A novel mRNA-based multi-cytokine strategy to reprogram the peritoneal tumor microenvironment in ovarian cancer

Abstract

Background

Peritoneal dissemination is a hallmark feature of advanced ovarian cancer and significantly contributes to its poor prognosis. An intraperitoneal mRNA-based immunotherapy delivering a combination of single-chain interleukins (IL), including IL-12, IL-15, pro-IL-18, and Caspase-1, encapsulated in lipid nanoparticles, is administered to reprogram the immunosuppressive tumor microenvironment (TME) in a syngeneic ID8-Fluc ovarian cancer mouse model. Results: This mRNA cocktail effectively suppresses tumor growth, reduces malignant ascites, and inhibits metastasis. Immune profiling reveals enhanced infiltration of effector CD8⁺ and CD4⁺ T cells, reduced regulatory T cells, and decreased expression of exhaustion markers such as TIM-3 and PD-1. Macrophage populations are shifted from immunosuppressive M2 to proinflammatory M1 phenotypes, with increased monocyte infiltration, indicating robust myeloid reprogramming. Although mRNA-expressed IL-12 monotherapy exhibits potent antitumor effects, the combination of IL-15, IL-18, and Caspase-1 elicits superior therapeutic efficacy. Although high-dose treatment induces hepatotoxicity and weight loss, a reduced-dose regimen maintains efficacy with improved safety. Conclusions: This study demonstrates the therapeutic potential of mRNA-based cytokine combinations to transform the ovarian cancer immune landscape. The modularity, localized delivery, and tunable expression of this platform provide a compelling framework for future mRNA immunotherapies targeting solid tumors with immunosuppressive TMEs.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04543-7.

Background

Ovarian cancer remains a significant clinical challenge in oncology, ranking among the most lethal gynecological malignancies due to its propensity for early peritoneal metastasis and the lack of effective methods for early detection [1, 2]. Despite advancements in chemotherapy, targeted therapies, and immune checkpoint inhibitors, survival rates for advanced-stage ovarian cancer, particularly in cases involving peritoneal dissemination, remain unfavorable [3]. The peritoneal cavity, frequently the primary site of metastasis, is a complex, immunosuppressive environment, presenting significant obstacles to effective treatment [4]. Myeloid cells, including monocytes, macrophages, and dendritic cells, are crucial components of the tumor microenvironment (TME), where they can either promote or suppress tumor progression depending on their activation state [5]. Similarly, T cells, both CD4⁺ helper and CD8⁺ cytotoxic T lymphocytes, play critical roles in immune surveillance, but are often rendered ineffective by the immunosuppressive signals present within the TME [6].

Given these challenges, immunotherapy has emerged as a promising strategy to reprogram the TME and restore effective anti-tumor responses. Among various approaches, cytokine-based immunotherapies have garnered attention for their ability to activate and direct immune effector functions at the tumor site [7]. Notably, interleukin (IL)-12, IL-15, and IL-18 have demonstrated potent immune-activating properties in preclinical studies. IL-12 promotes a robust Th1-type immune response and activates both T cells and natural killer (NK) cells, eliciting strong cytotoxic activity [8]. Likewise, IL-15 and IL-18 play pivotal roles in the expansion and activation of NK cells and cytotoxic T lymphocytes [9, 10]. Importantly, the biological activity of IL-18 depends on proteolytic processing by Caspase-1 (CASP1), a central component of the inflammasome that cleaves pro-IL-18 into its mature, active form. Although not a cytokine itself, CASP1 is essential for enabling the immunostimulatory function of IL-18, thereby representing a critical adjunct in cytokine-based immunotherapeutic strategies [11, 12].

Although cytokine-based therapies hold significant promise, their clinical application has been hindered by issues related to systemic toxicity and inefficient delivery to the tumor site [13]. Recent clinical trials have endeavored to deliver cytokines via mRNA, but with limited success. Notably, a phase I trial by Bechter et al. [14]. investigated a synthetic mRNA mixture encoding single-chain IL-12 (scIL-12), scIL-15, and interferon alpha 2b (IFNα2b), administered intramuscularly alone, or in combination with the PD-1 inhibitor, cemiplimab, in patients with advanced solid tumors including melanoma. In this study, scIL-12 encodes a single-chain fusion of IL-12α (p35) and IL-12β (p40) subunits joined by a flexible peptide linker, while scIL-15 comprises IL-15 fused to the IL-15 receptor alpha (IL-15Rα) sushi domain via a similar linker, thereby enhancing cytokine stability and receptor engagement in vivo. Despite demonstrating biological activity, the trial failed to achieve sufficient efficacy, primarily due to systemic toxicities and suboptimal cytokine delivery to the TME [14].

While previous intraperitoneal mRNA cytokine studies have primarily focused on IL-12 alone or on dual- and triple-cytokine combinations [14, 15], our approach is distinct in its incorporation of CASP1 alongside IL-12, IL-15, and IL-18. CASP1 activates the inflammasome, promoting the maturation of IL-1β and IL-18, which enhances innate immune stimulation and reprograms myeloid populations within the peritoneal TME. By augmenting IL-18–mediated Th1-type responses, the inclusion of CASP1 synergistically boosts cytotoxic T cell activation and overall antitumor immunity, representing a mechanistic advance over previously reported mRNA cytokine strategies. Our combinatorial mRNA strategy addresses these limitations by coordinating activation of multiple immune pathways while leveraging CASP1 to enhance both innate and adaptive immune responses within the peritoneal TME.

mRNA-based therapeutics represent an innovative and highly adaptable approach, facilitating localized, controlled expression of cytokines directly within the TME [16]. This strategy reduces off-target effects and enables more precise modulation of immune responses. Specifically, intraperitoneal administration of mRNA-encoding cytokines offers a promising option for ovarian cancer with peritoneal metastasis, wherein tumor cells are confined to the peritoneal cavity, enabling more effective targeting by immune cells [17, 18]. Consequently, this study sought to evaluate the therapeutic potential of mRNA-encoded scIL-12, scIL-15, pro-IL-18, and CASP1 in an ovarian cancer model with peritoneal metastasis. The objectives of the study include characterization of the in vivo expression of these cytokines, evaluation of their immune cell uptake in the peritoneal and determination of their capacity to transform immune responses and suppress tumor progression. To our knowledge, this is the first study demonstrating that localized, intraperitoneal delivery of an mRNA cytokine cocktail encoding IL-12, IL-15, pro-IL-18, and CASP1 can effectively reprogram the peritoneal TME in ovarian cancer. Our findings provide valuable insights into the potential for treating ovarian cancer using mRNA-based cytokine therapy, particularly in cases of peritoneal dissemination, and highlight new avenues for optimizing immunotherapy in cancer treatment.

Methods

Design of mRNA complexes

The DNA template of the pGH vector-based mRNA platform, designated CUK3-1, was engineered to include regulatory elements such as the 5′-untranslated region (UTR), 3′-UTR, and polyadenylation (poly-A) tail to enable efficient expression. Detailed sequence information of the platform is available in the patent KR10-2022-0115033. Genes of interest were inserted into the multi-cloning site (MCS) located between the 5′- and 3′-UTRs, including green fluorescent protein (GFP), Renilla luciferase (R/L), scIL-12, scIL-15, IL-18, and CASP1. Coding sequences were derived from their respective wild-type reference sequences (M86671.1, M86672.1, BC022705.1, U14332.1, NM_008360.1, and NM_009807.2).

In vitro transcription and RNA purification

The DNA templates were linearized using NotI. An EZ T7 High-Yield In Vitro Transcription Kit (Enzynomics) was used for in vitro transcription. Briefly, 4 µg of linearized DNA template was initially incubated with 40 µL of 5× T7 transcription buffer, 50 µL of 10× MgCl₂ (250 mM), 20 µL of 10× DTT (10 mM), 8 µL of ATP, CTP, and GTP (100 mM), 8 µL of N1-methyl-pseudo-UTP (100 mM), 6.4 µL of SC-101 (smart cap), 4 µL of 20× enhancer solution, 4 µL of 50× T7 polymerase (200 U), and nuclease-free water in a final volume of 200 µL overnight at 37 °C, followed by incubation with RNase-free DNase I (Promega, Madison, WI, USA) for 15 min at 37 °C. For RNA purification, RNA was precipitated by adding 100 µL of a 7.5 M LiCl solution and incubated for 30 min at 4 °C. After centrifugation for 15 min at 13,000 × g and 4 °C, the supernatant was discarded, and 600 µL of 70% ethanol was added to the pellet. The suspension was then centrifuged for 15 min at 13,000 × g and 4 °C for washing. Finally, the supernatant was discarded, and the pellet was resuspended in RNase-free water after air drying. To remove the dsRNA, cellulose purification was performed as previously described [19]. The purity and concentration of the DNA and RNA were evaluated using a NanoDrop-2000 spectrophotometer (Thermo Fisher Scientific, MA, USA).

