Autologous tumor-immune effusion co-cultures enable ex vivo functional profiling of radiotherapy-immunotherapy combinations

Autologous tumor-immune effusion co-cultures enable ex vivo functional profiling of radiotherapy-immunotherapy combinations | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Autologous tumor-immune effusion co-cultures enable ex vivo functional profiling of radiotherapy-immunotherapy combinations Rebecca Zirnbauer, Daphni Ammon, Berta Mosleh, Nora Speiser, Anna Theophil, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8516163/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 14 Apr, 2026 Read the published version in Journal of Experimental & Clinical Cancer Research → Version 1 posted 9 You are reading this latest preprint version Abstract Background Currently, radiotherapy-immunotherapy (RT-IO) combinations provide limited and heterogeneous benefit in human solid cancers and are frequently selected empirically, partly because human models that preserve native, autologous tumor-immune interactions are lacking. We developed patient-derived autologous tumor-immune effusion co-cultures (PATEC) as an ex vivo platform to functionally evaluate RT-IO regimens within an immunocompetent tumor microenvironment. Methods Malignant pleural and peritoneal effusions (n = 29) from patients with metastatic solid cancers were processed for biobanking and primary tumor culture. Expandable tumor cultures were established in six effusions and recombined with matched autologous immune cells to generate PATEC. PATEC were treated with radiotherapy (RT), innate immune agonists (STING, TLR7/8) and immune checkpoint inhibitors (CTLA-4, PD-1, PD-L1, TIGIT) in combinatorial regimens. Tumor cell death, T-cell activation, cytokine secretion and CD8⁺ T-cell checkpoint expression was assessed using multiparametric flow cytometry and multiplex immunoassays. Contact dependence of cytotoxicity was evaluated by comparing tumor monocultures, direct co-cultures and transwell separated co-cultures. Results Across conditions, regimens combining RT with a stimulator of interferon genes (STING) agonist were the most tumoricidal in PATEC, with marked interpatient variability and Bliss defined synergy in a subset of effusions (3/6). STING agonist-mediated cytotoxicity required immune cells and was attenuated by spatial separation of tumor and immune compartments, whereas RT alone produced similar cytotoxicity in monocultures and co-cultures, indicating a predominantly tumor-intrinsic effect. STING based RT-IO induced early T-cell activation and a type I interferon-rich cytokine milieu, followed by increased expression of multiple inhibitory checkpoints on CD8⁺ T cells. A composite CD8⁺ checkpoint co-expression score correlated with both overall and contact-dependent tumor cell death. Conclusions PATEC enables functional dissection of RT-IO combinations in a native effusion-derived tumor-immune microenvironment and shows that the additional tumor cell killing conferred by STING-based RT-IO depends on immune cells and direct tumor-immune contact and varies between patient samples. These findings support the use of PATEC as a functional ex vivo system for testing therapeutic combinations in a patient-specific setting. Pleural Effusion Malignant Ascites Neoplasms Tumor Microenvironment Coculture Techniques Radiotherapy Immunotherapy Immune Checkpoint Inhibitors Precision Medicine PATEC Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Background Advances in immunotherapy for solid cancers are perpetually evolving. Clinical use of immunomodulatory drugs to combat cancer has become more prominent over the last decade, with patients eligible for immune checkpoint inhibition (ICI) increased to 44% in 2018( 1 ). However, only 12% of patients are estimated to respond to ICI and most indications involve monotherapy ( 1 ). This has driven the inclination of combining immunomodulatory therapies. Within a couple of years several trials administering dual immunotherapy, including antibodies programmed death 1 (PD-1), programmed death ligand 1 (PD-L1), cytotoxic T lymphocyte associated protein 4 (CTLA-4), lymphocyte activation gene 3 (LAG-3) or T cell immunoreceptor with Ig and ITIM domains (TIGIT), as well as combinations with innate agonists targeting stimulator of interferon genes (STING) or Toll like receptor 7/8 (TLR7/8), have been performed( 2 – 6 ). Moreover, preclinical studies indicate that radiotherapy (RT) is a potent immunological adjuvant. Ionizing radiation induces immunogenic tumor cell death, enhances tumor antigen release and MHC class I expression, and activates cGAS-STING-mediated type I interferons (IFN-I) and inflammatory signaling within the tumor microenvironment (TME). These effects collectively promote recruitment and activation of effector lymphocytes and potentiate the efficacy of ICI and innate immune agonists ( 7 – 9 ). Clinical trials effectively combining RT with ICI in locally advanced solid tumors further support this rationale ( 10 – 14 ). Notably, systematic analyses indicate that up to 72% of combinations lacked an explicit biological rationale and indicated most combinations follow an empirical rather than rational design ( 15 , 16 ). One rational approach to design effective radiotherapy-immunotherapy (RT-IO) combinations would be to identify them through functional testing in patient-derived systems. However, hardly any approaches have been implemented, most likely because testing immunotherapies, and ICIs in particular, requires scalable models that preserve a genuine, autologous tumor-immune microenvironment (TIME). Conventional immunocompetent mouse models and other standard preclinical models do not adequately recapitulate the complexity and interpatient heterogeneity of the human immune system in the context of the TIME. Precision-oriented platforms that use patient-derived material offer the potential to gain insights into the local interactions of the TIME and to design combination therapies with a translational rationale. Organotypic cultures derived from malignant effusions appear to offer a promising approach, allowing the expansion of effusion-derived tumor cells alongside accessible autologous immune cells within a clinically, relevant immunosuppressive TIME ( 17 – 19 ). Published effusion-derived cultures have predominantly expanded tumor compartments for chemosensitivity and targeted therapy screening while the autologous immune compartment is not preserved or not functionally investigated. Furthermore, they do not maintain a native effusion TIME for systematic testing of immunotherapeutic combinations in autologous co-cultures ( 17 – 20 ). Malignant pleural and peritoneal effusions provide tumor cells, stromal fibroblasts and autologous enriched tumor-associated immune cells that are naturally adapted to the liquid environment, making them feasible for ex vivo experiments. T cells, B cells and myeloid cells have been characterized as being more similar to tumor-associated immune cells than peripheral blood mononuclear cells (PBMCs) in respect to their phenotypical and transcriptional states ( 21 ). Effusion-derived T cells display increased expression of inhibitory checkpoints and exhaustion markers on T cells are enriched, the transcriptional signatures for regulatory T cells are enhanced and macrophages have continuous phenotypes of M1/M2 with more prominent M2 polarization characteristics ( 21 – 23 ). Functional and TCR sequencing studies further indicate that malignant effusions harbor clonally expanded, memory phenotype and CD39 + CD8⁺ T cell subsets with features of tumor-reactive tumor-infiltrating lymphocytes, together with a cytokine and chemokine milieu enriched for Vascular Endothelial Growth Factor, Interleukin-6, Transforming Growth Factor-beta and C-C motif chemokine ligand 22 that recapitulates key immunosuppressive attributes of the solid TIME ( 24 – 27 ). Collectively, these phenotypic and functional similarities to the solid TIME support the use of malignant effusions as a platform to investigate therapeutics modulating the cancer-associated immune cell environment. In this study, we developed a patient-derived autologous tumor-immune malignant effusion co-culture (PATEC) and used it to assess the functional effects of RT-IO combinations. We systematically tested combinations of innate agonists, as well as ICI in PATEC and identified the STING agonist, particularly in combination with RT, as the most potent inducer of tumor cell death in a subset of patients. The STING agonist-mediated cytotoxicity is largely immune cell dependent and requires direct immune tumor contact. By condensing the distribution of multi checkpoint-positive CD8⁺ T cell phenotypes into a composite Checkpoint Co-expression Score (CES), we link these checkpoint co-expression states specifically to the magnitude of overall and contact-dependent tumor cell killing. Overall, these data provide proof‑of‑concept that malignant effusion-derived PATECs can serve as a patient‑specific ex vivo platform to functionally test and characterize RT-IO regimens within an autologous TIME. Methods Acquisition and processing of malignant effusions Malignant effusions specimens (ascites and pleural) were prospectively collected from cancer patients undergoing drainage procedures performed by the Department of Thoracic Surgery and Department of General Surgery. Only pathological confirmed (Department of Pathology) malignant effusions were included. All patients provided written informed consent form (ICF). The study was approved by the local Ethics Committee (EC) of the Medical University of Vienna (No. 2042/2019). Samples were collected aseptically in sterile containers or fluid bags (1–2 L) and transported to the laboratory. Processed within 3 h of