treatment. Xie et al 26

| Patients and clinical samples
Twenty-five osteosarcoma patients who underwent tumour resection in the department of orthopaedics of Wuhan Union Hospital were selected. All patients received a cisplatin-based chemotherapy regimen before surgery. Tumour tissues and corresponding adjacent normal tissues were frozen in liquid nitrogen within 30 minutes after isolation. The collected specimens were obtained with the informed consent of the patients or the patient's parents. The chemotherapeutic response of cancer tissue was assessed by an expert panel of pathologists. The good chemotherapy response (GR) was defined as the necrosis rate higher than 90% and the poor response (PR) as the necrosis rate <90%. This study was approved by the Ethics Committee of Tongji Medical College and complied with the Declaration of Helsinki.
For cisplatin-resistant cell lines generation, we prepared 50 mmol/L cisplatin solution and diluted it to 1~50 μmol/L for use. In vitro experiments, we used completed medium containing 0.01% DMF as control group and 1 μmol/L cisplatin treated cells for 48 hours as cisplatin treatment group.

| Generation of cisplatin-resistant cell lines
According to a previous study, 4

| Cell viability assay
Cell viability was determined by Cell Counting Kit-8 (CCK-8, Dojindo) assay and colony formation assay. For the CCK-8 assay, cells with different treatments were cultured in 96-well plates at 5000 cells per well. After the indicated times, exchanged for serum-free medium and added 10 μL CCK-8 solution into each well. The absorbance at 450 nm was measured by a microplate reader (Thermo Scientific) after incubation at 37℃ for 2 hours.
The transfected and control group cells were seeded in 6-well plates at 500 cells per well, treated with 1 μmol/L cisplatin for 48 hours upon cell adherence and then cultured under standard conditions, and the culture medium was changed every 3 days. Fourteen days later, the cells were fixed with polyformaldehyde and stained with 0.1% crystal violet (Beyotime). The number of colonies was counted with ImageJ software.

| Flow cytometry
An Annexin V-FITC apoptosis kit (BD Biosciences) was used to de-

| Western blot analysis
Total proteins were extracted from cells using lysis buffer supplemented with the complete protease inhibitor cocktail (Abcam). The following primary antibodies were used for Western blot analysis: anti-PTBP1 (ab133734, Abcam), anti-SLC31A1 (ab129067, Abcam), anti-Cleaved Parp antibody (ab32064, Abcam), anti-Cleaved Caspase-3 antibody (ab2302, Abcam), anti-Bax antibody (ab32503, Abcam), anti-Bcl-2 antibody (ab32124, Abcam) and anti-β-actin antibody (ab8226, Abcam). The detailed steps of the assay have been described in our previous paper. 33 Briefly, the protein concentration was determined by BCA method, and then, total protein lysates were fractionated using SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% non-fat milk and incubated with the above primary antibodies. Horseradish peroxidase-conjugated secondary antibody was used to develop blots. The grey value was measured by ImageJ software.

| Transcriptome sequencing and gene enrichment analysis
About 3 μg of high-quality total RNA from MG-63 CISR or U-2OS CISR cells transfected with sh-PTBP1 or sh-control was used as the sample preparations and sent to a commercial company for transcriptome sequencing (Genergy Inc). The DEGseq package was used to identify differentially expressed genes between the control group and the sh-PTBP1 group with parameters of |log 2 FoldChange| ≥ 1 and P value < .05. Gene Ontology analysis and KEGG Pathway analysis of differentially expressed genes were conducted with the clusterprofiler package in R.

| RNA stability
To evaluate the effect of PTBP1 on the stability of SLC31A1 mRNA, we detected the half-life of SLC31A1 mRNA. About 5 μg/mL actinomycin D was used to treat cells, which were subsequently collected at 0, 2, 4, 6 and 8 hours. Total RNA was extracted by TRIzol reagent as previously described. The level of SLC31A1 mRNA was measured by qRT-PCR, and the half-life of mRNA was obtained by non-linear regression analysis.

| Luciferase reporter assay
The human SLC31A1 3′-UTR luciferase reporter vector and mutant SLC31A1 3′-UTR vector containing a mutation in the predicted PTBP1 binding sequence (CUCUCU to AAAAA) were purchased from Genechem. Briefly, cells were seeded in 6-well plates, transfected with Renilla luciferase vector and pGL3 reporter for 48 hours, and the luciferase activity was measured by a Dual-Luciferase Reporter Assay System (Promega).

| Measurement of intracellular Pt accumulation
The level of intracellular Pt accumulation was detected by induc-

| TUNEL assay
Formalin-fixed osteosarcoma tissues were embedded in paraffin and cut into 4-μm sections. The TUNEL assay was performed with the One Step TUNEL Apoptosis Assay kit (Beyotime). Briefly, the paraffin-embedded sections were dewaxed and washed for 20 minutes with 20 μg/mL DNase-free protease K at 37°C and then washed three times with PBS. The TUNEL assay solution was allocated according to the instructions, and sections were incubated in TUNEL assay solution at 37°C for 60 minutes in the dark. After PBS washed three times, the nuclei were stained with DAPI. Fluorescence images were captured by the fluorescence microscopy (Olympus).