Formulation of mRNA-encapsulated lipid nanoparticles

Lipid nanoparticles (LNPs) were formulated using the following method. All lipid components, including SM-102, trehalose-6,6’-dioleate, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), butyl lithocholate, and 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000 (DMG-PEG2000), were dissolved in a chloroform/methanol (1/1, v/v) mixture at a concentration of 1 mg/mL. After mixing at a specific molar ratio (25:25:10:38.5:1.5, SM-102:trehalose-6,6’-dioleate: DOPE: butyl-lithocholate: DMG-PEG2000), the lipid mixture was concentrated under reduced pressure to remove the solvent. The lipid mixture was re-dissolved in ethanol, and each mRNA was dissolved in sodium citrate buffer (50 mM, pH 4.0) solution to have a specific charge ratio (N/P = 3). For each type of mRNA, LNPs were formulated using the NanoAssemblr^® Ignite™ system (Precision NanoSystems Inc., Vancouver, BC, Canada) at a total flow rate of 10 mL/min and a mixing ratio of 1:3 (v: v) of organic lipid to aqueous mRNA solution. The mRNA-encapsulated LNP solutions were washed with 1× Dulbecco’s phosphate-buffered saline (DPBS) and concentrated using Amicon^® Ultra-15 centrifugal filter units (30 kDa cutoff, UFC9030, Merck Millipore, Darmstadt, Germany) with a 624R centrifuge (LABOGENE, Seoul, South Korea). The particle size and zeta potential of the mRNA-encapsulated LNPs were measured using a Zetasizer Ultra (Malvern Panalytical Ltd., Worcestershire, UK).

Cell cultures

ID8-Fluc cells were obtained from the Korean Cell Line Bank (Seoul, Republic of Korea). The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; HyClone™, Cytiva, Buckinghamshire, UK), supplemented with 10% fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and 1% antibiotic–antimycotic solution (Gibco). Cultures were maintained under standard conditions at 37 °C in a humidified incubator with 5% CO₂.

Mice

Five-week-old female C57BL/6 mice were purchased from Dae-Han Bio-Link (Chungbuk, Republic of Korea). All mice were housed in the animal facility at the Catholic University of Korea under specific pathogen-free conditions at 21–22 °C and a 12-h/12-h light/dark illumination cycle. All experimental procedures conducted in animal followed the guidelines of and were approved by the Institutional Animal Care and Use Committee of the Catholic University of Korea (approval no. CUK-IACUC-2022-031 and CUK-IACUC-2022-044). The study was conducted in compliance with the ARRIVE guidelines.

Animal experiments

To establish a peritoneal dissemination model of epithelial ovarian cancer (EOC), 1 × 10⁷ ID8-Fluc cells were intraperitoneally injected into mice. Two weeks post-cell administration, the mice were randomized based on in vivo luciferase measurements. LNP-formulated mRNA-cytokines were administered intraperitoneally once per week for 3 weeks. Body weight and abdominal circumference were recorded weekly, and tumor progression was monitored using an in vivo imaging system. One week following the final administration, the mice were euthanized, and ascites and blood samples were collected. The abdominal cavity was rinsed with 5 mL of PBS, and immune cells were isolated from ascites, abdominal washings, spleen, and epididymal white adipose tissue (eWAT) for subsequent immune analysis. Blood samples were allowed to clot at 25 °C for 2 h before serum collection. Serum levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were analyzed at the Korean Pathology Technical Center (KP&T, South Korea).

In vivo bioluminescence imaging

To monitor tumor growth, the abdominal fur of mice was shaved, and VivoGlo™ Luciferin, In Vivo Grade (Promega, WI, USA), was dissolved in saline to a concentration of 3 mg/100 µL per mouse. Intraperitoneal injections of luciferin were administered, and mice were anesthetized for respiratory support during imaging using the LUCI Luminescence In Vivo Imaging System (CELLGENTEK, Daejeon, South Korea). Luminescence signals were captured with a 10-s exposure time. To assess organ-specific metastasis of ID8-Fluc cells, peritoneal wash fluid was collected at the time of sacrifice, and luciferin was introduced into the abdominal cavity with gentle agitation to ensure even distribution. The abdomen was then opened, and the organs were excised and placed on a black plate. Luminescence signals from each organ were measured. Data analysis was performed using the NEOimage software (CELLGENTEK).

Cytokine enzyme-linked immunosorbent assay

The cytokine levels in serum and abdominal lavage fluid were assessed using enzyme-linked immunosorbent assay (ELISA) kits to determine the concentrations of IL-12 (Cat. #88-7121-88), IL-15/IL-15R complex (Cat. #88-7215-88), and IL-18 (Cat. #88-50618-88) (Invitrogen; Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer’s instructions. Ninety-six-well plates were coated with capture antibodies for IL-12, IL-15/IL-15Rα complex, and IL-18 and incubated overnight at 4 °C. After incubation, the plates were washed three times with 0.05% PBST (0.05% Tween-20 in PBS) and blocked with 1× diluent for 1 h at room temperature. Following additional washes, 1:10 diluted serum and 50 µL of 1× abdominal lavage fluid were added and incubated for 2 h at room temperature. After washing, the plates were incubated with diluted detection antibodies for IL-12, IL-15/IL-15Rα complex, and IL-18 for 1 h. Subsequently, diluted avidin-HRP or streptavidin-HRP was added and incubated for 30 min. After washing five times, 100 µL of 1× TMB solution was added and incubated for 15 min. The reaction was quenched using 2 N H₂SO₄. The optical density was measured at 450 nm using a GloMax Explorer Multimode Microplate Reader (Promega, WI, USA). Cytokine concentrations were calculated according to standard curves, and the results are expressed as IL-12, IL-15/IL-15Rα complex, and IL-18 concentrations in pg/mL for serum and abdominal lavage fluid.

Flow cytometry

To characterize immune cell populations, cells were isolated from the spleen, peritoneal fluid, and eWAT. Fc receptors were blocked using anti-CD16/CD32 antibodies (eBioscience, San Diego, CA, USA) for 20 min at 4 °C, followed by staining for surface markers using the following fluorochrome-conjugated antibodies: CD45 (clone 30-F11, BioLegend), CD8a (clone 53 − 6.7, BioLegend), CD4 (clone GK1.5, BioLegend), CD19 (clone eBio1D3, BioLegend), CD11b (clone M1/70, BioLegend), CD11c (clone N418, BioLegend), F4/80 (clone BM8, BioLegend), Ly6G (clone 1A8, BioLegend), Ly6C (clone HK1.4, BioLegend), MHC II (clone M5/114.15.2, BioLegend), CD206 (clone C068C2, BioLegend), CD86 (clone GL-1, BioLegend), CD80 (clone 16-10A1, BioLegend), CD44 (clone IM7, Pharmingen), CD62L (clone MEL14, invitrogen), CD127 (clone A7R34, BioLegend), KLRG1 (clone 2F1, BioLegend), CD25 (clone PC61.5, BioLegend), PD-1 (clone J43, BioLegend), Tim-3 (clone RMT3-23, BioLegend), and EpCAM (clone G8.8, BioLegend). Dead cells were excluded using Fixable Viability Dye eFluor™ 520 (eBioscience). Staining was conducted for 30 min at 4 °C in the dark, followed by fixation with 4% paraformaldehyde (Biosesang, Seongnam, Gyeonggi, South Korea).

For intracellular cytokine analysis, a subset of splenocytes was permeabilized using the Fixation/Permeabilization Solution Kit (BD Biosciences, San Jose, CA, USA) and subsequently stained with anti-IFN-γ (clone XMG1.2). Samples were analyzed using a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, USA), and data were processed with CytExpert software (Beckman Coulter).

Enzyme-linked immunospot assay

Splenocytes were seeded at 5 × 10⁵ cells/well and incubated for 48 h at 37 °C. Enzyme-linked immunospot (ELISpot) assays to detect IFN-γ T cells were conducted using the mouse IFN-γ ELISpotBASIC kit (Mabtech, Stockholm, Sweden). Each step was performed according to the manufacturer’s protocols.

