Published online Oct 18, 2023.
https://doi.org/10.3802/jgo.2024.35.e13
Cisplatin-induced PANDAR-Chemo-EVs contribute to a more aggressive and chemoresistant ovarian cancer phenotype through the SRSF9-SIRT4/SIRT6 axis
Abstract
Objective
We previously elucidated that long non-coding RNA Promoter of CDKN1A Antisense DNA damage Activated RNA (PANDAR) as a p53-dependent oncogene to promote cisplatin resistance in ovarian cancer (OC). Intriguingly, high level of p53-independent PANDAR was found in cisplatin-resistant patients with p53 mutation. Here, our study probed the new roles and the underlying mechanisms of PANDAR in p53-mutant OC cisplatin-resistance.
Methods
A2780 and A2780-DDP cells were served as OC cisplatin-sensitive and cisplatin-resistant cells. HO-8910PM cells were subjected to construct chemotherapy-induced extracellular vesicles (Chemo-EVs). Transmission electron microscopy (TEM) and nanoparticle tracking analysis were employed to evaluate Chemo-EVs. Cell viability was assessed using cell counting kit-8 and colony formation assays. Cell apoptosis was assessed using Annexin V and propidium iodide staining. The relationships between PANDAR, serine and arginine-rich pre-mRNA splicing factor 9 (SRSF9) were verified by RNA immunoprecipitation and fluorescence in situ hybridization. Tumor xenograft experiment was employed to evaluate the effects of PANDAR-Chemo-EVs on OC cisplatin-resistance in vivo. Immunofluorescent staining and immunohistochemistry were performed in tumor tissue.
Results
PANDAR level increased in OC patients with p53-mutation. PANDAR efflux enacted via exosomes under cisplatin conditions. Additionally, exosomes from OC cell lines carried PANDAR, which significantly increased cell survival and chemoresistance in vitro and tumor progression and metastasis in vivo. During cisplatin-induced stress, SRSF9 was recruited to nuclear bodies by increased PANDAR and muted apoptosis in response to cisplatin. Besides, SRSF9 significantly increased the ratio of SIRT4/SIRT6 mRNA in OC.
Conclusion
Cisplatin-induced exosomes transfer PANDAR and lead to a rapid adaptation of OC cell survival through accumulating SRSF9 following cisplatin stress exposure.
Synopsis
PANDAR efflux enacted via exosomes between ovarian cancer (OC) cells under cisplatin conditions. PANDAR-chemotherapy-induced extracellular vesicles serve as a platform to control cisplatin sensitivity via regulating serine and arginine-rich pre-mRNA splicing factor 9 (SRSF9)-mediated apoptosis in p53-mutant OC cells. High level of SRSF9 indicates a poor prognosis in OC by upregulating the ratio of SIRT4/SIRT6 transcription.
Graphical Abstract
INTRODUCTION
Ovarian cancer (OC) is the deadliest gynecological malignancy worldwide [1]. The 5-year survival rate is less than 50% [2]. Treatment requires cytoreduction and chemotherapy. Although 80% of women have response to chemotherapy, cancer will recur with a median time to recurrence of 16 months [3]. The main reason for OC recurrence is drug resistance [4]. Chemotherapy-induced extracellular vesicles (Chemo-EVs) are considered a critical factor in promoting the spread of drug resistance in OC [5]. During cisplatin-induced stress, the tumor microenvironment (TME) favors the secretion of various functional materials that impact nearby cells via small vesicles called “exosomes,” which are endocytic in origin (40–150 nm in size) [6]. Exosomatic contents act as either tumor suppressors or promoters through their direct actions on a cell or by modifying the microenvironment [7]. Aberrant activity of export machinery leads to the expulsion of many proteins, RNAs, and microRNAs [8]. However, whether exosomes export long non-coding RNAs (lncRNAs) in the progression of chemoresistance remains unclear.
LncRNAs are recognized as fundamental regulators of gene expression in human [9]. During drug-induced DNA damage, a subset of p53-dependent RNAs function as oncogenes that are induced in response to specific stresses. PANDAR, the Promoter of CDKN1A Antisense DNA damage Activated RNA, was first reported as being the most upregulated p53-dependent lncRNA responding to drug-induced cell apoptosis [10]. We have previously demonstrated that PANDAR attenuates cisplatin sensitivity in OC chemotherapy depending on wild-type p53 [11]. Intriguingly, we found high level of p53-independent PANDAR in OC patients with p53 mutation [11]. These results raise the intriguing possibility that PANDAR transfers to p53-mutant OC cells and regulates stress-responsive recovery. To investigate where these p53-independent PANDAR come from and whether these PANDAR play a new role in OC chemoresistance in tumor environment, we performed a new work in this study. To our best knowledge, the function of PANDAR and its key regulator involved in the patient abdominal cavity in this process remains unknown.
