Targeting cancer stem cells predicts response and reverses chemoresistance in ascites-derived ovarian cancer organoids

Abstract

Background

Ovarian cancer (OC) is frequently diagnosed at an advanced stage, where tumor heterogeneity and rapid development of chemoresistance contribute to a poor prognosis. The lack of reliable predictive biomarkers further hinders the development of effective treatment strategies. Patient-derived organoids (PDOs) have recently emerged as promising preclinical models with the potential to predict therapeutic responses.

Methods

OC PDOs were generated from ascites samples representing diverse histological subtypes. Histological and genomic fidelity to parental tumors was confirmed through histopathological analysis and whole-exome sequencing. Drug sensitivity to cisplatin and poly (ADP-ribose) polymerase (PARP) inhibitors was evaluated and correlated with 1-year clinical outcomes. We also investigated the therapeutic efficacy of oncolytic herpes simplex virus 2 (OH2) both as a single agent and in combination with cisplatin. The expression of cancer stem cell (CSC) markers CD44 and ALDH1A1 under treatment conditions was analyzed using immunohistochemistry and flow cytometry.

Results

PDOs were successfully established with an 86.2% success rate. These PDOs faithfully recapitulated the histopathological and genomic features of their corresponding tumors, maintaining intratumoral heterogeneity, and were amenable to xenotransplantation. Drug sensitivity assays demonstrated that PDOs accurately predicted patient-specific responses to cisplatin and PARP inhibitors. OH2 exhibited direct cytotoxicity in both cisplatin-sensitive and cisplatin-resistant PDOs, reducing cell viability by 20–60%. Notably, the combination treatment with OH2 and cisplatin enhanced antitumor efficacy, resulting in a significant reduction of the CD44⁺CSC subpopulation.

Conclusions

Ascites-derived OC PDOs represent a robust platform for individualized drug testing. The combination of OH2 and cisplatin offers a novel and effective strategy for circumventing chemoresistance in OC.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12916-026-04730-1.

Background

Ovarian cancer (OC) is a highly heterogeneous malignancy, with a small fraction originating from germ cells in the ovary, primarily affecting women under 40. More than 90% of cases are epithelial ovarian cancer (EOC) [1–3]. Due to its insidious early symptoms and the lack of effective early screening methods, most patients are diagnosed at an advanced stage (FIGO III–IV), with widespread metastasis and dissemination, significantly increasing the difficulty of treatment. Currently, cytoreductive surgery, adjuvant platinum-based chemotherapy, and bevacizumab are used as standard treatments for OC. Despite good initial treatment responses, approximately 15% of patients exhibit primary resistance [4, 5], and 70–80% relapse after treatment, further developing into chemotherapy-resistant recurrent advanced ovarian cancer, with a low 5-year survival rate of 25–35% [6, 7]. Although PARP inhibitors and/or anti-angiogenic agents are now used for maintenance therapy, drug resistance remains a significant challenge. Therefore, overcoming chemotherapy resistance and developing more effective treatment strategies are crucial to improving the survival of OC patients.

Significant progress has been made in OC research over the past few decades. Genomic studies have revealed the complex molecular heterogeneity of OC, identifying mutations in key genes such as TP53, BRCA1/2, and PIK3CA, leading to the development of targeted therapies. However, these targeted drugs are only effective in a subset of patients and are prone to resistance. The OC tumor microenvironment (TME) typically exhibits an immunosuppressive state, characterized by significant inter- and intratumoral heterogeneity (ITH) [8]. The interaction between various components of the TME collectively influences tumor growth, metastasis, and response to treatment. Oncolytic virotherapy, as an emerging tumor biotherapy, selectively infects and replicates in tumor cells, inducing direct tumor cell lysis while simultaneously activating antitumor immunity, thereby representing a promising strategy to overcome TME-mediated immunosuppression. Both preclinical and clinical studies have reported encouraging results for OV therapy in OC [9–15]. For example, Basnet et al. demonstrated that treatment with the oncolytic adenovirus TILT-322 reduced the proportion of exhausted CD8⁺ T cells and enhanced T-cell activation in ascites samples from patients with OC [12].

