Reovirus resistance in tumors mediated by elevated ISG expression: overcoming therapeutic resistance via JAK/STAT pathway modulation

Abstract

Background

Oncolytic viruses such as reovirus hold significant promise for cancer therapy, but their efficacy varies across tumor types due to heterogeneous tumor cell sensitivity to reovirus. Elucidating the molecular mechanisms underlying reovirus resistance is therefore critical to improving therapeutic outcomes for resistant tumors.

Methods

We comprehensively assessed the oncolytic sensitivity of 12 diverse tumor cell lines to reovirus, followed by transcriptomic analysis of two ovarian cancer cell lines with divergent responses. Differentially expressed resistance-associated genes were verified, and the Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway was modulated using the type I interferon (IFN) receptor agonist RO8191 and JAK inhibitor ruxolitinib to evaluate impacts on reovirus replication and oncolysis. Additionally, patient-derived primary breast cancer cells were tested to explore cross-resistance mechanisms between chemotherapy and reovirus.

Results

Tumor cells were stratified into three groups: high sensitivity (HS, cell viability ≤ 50%), moderate sensitivity (MS, 50–80%), and low sensitivity/resistance (LS, ≥ 80%). Transcriptomic analysis revealed higher baseline IFN-stimulated gene (ISG) expression in reovirus-resistant cells, with 170 ISGs upregulated in LS cells (including 28 interferon-related DNA damage resistance signature (IRDS) genes). ISG overexpression correlated with reovirus resistance: ruxolitinib-mediated JAK-STAT inhibition reduced ISGs and enhanced oncolysis, while pathway activation suppressed reovirus replication. Unphosphorylated ISGF3 (U-ISGF3) may drive IRDS-related ISG expression, contributing to cross-resistance to doxorubicin and reovirus.

Conclusions

Tumor cell resistance to reovirus is associated with elevated baseline expression of specific ISGs, potentially mediated by U-ISGF3-induced IRDS upregulation. These insights provide a foundation for optimizing reovirus therapy in resistant tumors.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12985-026-03118-0.

Introduction

Oncolytic viruses, known for their groundbreaking contributions to cancer therapy, have undergone significant advancements, particularly following the FDA’s approval of talimogene laherparepvec (T-VEC) in 2015 [1]. Among these viruses, reovirus, characterized by a non-enveloped, double-stranded RNA (dsRNA) structure, holds prominence in the field [2]. Its unaltered strain, reovirus type 3 Dearing, is marketed as pelareorep^® and is currently being investigated in over 30 clinical trials for the treatment of various cancers [3–6]. Despite demonstrating considerable potential in cancer therapy, numerous clinical trials have revealed substantial variability in efficacy across different tumor types or even among patients with the same cancer type, with objective response rates ranging from 17% to 48% [7–9]. This variability in efficacy is likely due to varying levels of oncolytic sensitivity of tumor cells to reovirus across different tumor types. As a distinctive class of novel antitumor agents, the efficacy of reovirus heavily relies on its ability to extensively replicate within tumor cells, thereby exerting its oncolytic effects [10]. The mechanism by which reovirus selectively infects and replicates within tumor cells to induce oncolysis remains poorly understood. Previous reports have suggested that reovirus preferentially replicate in, and subsequently lysis cells with an activated Ras pathway [11–13]. Additionally, the expression of junctional adhesion molecule-A (JAM-A) or the presence of intracellular lysosomal cathepsins B/L are also significant factors that influence the oncolytic sensitivity of tumor cells to reovirus [14–16]. However, increasing evidence in recent years indicates that none of these factors play a decisive role.

Normal cells utilize a range of protective mechanisms to defend against pathogens, with type I interferons (IFN) serving as essential components of their antiviral defense systems. When a cell is infected by a virus, the secreted type I IFN bind to their receptor, activating the Janus kinase-signal transducer and activator of transcription (JAK-STAT) signaling pathway. Activated STAT1 and STAT2, in conjunction with interferon regulatory factor 9 (IRF9), form the IFN-stimulated gene factor 3 (ISGF3) complex. ISGF3 further associates with the IFN-stimulated response element (ISRE), thereby inducing the expression of a multitude of IFN-stimulated genes (ISGs). To date, hundreds of ISGs have been identified [17–19]. These ISGs function through various mechanisms to limit viral replication and spread, including blocking viral infection, promoting the degradation of viral RNA, and inhibiting the translation of viral proteins. Multiple previous studies have demonstrated that ISGs can modulate tumor cell sensitivity to oncolytic viruses through individual or synergistic actions, leading to varying levels of oncolytic resistance [20–22].

Our study identified a strong correlation between tumor cell resistance to reovirus and increased expression of ISGs. Importantly, it revealed a direct relationship between tumor cell resistance to both chemotherapy and reovirus, and the activity of the unphosphorylated IFN-stimulated gene factor 3 (U-ISGF3). This complex plays a pivotal role in modulating the expression of ISGs, particularly affecting the IFN-related DNA damage resistance signature (IRDS), which is implicated in tumor cell resistance to both chemotherapy and radiotherapy. Through the application of ruxolitinib, we achieved a significant reduction in ISG levels, thereby enhancing the effectiveness of reovirus therapy. This innovative approach suggests a promising avenue for improving the outcomes of reovirus therapy in tumors exhibiting resistance to the virus, offering potential benefits for patients.