collection under aseptic conditions in laminar flow cabinet samples were filtered through a 100 µm cell strainer (pluriStrainer®, pluriSelect) into 50 mL conical Falcon tubes (Sarstedt AG & Co.) removing fibrotic tissue and aggregates. After multiple washes with phosphate-buffered saline (PBS; Gibco) and multiple centrifugation steps at 400×rcf for 10 min, erythrocytes were lysed using ACK lysis buffer (Gibco™ A1049201). Cell counts were assessed using an automated Sysmex Cell Counter XN 350 (Sysmex). Specimens with fewer than 100×10 6 viable cells were excluded from cultivation of primary tumor cell cultures. Cells were resuspended in CryoStor® CS10 medium (MKCL1984, C2874) in addition to 10 µM Y27632 dihydrochloride (MedChemExpress, HY 10071/CS 0131) and aliquoted into sterile 2 mL cryotubes (VWR), and gradually frozen overnight using CoolCell LX containers (BioCision) at -80°C before transfer to liquid nitrogen storage until further use. Development and characterization of primary tumor cell cultures Primary tumor cell cultures were developed by seeding CD45-depleted cells into standard adherent culture flasks (25 cm²) using Advanced DMEM/F12 medium (Gibco, 12634 010) with 10% fetal bovine serum (FBS, Gibco), 1% Antibiotic Antimycotic solution (Sigma, A5955), and 10 µM Y 27632 dihydrochloride. Cryopreserved malignant effusion-derived cells were rapidly thawed in 100% FBS supplemented with 10 µM Y 27632, washed once with protocol medium and centrifuged at 400×rcf for 5 min. Leukocyte depletion was performed using the EasySep™ Human CD45 Depletion Kit (Stemcell Technologies, Cat#17898) according to the manufacturer’s protocol, using 5 mL polystyrene round bottom tubes (Stemcell Technologies, Cat#38007) and the EasySep magnetic cell isolation system. Tumor-enriched cells were cultured at 37°C in a humidified incubator with 5% CO₂, with medium changes every 2–3 days. Successful primary tumor culture development was defined as growth and proliferation beyond three passages. Immunohistochemical (IHC) analysis of primary tumor cell cultures Primary tumor cells seeded onto gelatin-coated slides were fixed in 2% formaldehyde for 20 min at room temperature after confluency. Antigen retrieval was performed using a citrate buffer in an autoclave at 121°C. Endogenous peroxidase was blocked (BLOXALL™, Vector Laboratories) for 10 min. Following blocking with 10% horse serum, slides were incubated for 60–90 minutes at room temperature with primary antibodies diluted in 10% horse serum: EpCAM (clone HEA 125, MACS Miltenyi, 1:80), pan cytokeratin (clone C11, Cell Signaling, 1:500), and STING (clone D2P2F, Cell Signaling, 1:400). Detection was performed with ImmPRESS™ Excel Polymer Reagent and visualized using DAB substrate (Vector Laboratories). Slides were counterstained with hematoxylin and mounted using Entellan™ (Merck). Negative controls without primary antibody were included. Imaging was performed by bright-field microscopy using the Vectra Polaris™(Akoya Biosciences). Flow cytometry (FCM) analysis of cellular composition Cells isolated from malignant effusions and primary tumor cultures were characterized using multiplex flow cytometry. Following thawing and washing, Fc receptor blockade was performed using human AB serum (diluted 1:1 with PBS containing 2% FBS) at 4°C. Cells were then stained with Zombie Violet™ or Zombie yellow™ viability dye (BioLegend) and incubated with fluorochrome conjugated monoclonal antibodies targeting CD45 (AF700), EpCAM (AF488), CD3 (FITC), CD4 (APC), CD8a (PE, AF700), CD56 (PE), CD14 (PE-Cy7), CD11c (PerCP-eFluor710), HLA-DR (APC-Cy7), CD20 (APC), CD163 (PerCP-eFluor710), CD66b (PE), CD11b (APC-Cy7), CD86 (BV650), CD107a (PE-Cy7), and CD69 (PE). For checkpoint receptor analysis, cells were stained with antibodies specific to CTLA-4 (PE-Dazzle), TIGIT (PerCP-eFluor710), LAG3 (PE-Cy7), PD-1 (BV510), PD-L1 (BV785), and TIM3 (APC-Cy7). Corresponding IgG isotype controls matched for each fluorochrome were included. Compensation was performed using UltraComp eBeads™ (Invitrogen). Data were acquired on a CytoFLEX flow cytometer (Beckman Coulter) and analyzed using Kaluza Analysis 2.1 (Beckman Coulter). Patient-derived Autologous Tumor-immune Effusion Co-culture (PATEC) experiments Primary tumor cells and matched immune cells derived from malignant effusions were co-cultured at a 1:5 ratio (5×10⁴ tumor cells to 2.5×10⁵ immune cells per well) in flat-bottom 24-well plates (Greiner Bio One). The cells were cultured in DMEM/F 12 medium (Gibco) in addition to 10% heat inactivated fetal bovine serum (FBS, Gibco) and 1% Antibiotic Antimycotic Solution (Sigma) in a humidified incubator at 37°C with 5% CO₂. After seeding tumor cells, they were incubated overnight prior to immune cell addition. Selected cultures were subjected to radiotherapy (8 Gy; YXLON Maxishot X-ray system, 200 kV, ~ 1 Gy/min). Immunotherapeutic treatments