RNA sequencing, data processing, and functional analysis

Peritoneal fluid cells, containing both immune and tumor cells, were collected from C57BL/6J mice bearing ID8-Fluc ovarian tumors. RNA was isolated from peritoneal fluid cells, and library preparation and sequencing were performed by Macrogen Inc. (Seoul, South Korea) using standard protocols. The resulting RNA-seq data were processed into a comprehensive expression matrix containing gene identifiers, symbols, functional annotations, biotype, protein IDs, and quantitative expression measures (FPKM and TPM) for each treatment group. Differential expression metrics—including fold change, log counts per million, raw p-values, and adjusted p-values (Benjamini–Hochberg correction)—were calculated for all pairwise comparisons, providing the basis for downstream analyses.

Visualization of transcriptional patterns employed TPM-normalized values. Heatmaps were generated using row-wise z-score standardized, log2-transformed TPM values to emphasize relative expression changes across genes while controlling for baseline differences. Gene-level clustering was performed using hierarchical clustering with average linkage and Euclidean distance, and heatmaps were annotated with treatment labels corresponding to MCS, IL-15, IL-18/CASP1, IL-12, and the combination therapy. This approach allowed identification of coordinated gene expression programs induced specifically by the combination therapy.

Principal component analysis (PCA) was performed on the same z-score–standardized TPM matrix to evaluate global transcriptional variance across samples. PCA plots enabled assessment of the relative positioning of combination-treated samples compared to single cytokine treatments, providing a quantitative measure of treatment-specific transcriptional reprogramming.

To characterize functional consequences of combination therapy–specific transcriptional changes, Gene Ontology (GO) enrichment analysis was performed focusing on the Biological Process (BP) category. Upregulated genes specific to the combination treatment were selected based on absolute log2 fold-change ≥ 0.5 and adjusted p-value < 0.05. Enrichment was determined by over-representation analysis comparing observed versus expected gene counts per GO term, with multiple testing correction via Benjamini–Hochberg procedure. GO terms with adjusted p-value < 0.05 were considered significant. The top 20 enriched BP terms were visualized using dot plots, with dot size representing gene ratio and dot color representing −log10 adjusted p-value. Hierarchical clustering of GO terms (average linkage, Euclidean distance) was applied to organize terms by functional similarity, and dendrograms were displayed along the y-axis to identify clusters of related biological processes. Treatment groups were mapped along the x-axis to highlight enrichment patterns specific to the combination therapy, revealing activation of interconnected networks regulating immune activation, environmental sensing, cellular homeostasis, and stress responses.

Volcano plots were generated using log2-transformed FPKM values (log2[FPKM + 1]) to stabilize variance. Differential expression was calculated by comparing combination therapy with the mean of single cytokine treatments, with genes meeting an absolute log2 fold-change ≥ 1 and an adjusted p-value < 0.05 highlighted. Key immune regulatory genes—including M1/M2 macrophage markers, NF-κB signaling components, STAT-mediated immune modulators, PDCD-related apoptosis regulators, and other cytokine-responsive genes—were annotated to facilitate mechanistic interpretation.

All data processing and visualization were performed using Python (including pandas, NumPy, Seaborn, Matplotlib, SciPy, and scikit-learn libraries), following best practices for reproducibility, statistical rigor, and clear representation of transcriptional and functional patterns.

Statistical analysis

Analyses were performed using one-way analysis of variance with Tukey’s post hoc test for multiple-group comparisons and unpaired t-tests to compare between two groups. Differences were considered significant at p < 0.05. Data are expressed as mean ± standard deviation. Statistical analyses were conducted using GraphPad Prism 9 (GraphPad Software Inc., San Diego, CA, USA).

Results

Design and in vivo expression and distribution of mRNA encoding IL-12, IL-15, pro-IL-18, and CASP1

We designed a cap-dependent mRNA expression platform incorporating a MCS between the 5’- and 3’-UTRs, with a poly(A)50 tail linked via a stabilizing sequence to enhance mRNA stability against RNase degradation [20]. An MCS-only mRNA lacking any coding sequence was used as a negative control. For cytokine expression, coding sequences for IL-12, IL-15, pro-IL-18, and CASP1 were inserted into the MCS (Fig. 1A). IL-12 was generated by fusing IL-12α (p35) and IL-12β (p40) via a flexible linker, while IL-15 was fused to the Sushi domain of IL-15 receptor α (IL-15Rα) to enhance stability and activity. Functional activation of IL-18 was achieved by co-administering pro-IL-18 mRNA with CASP1 mRNA.

Fig. 1.

Construction and intraperitoneal immune cell distribution of mRNAs encoding IL-12, IL-15, IL-18, and Caspase-1. C57BL/6 mice were intraperitoneally administered 20 µg of each cytokine-expressing mRNA-LNP for cytokine expression analysis, and 50 µg of GFP- expressing mRNA-LNP for immune cell distribution analysis. (A) Schematic diagrams of each cytokine mRNA construct. Cytokine expression was evaluated using ELISA, whereas flow cytometry was performed to analyze immune cell distribution based on GFP expression. (B) Levels of IL-12, IL-15/IL-15Rα complex, and IL-18 in serum and peritoneal lavage fluid of cytokine-administered mice were measured using ELISA (n = 3). (C) Percentage of immune cell types among total GFP⁺ cells was measured using flow cytometry with cells isolated from peritoneal lavage fluid (n = 5). Data are presented as mean ± SD. *p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.ELISA, enzyme-linked immunosorbent assay; GFP, green fluorescent protein; IL, interleukin; IL-15Rα, interleukin-15 receptor alpha; LNP, lipid nanoparticle; mRNA, messenger RNA; SD, standard deviation

To assess in vivo cytokine expression, serum and peritoneal lavage fluid were collected at 1 and 6 h after intraperitoneal administration and analyzed by ELISA (Fig. 1A, B). Serum IL-12 peaked at 1 h (approximately 20,000 pg/mL) and remained elevated at 6 h, with no statistically significant change. In peritoneal lavage fluid, IL-12 levels increased significantly only at 6 h, indicating delayed local accumulation compared to systemic circulation.

IL-15 serum levels increased significantly at 1 h and further at 6 h, demonstrating sustained systemic expression. In peritoneal fluid, IL-15 levels were elevated at both 1 and 6 h, with statistical significance achieved at 6 h.

Serum IL-18 levels increased at 1 h, with further enhancement upon CASP1 co-administration. At 6 h, IL-18 remained elevated, with CASP1 significantly increasing levels at both time points (Fig. 1B). In peritoneal fluid, IL-18 increased significantly only at 6 h when CASP1 was co-administered, highlighting its essential role in activating pro-IL-18 locally. Collectively, these data demonstrate effective systemic and local expression of IL-12, IL-15, and IL-18, with CASP1 critically enhancing IL-18 activation in the peritoneal cavity. Temporal differences between serum and peritoneal cytokine levels suggest distinct kinetic profiles in systemic versus local environments.

To evaluate mRNA uptake by peritoneal cells, GFP-expressing mRNA was administered intraperitoneally, and peritoneal cells were collected 24 h later for flow cytometry (Additional file 1, Figure S1; Fig. 1C). GFP-expressing cells constituted 4.2% of total peritoneal cells. Among immune cells, monocytes and dendritic cells showed the highest GFP expression (> 10%), whereas neutrophils, macrophages, B cells, and CD4⁺ T cells exhibited moderate expression (4%–5%), and CD8⁺ T cells displayed minimal expression (< 1%). Non-immune cells accounted for the largest fraction of GFP-positive cells. Within immune populations, monocytes were the predominant GFP-positive subset (approximately 1%), followed by neutrophils. A stacked bar graph (Fig. 1C, upper panel) illustrates total GFP percentages across immune subsets, while a pie chart (Fig. 1C, lower panel) depicts the relative distribution of GFP-expressing cells. These results indicate preferential mRNA uptake by myeloid cells, particularly monocytes and dendritic cells, over lymphocytes in the peritoneal cavity.

Co-administration of mRNA-expressing IL-12, IL-15, or pro-IL-18/CASP1 suppresses ovarian cancer growth in the peritoneal cavity

To evaluate the therapeutic potential of cytokine-encoding mRNAs in ovarian cancer, we employed a syngeneic murine model. C57BL/6J mice were intraperitoneally implanted with ID8-Fluc cells on Day − 14 to establish peritoneal tumors. Tumor progression was monitored via bioluminescence imaging immediately before treatment (Day − 1) and on Days 6, 13, and 20 after intraperitoneal mRNA administration (Fig. 2A). Mice were divided into six groups (Fig. 2B): untreated (Nil), MCS control, single cytokine mRNAs (scIL-15, scIL-12, or pro-IL-18/CASP1), and a combination of all three cytokines.