Materials and methods

Cell lines and viruses

We obtained human ovarian cancer cell lines (SKOV3, OV-90, and A2780), human leukemia cell lines (THP-1, HL-60), and lymphoma cell lines (Raji and Daudi) from the Chinese Centre for Typical Culture Collection (CCTCC). Previous research from our group has confirmed that the COC1/DDP ovarian cancer cell line is resistant to cisplatin [23]. Human colon cancer cell lines (SW620, SW480, HT29, and COLO205) were generously provided by Dr. Xiaowei Dou. In this study, the primary tumor cells BC4#, BC5#, BC10#, and BC13# derived from breast cancer patients have been previously confirmed by Dr. Xiaowei Dou’s research to exhibit varying sensitivities to doxorubicin [24]. SKOV3 cells were cultured in McCoy’s 5 A medium, while OV-90 and A2780 cells were cultured in DMEM-F12 medium. COC1/DDP, THP-1, Raji, HL-60, Daudi, SW620, SW480, HT29, and COLO205 cell lines were cultured in RPMI 1640 medium. BC4#, BC5#, BC10# and BC13# cells were cultured in DMEM medium. All media, obtained from Gibco, were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. The cells were incubated at 37 °C in a humidified atmosphere containing 5% CO₂ and 95% air.

The reovirus strain type 3 Dearing (T3D) was acquired from ATCC (VR824), propagated in L929 cells, and its titers were determined using the 50% tissue culture infectious dose (TCID₅₀) method.

Cell viability assay

SKOV3, OV90, A2780, COC1/DDP, THP-1, HL-60, Raji, Daudi, SW620, SW480, HT29, COLO205, BC4#, BC5#, BC10#, and BC13# cells were rinsed with medium and subsequently exposed to 1 and 5 multiplicity of infection (MOI) reovirus for 48 h. SKOV3 cells were pre-treated with 30 ng/mL IFN-β (PeproTech, USA) for 6 h before reovirus exposure [25]. SKOV3 and COC1/DDP cells were pre-treated with 3 µM RO8191 (MCE, USA) or 10 µM IFN-alpha-IFNAR-IN-1 (MCE, USA) for 24 h, followed by exposure to 5 MOI reovirus for 48 h [26, 27]. COLO205, COC1/DDP, and BC10# cells were pre-treated with 3 µM ruxolitinib (MCE, USA) for 48 h, followed by exposure to 5 MOI reovirus for 24 and 48 h [28]. Optical densities (ODs) were measured at 450 nm at designated time points using the Cell Counting Kit-8 (CCK-8) from Meilunbio(China), according to the manufacturer’s instructions. Cell viability was calculated using the following formula: viability rate (%) = (OD _experiment - OD _blank) / (OD _control - OD _blank) × 100%.

Flow cytometry

To assess the sensitivity of tumor cell lines to reovirus, SKOV3、COLO205 and COC1/DDP cells were stained with anti-dsRNA (clone J2, Nordic-Mubio, Netherlands) according to the manufacturer’s protocol. To detect JAM-A expression on the tumor cell surface, FITC-conjugated anti-JAM-A antibody (Abcam, UK) was used to stain the cells according to the manufacturer’s protocol. The stained cells were then collected using a flow cytometer (FC500, Beckman Coulter), and data were analyzed using FlowJo software.

Immunofluorescence

BC10# cells were seeded into culture dishes and treated with either 3 µM ruxolitinib alone, 5 MOI reovirus alone, or a combination of both for 48 h. The tumor cells on coverslips were fixed with 4% paraformaldehyde for 15 min at room temperature. After rinsing with PBS, the cells were permeabilized with 1% Triton X-100 for 30 min at room temperature. Subsequently, the cells were stained with anti-dsRNA (J2) antibody at a 1:100 dilution for 1 h. Following this, the cells were incubated with Alexa Fluor 488-conjugated anti-rabbit antibody at a 1:300 dilution in PBS containing 5% FBS at room temperature. After three PBS washes, the nuclei were stained with DAPI (Thermo Fisher, USA) in the dark. The dishes were then analyzed using a Zeiss LSM780 Laser Scanning Confocal Microscope.

RT-qPCR

Total RNA was extracted with Trizol reagent (Invitrogen, USA) and was subsequently reverse-transcribed into cDNA using a cDNA synthesis kit from Accurate Biology (China). mRNA expression levels of IFNAR1, MX2, ISG15, OAS1-3, T3D, and GAPDH were quantified via real-time PCR with SYBR Green on a Bio-Rad CFX96 system, using reagents from Accurate Biology (China). The PCR primers, detailed in Table S4, were procured from Shanghai Sangon Biotechnology. Data analysis was performed using the Bio-Rad CFX Manager Software version 2.1. Relative expressions of each gene were calculated using the ΔΔCt or 2^−ΔΔCt method and normalized against expression of GAPDH.

Enzyme-linked immunosorbent assay

After infecting SKOV3 and COC1/DDP cells with 10 MOI of reovirus for 24 h, we collected the supernatants. The concentrations of IFN-β in these supernatants were determined using the Human IFN-β Quantification ELISA Kit (Lianke Bio, China), following the manufacturer’s protocol.

Western blot analyses

We collected SKOV3 and COC1/DDP cells and rinsed them with PBS before lysing them using RIPA buffer (Beyotime, China) supplemented with 1mM PMSF (Solarbio, China). The lysates were centrifuged at 12,000 g at 4℃ for 15 min. Approximately 10 µg of protein samples were loaded onto SDS-PAGE and then electro-transferred onto PVDF membranes (Millipore, USA). These membranes were blocked with 5% BSA in TBST and then incubated with primary antibodies overnight at 4℃. The primary antibodies used were as follows: anti-STAT1 (1:1000, Abcam, UK), anti-p-STAT1 (phosphoS727, 1:1000, Abcam, UK), anti-STAT2 (1:5000, Abcam, UK), anti-p-STAT2 (phosphoY690, 1:1000, Abcam, UK), anti-IRF9 (1:1000, Abcam, UK), anti-IFNAR1 (1:5000, Abcam, UK), and anti-actin (1:50000, HuaBio, China). Subsequently, the blots were washed and incubated with secondary antibodies (1:5000, goat-anti-mouse IgG and goat-anti-rabbit IgG, HuaBio, China) at room temperature for 2 h. Signal detection was performed using enhanced chemiluminescence (Natural Biotechnology), followed by film development.