Furthermore, traditional in vitro cell line models and in vivo animal models fail to accurately mimic the complex biological characteristics of OC and cannot effectively predict patient response to treatment, leading to a high drug development failure rate. Patient-derived organoids (PDOs) can simulate tumor heterogeneity and faithfully recapitulate the genomic characteristics of the original tumor, and are widely used for individualized drug screening [16–20]. However, most OC organoids rely on fresh surgical specimens, which limits their clinical translation. This study established OC organoids of various subtypes derived from ascites samples and conducted drug sensitivity testing. Through a 1-year clinical follow-up, the consistency between PDO drug sensitivity results and clinical efficacy was verified. Simultaneously, we evaluated the antitumor activity of oncolytic herpes simplex virus 2 (OH2) in OC organoids for the first time and preliminarily explored the therapeutic potential of combining OH2 with chemotherapy. In addition, we used immunohistochemistry (IHC) to analyze the expression differences of tumor stem cell (CSC) markers CD44 and ALDH1A1 in platinum-sensitive and resistant patients, and evaluated the effect of OH2 combined with cisplatin treatment on the proportion of CD44⁺ CSC subsets. We hope that this study will provide a theoretical basis for individualized treatment and potential new strategies for OC.

Methods

Patient samples

From January 2023 to May 2024, ascites samples were collected from ovarian cancer patients at the Cancer Hospital, Chinese Academy of Medical Sciences. Inclusion criteria were as follows: (1) malignant effusions obtained using sterile drainage bags; (2) age ≥ 18 years; (3) availability of residual clinical specimens after diagnostic procedures; and (4) newly diagnosed ovarian malignancies with no prior treatment. The study was approved by the Ethics Committee of the Cancer Hospital, Chinese Academy of Medical Sciences (NCC2022C-478), and written informed consent was obtained from all participants.

Organoid generation and culture

Ascites samples were transferred to 50 mL centrifuge tubes and transported on ice. Cells were pelleted by centrifugation at 1500 rpm for 10 min. Red blood cell lysis was performed for 5–10 min if needed. The resulting cell pellets were resuspended in advanced DMEM/F-12+++ (advanced DMEM/F-12 supplemented with 1% penicillin–streptomycin (PS), 1 × Glutamax, and 1% HEPES). The cell suspension was then mixed with Culture Reduced Growthfactor Basement Membrane Extract (BME) (R&D Systems, cat. no. 3533–010-02) to a final BME concentration of 70% BME. Approximately 40,000 cells or cell clusters were embedded in each 40 μL BME droplet and plated into 6-well plates. After BME solidification, 1.5 mL of ovarian cancer organoid culture medium was added to each well.

The complete organoid culture medium was composed of advanced DMEM/F-12 (Thermo Fisher Scientific, cat. no. 12634028) supplemented with the following components: 1% PS (Thermo Fisher Scientific, cat. no. 15140–122), 1 × Glutamine (Thermo Fisher Scientific, cat. no. 12634–034), 1% HEPES (Thermo Fisher Scientific, cat. no. 15630–056), 250 ng/mL R-spondin1 (PeproTech, cat. no. 120–38), 100 ng/mL Noggin (PeproTech, cat. no. 120-10C), 5 ng/mL EGF (Novoprotein, cat. no. C029), 100 ng/mL FGF-10 (PeproTech, cat. no. 100–26), 25 ng/mL FGF-2 (PeproTech, cat. no. 100-18B), 1 × B27 (Thermo Fisher Scientific, cat. no. 17504–44), 5 mM nicotinamide (Sigma-Aldrich, cat. no. N0636), 1.25 mM N-acetyl-L-cysteine (Sigma-Aldrich, cat. no. A9165-5G), 10 μmol/L SB202190 (Sigma-Aldrich, cat. no. S7076), 500 nM A83-01 (Tocris Bioscience, cat. no. 2939), 5 μM Y-27632 (AbMole Bioscience, cat. no. M1817), 100 ng/mL Wnt3a (PeproTech, cat. no. 315–20), and 5 nM Neuregulin-1 (Univ, cat. no. 208628).

Whole-exome sequencing

Genomic DNA was extracted from ascites-derived cells, matched organoids, and patient-matched normal lymph node tissues. WES was conducted by Beijing Geneplus Technology Co., Ltd (Beijing, China). Somatic mutations were filtered based on quality metrics including base quality scores, sequencing depth, genomic position, and population allele frequency. SNVs and Indels were detected using MuTect2 [21]. Somatic CNVs were analyzed in matched samples using GATK software [22]. Finally, the R package maftools was employed to generate a comprehensive and personalized visualization of somatic mutations across samples [23].

Establishment of organoid-derived xenografts

Mature organoids were dissociated into single cells using TrypLE Express (Thermo Fisher Scientific, cat. no. 12605010). A total of 100 μL of the organoid suspension was mixed with BME and subcutaneously injected into the right flanks of 6-week-old BALB/c nude mice (SiPeiFu Biotechnology). Tumor growth was monitored twice per week using vernier calipers. All animal procedures were performed in accordance with protocols approved by the Experimental Animal Ethics Committee of the Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (NCC2024A536).