Fig. 2.

Anti-tumor effects induced by IL-12 and combined IL-12/IL15/IL18/Caspase-1 treatment. C57BL/6 mice were intraperitoneally injected with 1 × 10⁷ ID8-Fluc cells, and 2 weeks later, were intraperitoneally administered LNP-formulated mRNAs encoding IL-12, IL-15, and IL-18/CASP1 once weekly for 3 weeks. Mice were sacrificed one week after the final administration. (A-B) Schematic illustrating the experimental design. C57BL/6 mice were intraperitoneally administered 20 µg of each LNP-formulated mRNA-encoding cytokine. (C) Change in abdominal circumference measured as an indicator of ascites accumulation. (D) Luminescence of ID8-Fluc cells was measured weekly using an in vivo imaging system (n = 5). (E) Kaplan–Meier survival analysis during the experimental period. (F) Representative in vivo bioluminescence images of mice from each treatment group at the indicated time points.Data are presented as mean ± SD.*p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.X indicates a dead mouse.CASP1, caspase-1; IL, interleukin; IL-15, interleukin-15; IL-18, interleukin-18; IL-12, interleukin-12; LNP, lipid nanoparticle; mRNA, messenger RNA; SD, standard deviation

Abdominal circumference, an indirect measure of tumor burden and ascites, was tracked on Days 1, 2, 6, 9, 13, 15, and 20 (Fig. 2C). By Day 19, only mice receiving the IL-12/IL-15/IL-18/CASP1 combination showed a reduction in abdominal size relative to baseline, indicating effective suppression of ascites and tumor growth (Fig. 2C). In contrast, all other groups exhibited progressive abdominal enlargement. Among single agents, IL-12 and IL-18/CASP1 modestly reduced abdominal circumference compared to Nil, whereas IL-15 alone unexpectedly resulted in greater abdominal enlargement than other treated groups, including MCS. Statistical analysis confirmed that the combination treatment significantly reduced abdominal size relative to all other groups. Notably, mild treatment-related toxicity, including weight loss and elevated liver enzymes, was observed in the combination group (Additional file 1, Figure S2A, S2B), suggesting the need for dose optimization in future studies.

Tumor progression quantified via in vivo bioluminescence imaging showed that the IL-12/IL-15/IL-18/CASP1 combination achieved the greatest reduction in tumor signal intensity (Fig. 2D, F). IL-12 monotherapy also significantly decreased tumor burden, albeit to a lesser extent, while Nil, MCS, IL-15, and IL-18/CASP1 alone produced minimal effects. One mouse in the Nil group died around Day 19, likely due to tumor progression (Fig. 2E). A similar death occurred in the combination group despite the absence of a detectable tumor signal, suggesting potential treatment-related toxicity.

These results demonstrate that co-administration of IL-12, IL-15, and IL-18/CASP1 mRNAs elicits robust antitumor activity in the peritoneal cavity, surpassing the efficacy of individual cytokines while highlighting the importance of monitoring safety alongside therapeutic benefits.

Co-administration of mRNA-expressing IL-12, IL-15, and IL-18/CASP1 reduces metastatic dissemination in the peritoneal cavity

To assess effects on metastatic spread, mice from the therapeutic study were euthanized on Day 21. Major abdominal organs, including spleen, liver, kidney, uterus, pancreas, eWAT, mesentery, digestive tract, peritoneum, and diaphragm, were examined using bioluminescence imaging to visualize individual metastases (Fig. 3A) and quantify tumor burden via photon flux (photons s⁻¹ cm⁻²) (Fig. 3B). The average metastasis distribution per treatment group is presented as a stacked bar plot (Fig. 3C), with additional organ-specific comparisons shown in Additional file 1, Figure S3A and S3B.

Fig. 3.

Anti-metastatic effects induced by IL-12 and the combination of mRNA IL-12/IL-15/IL-18/Caspase-1. (A) Luminescence in individual organs was measured at the time of sacrifice using an in vivo imaging system. (B) Proportion of tumor intensity observed for each organ. (C) Proportion of total tumor luminescence signal distributed across organs.Data are presented as mean ± SD.X indicates a dead mouse. IL, interleukin; SD, standard deviation

In the Nil, MCS, and IL-15 groups, tumor dissemination was widespread across nearly all organs. IL-18/CASP1 treatment produced modest reductions in metastatic burden, particularly in the uterus and digestive tract. IL-12 monotherapy further decreased metastases, notably in the pancreas and mesenteric regions. Remarkably, the combination of IL-12, IL-15, IL-18, and CASP1 nearly eliminated detectable metastases in all examined sites, indicating substantially enhanced efficacy.

Across all groups, eWAT consistently exhibited the highest metastatic burden, followed by mesenteric fat, pancreas, and diaphragm. Tumor signal intensity in eWAT and other adipose tissues mirrored the overall treatment hierarchy: Nil > MCS > IL-15 > IL-18/CASP1 > IL-12 > combination treatment. Combination treatment markedly reduced detectable metastatic signals across most examined organs compared with individual cytokine treatments.

mRNA-expressing IL-12, IL-15, and IL-18/CASP1 remodels myeloid cell composition in the spleen and TME

To investigate the immunological mechanisms underlying the anti-tumor efficacy of cytokine-encoding mRNAs, we analyzed peritoneal fluid accumulation, tumor burden, and myeloid cell composition across spleen, ascites, and eWAT compartments using flow cytometry in the same cohort of treated mice.

Ascites accumulation served as a gross indicator of peritoneal disease progression. Nil mice developed extensive ascites, reflecting severe tumor burden. MCS and IL-15 treatments partially reduced ascites (approximately 50% of Nil), whereas IL-18/CASP1 and IL-12 further suppressed accumulation. The IL-12/IL-15/IL-18/CASP1 combination nearly abolished ascites formation, consistent with the anti-metastatic activity observed previously (Fig. 4A). Analysis of EpCAM⁺ cancer cells in ascites mirrored these results: percentages were highest in Nil and MCS groups, while IL-12 and combination treatments significantly reduced tumor cell content (Fig. 4B).

Fig. 4.

Cell population changes in spleen, ascites, and eWAT induced by IL-12 and mRNA IL-12/IL-15/IL-18/Caspase-1 treatment. Populations of tumor cells and immune cells were analyzed using flow cytometry (n = 5). (A) Volume of ascites measured at the time of sacrifice. (B) Percentage of tumor cells (EpCAM+) in ascites. (C) Percentage of immune cells (CD45⁺) in ascites and eWAT. (D) Percentage of macrophages (CD11b+, Ly6C-, Ly6G-, F4/80+) in spleen, ascites, and eWAT. (E) Distribution of macrophage subsets, defined as CD86 + CD206- (M0), CD86 + CD206- (M1), CD86 + CD206+ (transition), and CD86-CD206+ (M2), in ascites and eWAT. (F) Percentage of monocytes (CD11b+, Ly6G-, F4/80-, Ly6C+) in spleen, ascites, and eWAT.Data are presented as mean ± SD.*p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.eWAT, epididymal white adipose tissue; IL, interleukin; SD, standard deviation

Immune cell infiltration was next evaluated via CD45⁺ staining. In ascites, IL-12 and combination treatments significantly increased total immune cell frequencies compared to Nil, whereas MCS, IL-15, and IL-18/CASP1 groups showed no significant changes. In eWAT, only the combination therapy induced a statistically significant increase in CD45⁺ immune cells (Fig. 4C).

Macrophage populations were analyzed across compartments to assess therapy-induced remodeling. In the spleen, macrophage percentages were highest in Nil and were significantly reduced by MCS, IL-18/CASP1, IL-12, and combination treatments (Fig. 4D). In ascites, macrophage levels remained high in Nil, MCS, and IL-15 groups, whereas IL-12 and combination treatments significantly reduced macrophage abundance. In eWAT, macrophage levels were lowest in Nil and increased modestly in MCS, IL-15, and IL-18/CASP1 groups. IL-12 alone significantly elevated macrophage infiltration, and the combination treatment further increased macrophage abundance, indicating differential recruitment or retention within the TME.