TCID₅₀ assay

The TCID₅₀ assay was employed to measure viral titers in the cell supernatant post-infection. COLO205 and COC1/DDP cells were either untreated or pretreated with 3µM ruxolitinib (MCE, USA) before exposure to reovirus for 48 h. Plates were frozen at -80 °C for 6 h and then thawed with running water, this process was repeated three times. The cell suspension was collected by centrifuging at 1200 rpm for 5 min. Subsequently, L929 cells were seeded into a 96-well plate. The medium was discarded after overnight attachment. Cell supernatants were diluted in a gradient from 10^− 1 to 10^− 8 and then added separately to the culture wells. The cells were continuously observed for 5 days. TCID₅₀ was determined using the Reed and Muench method, as previously described. Finally, cells were stained with 0.1% crystal violet dye and photographed.

RNA-seq analysis

Total RNA extraction was carried out using Trizol reagent (Invitrogen, USA). RNA purification, reverse transcription, library construction, and sequencing were performed at Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China), following the manufacturer’s instructions (Illumina, San Diego, CA). The libraries were constructed using the VAHTS Universal V6 RNA-seq Library Preparation Kit and sequenced on the Illumina Novaseq6000 platform to generate 150 bp paired-end sequences. Functional-enrichment analysis, including Gene Ontology (GO), was conducted to identify significantly enriched differential expression genes (DEGs) in GO terms and metabolic pathways at a Bonferroni-corrected p-value ≤ 0.05 compared with the whole-transcriptome background.

Processing of public gene expression data

We downloaded public gene expression data (GSE28784) from the The Cancer Genome Atlas (TCGA) database. This dataset comprises gene expression profiles from three parental MDA-MB-231 cells and three paclitaxel-resistant MDA-MB-231 cells. To compare the log₂ fold changes between breast cancer cells and paclitaxel-resistant breast cancer cells, we conducted multiple t-test comparisons using the Welch t-test on different genes. We performed statistical analyses using R software (v3.0.2) and GraphPad Prism 9. A significance level of p < 0.05 was applied for all analyses.

Statistical analyses

All experiments were conducted in triplicate, and statistical analyses were performed using GraphPad Prism version 9. Data are presented as mean ± standard deviation (SD). We utilized the Student’s t-test to determine statistical significance, considering a p-value of less than 0.05 as significant. Levels of significance are denoted as follows: *p < 0.05, **p < 0.01, and ***p < 0.001. Additionally, we conducted analyses on the TCGA database using Welch’s t-test and conducted analyses on the TCID₅₀ database using Two-way ANOVA.

Data Availability

The data (PRJNA1068857, GSE28784) utilized in this study are available within the article and its supplementary data files. Datasets and materials can be provided upon request.

Result

Evaluating reovirus oncolytic efficacy and tumor cell sensitivity across varied human cancer cell lines

To assess the oncolytic potential of reovirus, we examined its impact on the viability of various human tumor cell lines cultured in vitro, encompassing hematologic malignancies (THP-1, HL-60, Raji, Daudi) and solid tumor (SKOV3, A2780, OV90, COC1/DDP, SW620, SW480, HT29, COLO205) cell lines. Our previous investigations have identified distinct sensitivities to reovirus among specific cell lines, including SKOV3 and OV90 ovarian cancer cell lines, as well as THP-1 and HL-60 leukemia cell lines [29, 30]. Moreover, the COC1/DDP cell line has been characterized as resistant to cisplatin in ovarian cancer [23]. As depicted in Fig. 1A, these 12 tumor cell lines displayed a dose-dependent response to reovirus, evident 48 h post-infection with 1 and 5 MOI of reovirus. However, sensitivity to reovirus varied considerably among these cell lines. Illustrated in Fig. 1B, based on cell viability 48 h post-infection with 5 MOI of reovirus, the 12 tumor cell lines were categorized into groups of high sensitivity (HS, ≤ 50% viability), moderate sensitivity (MS, 50 ~ 80% viability), and low sensitivity/resistance (LS, ≥ 80% viability). We also evaluated the relative infectivity of cell lines and reovirus. Fig.S1 showed that reovirus infection of HS cells (SKOV3) was stronger than LS cells (COC1 / DDP) within 12 h. When reovirus infection reached 48 h, SKOV3 infected with more reovirus caused cell disintegration and released a large number of progeny viruses. At the same time, there are still a large number of reovirus in COC1/DDP. Consistent with findings from other studies, COLO205 cells are mainly KRAS non-mutated colon cancer cell lines [31]. Our investigation involving 12 tumor cell lines demonstrates that the RAS mutation status and JAM-A expression do not consistently predict tumor cell sensitivity to reovirus oncolysis (Fig. S2).

Fig. 1.