Drug sensitivity assays

Organoids were digested using TrypLE Express and resuspended in advanced DMEM/F12+++ medium. Cell concentrations were measured using Cellometer® Auto 2000. For OH2 infection, cells were incubated in Advanced+++ medium with OH2 (MOI = 1) for 2 h, then centrifuged at 1200 rpm to collect cell pellets. Cells were mixed with BME at a density of 5000 cells/10 μL and seeded into pre-warmed 96-well plates. After matrix solidification (37 °C, 5% CO₂, 30 min), 200 μL of complete ovarian cancer organoid medium treated with various agents (cisplatin, niraparib) were added. Organoids were treated with cisplatin or niraparib for 72 h to determine their IC50 values. For OH2 monotherapy, cell viability was evaluated at 0, 24, and 48 h post-infection, whereas for combination treatment, organoids were exposed to OH2, cisplatin, or their combination for 48 h before viability assessment. Following drug exposure, the medium was replaced with fresh medium containing 20% CellTiter-Blue® reagent (Promega, cat. no. G8080), and plates were incubated at 37 °C, 5% CO₂ for 2 h in the dark. Fluorescence was measured at 560/590 nm using a Synergy H1MFD microplate reader (BioTek).

H&E staining, immunohistochemistry, and immunofluorescence

Ascites-derived cells and organoids were fixed in 4% paraformaldehyde, processed through dehydration, paraffin embedding, sectioning, followed by H&E staining. For IHC, markers were selected based on histologic subtype. Ready-to-use solutions were used throughout: EpCAM (Genetech, cat. no. GM080404) to label epithelial cells; PAX8 (Zhongshan Golden Bridge Biotechnology, cat. no. ZM-0468) and WT1 (Zhongshan Golden Bridge Biotechnology, cat. no. ZA-0559) to identify high-grade serous ovarian carcinoma (HGSOC); Napsin A (Fuzhou Maixin Biotech, cat. no. MAB-0704) and HNF-1β (Fuzhou Maixin Biotech, cat. no. MAB-0884) for ovarian clear cell carcinoma (OCCC); and SALL4 (Genetech, cat. no. GT2283) and AFP (Genetech, cat. no. GA0008) for yolk sac tumors.Stemness markers included CD44 markers (1:100, Nature Biosciences, cat. no. A63239) and ALDH1A1 (1:100, Nature Biosciences, cat. no. A15511). For immunofluorescence, samples were stained with anti-CD44 antibodies, followed by Alexa Fluor 488- and 568-conjugated secondary antibodies.

Flow cytometric analysis of CD44 expression

Following 48h treatment with cisplatin, OH2, or their combination, organoids were harvested using TrypLE Express. Cells were washed with PBS and stained with APC-conjugated anti-CD44 antibody (BioLegend, cat. no. 103012) for 15 min on ice in the dark. Flow cytometry was performed using a BD LSRII instrument, and results were analyzed with FlowJo software.

Statistical analysis

All statistical analyses were conducted using R software (version 4.3.2; R Foundation for Statistical Computing). Intergroup comparisons were performed using appropriate statistical tests, including multiple comparison tests and ANOVA. Experiments were independently repeated at least three times. Statistical significance was defined as p < 0.05, with significance levels indicated as follows: ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Results

Characteristics of patients with ascites-derived OC organoids

Between January 2023 and May 2024, we collected ascitic fluid samples from 58 patients with newly diagnosed ovarian cancer. Among them, 51 were HGSOC, 2 had poorly differentiated endometrioid carcinomas, 2 had OCCC, 1 was a yolk sac tumor, and 2 were low-grade serous ovarian carcinomas (LGSOC), as illustrated in Fig. 1. Organoids were successfully established from 50 of the 58 cases, including 47 HGSOC, 2 OCCC, and 1 yolk sac tumor, resulting in an overall establishment success rate of 86.2% (50/58). Drug sensitivity testing to cisplatin was conducted on the successfully established organoids. In parallel, a 1-year clinical follow-up was performed for the corresponding patients. Based on the platinum-free interval (PFI), patients were classified into three categories: platinum-sensitive (PFI > 12 months), partially platinum-sensitive (PFI 6–12 months), and platinum-resistant (PFI < 6 months) [24]. By the end of the follow-up period, 3 patients were identified as platinum-sensitive and 7 as platinum-resistant. Among the platinum-resistant cases, 2 were primary platinum-refractory, experiencing disease progression during the initial course of platinum-based chemotherapy. The remaining 5 patients were platinum-resistant, with disease relapse occurring within 6 months after completion of their last platinum-based regimen. Detailed clinical information of the patients is presented in Table 1.

Fig. 1.