Macrophage polarization was evaluated using canonical M1 (CD86) and M2 (CD206) markers (Fig. 4E). Although differences were not statistically significant, the M1:M2 ratio varied across treatments and tissues. In peritoneal fluid, Nil, MCS, and IL-15 groups displayed a roughly balanced M1:M2 ratio, whereas IL-12 and combination treatments exhibited a higher proportion of CD86⁻/CD206⁻ macrophages, consistent with a mixed or transitional activation state rather than strict polarization. In eWAT, M2 macrophages predominated in Nil and MCS groups, while IL-15 and IL-18/CASP1 treatments showed slight shifts toward M1 dominance. IL-12 and combination treatments further reinforced M1-like profiles.

Monocyte abundance was also affected by treatment. In the spleen, only the combination group exhibited a significant increase in monocyte frequency. In ascites, monocytes were lowest in Nil and significantly elevated in MCS and IL-12 groups. The combination group displayed similar levels to IL-12, reaching statistical significance relative to Nil (Fig. 4F). These changes suggest IL-12–driven recruitment of monocytes to the peritoneal cavity, which may contribute to macrophage turnover and functional activation within the tumor microenvironment.

Co-administration of mRNA-expressing IL-12, IL-15, and IL-18/CASP1 promotes T cell reprogramming in the spleen and TME

To examine T cell–mediated immune changes beyond myeloid remodeling, we performed flow cytometric analyses on immune cells isolated from spleen, ascites, and eWAT, representing systemic and local tumor-associated compartments in the same cohort of tumor-bearing mice.

Total T cell frequencies are shown in Fig. 5A. In the spleen, the combination treatment (IL-12/IL-15/IL-18/CASP1) significantly increased both CD8⁺ and CD4⁺ T cell frequencies compared to Nil, MCS, and IL-18/CASP1 groups. Single IL-12 treatment also significantly elevated these populations relative to Nil. In ascites, CD4⁺ T cells were significantly higher in the combination group than in Nil, MCS, IL-15, and IL-18/CASP1 groups, whereas CD8⁺ T cell frequencies showed no significant differences. In eWAT, the combination group exhibited significantly increased CD8⁺ T cells compared to the Nil, MCS, and IL-15 groups, while CD4⁺ T cell frequencies appeared elevated but did not reach statistical significance.

Fig. 5.

T cell responses induced by IL-12 and mRNA IL-12/IL-15/IL-18/Caspase-1 treatment in spleen, ascites and eWAT. The population of T cells was analyzed using flow cytometry (n = 5). (A) Percentage of CD8 + and CD4 + T cells in spleen, ascites and eWAT. (B) Percentage of effector T cells (CD8 + or CD4+, CD127-, and KLRG1+) in spleen, ascites and eWAT. (C) IFN-γ–producing cells were quantified using the ELISpot assay. (D) IFN-γ–producing CD8 + and CD4 + T cells in splenocytes. (E-F) Percentage of Tregs (CD4+, CD25+, CD127-) in spleen and ascites, and (G-H) Percentage of exhausted CD8 + and CD4 T cells in ascites (CD8 + or CD4+, Tim3+) and eWAT (CD8 + or CD4+, PD-1+).*p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001. eWAT, epididymal white adipose tissue; ELISpot, enzyme-linked immunosorbent spot assay; IFN-γ, interferon gamma; IL, interleukin; KLRG1, killer cell lectin-like receptor G1; PD-1, programmed cell death protein 1; SD, standard deviation

Effector T cell populations (Fig. 5B) were next evaluated. In the spleen, both IL-12 single and combination treatments significantly increased effector CD8⁺ and CD4⁺ T cells relative to other groups. In ascites, IL-12 monotherapy and combination treatments significantly elevated effector CD8⁺ T cells compared to the Nil and IL-15 groups, whereas effector CD4⁺ T cells showed no significant differences. In eWAT, combination treatment significantly increased both effector CD8⁺ and CD4⁺ T cells, while IL-12 monotherapy treatment also enhanced effector CD4⁺ T cells.

Functional cytokine output was assessed by IFN-γ ELISpot in the spleen (Fig. 5C, D). Both IL-12 monotherapy and combination treatments significantly increased total IFN-γ-producing cells compared to the Nil, MCS, IL-15, and IL-18/CASP1 groups, with IL-12 monotherapy showing the highest overall IFN-γ response. Subset analysis indicated that IL-12 alone significantly increased IFN-γ production in both CD8⁺ and CD4⁺ T cells, whereas combination treatment enhanced CD8⁺ IFN-γ production and moderately increased CD4⁺ IFN-γ compared to controls.

Regulatory T cells (Tregs) were also examined. In the spleen, IL-12 monotherapy significantly reduced Treg percentages compared to all groups except the combination, which also showed a significant decrease relative to IL-18/CASP1 (Fig. 5E). IL-15 treatment resulted in higher Treg frequencies than the Nil and MCS gropus. In ascites, combination treatments decreased Treg percentages relative to the Nil, MCS, IL-15, IL-18/CASP1, and IL-12 groups, with the combination group displaying the lowest Treg frequency and the highest ratio of Tregs to CD8⁺ effector T cells (Fig. 5F).

Exhaustion markers were next evaluated. TIM-3 expression on CD8⁺ and CD4⁺ T cells in ascites was lower in IL-12 single and combination groups than in the Nil, MCS, and IL-15/IL-18/CASP1 groups, with combination treatment showing the lowest expression, although differences between IL-12 monotherapy and combination were not significant (Fig. 5G). PD-1 expression on eWAT T cells was highest in MCS-treated mice. For CD4⁺ T cells, the combination treatment significantly reduced PD-1 levels relative to MCS, while CD8⁺ T cells showed no significant differences across groups (Fig. 5H).

IL-12/IL-15/IL-18 + CASP1 treatment induced distinct tissue-specific redistribution of T cell subsets, characterized by increased Tem populations in the spleen, limited MPEC accumulation in ascites, and marked depletion of both CD4⁺ and CD8⁺ MPECs in eWAT.

These findings suggest that the cytokine cocktail differentially shapes local T cell fate depending on the tissue microenvironment (Additional file 1, Figure S3).

Overall, these results indicate that co-administration of IL-12, IL-15, and IL-18/CASP1 mRNA enhances effector T cell frequencies and functional activity, while reducing Treg populations and expression of exhaustion markers, with effects observable across systemic and tumor-associated compartments.

Reduced-dose mRNA cytokine therapy maintains antitumor efficacy with improved safety profile

Although co-administration of high-dose mRNA-expressing IL-12, IL-15, and IL-18/CASP1 showed strong antitumor activity, a mortality event occurred in this treatment group during the study, prompting evaluation of treatment-related toxicity.

To mitigate toxicity while maintaining therapeutic efficacy, a reduced cytokine mRNA dose (5 µg total per cytokine) was evaluated, preserving the same proportional composition of IL-12, IL-15, IL-18, and CASP1 (Fig. 6A–B). Abdominal circumference, as a surrogate measure of ascites accumulation, was significantly lower in the reduced-dose IL-12/IL-18/CASP1 and IL-12/IL-15/IL-18/CASP1 groups than in the Nil, R/L groups, and most other treatment arms, consistent with longitudinal abdominal measurements (Fig. 6C). Quantification of tumor burden, calculated as the product of tumor area and photon flux per mouse, is displayed as a longitudinal tumor intensity data (Fig. 6D). In vivo bioluminescence imaging provided visual assessment of tumor burden across all treatment groups and time points (Fig. 6F). Notably, reduced-dose IL-12/IL-18/CASP1 and IL-12/IL-15/IL-18/CASP1 combinations achieved tumor control comparable to the initial high-dose regimen, without evidence of significant toxicity (Additional file 1, Figure S4). Body weight remained stable, and AST and ALT levels were within normal ranges (Additional file 1, Figure S4A and S4B). No mortality events were observed in the reduced-dose IL-12/IL-18/CASP1 or IL-12/IL-15/IL-18/CASP1 groups. Isolated deaths occurred in other groups: one mouse in the R/L group and one in the IL-15/IL-18/CASP1 group died around day 19, and one in the IL-12 group died around day 13; no other groups experienced mortality (Fig. 6E). Consistent with the strong antitumor activity observed in the IL-12/IL-15/IL-18/CASP1 group in Fig. 2E, the reduced-dose regimen maintained measurable therapeutic benefit while notably improving tolerability. Importantly, no treatment-related mortality was observed in the reduced-dose IL-12/IL-15/IL-18/CASP1 group, in contrast to the higher-dose combination. Although the lower-dose regimen did not further enhance antitumor efficacy compared with the higher-dose treatment, these findings indicate that dose reduction can mitigate toxicity while preserving meaningful antitumor effects.