Assessment of reovirus oncolytic sensitivity and IFN signaling in tumor cell lines. A Cell viability was evaluated across a range of tumor cell lines 48 h after treatment with reovirus at 1 and 5 MOI. B Classification of tumor cells based on oncolytic sensitivity: Tumor cells exhibiting cell viability below 50% were categorized as highly sensitive (HS); those with cell viability between 50% and 80% were categorized as moderately sensitive (MS); and those with cell viability above 80% were categorized as low sensitivity (LS). C Relative expression of IFNAR1 mRNA in SKOV3 and COC1/DDP cell lines was measured by RT-qPCR. D Western blot analysis for IFNAR1 and β-actin proteins was conducted in SKOV3 and COC1/DDP cell lines. E SKOV3 cells were pretreated with 30 ng/mL IFN-β for 6 h before treatment with 1 or 5 MOI reovirus for 48 h. Cell viability was assessed using the CCK-8 assay. F SKOV3 and COC1/DDP cells were treated with 10 MOI reovirus for 24 h. IFN-β levels in the supernatant were measured by ELISA. G SKOV3 and COC1/DDP cells were pretreated with 10 µM IFN alpha-IFNAR-IN-1 for 24 h before being treated with 5 MOI reovirus for 48 h. Cell viability was assessed using the CCK-8 assay. Statistical analyses were performed using the Student’s t-test (*p < 0.05, **p < 0.01, ***p < 0.001). Data are presented as mean ± SD

Given the critical role of type I IFN as innate antiviral defenses, our study explored the relationship between type I IFN and tumor cell sensitivity to reovirus-induced oncolysis. We focused on two ovarian cancer cell lines: SKOV3, categorized as highly sensitive, and COC1/DDP, which showed low sensitivity/resistance. Initial tests confirmed consistent expression levels of the type I IFN receptor (IFNAR1) in both cell lines (Fig. 1C-D). Pre-treatment with IFN-β weakly reduced the oncolytic effect of reovirus on SKOV3 cells, as illustrated in Fig. 1E. Following reovirus infection, SKOV3 cells released high levels of IFN-β, whereas COC1/DDP cells released almost none (Fig. 1F). Consequently, the IFN-β produced by SKOV3 cells post-reovirus infection did not restrict extensive killings within the cells. In contrast, despite the deficiency in IFN-β production post-infection, COC1/DDP cells resisted reovirus infection. Moreover, treatment with an IFNAR1 inhibitor (IFN alpha-IFNAR-IN-1) during reovirus infection enhanced the oncolytic efficacy of reovirus in both COC1/DDP and SKOV3 cells (Fig. 1G). These results indicate that elevated levels of type I IFN post-reovirus infection do not hinder its oncolytic effectiveness. We hypothesize that the IFN-triggered JAK-STAT signaling cascade is a key determinant of tumor cell susceptibility to reovirus.

RNA sequencing reveals differential baseline expression of ISGs in tumor cells with varying oncolytic sensitivity

To further investigate, we analyzed RNA-sequencing (RNA-seq) data from cells grouped as HS and LS to understand their different responses to reovirus treatment. The RNA-seq data included samples from SKOV3 and COC1/DDP cells. Employing a differential expression analysis using the R package, we identified differences in gene expression between these groups exhibiting varied oncolytic responses. Principal Component Analysis (PCA) was then utilized to analyze mRNA expression profiles, effectively highlighting the DEGs and their ability to differentiate gene variations between the groups.

Our analysis identified a total of 9,366 DEGs, with 4,716 genes upregulated and 4,650 downregulated. Subsequent GO enrichment analysis of these DEGs revealed their involvement in defense response to virus, type I IFN signaling pathway, and JAK-STAT cascade. Notably, the list of significantly varied genes included key members of the JAK-STAT signaling pathway, such as JAK-1, STAT 1/2/3, and IRF9, as well as other well-recognized antiviral ISGs like IFITM2/3, MX1/2, ISG15, USP18, and OAS1/2/3. Our findings suggest that, in the absence of viral infection, baseline ISG expression levels in COC1/DDP cells are elevated compared to those in SKOV3 cells. As such, we believe that the variation in baseline ISG expression within tumor cells is a significant determinant of their differential oncolytic sensitivity to reovirus. Figure 2 depicts the highlighted results of our analysis.

Fig. 2.

RNA sequencing reveals differential ISGs baseline expression in ovarian cancer cells. RNA-seq was performed on the total RNA that was extracted from SKOV3 and COC1/DDP cells. A PCA analysis of HS cells (SKOV3) and LS cells (COC1/DDP). B Volcano plot of DEGs. Cut-off criteria for the DEGs are when absolute log₂ fold change > 1 and when adjusted p-value < 0.001. C GO pathway enrichment analyses. D Expression of ISG panel. Results are presented as log₂ fold change ± SD of gene expression. ISGs showing more than a 4-fold change in COC1/DDP compared to SKOV3 are presented. E Expression of the JAK-STAT gene presented as log₂ fold change ± SD of gene expression

Tumor cell resistance to reovirus correlates with elevated baseline ISG expression

To validate the results of the RNA-seq analysis, we used RT-qPCR to assess the expression levels of ISG genes, including OAS1, OAS2, OAS3, MX2, and ISG15 in HS and LS tumor cell lines. These selected ISGs, known to possess potent antiviral functions [23–25], showed a significant increase in baseline expression levels in LS tumor cells (COC1/DDP, COLO205, and Daudi) compared to HS tumor cells (SKOV3, SW480, and THP-1), as demonstrated in Fig. 3A.

Fig. 3.