Flowchart of clinical follow-up and OC organoids drug sensitivity tests in this study

Table 1.

Clinical characteristics of patients

ID	Pathological subtype	Age (years)	Clinical stagea	CA125 (U/ml)	Neoadjuvant chemotherapy regimen	PFI (month)	Platinum sensitivity	PARPi maintenance regimen	PARPi sensitivity	Duration of PARPi response (month)
P01	HGSOC	64	IIIC	5111	Liposomal doxorubicin + carboplatin	> 12	Sensitive	Olaparib	Sensitive	24
P07b	HGSOC	62	IIIC	2893	Liposomal doxorubicin + carboplatin	> 12	Sensitive	Niraparib	Sensitive	20
P10	HGSOC	65	IIIC	1295	Abraxane + carboplatin	5	Resistant	Niraparib	Resistant	2
P11c	HGSOC	65	IVA	723	Paclitaxel + nedaplatin	5	Resistant	-	-
P23d	OCCC	57	IIIC	19.8	-	Primary resistance	Resistant	-	-
P26	HGSOC	50	IIIC	6169	Liposomal paclitaxel + carboplatin	5	Resistant	Niraparib	Resistant	1
P32e	OCCC	54	IIIC	306	-	4	Resistant	-	-
P37f	HGSOC	69	IIIC	1450	Liposomal paclitaxel + carboplatin	> 12	Sensitive	Niraparib	Sensitive	12
P51	HGSOC	69	IVA	5884	Liposomal paclitaxel + carboplatin	primary resistance	Resistant	Niraparib	Resistant	3
P52	HGSOC	41	IIIC	1804	Liposomal paclitaxel + carboplatin	1	Resistant	Niraparib	Resistant	2

HGSOC high-grade serous ovarian carcinoma, OCCC ovarian clear cell carcinoma, PARPi poly (ADP-ribose) polymerase inhibitor, PFI platinum-free interval

^aClinical stagedetermined according to the surgical-pathological staging system of the International Federation of Gynecology and Obstetrics (FIGO)

^b,fThe patient remained progression-free on PARPi at the end of follow-up

^cThe patient was not receive maintenance therapy with a PARPi

^d,eNeoadjuvant therapy was not administered before surgery. After the operation, the patient received platinum-based chemotherapy in combination with targeted therapy

Morphological and histological characterization of OC organoids

We obtained clinical diagnostic residual ascitic fluid samples from OC patients undergoing initial treatment with ascitic drainage, which were then used for organoid culture. The schematic of the culture process is shown in Fig. 2A. To assess the growth dynamics of the organoids, we conducted a longitudinal observational study (Fig. 2B). We observed that tumor cell aggregates from the ascitic fluid transformed into spherical structures within 2 days, and by day 12 of cultivation, the longest diameter of individual organoids reached up to 100 μm. HGSOC organoids (P37, P51) exhibited both cystic and solid structures, while OCCC organoids (P23) displayed dense structures with prominent lumens. Additionally, we observed epithelial folds and invaginations in both HGSOC and OCCC organoids. To compare the organoids with their corresponding parental tumors, we performed H&E staining on 2 cases of HGSOC, 1 case of OCCC, and 1 case of yolk sac tumor, and assessed the expression of OC protein biomarkers. Notably, the tumor organoids were composed of transformed tumor epithelial cells, without immune, vascular, or stromal components. The H&E staining of the OC organoids revealed a morphology consistent with that of the original tumors, showing marked nuclear pleomorphism, prominent nucleoli, and dense chromatin. We used EpCAM to label epithelial-derived tumor cells, all of which showed robust EpCAM positivity (Fig. 2C–E). IHC analysis of HGSOC organoids matched that of the parental tumors. with retention of PAX8 and WT1 expression (Fig. 2C). The OCCC organoids (Fig. 2D) were positive for HNF-1β and negative for Napsin A, a result consistent with the relatively lower sensitivity of Napsin A in diagnosing OCCC [25]. The yolk sac tumor organoids exhibited high expression of SALL4 and AFP (Fig. 2E), confirming their origin as germ cell-derived tumors. Thus, the histological analysis of the OC organoids demonstrated their similarity to the cancerous regions of the corresponding primary tumors.

Fig. 2.