Fig. 6.

Anti-tumor effects induced by a reduced dose of cytokines and different combination strategies. C57BL/6 mice were intraperitoneally injected with 1 × 10⁷ ID8-Fluc cells, and two weeks later intraperitoneally administered once per week with LNP-formulated mRNA combinations encoding scIL-12, scIL-15, and IL-18/CASP1 for three weeks. (A-B) Schematic illustrating the experimental design. C57BL/6 mice were intraperitoneally administered a total of 15 µg LNP-formulated mRNA, comprising a combination of cytokines encoding scIL-12, scIL-15, and IL-18/CASP1. (C) Change in abdominal circumference measured as an indicator of ascites accumulation. (D) Luminescence of ID8-Fluc cells was measured weekly using an in vivo imaging system (n = 5). (E) Kaplan–Meier survival analysis during the experimental period. (F) Representative in vivo bioluminescence images of mice from each treatment group at the indicated time points. Data are presented as mean ± SD.*p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.X indicates a dead mouse. ID8-Fluc, luciferase-expressing ID8 ovarian cancer cells; IL, interleukin; LNP, lipid nanoparticle; mRNA, messenger RNA; scIL, single-chain interleukin; SD, standard deviation

Metastatic suppression under reduced-dose conditions mirrored the high-dose regimen, particularly in eWAT, mesentery, and pancreas (Additional file 1, Figure S5). Immune profiling of ascites fluid revealed broadly consistent findings with those observed under the high-dose treatment, including increased immune cell infiltration and modulation of effector and suppressor cell subsets (Additional file 1, Figures S6–S7). These results indicate that dose-adjusted cytokine mRNA therapy can preserve antitumor efficacy while improving the safety profile.

Combination cytokine therapy coordinates multilevel transcriptional programs in the peritoneal microenvironment

To explore potential molecular mechanisms underlying macrophage modulation and broader immune remodeling, RNA sequencing was performed on peritoneal fluid cells from mice treated with individual cytokines (IL-12, IL-15, IL-18/CASP1) or the combination therapy (IL-12/IL-15/IL-18/CASP1). Genes exhibiting moderate to high expression changes (absolute log2 fold-change ≥ 0.5) with an adjusted p-value < 0.05 were included. TPM-normalized values were used for heatmap visualization and PCA, as TPM accounts for sequencing depth and gene length, enabling robust comparison of relative transcriptional patterns across groups. For volcano plots, FPKM values were used to quantify differential expression between the combination and the mean of single cytokine treatments, with log2 transformation applied to stabilize variance and improve interpretability of fold changes.

The heatmap illustrates that combination-treated samples display a transcriptional signature clearly distinct from all individual cytokine treatments (Fig. 7A). Expression values were log2-transformed and row-standardized (z-score), allowing visualization of relative up- and downregulation across genes. The IL-12/IL-15/IL-18/CASP1 group clusters separately, reflecting a coordinated transcriptional program rather than a simple additive effect of the individual cytokines. Notably, this unique signature encompasses hundreds of genes involved in transcriptional regulation, apoptotic control, metabolic processes, and cytokine responsiveness.

Fig. 7.

Transcriptional programs induced by combination cytokine therapy in peritoneal fluid cells. (A) Heatmap of differentially expressed genes (log2 fold-change ≥ 0.5, adjusted p < 0.05) using row-wise z-score–standardized, log2 TPM values. Hierarchical clustering (average linkage, Euclidean distance) was applied to genes; columns represent MCS, IL-15, IL-18/CASP1, IL-12, and IL-12/15/18/CASP1. The combination therapy induces a distinct transcriptional signature, with coordinated gene regulation. (B) PCA of z-score standardized TPM data showing clear separation of combination-treated samples from single-cytokine groups, indicating global transcriptional reprogramming (PC1–PC2). (C) GO enrichment analysis of upregulated genes in the combination group (top 20 significantly enriched Biological Process terms, adjusted p < 0.05). Dot size represents the gene ratio (number of upregulated genes in the term relative to the total number of upregulated genes), and color represents −log10 adjusted p-value. Hierarchical clustering of GO terms highlights functional clusters, revealing engagement of processes related to cellular responses to external stimuli, metabolic and cellular regulation, stress responses, and signal transduction. (D) Volcano plot comparing the combination therapy versus the mean of single cytokine treatments (|log2 fold-change| ≥ 1, adjusted p < 0.05), showing globally altered gene expression. (E) Selected key immune-regulatory and macrophage-associated genes differentially expressed in the combination group. Categories include PDCD-related apoptosis regulators, NF-κB signaling components, STAT-mediated immune modulators, and macrophage/cytokine response genes. Log2 fold-change, adjusted p-values, and expression direction are indicated. Upregulated genes in the combination therapy include ARG1, CHIL3, MRC1, NKRF, NFRKB, and SOCS3, reflecting activation of macrophage-associated and immune-modulatory pathways, whereas downregulated genes include PDCD4, PDCD6, NFIB, and NFAT5, indicating suppression of apoptosis and certain transcriptional regulators. Notably, classical M1-associated genes were not significantly detected, consistent with preferential activation of alternative macrophage and immunoregulatory programs.GO, Gene Ontology; IL, interleukin; MCS, mock control system; PCA, principal component analysis; TPM, transcripts per million

PCA confirmed the global distinction of combination-treated samples (Fig. 7B). Projection onto the first two principal components derived from row-standardized TPM values revealed substantial separation of the combination therapy from individual cytokine treatments, with pairwise PC1–PC2 distances markedly larger than those among the single treatments.

GO analysis of upregulated genes in the combination group revealed the top 20 “up-biased” BP terms (Fig. 7C), representing processes most strongly associated with genes selectively increased in expression relative to single cytokine treatments. These enriched terms include cellular responses to external stimuli, regulation of metabolic and cellular processes, stress responses, and signal transduction, reflecting the functional programs preferentially engaged by the combination therapy. In this context, the up-biased BP terms indicate the molecular pathways through which the combination can exert coordinated effects on the peritoneal microenvironment.

To further explore these effects at the level of individual genes, volcano plot analysis comparing the combination therapy with the mean of single cytokine treatments was performed (Fig. 7D). This analysis highlights transcriptional features underlying the functional programs identified by the GO terms. M2-associated markers—including ARG1, CHIL3, and MRC1—were upregulated, suggesting induction of alternative macrophage programs. Apoptosis-related genes were differentially regulated, with PDCD4 and PDCD6 downregulated and PDCD10 upregulated, reflecting selective modulation of cell-survival pathways. Transcriptional regulators associated with NF-κB (RELA, NKRF) and STAT-mediated signaling (NFIB, NFAT5, STAT1) were also differentially expressed, indicating adjustments in transcriptional and stress-response networks. Classical M1 markers were not significantly detected, consistent with a lack of strong classical M1 polarization. Key immune regulatory genes, macrophage markers, apoptosis regulators, and transcriptional modulators are summarized in Fig. 7E to facilitate mechanistic interpretation of the combination therapy’s transcriptional impact.

Combination therapy clustered separately from individual cytokine treatments in heatmap analysis and PCA. GO enrichment and volcano plot analyses identified differentially expressed genes associated with immune regulation, metabolic processes, and stress-response pathways in peritoneal fluid cells.

Discussion

Ovarian cancer is often diagnosed at an advanced stage and is characterized by peritoneal dissemination and the accumulation of malignant ascites (MA). MA contributes to the transport of tumor cells to the peritoneum, creating a unique microenvironment comprising tumor cells, immune cells, and various other factors. This environment plays a critical role in tumor growth and immune evasion, promoting immunosuppression that interferes with antitumor immunity. Consequently, a tumor-supportive milieu that facilitates tumor progression and metastasis is established [21].

IL-12 is a potent pro-inflammatory cytokine that induces Th1-type T cell responses and has shown synergistic effects when combined with IL-15 and IL-18 [22–26]. Previous studies have predominantly focused on IL-12/IL-15/IL-18 treatment strategies aimed at stimulating NK cells ex vivo or delivering recombinant proteins to induce antitumor effects [23, 27]. More recently, mRNA-LNP–mediated delivery of IL-12 has progressed into early clinical development, including STX-001, a programmable self-replicating mRNA-LNP designed to drive IL-12 expression in defined cellular and microenvironmental contexts [28]. Preclinical studies have likewise demonstrated the feasibility of in vivo IL-12 expression using mRNA- or nanoparticle-based platforms [29].