Here, we performed a clinicopathologic analysis of PANDAR expression in patients with OC upon cisplatin treatment after surgery. We found that lncRNA PANDAR expression is related to a poor prognosis. Further investigation revealed PANDAR is transferred via exosomes between OC cells and finally showed a muted response to cisplatin in vivo and in vitro. RNA-binding protein analysis and RNA immunoprecipitation (RIP) assay identified 2 splicing factors (SFs), SRSF2 and SRSF9, as stable core components of lncRNA PANDAR. SRSF2 has been previously shown to be a key regulator contributing to the PANDAR-p53 feedback loop to promote cisplatin resistance in p53-wildtype OC cells [11]. Here, we focused on SRSF9, a serine/arginine-rich (SR) protein that is specifically recruited by PANDAR during cisplatin-induced stress recovery upon p53 mutation in OC cells. Clinicopathologic analysis showed high level of SRSF9 suggests a worse clinical outcome. Notably, SRSF9-knockdown cells accelerated the rate of cell apoptosis. A higher ratio of SIRT4/SIRT6 (2 sirtuins of silent information regulator two) mRNA [12] was observed in SRSF9-overexpressed OC cells.
Based on our findings, we propose that PANDAR-Chemo-EVs serve as a platform to control cisplatin sensitivity via regulating SRSF9-mediated apoptosis in p53-mutant OC cells. The clinicopathologic and RNA-binding analysis described here unveils the molecular events occurring in SRSF9, which have remained enigmatic since its discovery.
MATERIALS AND METHODS
1. Reagents
Cisplatin used for this study was provided by Shanghai General Hospital.
2. Patients
All biospecimens used in this study were provided by Shanghai General Hospital. All specimens were obtained with informed consent under the approval of the Shanghai General Hospital Institutional Ethics Review Board protocol.
3. Mouse colonies
Four-week-old female BALB/c nude mice used in this study were purchased from SLAC Laboratory Animal CO. LTD (Shanghai, China). Mice were housed under pathogen-free at the animal facility of Shanghai General Hospital. All study protocols were approved by the Institutional Animal Care and Use Committee at Shanghai General Hospital.
4. Tumor xenografts
The previously described HO-8910-Vector and HO-8910-PANDAR cells expressing green fluorescent protein (GFP) [13] were treated with cisplatin after 48-hour. Then the cultural exosomes labeled with pLVX-GFP vector (Vector-Chemo-EVs) and pLVX-GFP PANDAR (PANDAR-Chemo-EVs) were derived and co-cultured with HO-8910PM cells for a week respectively. The labeled cells were injected subcutaneously for orthotopic implantation in the ovary of BALB/c nude mice. The according protocol was repeated as previous study [11] did.
5. Cell culture
Human ovarian cancer cell lines HO-8910PM, A2780, and cisplatin-resistant cell line A2780-DDP were obtained from the Chinese Academy of Sciences Committee on Type Culture Collection Cell Bank (Shanghai, China). For co-culture of cells, the exosomes were extracted from cisplatin-resistant cell line (A2780-DDP) when cisplatin-treated for 48 hours, then put these exosomes into cisplatin-sensitive cell lines (HO-8910PM, A2780) to co-culture for 72 hours. All cells were cultured in Roswell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum (Corning Cellgro; Mediatech, Inc. Manassas, VA, USA) and 1% penicillin/streptomycin (Beyotime, Shanghai, China) in a humid incubator containing 5% CO2 at 37°C.
6. Lentivirus infection and stable cell line generation
Lentivirus containing overexpressed lncRNA PANDAR/SRSF9 or shRNA that knock down human PANDAR/SRSF9 were constructed by sub-cloning the synthesized PANDAR/SRSF9 open reading frame (ORF) into the pLVX or pLVX-GFP vector. Cell line generation was repeated as previous study [11] did. PANDAR/SRSF9 shRNA sequences are provided in Table S1. Lentivirus concentrations were chosen based on preliminary studies.
7. Quantitative real-time polymerase chain reaction (qRT-PCR)
The protocol involved qRT-PCR in this article was repeated as previous study [11] did. Primer sequences are shown in Table S1.
8. Western blotting
The protocol in this article was repeated as previous study [11] did.
9. Apoptosis assay
Apoptosis was examined using FITC Annexin V Apoptosis Detection Kit I (Cat. No. 556547; BD Biosciences, Franklin Lakes, NJ, USA) or PE Annexin V Apoptosis Detection Kit I (Cat. No. 559763; BD Biosciences) according to the manufacturer’s protocol.
10. Cell Counting Kit-8 (CCK-8) assay
Cell viability was measured using the Cell Counting Kit (CCK-8/WST-8) (Cat. No. CK04; DOJINDO, Kumamoto, Japan). The protocol was repeated as previous study [11] did.
11. Clonogenic assay
After indicated treatments, 100 number of tumor cells were plated in 12-well-plates to generate single colonies. After incubated 10 days at 37°C, cells were fixed in 4% paraformaldehyde for 30 minutes, followed by 1% crystal violet staining for 15 minutes. After washed by water for 3 times, samples were photographed, and the number of visible colonies was counted by Image J software.
12. Immunohistochemistry (IHC), LNA-based in situ hybridization (ISH), fluorescence ISH, and immunocytochemistry
The protocol involved IHC, ISH assay in this article was repeated as previous study [11] did. Immunostained tissues were scored by multiplying the intensity (0 to 3) and extent (0 to 1) of staining for each tissue point as previously described [14].