Bright-field imaging and histological features of OC organoids. A Schematic diagram illustrating the culture model of organoids derived from OC ascites. B Representative morphology of long-term cultured organoids from HGSOC (P37, P51) and OCCC(P23). Scale bars, 100 μm. C Histological comparison between HGSOC organoids and their corresponding parental tumors (P37: top; P51: bottom). H&E staining of the P37 organoid line shows papillary-like structures protruding into the lumen, forming a dense phenotype. Scale bars, 50 μm. D H&E and IHC staining with HNF-1β, Napsin A and EpCAM demonstrate histological consistency between cultured OCCC (P23) organoids and their corresponding parental tumors. Scale bars, 50 μm. E H&E and IHC staining with SALL4, AFP and EpCAM show similar histological features between YST (P41) organoids and their parental tumors. Scale bars, 50 μm. Figure 2A was created using BioRender

Ovarian cancer organoids retain the genetic characteristics of ascites-isolated tumor cells

To assess whether PDOs faithfully recapitulate the genomic landscape of their parental tumors, we performed WES on organoids derived from HGSOC (P26, P51, P52, P58) and OCCC (P23), along with their matched ascites tumor cells and the normal lymph node FFPE tissues. SNV and CNV analyses revealed a high degree of concordance between the organoids and their parental tumors. CNV profiles demonstrated strong consistency in chromosomal breakpoints, chromothripsis-like rearrangements, and ploidy changes between organoid–tumor pairs (Fig. 3A–D, Additional file 1: Fig. S1). For somatic SNV analysis, we used MuTect2 and annotated variants via ANNOVAR. After filtering, mutation frequencies were calculated, and representative mutations commonly observed in HGSOC and OCCC were visualized using oncoplots. Basic SNV information is provided in Additional file 1: Table S1. Detected mutation types included frameshift deletions, splice-site alterations, multi-hit mutations, and missense variants. TP53 mutations were detected in the HGSOC samples we analyzed, while OCCC-derived organoids retained recurrent ARID1A mutations, often presenting multi-hit patterns (Fig. 3E). Base substitution analysis (Fig. 3F) revealed conserved mutational signatures across organoid–tumor pairs, including C > A, C > G, C > T, T > C, T > A, and T > G transitions, with C > T substitutions being the most predominant. Furthermore, we evaluated the tumorigenic potential of ovarian cancer organoids in immunodeficient mice. Organoids were subcutaneously transplanted into BALB/c nude mice, and their growth was monitored. As shown in Fig. 3G, tumors formed successfully approximately 4 weeks post-transplantation. Collectively, these results indicate that short-term cultured OC organoids retain the somatic genomic features of their parental tumors without acquiring additional mutations, supporting their value as ex vivo models for functional and therapeutic studies.

Fig. 3.

The concordance for genomic profiling between ascites and ascites-derived organoids. A–D CNV analysis of organoids and matched ascites samples using GATK software. E The representative mutant genes are obtained by analyzing the SNV of organoids and their parental tumors. F Distribution and classification of mutation types and their proportions. G Tumor formation from the OC organoids was observed in three BALB/c nude mice. P26-A: ascites of P26; P26-O: organoids of P26; P51-A: ascites of P51; P51-O: organoids of P51; P52-A: ascites of P52; P52-O: organoids of P52. P58-A: ascites of P58; P58-O: organoids of P58; P23-A: ascites of P23; P23-O: organoids of P23

Drug sensitivity testing in OC organoids

We next assessed the sensitivity of OC organoids to platinum-based chemotherapy. Based on clinical follow-up data, organoids were categorized into platinum-sensitive, partially platinum-sensitive groups, and platinum-resistant groups. For the cisplatin sensitivity assay, we specifically included organoids derived from platinum-sensitive and platinum-resistant patients. Cisplatin IC50 values demonstrated a significant difference between the platinum-sensitive group and the platinum-resistant group (Fig. 4A). In the platinum-sensitive group (P01, P07, and P37), IC50 values were 1.708 μg/mL, 2.634 μg/mL, and 5.586 μg/mL, respectively. In contrast, all organoids from platinum-resistant patients, except P11, had IC50 values exceeding 10 μg/mL.

Fig. 4.

Drug sensitivity and CSC marker expression in OC organoids. A Dose–response curves of organoids treated with cisplatin for 72 h. The dashed line indicates 50% cell viability, illustrating the IC₅₀ value determined for each patient-derived organoids. B Dose–response curves of organoids treated with niraparib for 72 h. The dashed line represents 50% cell viability, indicating the IC₅₀ value for each patient-derived organoid. C IHC staining of CSC markers CD44 and ALDH1A1 in organoids derived from cisplatin-sensitive patients (P01, P07, and P37) and cisplatin-resistant patients (P10, P26, P51, P32, P11, P23, and P52). Scale bars, 20 μm. D Immunofluorescence staining of CD44 in organoids from a cisplatin-sensitive patient (P07) and cisplatin-resistant patients (P23, P52). Scale bars, 20 μm. E Time-dependent response curve of OC organoids infected with OH2 at a MOI of 1. Each data point represents the mean of three biological replicates, with error bars indicating the standard deviation (mean ± SD). Cisplatin-S: cisplatin-sensitive; Cisplatin-R: cisplatin-resistant