The Impact of JAK-STAT Signaling Pathway-Induced ISGs on Reovirus Replication and Oncolytic Activity. A Relative expression of OAS1/2/3, ISG15, and MX2 mRNA in HS and LS cells measured by RT-qPCR. B COC1/DDP cells were pretreated with RO8191 for 24 h, followed by treatment with 5 MOI reovirus for an additional 24 h. Subsequently, the relative expression of OAS1/2/3, ISG15, and MX2 mRNA in COC1/DDP cells was measured using RT-qPCR. C Quantification of viral RNA levels by RT-qPCR. SKOV3 and COC1/DDP cells were treated as described in (B). D Assessment of cell viability using the CCK-8 assay. SKOV3 and COC1/DDP cells were pretreated with RO8191 for 24 h, followed by treatment with 5 MOI reovirus for 48 h. Statistical analyses were performed using the Student’s t-test (n = 3 per group; *p < 0.05, **p < 0.01, ***p < 0.001). Data are presented as mean ± SD

To delve deeper into the association between ISGs expression and oncolytic resistance, we employed RO8191, a type I IFN receptor agonist. This agonist acts by binding to IFNAR and triggering ISGs expression via the JAK-STAT signaling pathway. As depicted in Fig. 3B, RO8191 treatment significantly increased the expression of OAS1, OAS2, OAS3, MX2, and ISG15 in COC1/DDP cells. However, it’s noteworthy to mention that the expression levels of the aforementioned ISGs in SKOV3 cells remained very low even post-RO8191 treatment, impeding our ability to detect significant expression changes and thus precluding presentation of this data. Therefore, these results further confirm that SKOV3 cells have defects in the IFN-triggered JAK-STAT signaling pathway, resulting in impaired ISG expression and increased sensitivity to reovirus.

However, as shown in Fig. 3C and D, treatment with RO8191 reduced reovirus replication and weakened its ability to kill cancer cells in COC1/DDP cells, though it did not entirely halt the virus’s oncolytic activity. Therefore, we suggest that the increased baseline expression of ISGs in cancer cells, induced by activation of the JAK-STAT signaling pathway, is linked to the cells’ resistance to the oncolytic effects of reovirus, although other mechanisms may also be involved.

The JAK inhibitor ruxolitinib reduces ISG expression and enhances reovirus effects in oncolytic-resistant tumor cells

Previous studies have indicated an association between the elevated baseline expression of ISGs in tumor cells and their resistance to reovirus oncolysis. Therefore, we hypothesized that inhibiting ISG expression might potentially enhance tumor cell sensitivity to reovirus oncolysis. To validate this hypothesis, we utilized a JAK inhibitor, ruxolitinib, which has been reported to inhibit type I IFN signaling and downregulate ISG expression [26–28]. In this study, ruxolitinib effectively reduced the baseline expression levels of OAS1, OAS2, OAS3, MX2, and ISG15 in COC1/DDP cells (Fig. 4A). Consistent with our hypothesis, pretreatment with ruxolitinib for 48 h increased the oncolytic effect of reovirus against COC1/DDP (representing the LS group) and COLO205 (representing the MS group) cells, demonstrating the potential of this approach to enhance tumor cell oncolytic response (Fig. 4B-C).

Fig. 4.

Ruxolitinib enhanced reovirus replication and oncolytic effects in COC1/DDP and COLO205 cells. A Relative expression of OAS1/2/3, ISG15, and MX2 mRNA in COC1/DDP cells measured by RT-qPCR. COC1/DDP cells were pretreated with 3 µM ruxolitinib (Ru) for 48 h and subsequently treated with 5 MOI reovirus for 24 h. B-C Cell viability assessed using the CCK-8 assay. COC1/DDP and COLO205 cells were pretreated with 3µM ruxolitinib (Ru) for 48 h and subsequently treated with 5 MOI reovirus for 48 h. D-E Flow cytometric analysis of dsRNA in COC1/DDP and COLO205 cells. Treatment was the same as in (B-C). After treatment, samples were stained with J2/488 for flow cytometry. Statistical analyses were conducted using the Student’s t-test (n = 3 per group; *p < 0.05, **p < 0.01). Data are presented as mean ± SD. F-G TCID50 assay to determine the titer of reovirus. COC1/DDP and COLO205 cells were treated as in (B-C). The supernatant was then collected for the TCID₅₀ assay. Statistical analyses were conducted using the Student’s t-test (n = 3 per group; *p < 0.05). Data are presented as mean ± SD

Additionally, we investigated the effects of ruxolitinib on reovirus replication and propagation to adjacent cells in COC1/DDP and COLO205 cells. Our approach, involving dsRNA staining and viral titer assessment, provided insights into ruxolitinib’s effects on reovirus replication and dissemination in cell lines. Figure 4D-E illustrates increase in the proportion of dsRNA⁺ cells in COC1/DDP and COLO205 cells post-ruxolitinib treatment, with percentages increasing from 21.57 ± 1.91% to 31.53 ± 2.55% and from 16.37 ± 3.44% to 23.57 ± 4.24% (p < 0.01), respectively. Furthermore, Fig. 4F-G reveals increase in reovirus titers within these cells over time following ruxolitinib treatment. Overall, the observed reduction in basal ISG expression and the subsequent increase in the replication and spread of reovirus highlight the JAK inhibitor ruxolitinib’s potential as a reovirus therapy enhancer.

Exploring the link between ISGs expression and reovirus oncolysis resistance in doxorubicin-resistant patient-derived primary breast cancer cells

The cisplatin-resistant ovarian cancer cell line COC1/DDP exhibits marked resistance to reovirus oncolysis. This resistance is closely associated with the elevated baseline expression of ISGs in these cells. To further clarify this relationship between ISG expression in tumor cells and their resistance to reovirus oncolysis, we aim to validate our results further using clinical tumor samples. This would help provide a closer approximation of actual tumor conditions, which is especially important in examining the potential overlap between chemotherapy resistance and oncolytic resistance. In this study, we selected primary breast cancer cells derived from patients, categorized based on their sensitivity to doxorubicin into doxorubicin-sensitive cells (BC4# and BC5#) and resistant cells (BC10# and BC13#), to enhance our understanding of these relationships [24].