13. RIP assays
RIP experiments were performed using a Magna RIPTM RNA-binding Protein Immunoprecipitation Kit (Millipore, Burlington, MA, USA) according to the manufacturer’s instructions. The SRSF9 antibody for RIP assays was obtained from Abcam Biotechnology (Cambridge, UK).
14. Transmission electron microscopy (TEM)
Cells were fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) and then post-fixed with 1% OsO4 for 2 hours. Cells were dehydrated using a gradient series of ethanol (30%, 50%, 70%, 90%, and 100%). Cells were then incubated with LR White resin (62661; Sigma-Aldrich, St. Louis, MO, USA) twice for 1 hour, and subsequently embedded in LR White resin. The solidified blocks were cut into 60-nm sections and stained with uranyl acetate and lead citrate. Samples were observed and imaged under a transmission electron microscope (Hitachi H-7600; Hitachi High-Technologies Corporation, Tokyo, Japan). Ten fields were selected by the presence of cytoplasm shrinkage, nuclear membrane shrinkage and/ or nuclear chromatin in the outer nuclear layer gathered towards the center with uneven distribution, and the results were averaged.
15. Immunofluorescent staining
The protocol was repeated as previous study [11] did.
16. Statistical analyses
All data were exported to GraphPad Prism v9.0 (GraphPad Software) for statistical analyses, and to x-tile program for immunostaining score for ovarian cancer tissues. Values represent the mean ± standard deviation (SD). Statistical significance was determined based on p-values obtained from an unpaired 2-tailed Student’s t-test or 2-way analysis of variance.
RESULTS
1. Isolation, characterization, and quantitative analysis of PANDAR-Chemo-EVs derived from cisplatin-sensitive and cisplatin-resistant OC cells
To investigate whether PANDAR draws clinicopathologic features of OC, we examined the PANDAR expression profile in exosomes derived from normal ovarian epithelium and OC tissues. The PANDAR level was higher in exosomes of OC (Fig. 1A). Patients with OC at a III-IV stage showed an increased PANDAR level compared to those at a I–II stage (Fig. 1B, Table S2). Tissue chip data from a cohort of 160 OC patients showed high level of PANDAR indicating a poor prognostic overall survival (OS) (p<0.001) of patients with OC (Fig. 1C). These results suggest that PANDAR may be a biomarker indicating poor prognosis of patients with OC.
Fig. 1
Data are presented as the mean ± SD. Two-tailed Student’s t-test. Three independent experiments. Chemo-EV, chemotherapy-induced extracellular vesicle; FIGO, International Federation of Gynecology and Obstetrics; GO, Gene Ontology; NTA, nanoparticle tracking analysis; OC, ovarian cancer; OS, overall survival; PANDAR, Promoter of CDKN1A Antisense DNA damage Activated RNA; qRT-PCR, quantitative real-time polymerase chain reaction; SD, standard deviation; TEM, transmission electron microscopy. *p<0.05, **p<0.01, ****p<0.005.
Isolation, characterization, and quantitative analysis of PANDAR-Chemo-EVs derived from cisplatin-sensitive and cisplatin-resistant OC cells.
(A) qRT-PCR analysis of PANDAR expression in exosomes from normal ovarian epithelium in patients without ovarian disease and cancer tissues in patients with OC. (B) qRT-PCR analysis of PANDAR expression in tissues of patients with OC between stage I–II and stage III–IV according to the science and practice of global basis of the FIGO stage. (C) OS of patients with OC with high and low expression of PANDAR based on tissue chips from 150 patients with OC. (D) GO analysis of OC based on molecular function, cellular component and biological process. (E) NTA and TEM analysis of exosomes from cancer tissues of patients with OC. (F) RT-PCR assay of PANDAR expression in OC tissues of newly diagnosed patients and chemoresistant patients. (G) Protein expression as shown by western blotting assay, showing an increased level in CD63 and CD81 in exosomes derived from A2780 and A2780-DDP cells. (H) qRT-PCR assay of PANDAR expression in cellular and culture medium of A2780 and A2780-DDP cells. (I) qRT-PCR analysis of PANDAR expression in the cellular and culture medium of A2780-DDP cells with or without a 24-hour treatment of 20 μM cisplatin.
We next analyzed the biological pathway of PANDAR in OC. Gene Ontology analysis showed PANDAR may play a role through the mass of extracellular space, extracellular region and extracellular exosome (Fig. 1D). Focusing on the extracellular component, we isolated and analyzed extracellular exosomes from OC cell medium upon cisplatin treatment and observed exosomes through TEM assay and nanoparticle tracking analysis (Fig. 1E). We also isolated exosomes from primary cancer cells derived from newly diagnosed and recurrent/chemo-resistant OC samples with p53 mutation, and then measured the PANDAR level in their extracellular exosomes. A significant increase in PANDAR level was observed in exosomes from chemo-resistant OC primary cells compared to those from newly diagnosed OC primary cells (Fig. 1F). We next examined the characteristic proteins of exosome membranes in OC cell lines (cisplatin-sensitive A2780 cells and cisplatin-resistant A2780-DDP cells). We observed higher expression of CD63 and CD81 in extracellular exosomes derived from cell lines compared to intracellular contents (Fig. 1G). Upon cisplatin treatment, PANDAR expression was higher in extracellular space compared to the intracellular space, especially in A2780-DDP cells (Fig. 1H). However, PANDAR expression was muted in the absence of cisplatin treatment in extracellular space (Fig. 1I). The results above indicate PANDAR exists in Chemo-EVs in OC tumor microenvironment.