Among eight HGSOC patients with known responses to platinum therapy, seven received maintenance treatment with PARP inhibitors. Their clinical responses to PARP inhibitors correlated with their in vitro cisplatin sensitivity. Organoid-based PARP inhibitors testing further corroborated these observations. For instance, P01, who relapsed after 24 months of olaparib treatment, exhibited a niraparib IC50 of 59.91 μM. In contrast, P07 and P37, who remained progression-free after 20 and 12 months of niraparib maintenance, showed IC50 values of 59.91 μM and 114.7 μM for niraparib, respectively. By comparison, platinum-resistant patients receiving niraparib all progressed within 3 months (Fig. 4B). These results highlight the potential of organoids to predict PARP inhibitor response, although the samples were derived from treatment-naïve ascites, which may not fully reflect tumor biology at the time of PARP inhibitor administration.

Given the critical role of cancer stem cells (CSCs) in metastasis, chemoresistance, and recurrence, we investigated CSC marker expression in OC organoids using immunohistochemistry. CD44 expression was ubiquitous but significantly higher in organoids derived from platinum-resistant patients compared to sensitive ones. ALDH1A1 expression was observed in only a subset of resistant organoids (P11, P32, and P52). Notably, P11-derived organoids displayed marked intra-sample heterogeneity in CD44 expression, with subsets of cells exhibiting low versus high expression levels (Fig. 4C), suggesting that organoids partially retain the intratumoral heterogeneity of HGSOC. Figure 4D presents the immunofluorescence results of CD44 from P07, P23, P52.

Given the limited therapeutic efficacy observed in platinum-resistant organoids and the persistence of cancer stem cell features, we sought to explore alternative therapeutic strategies. Oncolytic virotherapy has recently emerged as a promising approach. We next evaluated the antitumor activity of the oncolytic herpes simplex virus OH2 in OC organoids. At MOI = 1, OH2 treatment resulted in a 20–40% reduction in cell viability at 48 h post-infection, even in organoids derived from platinum-resistant patients (Fig. 4E), indicating its potential therapeutic effect in resistant disease contexts.

Combined OH2 and cisplatin therapy enhances the suppression of organoid viability

To further explore the therapeutic potential of OH2 in combination with cisplatin, we evaluated their combined effects in ovarian cancer organoids derived from both platinum-sensitive and platinum-resistant patients. Organoids were treated for 48 h with OH2 (MOI = 1), cisplatin, or their combination, followed by assessment of cell viability. The results demonstrated a significant enhancement in antitumor efficacy upon combination therapy. In cisplatin-sensitive P07-derived organoids, the viability of cells treated with cisplatin alone was 64.67% ± 4.86%, whereas the combination group exhibited a further reduction to 42.17% ± 3.06% (Fig. 5A). This suggests that OH2 markedly enhances cisplatin sensitivity. Similarly, in cisplatin-resistant P26-derived organoids, cisplatin monotherapy resulted in a viability of 93.11% ± 0.36%, which dropped to 66.28% ± 2.09% upon co-treatment with OH2, indicating that OH2 may partially reverse cisplatin resistance (Fig. 5B). These findings support the potential of OH2 in combination with platinum-based chemotherapy as a promising therapeutic strategy, particularly for patients with cisplatin-resistant ovarian cancer.

Fig. 5.

Combined treatment of OC organoids with OH2 and cisplatin, and detection of the CSC marker CD44. A Cell viability assays following 48h treatments with OH2 (MOI = 1), cisplatin, or the combination therapy in cisplatin-sensitive organoids (P01, P07, P37). For cisplatin, the concentration was 0.8 µg/mL for P01 and P37, and 0.1 µg/mL for P07. B Cell viability assays following 48h treatments with OH2 (MOI = 1), cisplatin, or the combination therapy in cisplatin-resistant organoids (P10, P26, P51, P32, P11, P23, and P52). Cisplatin was administered at 2 µg/mL for P10, P32, and P52, and at 0.8 µg/mL for P11, P23, P26, and P51. C Ridgeline plots showing CD44 expression in representative cisplatin-sensitive and cisplatin-resistant patient-derived organoids after various drug treatments. D CD44 expression levels in cisplatin-sensitive organoids after 48h treatment with OH2 (MOI = 1), cisplatin, or their combination. E CD44 expression levels in cisplatin-resistant organoids after 48h treatment with OH2 (MOI = 1), cisplatin, or their combination. Data are presented as mean ± SD. ns: not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001