Based on their chemotherapy sensitivity profiles, BC4# and BC5# cells demonstrated higher susceptibility to reovirus oncolysis compared to the resistance observed in BC10# and BC13# cells (Fig. 5A). Importantly, BC10# cells exhibited significant overexpression of ISGs such as OAS1, OAS2, and OAS3, while ISG15 was expressed at lower levels (Fig. 5B). This suggests that ISG15 may not play a crucial role in cellular resistance to reovirus oncolysis in BC10# cells. Additionally, pretreatment with ruxolitinib enhanced the oncolytic effect of reovirus in BC10# cells, as evidenced by increased viral replication within the cells (Fig. 5C-D).

Fig. 5.

Doxorubicin-Resistant Patient-Derived Primary Breast Cancer Cells Exhibit Resistance to Reovirus and High Expression of ISGs. A Viability of patient-derived primary breast cancer cells (BC4#, BC5#, BC10#, and BC13#) treated with 5 MOI reovirus for 48 h, assessed using the CCK-8 assay. B Relative expression levels of OAS1-3 and ISG15 in BC5# and BC10# cells, measured by RT-qPCR. C Viability of BC10# cells pre-treated with ruxolitinib (Ru) for 48 h, then exposed to 5 MOI reovirus for an additional 24 h, assessed using the CCK-8 assay. D Immunofluorescence analysis of dsRNA in BC10# cells treated as described in (C), followed by staining with the J2/488 antibody. The scale bar represents 12.5 μm. Statistical analyses were conducted using the Student’s t-test (n = 3 per group; *p < 0.05, **p < 0.01, ***p < 0.001). Data are presented as mean ± SD

Investigating tumor cells’ cross-resistance to reovirus and chemotherapeutic agents

The observed parallel between resistance to chemotherapeutic agents and reovirus in tumor cells, indicative of ‘cross-resistance,’ suggests common underlying mechanisms for resistance to both chemotherapeutic agents and oncolytic viruses. This connection likely stems from the high baseline expression of ISGs in tumor cells, underscoring a profound interconnectedness in their resistance mechanisms. Previous studies have reported that the interferon-related DNA damage resistance signature (IRDS), identified as a subset of ISGs, plays a crucial role in the cross-resistance to both chemotherapy and oncolytic viruses. Therefore, this suggests that IRDS may be involved in the cross-resistance of tumor cells to reovirus and chemotherapeutic agents [32, 33].

Analysis of RNA-seq data showed that, compared to HS cells, 170 ISGs were upregulated in LS cells (see Table S1). Among these, 28 genes were found to be part of the IRDS (Fig. 6A-C), including genes such as OAS1, OAS2, OAS3, MX2, and ISG15.These genes are known to be highly expressed in cisplatin-resistant COC1/DDP cells and patient-derived primary breast cancer cells.

Fig. 6.

The cross-resistance of tumor cells may be related to the increased expression of IRDS promoted by U-ISGF3. A Previously reported IRDS Genes. B 28 ISGs identified to be significantly upregulated and intersected with previously reported IRDS genes. C Identified 28 ISGs. Results are presented as log₂ fold change ± SD of gene expression. D Expression of STAT1, STAT2, and IRF9 in SKOV3 and COC1/DDP cells assessed through immunoblotting. E The graph represents the quantification of protein levels relative to the control, with normalization against actin (mean ± SD, n = 3, *p < 0.05, **p < 0.01). F Expression of P-STAT1, P-STAT2, STAT1, STAT2 and IRF9 in COC1/DDP cells assessed through immunoblotting.The COC1/DDP cells were pretreated with 5µM NSC87887 for 48 h. G The graph represents the quantification of protein levels relative to the control, with normalization against actin (mean ± SD, n = 3, *p < 0.05, ***p < 0.01). H Expression levels of OAS1-3, ISG15, STAT1, STAT2, and IRF9 in COC1/DDP cells as measured by RT-qPCR. The COC1/DDP cells were pretreated with 5µM NSC87887 for 48 h. Relative gene expression levels were then quantified. I Analysis of gene expression in taxol-resistant MDA-MB-231 cells vs. sensitive MDA-MB-231 cells from the GEO database (GSE28784)

Considering that ISGF3, composed of IRF9 ,tyrosine-phosphorylated STAT1 and STAT2, drives the expression of the full set of ISGs, and that IRDS is believed to be induced by U-ISGF3 (comprising unphosphorylated STAT1, STAT2, and IRF9) [19], we evaluated the expression levels of U-ISGF3 in LS and HS tumor cells. As shown in Fig. 6D-E, compared to SKOV3 cells, the expression levels of STAT1, STAT2, and IRF9 were elevated in COC1/DDP cells, with significant increases in STAT1 and IRF9. Phosphatase inhibitors NSC87877 can specifically inhibit the key phosphatases of tyrosine and serine dephosphorylation [19].We used the phosphatase inhibitor NSC87877 to specifically target the key phosphatases responsible for dephosphorylating p-STAT1 and p-STAT2 at both tyrosine and serine residues. Figure 6F demonstrated that pretreatment with the phosphatase inhibitor NSC87877 specifically elevated the protein levels of p-STAT1 and p-STAT2. However, we observed that NSC87877 did not affect the basal expression levels of the U-ISGF3 and IRDS(Fig. 6F-H).