Building upon these advances, the present study evaluates a distinct immunomodulatory strategy based on the coordinated in vivo delivery of IL-12, IL-15, and IL-18 together with CASP1, with the goal of broadly reprogramming the tumor immune microenvironment in peritoneally disseminated ovarian cancer. Rather than focusing on single-cytokine activity, our approach leverages cytokine synergy to simultaneously enhance effector T cell function, promote macrophage repolarization, and alleviate immune exhaustion within the peritoneal TME. Intraperitoneal administration enables localized yet measurable systemic cytokine expression and addresses several limitations associated with protein-based therapies, including short half-life and inefficient tumor-site delivery. Monocytes and neutrophils, key components of the myeloid cell population, have been identified as the primary contributors to mRNA-LNP uptake, suggesting that these cells play a pivotal role in the initial immune response to mRNA-based cytokine delivery.

Our GFP reporter experiments demonstrated preferential uptake of mRNA-LNPs by myeloid populations, including macrophages and dendritic cells, whereas T cells exhibited minimal direct uptake. This distribution is consistent with previous reports showing that LNPs are efficiently internalized by phagocytic and antigen-presenting cells following systemic or intraperitoneal administration [30–32]. Importantly, this uptake pattern aligns with the intended mechanism of our therapeutic strategy. Rather than requiring direct transfection of effector lymphocytes, cytokine expression within myeloid cells enables these cells to act as local cytokine producers, amplifying paracrine activation of CD8⁺ T cells, CD4⁺ T cells, and NK cells within the peritoneal TME. Such indirect immune activation leverages the intrinsic immunoregulatory and antigen-presenting functions of myeloid cells and may contribute to enhanced immune coordination while limiting systemic cytokine exposure.

CASP1 was included to promote inflammasome-dependent processing of IL-18 and thereby enhance downstream Th1-type immune responses. Although systemic CASP1 activation could theoretically induce excessive release of IL-1β or IL-18 and associated toxicity, we did not observe a pathological increase in serum IL-1β following CASP1–containing mRNA treatment. These data, now included in Additional file 1 (Figures S9 and S10), suggest that overt systemic inflammasome overactivation did not occur under the dosing conditions used. The transient and localized nature of mRNA-mediated expression may further limit sustained CASP1 activity and systemic cytokine release. Nevertheless, future studies incorporating direct measurement of circulating CASP1 activity and broader inflammasome-associated cytokine profiling will be important for translational development.

To further clarify the functional contribution of CASP1, our findings provide mechanistic insight into its role within this combinatorial cytokine strategy. CASP1 is not considered to exert direct antitumor effects as a standalone component; however, its inclusion likely enhances cytokine-mediated immune responses when combined with IL-12, IL-15, and IL-18. This is consistent with its established role in inflammasome-dependent proteolytic processing of pro-IL-18 into its mature, biologically active form [11, 12]. Given that IL-18 activity depends on CASP1–mediated cleavage, its incorporation likely facilitates efficient IL-18 maturation and amplifies downstream Th1-type immune responses, including IFN-γ production and cytotoxic effector activation [8–10]. Accordingly, a cytokine combination lacking CASP1 may not fully recapitulate functional IL-18 signaling. Overall, these findings support a model in which CASP1 acts as a critical upstream regulator that enables and potentiates cytokine-driven immune activation within the peritoneal TME.

In our study, the therapeutic potential of mRNA-expressed IL-12, IL-15, IL-18, and CASP1 was evaluated in the ID8-Fluc mouse model of ovarian cancer. Notably, both scIL-12-mRNA monotherapy and the combination treatment group (IL-12-, IL-15-, IL-18-, and CASP1-mRNA) demonstrated significant tumor suppression. This was accompanied by reduced ascites formation and metastasis to distant organs.

Within the TME of ascites, we observed a notable shift in the macrophage-to-monocyte ratio. The IL-12 and IL-12/IL-15/IL-18/CASP1 groups, both of which demonstrated tumor suppression, exhibited a decrease in the macrophage population, coupled with an increase in the monocyte population. This suggests that the reduction of immunosuppressive macrophages and the influx of monocytes may contribute to the observed therapeutic effects. Macrophages in ascitic fluid were predominantly of the immunosuppressive M2 phenotype. Interestingly, analysis of macrophage polarization in metastatic sites (ascites and eWAT) revealed a shift toward the M1 phenotype in both scIL-12 and IL-12/IL-15/IL-18/CASP1 treatment groups. This transition from M2 to M1 macrophages, known to be associated with antitumor immunity, further substantiates the concept that mRNA-based cytokine therapy can reprogram the TME to favor tumor suppression.

IL-12, recognized for its antitumor effects primarily through T cell activation, particularly Th1 cells, is central to our therapeutic strategy [33]. Our findings confirm that both CD8⁺ and CD4⁺ T cell populations expanded in the spleen, ascites, and eWAT in response to IL-12 and IL-12/IL-15/IL-18/CASP1 treatments. Notably, IL-12 treatment elicited an enhanced IFN-γ response from both CD4⁺ and CD8⁺ T cells, indicative of their cytotoxic potential. These findings are consistent with the hypothesis that IL-12 can alleviate immunosuppression in the peritoneal cavity, facilitating the activation of CD8⁺ and CD4⁺ T cells, which are critical for antitumor responses.

Importantly, this expansion was accompanied by a significant increase in effector T cells and a concurrent reduction in Tregs, particularly in the ascites and eWAT. As Tregs are known to suppress effector T cell function and contribute to immune evasion in the TME, their depletion suggests a favorable rebalancing of immune populations toward antitumor immunity. Additionally, the expression of T cell exhaustion markers Tim-3 and PD-1 was significantly decreased in T cells from treated mice, particularly in the IL-12/IL-15/IL-18/CASP1 group. This indicates that the cytokine cocktail expands cytotoxic T cells and restores their functional competence by reversing exhaustion. Collectively, these findings highlight the dual immunomodulatory role of this therapy in enhancing effector function while alleviating immunosuppression, thereby reprogramming the TME into a more immune-permissive state.

To further elucidate the molecular mechanisms underlying this immune reprogramming, we performed RNA sequencing analysis of peritoneal immune cells following IL-12/IL-15/IL-18/CASP1 combination therapy. The combination treatment induced broad transcriptional and molecular changes, characterized by coordinated engagement of multiple functional programs rather than amplification of a single inflammatory pathway. GO analysis of the top 20 up-biased BP terms revealed enrichment of cellular responses to external stimuli, regulation of metabolic and cellular processes, stress responses, and signal transduction pathways. These transcriptional signatures suggest that the cytokine combination modulates the peritoneal microenvironment by orchestrating diverse immune-regulatory programs rather than driving excessive inflammation.

At the level of individual genes, volcano plot analysis comparing combination therapy with the mean response of single-cytokine treatments revealed differential regulation of macrophage-associated and immune-modulatory programs. Notably, transcripts associated with alternative macrophage activation, including ARG1, CHIL3, and MRC1, were increased, indicating engagement of non-classical macrophage programs [34, 35]. In parallel, apoptosis-related regulators PDCD4 and PDCD6 were downregulated, whereas PDCD10 was upregulated, suggesting selective modulation of cell survival pathways [36].

Transcriptional regulators linked to STAT signaling (NFIB and NFAT5) [37, 38] and NF-κB signaling were also differentially represented. Importantly, RELA, a core transcriptional subunit of the canonical NF-κB complex, was downregulated, while NKRF [39], a negative regulator of NF-κB–dependent transcription, was upregulated. Together, these changes indicate that the combination therapy fine-tunes NF-κB–mediated inflammatory signaling rather than inducing unchecked canonical NF-κB activation. Consistent with this interpretation, classical M1 macrophage markers were not prominently enriched, suggesting that the therapy promotes immune reprogramming through balanced transcriptional modulation rather than polarized macrophage activation alone.

Although canonical dendritic cell or NK cell marker genes were not strongly detected in the RNA-seq dataset, these transcriptional profiles complement the flow cytometry findings by providing molecular evidence for altered immune regulatory states induced by the cytokine cocktail. Collectively, the RNA-seq data support the concept that IL-12/IL-15/IL-18/CASP1 therapy reshapes the peritoneal immune landscape through coordinated transcriptional tuning of macrophage function, survival pathways, and inflammatory signaling, thereby reinforcing functional immune reprogramming observed at the cellular level.