| HE staining and immunohistochemical staining
For haematoxylin and eosin (HE) staining, the sections were stained with haematoxylin aqueous solution for 10 minutes, rinsed with running water for 1 hour, dehydrated in 70% and 90% alcohol and finally

| Statistical analysis
All experiments were independently repeated at least three times.

| PTBP1 expression is up-regulated in several common cancers
To explore the expression level of PTBP1 in several common cancers, we initially conducted bioinformatics analysis of the TCGA data set. We found that PTBP1 expression was increased in sev-

| PTBP1 up-regulation is associated with poor response to chemotherapy and serves as a prognostic factor in osteosarcoma
Because PTBP1 expression exhibits an increasing tendency in some common cancers, we investigated whether PTBP1 has a role in osteosarcoma, especially in the chemoresistance of osteosarcoma.
We first assessed the expression of PTPB1 in 25 pairs of osteosarcoma tissues and adjacent normal tissues by qRT-PCR. As shown in Figure 1A,B, the PTBP1 expression level was higher in 19 osteosarcoma tissues than that in the matched normal tissues. In the remaining 6 cases, there was no significant difference in PTBP1 levels between tumours and normal tissues. According to the pathologi-  (Table 1). Additionally, Kaplan-Meier analysis revealed that patients with high expression levels of PTBP1 had a poor overall survival rate ( Figure 1E). Taken together, these results demonstrate that PTBP1 up-regulation is associated with poor response to chemotherapy of osteosarcoma, and higher PTBP1 expression predicts worse prognoses in osteosarcoma patients.

| PTBP1 is up-regulated in osteosarcoma cell lines and cisplatin-resistant cell lines
Aberrant expression of PTBP1 has been observed in osteosarcoma tissues, and it is associated with poor response to chemotherapy.

| PTBP1 deficiency enhances cisplatin efficacy in CISR osteosarcoma cells
To further investigate the role of PTBP1 in CISR osteosarcoma cells, we silenced PTBP1 expression in MG-63 CISR and U-2OS CISR cells using shRNA specific for PTBP1. First, we validated the transfection efficiency of sh-PTBP1 in CISR cells by qRT-PCR and Western blot ( Figure S3A,B). According to the results, we chose sh-PTBP1 #1 for subsequent experiments. The CCK-8 assay and colony formation assay were used to detect the sensitivity of MG-63 CISR and U-2OS CISR cells to cisplatin after PTBP1 silenced. The results indicated that PTBP1 knock-down alone had no obviously effect on cell viability and colony formation ability but enhanced the effect of cisplatin on CSIR cells (Figure 2A-C). In addition, cell apoptosis was not affected by the suppression of PTBP1, whereas the number of apoptotic cells induced by cisplatin increased significantly upon PTBP1 knockdown ( Figure 2D). To determine whether the effect of sh-PTBP1 F I G U R E 1 PTBP1 up-regulation is associated with poor response to chemotherapy and serves as a prognostic factor in osteosarcoma. A and B, Detection of PTBP1 expression in 25 pairs of osteosarcoma tissues and matched normal tissues by real-time quantitative polymerase chain reaction (qRT-PCR). C, In cases with poor response (PR) to chemotherapy and cases with good response (GR) to chemotherapy, PTBP1 mRNA level was measured by qRT-PCR. D, PTBP1 expression in normal tissues, GR tissues and PR tissues was measured by immunohistochemistry (IHC). Scale bars: 500 μm (25×) and 50 μm (200×). IOD is integrated optical density. E, Kaplan-Meier plot survival time in patients with low and high PTBP1 expression. *P < .05, **P < .01 on cisplatin sensitivity is specific to the CISR cells, we transfected MG-63 and U-2OS cells with sh-PTBP1 and performed CCK-8 assay.
We found that sh-PTBP1 has no significant effect on cell viability and cisplatin sensitivity of MG-63 and U-2OS cells ( Figure 2E,F). Overall, these results reveal that inhibiting PTBP1 expression could enhance the killing effect of cisplatin on CISR osteosarcoma cells.