2. Secretory PANDAR-Chemo-EVs phenotype favors cell viability in response to cisplatin
We next investigated whether secretory PANDAR-Chemo-EVs mediate cisplatin sensitivity in OC. To address this question: i) we isolated exosomes derived from PANDAR-knockdown A2780 cells and PANDAR-overexpressed HO-8910PM cells with or without cisplatin treatment, ii) we used PCR assay to analyze the PANDAR level in these exosomes (Fig. 2A and B), and iii) we derived exosomes with different PANDAR levels and co-cultured them with A2780 cells upon cisplatin treatment. CCK-8 assay was used to investigate the cell half maximal inhibitory concentration (IC50) by co-culturing the highest PANDAR-level exosomes with A2780 cells. We found that the PANDAR-Chemo-EVs significantly increased cell viability (Fig. 2C) and the IC50 (Fig. 2D) to cisplatin. Additionally, co-culture of exosomes with the highest level of PANDAR with A2780 cells was associated with a significant elevation in cell colony formation potential (Fig. 2E) and with a dramatical elimination of apoptosis (Fig. 2F). These changes were reversed following co-culture of PANDAR-knockdown exosomes with OC cells treated with cisplatin. Taken together, these results converge on the idea that PANDAR-exosome secretion favors OC cell survival potential in response to cisplatin.
Fig. 2
Data are presented as the mean ± SD. Two-tailed Student’s t-test. Three independent experiments. CCK-8, Cell Counting Kit-8; Chemo-EV, chemotherapy-induced extracellular vesicle; IC50, half maximal inhibitory concentration; OC, ovarian cancer; PANDAR, Promoter of CDKN1A Antisense DNA damage Activated RNA; PBS, phosphate-buffered saline; qRT-PCR, quantitative real-time polymerase chain reaction; SD, standard deviation. *p<0.05, **p<0.01, ****p<0.005.
Secretory PANDAR-Chemo-EVs phenotype favors cell viability in response to cisplatin.
(A) qRT-PCR analysis of PANDAR expression in the exosomes of extracellular medium in A2780-PANDAR-knockdown cells (shPANDAR) after treatment with indicated doses of cisplatin or phosphate-buffered saline (PBS) for 24 hours, ctrl shRNA cells serve as controls. (B) qRT-PCR analysis of PANDAR expression in the exosomes of extracellular medium in HO-8910PM-PANDAR-overexpressed cells after treatment with indicated doses of cisplatin or PBS for 24 hours, vector cells serve as controls. (C) CCK-8 assay of cell viability in A2780-shPANDAR cells after treated with indicated doses (10–40 μM) of cisplatin for 24 hours with or without Chemo-EVs from cultured medium of resistant cells. The cell survival rate was calculated through three independent experiments. (D) The cell survival rate and the subsequent IC50 were calculated from D using GraphPad 9.0 software. (E) Colony formation and quantification in A2780 cells cultured with Chemo-EVs from cultured medium of PANDAR-knockdown (shPANDAR) OC cells and PANDAR-overexpressed OC cells after incubated with cisplatin for 10 days. (F) Apoptosis of A2780 cells cultured with Chemo-EVs of PANDAR-knockdown (shPANDAR) OC cells and PANDAR-overexpressed OC cells after incubated with 20 μM cisplatin for 24 hours.
3. PANDAR interacts with SRSF9 in the OC cell nucleus
Having confirmed that PANDAR-Chemo-EVs muted cell sensitivity to cisplatin, we next focused on investigating the key modulator of PANDAR in this process in the recipient OC cells. RBP analysis showed SRSF9 as top predicted binding protein of PANDAR sequence (relative score=100%), which is located in chromosome 12 (Fig. S1A). We then use SWISS MODOL software to predict the interaction domain between SRSF9 and PANDAR to investigate the protein structure (Fig. 3A) and sequence (Fig. 3B). The result of QMEAN4 evaluation showed a good match between SRSF9 and the template protein (Fig. 3C). Additionally, evaluation of the Cβ score, atom score, solvation score, and torsion score showed a good suitability of SRSF9 structure to match the PANDAR transcriptional regulation base of 832–836 bp (Fig. 3D). We also determined the number of post-transcriptional modification sites of SRSF9 (Fig. 3E). To confirm the predictions above, we detected endogenous PANDAR RNA co-immunoprecipitated with SRSF9 in cellular lysates from OC cells. The interaction was not detected when an isogenic immunoglobulin G antibody was used, whereas a robust and specific interaction between SRSF9 and PANDAR was read in the isogenic SRSF9 antibody group (Fig. 3F and G). Moreover, an increasing co-localization of PANDAR and SFRS9 in the nucleus over cisplatin treatment (Fig. 3H). To further investigate the role of SRSF9 in OC prognosis, we drew Kaplan–Meier curves of SRSF9 in OS and progression-free survival (PFS) probability. Low expression of SFRS9 was shown in the OS of OC patients treated with cisplatin (p=0.017) (Fig. S1B), as well as PFS in patients with OC (p=0.003) (Fig. S1C). These results indicate SRSF9 is a PANDAR-target gene during cisplatin-treated OC therapy.