Given the critical role of CSCs in chemoresistance and tumor relapse, we further assessed the impact of OH2 on CSC populations using CD44-based flow cytometry. OH2 monotherapy led to either stable or slightly reduced proportions of CD44⁺ cells, suggesting that OH2 may directly or indirectly impair CSC viability. In contrast, cisplatin monotherapy, as expected, significantly enriched the CD44⁺ subpopulation, consistent with chemotherapy-induced CSC selection. Importantly, the combination of OH2 and cisplatin reduced CD44⁺ cell proportions (Fig. 5C). For example, in organoids derived from a platinum-refractory P51 patient, the baseline CD44⁺ expression was 90.33% ± 3.22%, which decreased to 78.60% ± 6.06% following combination therapy (Fig. 5D–E). These findings indicate that the combined use of OH2 and cisplatin may simultaneously target both rapidly proliferating differentiated cancer cells and the slowly cycling, chemoresistant CSC population. This dual-targeting approach represents a novel and promising therapeutic strategy for overcoming cisplatin resistance in OC.

Discussion

This study successfully established organoids of different pathological subtypes of OC derived from OC ascites, including HGSOC, OCCC, and yolk sac tumors. Although we also collected samples of other pathological subtypes, such as poorly differentiated endometrioid carcinomas and LGSOC, organoid construction was unsuccessful. Researchers like Oded Kopper et al. successfully established organoids of differentiated endometrioid carcinomas and LGSOC using a culture medium supplemented with hydrocortisone, forskolin, and heregulin β−1 [26], which may be one of the reasons for our failure. However, due to the limited sample size, with only two cases each of poorly differentiated endometrioid carcinoma and LGSOC, no definitive conclusion can be drawn about the reasons for the organoid culture failure. In this study, the overall success rate of constructing ascites-derived organoids was 86.2% (50/58), significantly higher than the reported success rate of tissue-derived ovarian cancer organoids, which is approximately 23–65% [27, 28]. This suggests that the ascites organoid model can, to some extent, compensate for the low construction efficiency of tissue organoids. Notably, ascites drainage is a standard care treatment for OC patients, and specimen acquisition is convenient and minimally invasive, with good prospects for clinical application. Ascites-derived OC organoids maintain the histopathological characteristics of the tumor tissue, such as nuclear morphology and atypia, the expression of key biomarkers, and the ascites-derived organoids and corresponding tumors maintain high similarity at the genomic level, proving that they can serve as a reliable model for predicting drug sensitivity. We used ascites-derived organoids to test the sensitivity to cisplatin and PARP inhibitors, and conducted long-term follow-up of patients to evaluate the consistency between the organoid model and the real treatment response. The results showed that the drug sensitivity test results were highly consistent with the patient’s clinical treatment response. Patients sensitive to cisplatin and PARP inhibitors had significantly lower IC50 values than drug-resistant patients, further confirming the predictive ability of the ascites-derived organoid model. A large body of research evidence supports the role of ovarian cancer CSCs in intraperitoneal metastasis, chemotherapy resistance, and tumor recurrence. OC ascites is an environment rich in CSCs, which express a variety of stem cell-related markers, such as ALDH1A1, CD44, CD133, CD24, NANOG, OCT4, and SOX2 [29–34]. In this study, we assessed via IHC that platinum-resistant PDOs showed significantly higher expression of CD44, which further supports the role of CSCs in ovarian cancer chemotherapy resistance.

Immunotherapy has limited overall efficacy in OC, mainly due to the immunosuppressive microenvironment of OC tissue [35, 36]. However, several studies have shown that the immune microenvironment of ascites differs from that of tumor tissue. The ascites has an inflammatory microenvironment, rich in a large number of immune cells such as CD8⁺ T cells, macrophages, and dendritic cells. The T cell population is mainly composed of effector T cells (Teff) and memory T cells (Tcm/Tem), while the tumor tissue is mainly composed of effector T cells (Teff) and memory T cells (Tcm/Tem) [37]. An emerging immunotherapy strategy, OV therapy, has been proven to induce specific antitumor immune responses through the interaction between viruses, tumor cells, and the immune system. Preclinical and clinical studies show promising results for OV therapy in OC [9–15]. Jacob PV et al. demonstrated that an oncolytic poxvirus, combined with tumor debulking surgery, effectively controlled ascites accumulation [11]. Tian L et al. found that combining oncolytic herpesvirus with antibody therapy significantly enhanced therapeutic efficacy in advanced ovarian tumors [13]. Notably, the PROTA 1a trial, using the oncolytic adenovirus (TILT-123) with pembrolizumab, achieved a 64% disease control rate in platinum-resistant ovarian cancer, while enadenotucirev combined with paclitaxel resulted in a 30% overall response rate and a 40% clinical benefit rate [14].