Additionally, we analyzed the GSE28784 dataset from the GEO database, which includes data on sensitive, docetaxel- and paclitaxel-resistant MDA-MB-231 cells. Our analysis revealed that compared to the sensitive MDA-MB-231 cells, the paclitaxel-resistant MDA-MB-231 cells exhibited higher expression of the genes OAS1, OAS2, OAS3, ISG15, MX2, STAT1, STAT2, and IRF9 (Fig. 6I).

In conclusion, these findings suggest that U-ISGF3 may drive the expression of IRDS, potentially leading to cross-resistance against both chemotherapeutic agents and reovirus.

Discussion

Tumor cell sensitivity to OVs is critical for the success of virotherapy. This sensitivity is shaped by multiple factors across the viral replication cycle, which encompasses four key phases: entry, viral protein and nucleic acid synthesis, new virion assembly, and eventual release. For efficient replication, tumor cells must fulfill specific requirements [34], including surface expression of oncolytic virus-specific receptors to facilitate viral entry. Additionally, certain oncolytic viruses require tumor cells to produce specific enzymes that modify or cleave viral proteins—critical for generating mature, infectious virions [16, 35–37].

In our previous research, we found that THP-1 cells are more sensitive to reovirus than HL-60 cells. Although we believe this difference is linked to the higher expression of JAM-A on THP-1 cells [29], the current study indicates that JAM-A expression does not reliably predict tumor cell sensitivity to reovirus [38–40]. This implies that other factors might affect susceptibility to reovirus. Type I IFN, including IFN-α and IFN-β, can induce the expression of hundreds of ISGs, forming the primary defense against viral infections. The IFN response has two stages: an early phase that produces IFN-β, and a later phase involving IFN signaling [41]. IFN-β, once released, can bind to the IFNAR receptor through autocrine or paracrine pathways, activating the JAK-STAT signaling cascade, resulting in ISGs expression. Tumor cells with defects in their capacity to produce or respond to type I IFN tend to be more susceptible to OVs than normal cells [42–45].Impaired interferon signaling is common in various cancer types, with many cancer patients showing low expression of IFN receptors or genes, leading to unchecked tumor proliferation [46]. Many RNA viruses, like newcastle disease virus, measles virus and vesicular stomatitis virus, are particularly effective in targeting IFN-deficient tumor cells [47–50]. Interestingly, our research showed that despite producing high IFN-β levels, SKOV3 cells remain highly susceptible to reovirus—suggesting their defect lies in IFN signaling rather than IFN-β production. In contrast, COC1/DDP cells are reovirus-resistant despite defective IFN-β production. Collectively, these findings indicate that type I IFN responsiveness—including activation of the IFN-induced JAK-STAT signaling cascade and subsequent ISG expression—may be the key determinant of tumor cell susceptibility to reovirus oncolysis.

RNA-seq analysis revealed substantial baseline differences in ISG expression across tumor cell lines. In reovirus-resistant or low-sensitivity cells (e.g., COC1/DDP), we detected marked upregulation of ISGs, including OAS1, OAS2, OAS3, ISG15, and MX2. This upregulation was further enhanced by JAK-STAT pathway activation with RO8191, leading to reduced reovirus replication and impaired oncolytic efficacy. Notably, RO8191-mediated JAK-STAT activation alone was insufficient to fully abrogate reovirus oncolysis, indicating that ISG-upregulating mechanisms independent of JAK-STAT activation may also drive reovirus resistance. The type I IFN response comprises two distinct phases, reflecting the intricate regulatory network governing ISG expression in the context of oncolytic sensitivity and resistance. Initially, type I IFN activates JAK-STAT signaling, leading to formation of the ISGF3 complex—composed of IRF9 and tyrosine-phosphorylated STAT1/STAT2—which induces the expression of a diverse set of ISGs. To prevent cellular damage from prolonged antiviral activity, robust negative feedback mechanisms typically restrict sustained STAT1/STAT2 phosphorylation within 24 h. Importantly, STAT1, STAT2, and IRF9 are themselves ISGs; in their non-phosphorylated state, they form the U-ISGF3 complex, which sustains expression of a subset of ISGs (termed IRDS) linked to resistance to radiotherapy and chemotherapy [33, 51–53].

In this study, we found that the cisplatin-resistant ovarian cancer cell line COC1/DDP is also reovirus-resistant. We identified 28 highly expressed ISGs—including OAS1, OAS2, OAS3, ISG15, and MX2—previously reported as components of the IRDS [52]. Notably, COC1/DDP cells showed elevated levels of STAT1, STAT2, and IRF9 (components of the U-ISGF3 complex) [54], and IRDS expression was unaffected by treatment with the phosphatase inhibitor NSC87877. To further validate IRDS’s role in mediating tumor cell cross-resistance to chemotherapy and reovirus, we analyzed patient-derived primary breast cancer cells—both doxorubicin-sensitive and doxorubicin-resistant. Notably, OAS1, OAS2, and OAS3 expression was significantly upregulated in doxorubicin-resistant cells.