Fig. 3
Data are presented as the mean ± SD. Two-tailed Student’s t-test. Three independent experiments. FISH, fluorescence in situ hybridization; IgG, immunoglobulin G; IP, immunoprecipitation; lncRNA, long non-coding RNA; OC, ovarian cancer; PANDAR, Promoter of CDKN1A Antisense DNA damage Activated RNA; qRT-PCR, quantitative real-time polymerase chain reaction; RIP, RNA immunoprecipitation; SD, standard deviation; SRSF, serine and arginine-rich pre-mRNA splicing factor. **p<0.01, ****p<0.005.
PANDAR interacts with SRSF9 in the OC cell nucleus.
(A) SWISS MODOL analyzation the structure of SRSF9 and the protein domain of SRSF9 binding with lncRNA PANDAR. (B) Sequences of SRSF9 and lncRNA PANDAR binding motif by Systematic Evolution of Ligands by Exponential Enrichment. (C) QMEAN4 estimation of whether or not SRSF9 would be a good fit for the model. The score is characterized for the match-degree of SRSF9 with the model. (D) Evaluation of the QMEAN score, Cβ, atom, solvation, and torsion to assess the structure of the predicted protein SRSF9. (E) Analysis of the number of the modification sites of SRSF9. Cellular extracts from A2780 and HO-8910PM cells treated with 20 μM cisplatin for 24 hours, and RIP assay showing IP with control IgG or SRSF9 antibody by western blotting (F) and qRT-PCR (G) performed with isolated RNA and the following reverse transcription using primers for lncRNA PANDAR. (H) FISH assay of PANDAR RNA (green) in A2780 cells showing distribution in discrete foci through the nucleus and cytoplasm at the beginning of treatment with or without cisplatin (20 μM). The foci increased and focused in the nucleus after cisplatin treatment for 24 hours. Co-localization of SRSF9 protein (red) and PANDAR RNA (green) in A2780 cells via immunocytochemistry performed with SRSF9 antibodies and showed an increase in discrete foci through the nucleus after cisplatin treatment, and is largely detected in nucleus. Scale bar: 500 μm. (I) Quantification of the FISH assay of PANDAR RNA and SRSF9 in A2780 cells with or without cisplatin treatment.
4. Effect of PANDAR-Chemo-EVs on OC cisplatin-sensitivity in vivo
We next used an orthotopic ovarian tumor model to evaluate the aforementioned results in vivo. This model consisted of HO-8910PM OC cells co-cultured with Chemo-EVs isolated from A2780-PANDAR cells upon cisplatin treatment (20 μM) on alternate days for 3 weeks (Fig. 4A), followed by injection of the exosome treated HO-8910PM cells (sensitive to cisplatin) into the ovarian bursa and observed for the tumors in vivo for the next 5 weeks (Fig. 4B). The mice injected with HO-8910PM cells co-cultured with PANDAR-Chemo-EVs had a significant increase in ascites (Fig. 4C) and the number of metastatic nodules in the liver, renal, mesentery, and abdominal wall compared to the same cells cultured with PLVX-Chemo-EVs as control. The expression signature of PANDAR and SRSF9 in the metastatic tumor of PANDAR-Chemo-EVs co-cultured group was shown in Fig. 4D. The results above indicate that exosomes are carriers of oncogenic lncRNAs and can alter the signaling pathways in the recipient cells by transfer of PANDAR, and that cisplatin induces changes in PANDAR-exosomes that can: i) enhance the malignant phenotype of recipient cancer cells and ii) promote cancer-like behavior of sensitive OC cells.
Fig. 4
Data are presented as the mean ± SD. Two-tailed Student’s t-test. Three independent experiments. Chemo-EV, chemotherapy-induced extracellular vesicle; GFP, green fluorescent protein; OC, ovarian cancer; PANDAR, Promoter of CDKN1A Antisense DNA damage Activated RNA; SD, standard deviation. ****p<0.005.
Evaluating the effect of PANDAR in OC cisplatin-sensitivity in vivo.
(A) PANDAR transfer efficiency in HO-8910PM cells. HO-8910PM cells were co-cultured with Chemo-EVs from medium of PANDAR-overexpressed OC cells, and the PANDAR gene was labeled with GFP. (B) Schematic diagram showing the process of orthotopic transplantation in mice with HO-8910PM cells pre-cultured with PANDAR-overexpressed OC cells. (C) The hemorrhagic ascites from mice in both control group (HO-8910pm-PLVX) and PANDAR-over-expressed group (HO-8910PM-PANDAR) were extracted and measured by syringe after 8 weeks. (D) Tumor cells derived from ascites, orthotopic transplantation of ovary and its nearby organ planter is shown and stained by PANDAR probe and anti-SFRS9 reagent. Scale bar: 500 μm.