Conventional intraperitoneal chemotherapy, such as doxorubicin or paclitaxel, effectively reduces tumor burden in ovarian cancer peritoneal metastasis models but primarily targets tumor cell proliferation and exhibits limited capacity for immune reprogramming [40, 41]. In contrast, cytokine-based immunotherapies and mRNA-based approaches have demonstrated the ability to reduce immunosuppressive macrophage populations, restore T cell effector function, and reshape the immune landscape [42, 43]. These observations suggest that immunologically driven strategies may complement conventional chemotherapy, warranting future comparative and combinatorial studies.

Dose optimization is a critical consideration in cytokine-based immunotherapy, particularly in ovarian cancer with peritoneal dissemination, where systemic exposure is closely linked to dose-limiting toxicity. In intraperitoneal ovarian cancer models, localized or optimized dosing of IL-12 induces robust immune-mediated tumor control while minimizing systemic adverse effects, whereas higher or systemic doses provide limited additional benefit and increased toxicity [44]. Similar dose-dependent relationships have been reported for IL-15 and IL-18 in peritoneal metastasis models, highlighting the narrow therapeutic window of cytokine therapies and the necessity of controlled biodistribution to balance efficacy and safety [45, 46].

To rationalize the use of multiple cytokines rather than dose escalation of a single factor, it is important to consider the multifaceted nature of immune suppression within the ovarian cancer TME. Effective antitumor immunity requires coordinated activation of both innate and adaptive immune compartments, as well as sustained effector function over time. In this context, IL-12, IL-15, and IL-18 serve distinct yet complementary roles. IL-12 is a key driver of Th1 polarization and cytotoxic T cell activation, IL-15 supports the survival and persistence of NK cells and memory CD8⁺ T cells, and IL-18 cooperates with IL-12 to amplify IFN-γ–mediated effector responses [47, 48]. Although IL-12 plays a central role in initiating T cell–mediated antitumor immunity, extensive clinical and preclinical studies have demonstrated that dose escalation of IL-12 alone is constrained by systemic toxicity and does not necessarily translate into durable antitumor responses [47]. In contrast, combinatorial cytokine strategies distribute immune activation across multiple signaling pathways, thereby enhancing efficacy while potentially mitigating cytokine-specific toxicities. The inclusion of IL-15 and IL-18 complements IL-12–driven T cell priming by sustaining effector cell function and reinforcing innate immune activation, which may reduce the likelihood of rapid T cell exhaustion [48]. Furthermore, CASP1 facilitates the maturation of IL-18 and other inflammasome-associated mediators, strengthening innate immune signaling and promoting effective crosstalk between innate and adaptive immunity [49]. Collectively, these mechanistic considerations support a combinatorial cytokine strategy as a more balanced and effective approach for TME reprogramming than increasing the dose of a single cytokine.

Although IL-12 has demonstrated promising antitumor effects, it is not without risks. Clinical trials have reported significant adverse events, including toxicity and mortality, associated with high-dose IL-12 administration [50]. Co-administration of IL-12 and IL-18, in particular, has been linked to severe side effects such as weight loss, hemorrhagic colitis, and damage to the lungs and liver, primarily due to excessive systemic IFN-γ production [51–53]. In our study, similar systemic toxicity was observed, including weight loss and elevated liver enzyme levels (AST and ALT) in the IL-12 and IL-12/IL-15/IL-18/CASP1 treatment groups. One case of mortality occurred in the IL-12/IL-15/IL-18/CASP1 group, consistent with severe treatment-associated toxicity. These findings underscore the importance of minimizing off-target cytokine exposure in translational applications. mRNA-based cytokine delivery enables transient and localized expression of IL-12 within the TME, reducing systemic exposure while allowing temporal control over cytokine activity [54]. In addition, complementary strategies such as tumor-restricted promoters, biomaterial-anchored cytokine delivery systems, and nanoparticle-mediated targeting have been proposed to further confine IL-12 activity to the tumor site while preserving immune activation [55, 56]. Together, these approaches highlight that spatial and temporal restriction of cytokine signaling, rather than dose escalation alone, is central to improving the therapeutic window of IL-12–based immunotherapy.

Previous studies have demonstrated that tumor-restricted IL-12 gene therapy strategies, including intratumoral plasmid electroporation and mRNA-based delivery, can elicit robust antitumor immunity while limiting systemic cytokine exposure [57, 58]. These approaches underscore the importance of spatial control of cytokine activity, rather than dose reduction alone, for improving the therapeutic window of cytokine-based immunotherapies. Accordingly, our reduced-dose regimen should be viewed as a pragmatic preclinical optimization rather than a definitive safety solution. To mitigate toxicity in the current model, we conducted a follow-up experiment in which each cytokine dose was reduced to 5 µg and the combination regimen was adjusted, resulting in improved tolerability while maintaining antitumor activity. Nevertheless, we emphasize that murine tolerability does not directly predict human safety, and that future development of this platform will require incorporation of tumor-targeting or spatially restricted delivery strategies to further minimize systemic exposure.

From a safety standpoint, the transient expression of mRNA-expressed cytokines is advantageous, as they exert necessary immunological functions briefly before degrading, unlike persistently active protein-based cytokines. Moreover, our data suggest that cytokine synergy is maintained even at reduced doses, potentially expanding the therapeutic window and facilitating safer combination immunotherapy.

Conclusions

This study demonstrates that IL-12- and IL-12/IL-15/IL-18/CASP1-mRNA therapies can effectively reprogram the TME in ovarian cancer, promoting a shift in immune cell populations promoting a shift in immune cell populations toward a more proinflammatory and cytotoxic phenotype. This mRNA-based cytokine approach serves as a potent and tunable platform for inducing robust antitumor immunity.

Importantly, RNA-seq analysis further revealed that the IL-12/IL-15/IL-18/CASP1 combination therapy induces coordinated transcriptional reprogramming of peritoneal immune cells, fine-tuning inflammatory, metabolic, and immune-regulatory pathways rather than amplifying a single proinflammatory program. Going forward, such strategies hold significant potential for ovarian cancer and for other malignancies characterized by immunosuppressive microenvironments. The optimization of delivery methods, dosing regimens, and tumor-specific targeting will be crucial for maximizing therapeutic efficacy while minimizing off-target effects, thereby enhancing the clinical applicability of mRNA immunotherapy platforms.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

ALT

Alanine aminotransferase

AST

Aspartate aminotransferase

CD4

Cluster of differentiation 4

CD8

Cluster of differentiation 8

Caspase-1

Cysteine-aspartic acid protease 1

ELISA

Enzyme-linked immunosorbent assay

FACS

Fluorescence-activated cell sorting

IFN-γ

Interferon gamma

IHC

Immunohistochemistry

IL

Interleukin

LNPs

Lipid nanoparticles

LPS

Lipopolysaccharide

M1

Classically activated macrophage

M2

Alternatively activated macrophage

PBS

Phosphate-buffered saline

PD-1

Programmed cell death protein 1

Scanpy

Single-cell analysis in Python

TME

Tumor microenvironment

TIM-3

T cell immunoglobulin and mucin-domain containing-3

TNF-α

Tumor necrosis factor alpha

UMAP

Uniform manifold approximation and projection

Author contributions

Y.L, and J.N conceived the idea of study. Y.L, J.L, Y.L, H.Y, S.B, S.Y, S.L, G.R, Y.C, S.C, D.H, A.O, S.L, E.C, H.C, S.J, J.K, S.L, H.P, S.J, S.P, W.Y, J.K and T.C performed the experiments. Y.L, J.L, Y.L, H.Y and J.N conducted the data analysis and contributed to the writing of the manuscript.

Funding

This research was supported by grants from the Ministry of Food and Drug Safety (grant numbers: 22213MFDS421 and RS-2025-02213409) and partially supported by the Brain Korea 21 Four Program. Schematics were created with the support of BioRender (biorender.com). Yu-Sun Lee, Jisun Lee, Yeeun Lee, Hyunho Yoon contribuhrted equally to this work.

Data availability

Source data are available for this study. Data supporting the findings of this study are available from the corresponding authors upon reasonable request.

Declarations

Ethics approval and consent to participate

All experimental procedures involving animals were conducted in accordance with institutional guidelines for animal care and use and were approved by the Institutional Animal Care and Use Committee of the Catholic University of Korea (approval nos. CUK-IACUC-2022-031 and CUK-IACUC-2022-044). The study was conducted in compliance with the ARRIVE guidelines. No human participants, human data, or human tissues were used; therefore, informed consent was not required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.