| PTBP1 interacts with SLC31A1 mRNA in CISR osteosarcoma cells
Since PTBP1 is a typical RNA-binding protein, the post-transcriptional regulation of target genes is likely to be a potential molecular mechanism for its function. To identify the downstream targets of PTBP1, we performed transcriptome sequencing using total RNA from PTBP1 knock-down MG-63 CISR and U-2OS CISR cells. With cutoff criteria (|log 2 FoldChange| ≥ 1 and P value < .05), the results exhibited that the expression of 386 genes was up-regulated and 245 genes was down-regulated in PTBP1 knock-down MG-63 CISR cells compared with the control MG-63 CISR cells ( Figure 3A). Moreover, the expression of 367 genes was up-regulated and 151 genes was downregulated in PTBP1 knock-down U-2OS CISR cells compared with the control U-2OS CISR cells ( Figure 3B). All differentially expressed genes were listed in Tables S1 and S2 and subsequently performed Gene Ontology (GO) analysis ( Figure S4A,B) and KEGG Pathway analysis ( Figure S4C,D). According to previous studies on PTBP1 in cancer, we found that PTBP1 could reduce the stability of mRNA by binding to it in a variety of cancer cells, which further negatively regulates the expression of the target gene. 28,30,34 Therefore, we hypothesized that PTBP1 functions by reducing the stability of downstream target RNA and inhibiting the expression of them. Through analysis of the data, we found that 31 genes overlapped within up-regulated expression group, and among them, the expression of SLC31A1 was most significantly affected by PTBP1 (the P value was the lowest) ( Figure 3C).
To further verify whether PTBP1 directly binds to SLC31A1 mRNA, RNA immunoprecipitation assay was performed in MG-63 CISR and U-2OS CISR cells. The results confirmed that SLC31A1 mRNA was significantly enriched in anti-PTBP1 antibody co-precipitated RNA fragments ( Figure 3D), suggesting that PTBP1 binds to SLC31A1 mRNA in these cells. The results of the dual-luciferase reporter assay showed that compared with the sh-control, we detect a significant increase in luciferase activity of wild-type SLC31A1 3′-UTR reporter when transfected with sh-PTBP1 in CISR cells. However, down-regulation of PTBP1 had no repressive effect on the mutant SLC31A1 3′-UTR construct ( Figure 3E). Thereafter, we carried out the RNA stability assay to analysis the half-life of SLC31A1 mRNA. As shown in Figure 3F, the half-life of SLC31A1 mRNA was 4.21 and 4.54 hours in MG-63 CISR and U-2OS CISR cells, respectively, while the half-life of SLC31A1 was longer than 8 hours in PTBP1 knock-down cells. Taken together, these results indicate that PTBP1 can bind to the messenger RNA of SLC31A1, which further accelerates its degradation.

| SLC31A1 is down-regulated in chemotherapyinsensitive osteosarcoma tissues and CISR osteosarcoma cells
To further explore the role of SLC31A1 in osteosarcoma, we assessed the level of SLC31A1 in osteosarcoma tissues and CISR osteosarcoma cells. The results of qRT-PCR analysis showed that the expression of SLC31A1 mRNA was down-regulated in osteosarcoma tissues, especially lower in chemotherapy-insensitive osteosarcoma tissues, and there was a negative correlation between the expression of SLC31A1 and PTBP1 in osteosarcoma tissue ( Figure 4A,B).
Moreover, IHC staining showed that the positive rate of SLC31A1 in good responders was significantly higher than that in poor responders ( Figure 4C). qRT-PCR and Western blot analysis demonstrated that the levels of SLC31A1 in MG-63 CISR and U-2OS CISR cells were significantly lower than that in their original cells ( Figure 4D,G).
Moreover, PTBP1 knock-down increased the expression level of SLC31A1 in MG-63 CISR and U-2OS CISR cells ( Figure 4E,H)

| SLC31A1 knock-down abrogates the effect of PTBP1 silence on the chemosensitivity of CISR osteosarcoma cells
To evaluate whether SLC31A1 mediates PTBP1 functions, we car- knock-down ( Figure 5D). As expect, these data suggest that SLC31A1 mediates the effect of PTBP1 knock-down on the chemosensitivity of CISR osteosarcoma cells by regulating the uptake of cisplatin.

| SLC31A1 mediates the regulation of PTPB1 on cisplatin-induced apoptosis-related genes expression
The mitochondrial pathway is one of the main ways of cisplatin-induced cancer cell apoptosis. 5,9 As mentioned above, PTBP1 knockdown results in the up-regulation of SLC31A1 and the accumulation of intracellular cisplatin. Therefore, we hypothesized that PTBP1

| D ISCUSS I ON
Since adjuvant chemotherapy has been widely used in the treatment of osteosarcoma, the prognosis of osteosarcoma patients has been greatly improved. However, one of the major barriers in curing osteosarcoma is the emergence of chemoresistance. Therefore, over- cells, but PTBP1 silencing did not enhance the response of these sensitive cell lines to cisplatin. We think the possible reason is that there are already high levels of SLC31A1 in MG-63 and U-2OS cells.
These channels are sufficient for cells to take up external cisplatin.
Besides, the cisplatin uptake by SLC31A1 accounts for 60%~70% of the total uptake. 17 Therefore, even if the expression level of SLC31A1 is increased, more cisplatin cannot be taken up into cells when the external cisplatin concentration is unchanged. Of course, this requires our further verification and exploration.
In conclusion, our study reveals the role of PTBP1 in CISR osteosarcoma cells and its relevance with prognosis of osteosarcoma patients. PTBP1 affects the expression level of SLC31A1 by binding to its mRNA, thereby affecting the uptake of cisplatin by CISR osteosarcoma cells and regulating the sensitivity of CISR osteosarcoma cells to cisplatin. Our evidence provides a new perspective for the treatment of osteosarcoma chemoresistance. Shao designed the research study.

CO N FLI C T O F I NTE R E S T S
All authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.