5. Role of SRSF9 and downstream gene expression in OC cell apoptosis in response to cisplatin
Univariate analysis showed that a high level of SFSR9 expression was related to poor prognosis of OC (p=0.002) (Table S3). To establish the reason for this finding, we employed RBP-mRNA analysis and identified SIRT4 as the most relevant protein (r=0.468, p<0.001) that was predicted to be activated downstream of the SRSF9 network in 379 cases of patients with OC (Fig. S2A). Given that the high ratio of SIRT4/SIRT6 is a predictor of poor prognosis in OC [15], we investigated whether SRSF9 was inversely associated with SIRT6 in these OC samples. Indeed, a significant relationship was observed between SRSF9 and SIRT6 (r=−0.112, p<0.001) (Fig. S2B). IHC assay in OC tissue chips (n=150) (Fig. 5A) showed high SRSF9 level in immunostaining arrays indicates a poor prognosis of OC (Fig. 5B). SRSF9 and SIRT4 level was confirmed as positive relationship (r=0.32, p<0.001) (Fig. 5C). More of the ratio of SIRT4/SIRT6 mRNA was observed in PANDAR-overexpressed and/or SRSF9-overexpressed OC cells (Fig. 5D). Conversely, SIRT4/SIRT6 ratio was dramatically downregulated when SRSF9 knockdown (Fig. 5D). After knocking down SIRT4 and upregulating SIRT6 in OC cells, PANDAR or SRSF9 level were not altered (Fig. 5H). This result suggests SIRT4 and SIRT6 are the downstream genes of SRSF9. We next investigated whether the SRSF9-mediated SIRT4/6 ratio played a role in apoptosis in OC. Flow cytometry assay was used to assess apoptosis in SRSF9 overexpressed cells after co-cultured with a high level of PANDAR-Chemo-EVs. A significant decrease in cell apoptosis was observed in SRSF9-overexpressed cells, especially in cells co-cultured with PANDAR-Chemo-EVs (Fig. 5F). A higher rate of apoptosis was also showed in OC cells with a low level SIRT4/SIRT6 ratio (Fig. 5G). Further, BAX/Bcl-2 level was upregulated in OC cells with SIRT4-knockdown and SIRT6-overexpression (Fig. 5E). Our results suggest that SIRT4/SIRT6 mRNA level increase the accumulation of BAX and the elimination of Bcl-2, favoring the initiation of apoptosis in response to cisplatin in OC. To further confirm this suggestion, we derived OC tissues from 5 cisplatin-resistant patients and mice analyzed the levels of PANDAR, SRSF9, SIRT4, and SIRT6 separately. We found an increase in SIRT4 level and a decrease in SIRT6 level in tissues of all five patients with cisplatin-resistance (Fig. 5I), as well as in tissues derived from mice injected with PANDAR-overexpressed OC cells (Fig. S2C and D). Three patients with mutant-p53 exhibited higher levels of PANDAR, SRSF9, and SIRT4 and lower level of SIRT6 in their resistant tissues, compared with their sensitive tissues from newly-diagnosed periods (Fig. 5J). Kaplan–Meier curve showed a poor prognosis of OC patients with high-level of SIRT4 and low-level of SIRT6 (Fig. S2E and F). These results indicate the significance of PANDAR-mediated SRSF9-regulated apoptosis via SIRT4/6 ratio alteration in clinical chemoresistance. Accordingly, we propose a schematic model of underlying mechanism regarding cisplatin resistance initiated by lncRNA PANDAR transferred via exosomes in the ovarian TME (Fig. 6).
Fig. 5
Data are presented as the mean ± SD. Two-tailed Student’s t-test. Three independent experiments. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IHC, immunohistochemistry; ISH, in situ hybridization; lncRNA, long non-coding RNA; ns, not significant; OC, ovarian cancer; PANDAR, Promoter of CDKN1A Antisense DNA damage Activated RNA; qRT-PCR, quantitative real-time polymerase chain reaction; SD, standard deviation; SIRT, sirtuin; SRSF, serine and arginine-rich pre-mRNA splicing factor. ****p<0.005.
Evaluating the role of SRSF9 and downstream gene expression in OC cell apoptosis response to cisplatin.