Building on these findings, we first evaluated the therapeutic potential of OH2, an HSV-2–based oncolytic virus, in ascites-derived OC organoids. OH2 has been genetically modified by deleting the ICP34.5 and ICP47 genes, a well-established modification that restricts viral replication to cancer cells with defective antiviral signaling, ensuring treatment safety [38, 39]. OH2 has previously demonstrated promising antitumor activity in patients with metastatic rectal and esophageal cancers, further supporting its potential as a therapeutic agent in OC [40, 41]. OH2 can not only directly lyse tumor cells but also induce a sustained antitumor immune response [42]. This study is the first to use an ascites-derived OC organoid model to evaluate the direct killing effect of OH2. Even in platinum-resistant patients, OH2 inhibited tumor growth by 20–40% within 48 h of infection, which may be sufficient to elicit a meaningful clinical response. Consistent with our previous findings in colorectal cancer [43], this study supports the effectiveness of OH2 against both cancer cells and cancer stem cells, suggesting that OH2 could serve as a promising therapeutic strategy for cancer patients.

In further studies, we evaluated the efficacy of OH2 in combination with cisplatin. The results showed that combination therapy significantly enhanced the killing effect in several cisplatin-resistant patients (such as P23 and P51), reducing cell viability by 30.56% and 27.99% compared to the cisplatin monotherapy group. Consistent with previous studies [44, 45], we observed a significant increase in the proportion of CSCs in residual cells after cisplatin treatment. In contrast, treatment combined with OH2 not only effectively killed rapidly proliferating differentiated cancer cells but also weakened the drug-resistant CSC subset, providing a new therapeutic strategy to overcome chemotherapy resistance.

This study, as a real-world prospective study, also has certain limitations. Due to the relatively relaxed inclusion criteria for patients, the loss-to-follow-up rate was high. Ultimately, only 16 cases with complete treatment response records were followed up, including 6 cases of platinum-partially sensitive patients. This highlights the need for improvement, and we plan to establish a collaborative, multi-center follow-up system to enhance data completeness and reduce attrition bias. In subsequent studies, we aim to expand the patient cohort and focus on patients who are partially sensitive to platinum, aiming to establish a more accurate drug sensitivity testing system to better distinguish between platinum-sensitive, partially sensitive, and resistant patients. We also plan to further investigate the efficacy of OH2 in ovarian cancer in combination with the ascites immune microenvironment, with an in-depth exploration of the immunomodulatory effect of OH2 in OC treatment. Furthermore, it is crucial to consider biological differences in the organoids, such as growth rate and stemness, and their potential correlation with drug resistance. These factors may serve as important biomarkers for resistance.

Conclusions

We successfully established PDOs from ascites of OC patients, faithfully preserving the histological architecture and genetic characteristics of the original tumors. These organoids reliably predicted individual drug responses, including to platinum-based chemotherapies and PARP inhibitors, with in vitro sensitivity correlating strongly with clinical outcomes. Notably, PDOs derived from platinum-resistant cases exhibited elevated expression of CSC markers such as CD44, underscoring the pivotal role of CSCs in driving chemoresistance and disease recurrence. Importantly, OH2 demonstrated robust antitumor efficacy even in platinum-resistant organoids, highlighting its potential as a therapeutic agent against refractory OC.

Abbreviations

CSC

Cancer stem cell

EOC

Epithelial ovarian cancer

HGSOC

High-grade serous ovarian carcinoma

IHC

Immunohistochemistry

ITH

Intratumoral heterogeneity

LGSOC

Low-grade serous ovarian carcinoma

OC

Ovarian cancer

OCCC

Ovarian clear cell carcinoma

OH2

Oncolytic herpes simplex virus 2

PDO

Patient-derived organoid

PFI

Platinum-free interval

PARP

Poly (ADP-ribose) polymerase

TME

Tumor microenvironment

Authors’ contributions

H.G., K.Z., and N.L. conceived and designed the study. X.Z. and W.Z. performed the experiments. Y.C. collected the ascites samples and conducted clinical follow-up. X.Z. drafted the original manuscript. Q.Z., and D.W. assisted with data analysis. W.Z. revised the manuscript. All authors read and approved the final version of the manuscript.

Funding

This project was supported by funding from the Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (grant no. 2022-I2M-C&T-B-082) and the National Natural Science Foundation of China (grant No. 81972804).

Data availability

The data generated in this study are available within the article and its supplementary data files. For any data not included, please contact the corresponding author.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Cancer Hospital, Chinese Academy of Medical Sciences (NCC2022C-478), and written informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.