Based on these findings, we hypothesize that ISG expression suppression increases tumor cell sensitivity to reovirus, thereby enhancing its oncolytic efficacy. Previous studies show that pretreatment with ruxolitinib—an FDA-approved JAK inhibitor for myeloproliferative neoplasms (e.g., polycythemia vera, myelofibrosis) [55–57], reduces ISG expression and enhances susceptibility of malignant peripheral nerve sheath tumors to oncolytic HSV [58]. In vitro, ruxolitinib-mediated type I IFN pathway inhibition overcame resistance and enhanced viral replication in primary patient-derived GBM xenograft cells [28]. Consistent with these reports, our study demonstrated that ruxolitinib pretreatment markedly reduced ISG levels in both COC1/DDP cells and doxorubicin-resistant primary patient-derived breast cancer cells, thereby boosting reovirus replication and oncolytic activity. We investigated whether U-ISGF3 mediates tumor cell resistance to reovirus by selectively regulating IRDS (a subset of ISGs), which may impair reovirus infection or replication. Our findings demonstrate that ruxolitinib effectively reduces ISG expression, thereby enhancing reovirus infection, replication, and oncolytic efficacy in tumor cells (Fig. 7).

Fig. 7.

Schematic depicting the relationship between ISG expression and reovirus resistance in tumor cells, and strategies to improve oncolytic resistance. The expression of the ISG subgroup, IRDS, is associated with cross-resistance of tumor cells to reovirus and chemotherapeutic agents. U-ISGF3 may have promoted IRDS expression. JAK inhibition by ruxolitinib can reduce ISG expression in tumor cells, enhance reovirus replication and oncolysis, and improve the oncolytic resistance of tumor cells

Notably, our study has limitations. While ruxolitinib combination therapy boosts reovirus replication and oncolytic activity, its efficacy in animal models remains to be evaluated. Given tumor cells’ cross-resistance to reovirus and chemotherapeutics, future studies should explore whether combining chemotherapeutics with ruxolitinib yields similar therapeutic benefits. Additionally, PD-L1 is an ISG, and RNA-seq analysis revealed high its expression in COC1/DDP cells—future work should assess whether adding immune checkpoint inhibitors (e.g., anti-PD-1 therapy) to reovirus plus JAK inhibitors induces more robust antitumor effects. Beyond ruxolitinib, other interferon modulators are widely used in oncolytic virotherapy. For example, histone deacetylase (HDAC) inhibitors suppress type I IFN signaling to reduce cellular antiviral immunity, enhancing vesicular stomatitis virus (VSV) oncolysis in multiple cancer cell types [59]. Valproic acid pretreatment of glioma-derived cells inhibits antiviral genes (e.g., STAT1, PKR, PML), indirectly augmenting oncolytic herpesvirus efficacy [60]. Specific STAT1 and NF-κB inhibitors reduce baseline ISG expression in malignant peripheral nerve sheath tumor cells, improving HSV oncolytic therapy [61]. Notably, in rats with malignant glioma, rapamycin combined with VSV-DM51 reduces tumor volume and prolongs survival [48]. Collectively, combining oncolytic viruses (OVs) with various IFN modulators enhances viral therapy efficacy by inhibiting IFN production or cellular antiviral responses.Additionally, targeted inhibition of individual components within the JAK/STAT pathway using small interfering RNA (siRNA) allows for a more precise investigation into the impact of pathway-induced ISGs on the oncolytic viral effect. Currently, there is no universal consensus on the optimal siRNA targets for ISG suppression: knockdown of a single component only partially reduces ISG expression levels, whereas combinatorial knockdown carries the risk of off-target effects or cytotoxicity. This stands in contrast to the mode of action of ruxolitinib, which enables potent and specific inhibition of the JAK/STAT pathway. Nevertheless, siRNA holds substantial application value for follow-up studies, and in our future research, we intend to employ siRNA to further explore the roles of JAK/STAT pathway components and ISGs in modulating the oncolysis effect.

Conclusions

In summary, our study demonstrates that tumor cross-resistance to reovirus and chemotherapeutics is associated with elevated baseline ISG expression, likely mediated via U-ISGF3. Future work will focus on identifying key ISGs driving reovirus resistance and exploring whether ISG inhibition enhances reovirus therapeutic efficacy. Additionally, we aim to assess the feasibility and clinical relevance of using ISGs as predictive biomarkers of oncolytic sensitivity to identify patients most likely to benefit from reovirus therapy.

Abbreviations

PCA

Principal Component Analysis

ISGs

Interferon-stimulated genes

JAM-A

Junctional adhesion molecule A

T-VEC

Talimogene laherparepvec

GO

Gene ontology

IRDS

Interferon-related DNA damage resistance signature

JAK-STAT

Janus tyrosine kinases-signal transducers and activators of transcription

U-ISGF3

Unphosphorylated ISGF3

type I IFN

Type I interferons

DEG

Differential expression gene

TCGA

The Cancer Genome Atlas

Author contributions

Conceptualization, Y.X.Y., X.J.L, Y.T.F and X.Z.; methodology, Y.X.Y, X.J.L and D.L; formal analysis, Y.X.Y, X.J.L, D.L. and X.Z; data curation, Y.X.Y; writing original draft preparation, Y.X.Y; writing—review and editing, X.Z.; visualization, L.C., J.Z. and Y.C.Z.; supervision, X.W.D, L.C., J.Z. and Y.C.Z.; funding, L.C., X.Z.; All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant Number: 82060564), Key Program for Science and Technology of Guizhou Province (Grant Number: ZK (2021)012) and Program for Tumor Immunotherapy Technology Engineering Research Center (University Engineering Center 2024 [001]) provided to X. Zhao. Funded by the Guizhou Provincial Health Commission Science and Technology Fund Project (gzwkj2022-087) provided to L. Chen.

Data availability

Data supporting the findings of this study are included in the published article and its supplementary materials. Additional data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors have approved the final paper for publication.

Competing interests

The authors declare no competing interests.