(A) IHC assay of SRSF9 protein with SRSF9 antibody in matched ovarian cancer tissues (n=150 in tissue chips). (B) Based on the immunostaining score of OC tissue chips (n=150) and the overall survival statistics, Kaplan–Meier analysis of overall survival curve was performed in patients with OC with high SRSF9 expression versus low SRSF9 expression. One hundred fifty in OC cohort. (C) Correlation between SRSF9 and cisplatin treatment signature gene SIRT4 (150 OC tissue chips analyzed by Pearson correlation). (D) qRT-PCR quantification of SIRT4 and SIRT6 gene expression in PANDAR-overexpressed A2780 cells with SRSF9-downregulation or overexpression. (E) Western blotting assay showed the protein expression of Bcl-2 and BAX in SRSF9-overexpressed and SIRT4-knockdown or SIRT6-overexpressed cells upon cisplatin treatment (20 μM). GAPDH as control protein in these groups. (F) Flow cytometry assay showing the apoptosis rate in SRSF9-overexpressed cells cultured with exosomes containing PANDAR during cisplatin treatment or cultured without exosomes upon cisplatin treatment. (G) Flow cytometry assay showing the apoptosis rate upon SIRT4-knockdown or SIRT6-overexpressed OC cells. (H) qRT-PCR quantification of PANDAR expression upon SIRT4-knockdown and/or SIRT6-overexpression in A2780 cells. (I) IHC assay showing SIRT4 and SIRT6 protein staining in patients with OC with resistance to cisplatin. Scale bar: 800 μm. (J) LNA ISH analysis of lncRNA PANDAR with LNA probes and IHC assay of SRSF9 protein with SRSF9 antibody, SIRT4 antibody, and SIRT6 antibody in matched OC tissues before platinum-based therapy (Sensitive) and after disease progression during platinum-based treatment (Resistance). Scale bar: 100 μm.
Fig. 6
lncRNA, long non-coding RNA; OC, ovarian cancer; PANDAR, Promoter of CDKN1A Antisense DNA damage Activated RNA; SFRS2, splicing factor, arginine/serine-rich 2; SIRT, sirtuin; SRSF, serine and arginine-rich pre-mRNA splicing factor.
Graphic illustration demonstrating cisplatin-induced lncRNA PANDAR triggers muted response to cisplatin in OC.
A certain amount of p53-dependent PANDAR is upregulated with the sequential stimulation of cisplatin. In one way, PANDAR interacts with SFRS2 in wildtype-p53 OC cells, leading to a reduction of p53 and its phosphorylation at Ser15, which in turn inactivates p53-mediated gene expression, including PUMA and BAX, and downregulated apoptosis [6]. However, PANDAR was secreted to nearby OC cells with p53 mutation, interacting with SRSF9 in nucleus and leading to a high ratio of SIRT4/SIRT6 mRNA level. As a result, a muted response to cisplatin in OC cells was observed.
DISCUSSION
TP53 mutation leads to a poor prognosis and a high rate of chemoresistance in OC [16]. However, the addition of p53 gene therapy did not improve therapeutic effectiveness in OC patients [17]. Hence, more of p53-involved mechanisms in cancer therapy need to be investigated. We previously showed TP53-dependent lncRNA PANDAR as an oncogene in OC [11]. In this study, the role of p53-independent PANDAR in OC chemoresistance upon p53 mutation was investigated. We identified cisplatin-induced lncRNA PANDAR transfer via exosomes between OC cells, which was responsible for the transition from chemosensitivity to chemoresistance of the recipient OC cells. The mechanism involved in this transition bases on the regulation of SRSF9 activity in the recipient cell nucleus and the corresponding alteration of its downstream genes SIRT4/SIRT6 expression. From the sight of lncRNA-mRNA network [18], we elucidated that SRSF9 acts as a downstream modulator of lncRNA PANDAR to regulate SIRT4/SIRT6 mRNA ratio. This regulation is involved in cisplatin resistance, and may serve as a promising target for future advances in OC chemotherapy.
It is important to confirm how massive the implication of lncRNA PANDAR and SRSF9 expression pattern in real world prior to investigating the lncRNA-mRNA network. We followed 160 patients with ovarian cancer for 9 years, rarely without recurrence. Most of these malignancies are nearly always associated with mutations in the TP53 gene in the sight of clinical pathology. We analyzed the test of equality of survival distributions for the different levels of PANDAR and SRSF9 in matched ovarian cancer tissues. The OC patients with poor OS showed high expression level of PANDAR and SRSF9 among those patients. Metastasis in ovarian cancer patients showed significant association with high-expression level of SRSF9 compared to those patients without metastasis (p=0.009; Table S4). The histologic, molecular, and genetic evidence shows high expression levels of lncRNA PANDAR and SRSF9 in patients with ovarian cancer, especially in recurrent OC patients.
Non-coding RNAs have been revealed to regulate alternative splicing (AS) in cancer progression through posttranscriptional regulation of SFs [19, 20]. For example, lncRNA TPM1-AS directly block the binding between the SF and TPM1 pre-mRNA and regulate AS, inhibiting the progression of cancer [21]. Recently, lncRNA MALAT1 was found to contribute to SRSF1 phosphorylation and finally exacerbate colorectal cancer [22, 23]. Similar to SRSF1, SRSF9 is a SRSF involved in the regulation of alternative splicing, the process of which plays an important role in DNA damage [24] or carcinogenesis and anticancer therapy [25, 26]. However, the importance of SRSF9 interaction with lncRNAs in cancerous deregulation has been poorly investigated. Here, we found SRSF9 was recruited by lncRNA PANDAR in the nucleus under cisplatin conditions. During this recruitment, the binding motif between lncRNA PANDAR and SRSF9 lies out of the splicing sites of exon splicing enhancer, exon splicing silencer, intro splicing enhancer, and intro splicing silencer, this may partly because lncRNAs lack the cooperative interaction network of positive signals that efficiently navigate the splicing machinery to splice sites [27]. More of investigation in non-splicing sites of SRSF9 will be studied in the future.
SR proteins were found to regulate all levels of gene expression via transcription initiation and elongation, alternative splicing, mRNA stability, translation, and protein degradation [28]. The roles of SRSF9 were different in various pathologic-clinical subtype in cancer [29, 30, 31]. Recent pan-cancer analysis revealed that SRSF9 was upregulated in the proliferative subtype of OC and downregulated in the mesenchymal subtype of OC [32]. Data also showed that the higher the expression of SRSF9, the greater the purity of tumor cells. The above analysis is in accord with our data that SRSF9 upregulation in OC cells favors tumor cell proliferation in response to cisplatin. As is known, the multistep process of malignant transformation involves changes in several genes, the protein products of which are generally not final effectors but intermediate steps in signaling cascades [33]. Our study also confirmed this from the results of SRSF9 regulating the expression of multiple genes, including BAX, Bcl-2, SIRT4, and SIRT6. SIRT4 mRNA encodes bases from chromosome 12 and its expression participates in DNA damage repair and favors the suppression of cancer proliferation [34]. SIRT6 mRNA encodes bases from chromosome 19 and its protein regulates poly (ADP-ribose) polymerase (PARP), P53, E2F1, and SMAD4 in cancers with opposing phenotypes [35]. The above regulation involved in SRSF9 alteration may serve a critical role in tumorigenesis and chemoresistance to cisplatin or PARP inhibitor in epithelial OC treatment.
SRSF9 expression is also positively associated with immune function activity, depending on its splicing characteristics. Researchers have revealed that SRSF9 regulates CD44 splicing via competitively binding to CD44 exon v10 compared to Tra2β [36]. The above research provides SRSF9 as a potential target in OC immunotherapy [37]. SRSF9 phosphorylation depends on stress-induced lncRNAs [38]. For example, upon thermal stress exposure, highly repetitive satellite III lncRNAs induced nuclear stress bodies, resulting in CDC like kinase 1 recruitment and acceleration to re-phosphorylate SRSF9, thereby promoting target intron retention [39]. Similarly, we elucidated that SRSF9 activity was triggered by lncRNA PANDAR delivered from resistant OC cells upon cisplatin-induced stress, resulting in alteration in the ratio of SIRT4/6 mRNA and promoting poor prognosis of OC. Further elucidation of the SRSF9 role of splicing characters in immune function activity will provide important clues to understand the significance of the SRSF9 active expansion in the TME that specifically occurs in anticancer drug resistance.
SUPPLEMENTARY MATERIALS
Primer sequences and gene related sequencesTable S1
Correlation between lncRNA PANDAR expression and clinicopathological characteristicsTable S2
Univariate and multivariate analyses of the factors correlated with OS of ovarian cancer patientsTable S3
Correlation between SRSF9 expression and clinicopathological characteristicsTable S4
SRSF9 gene location and the correlation with the prognosis of patients with OC.Fig. S1
SRSF9-SIRT4/SIRT6 axis in the prognosis of patients with OC.Fig. S2
Funding:This work was supported by the National Natural Science Foundation of China (grant No. 81972425), the National Natural Science Foundation of Shanghai (grant No. 20ZR1444200), the Clinical Project of Shanghai Municipal Health Commission (202240217), and the Foundation of Shanghai Sailing Program (grant No. 21YF1451700).
Conflict of Interest:No potential conflict of interest relevant to this article was reported.
Author Contributions:
Conceptualization: W.H., B.W.
Data curation: L.Y., W.Y., S.X., Y.Z., L.S.
Formal analysis: W.H., L.Y., W.Y., L.S.
Funding acquisition: W.H., B.W.
Investigation: W.H., S.X.
Methodology: W.H., L.Y., W.Y.
Resources: W.H., L.S.
Software: W.H., L.Y., Y.Z.
Supervision: L.Y., B.W.
Validation: B.W.
Visualization: B.W.
Writing - original draft: W.H.
Writing - review & editing: L.S., B.W.
ACKNOWLEDGEMENTS
Special thanks to Professor Xiaoyan Zhou and Dr. Cong Qiao for their specialized guidance of pathology. Many thanks to Dr. Lei Yu for his kind assistance of statistical analyses. Also, thanks to LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.
References
-
Sun T, Yang Q. Chemoresistance-associated alternative splicing signatures in serous ovarian cancer. Oncol Lett 2020;20:420–430.
-
-
Krchnáková Z, Thakur PK, Krausová M, Bieberstein N, Haberman N, Müller-McNicoll M, et al. Splicing of long non-coding RNAs primarily depends on polypyrimidine tract and 5′ splice-site sequences due to weak interactions with SR proteins. Nucleic Acids Res 2019;47:911–928.
-
-
Kędzierska H, Piekiełko-Witkowska A. Splicing factors of SR and hnRNP families as regulators of apoptosis in cancer. Cancer Lett 2017;396:53–65.
-
-
Liu J, Wang Y, Yin J, Yang Y, Geng R, Zhong Z, et al. Pan-cancer analysis revealed SRSF9 as a new biomarker for prognosis and immunotherapy. J Oncol 2022;2022:3477148
-
MeSH Terms
Figures
1 / 6