included STING agonist ADU-S100 (10 µM, InvivoGen), TLR7/8 agonist Resiquimod (R848; 10 µg/mL, InvivoGen), and checkpoint inhibitors ipilimumab (anti CTLA-4; Yervoy, 10 µg/mL), atezolizumab (anti PD-L1; Tecentriq, 10 µg/mL), pembrolizumab (anti PD-1; Keytruda, 10 µg/mL), and tiragolumab (anti-TIGIT; eubio, 10 µg/mL), alone or in combinations as indicated. After 72 h, cells were detached using Accutase (Sigma) for flow cytometric analysis. Experiments were performed in duplicate. Spatial separation of PATEC In order to determine cell contact dependence, tumor cells and immune cells were cultured either in direct contact or spatially separated implementing ThinCert well inserts with 0.4 µm pore size (Greiner Bio One, 662640). The tumor cells were seeded at 5×10 4 cells per well in 24-well plates, while immune cells were seeded into inserts. Immunotherapeutic treatment with the STING agonist (ADU-S100, 10 µM) and RT (8 Gy) was applied as described above. After 72 h incubation at 37°C and 5% CO₂, cells were harvested using Accutase, stained with fluorochrome-conjugated monoclonal antibodies, and analyzed by FCM. FCM analysis of treatment outcomes: tumor cell death and immune checkpoint receptor expression Tumor cell death was quantified after 72 h incubation post-treatment by FCM using Zombie Violet™ viability dye (BioLegend) and Calcein AM (BioLegend). Positive controls were generated through repeated freeze-thaw cycles. Viable (Calcein AM⁺/Zombie Violet⁻) and dead (Zombie Violet⁺) cell populations were quantified, and results were expressed as fold change relative to untreated controls. Additionally, immune checkpoint receptor expression was analyzed on CD8⁺ and CD4⁺ T cells. Assessment of T cell activation Immune cells isolated from malignant effusions were thawed and resuspended in assay medium (DMEM/F12, 10% FBS, antibiotic antimycotic solution). Cells were seeded at 1×10 5 cells per well into 96-well U bottom plates (Greiner Bio One) and treated with STING agonist ADU-S100 (10 µM, InvivoGen) and/or RT (8 Gy; YXLON Maxishot, 200 kV X-ray, ~ 1 Gy/min). Following 24 h of incubation at 37°C and 5% CO₂, cells were stained with fluorochrome-conjugated antibodies against CD3 FITC (BioLegend), CD4 APC (BioLegend), CD8 PE (BioLegend), CD45 eFluor450 (Invitrogen), CD69 PE (Invitrogen), and CD107a PECy7 (BioLegend). Compensation controls were prepared using UltraComp eBeads™ (Invitrogen). Samples were acquired on a DxFLEX flow cytometer (Beckman Coulter), and expression levels of activation markers CD69 and CD107a on CD4⁺ and CD8⁺ T cells were analyzed using Kaluza software (Beckman Coulter). Multiplex cytokine profiling of culture supernatants The culture supernatants were harvested 24 h post-treatment and analyzed regarding cytokine secretion using the Luminex® Human Discovery Assay (13-Plex) LXSAHM-13 (Bio-Techne Ireland Limited) following the manufacturer's protocol. The cytokine panel included TNF-α, IL-6, IP-10, IFN-α, MIP-1α, MIP-1β, MCP-1, IL-1ra, IL-8, RANTES, IFN-β, IFN-γ, and IL-10. Data acquisition was performed on a Luminex® FlexMap 3D, and analyses were conducted using the xPONENT® 4.3 analysis software (Luminex®). Statistical analysis Statistical analyses were performed in R (version 2025.05.1 + 513) using the lme4 and lmerTest packages for linear mixed‑effects models, emmeans or multcomp for multiple comparisons, and ggplot2 together with dplyr and tidyr for data handling and visualization. Treatment effects across multiple patients were evaluated using linear mixed effects models, adjusting for patient-specific random effects. Post hoc pairwise comparisons between treatment groups were conducted using Tukey’s multiple comparisons test. The cytokine concentrations were normalized using z-score transformation for comparative analyses. Additionally, intra-patient analyses were performed using one way ANOVA followed by Tukey’s post hoc test, and, where appropriate, by two-way ANOVA with factors Treatment and Culture condition followed by Tukey’s post hoc test. For checkpoint expression panels with multiple CD8⁺ subsets, family-wise multiplicity across the panel was controlled (Holm or Benjamini-Hochberg false discovery rate) in addition to the within-panel Dunnett adjustments. For cytokines, p-values were adjusted within the cytokine family using the Benjamini-Hochberg FDR. Statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. CD8⁺ checkpoint panels and CES-killing analyses were conducted on the logit scale. Percent positive values were logit-transformed and analyzed using Gaussian linear mixed effects models with Condition as a fixed effect and patient as a random intercept; a random slope for Condition was included when supported by the split-plot design. For display, values are shown as fold change versus the patient matched control, while inference used the logit model. Omnibus p-values for Condition derive from the mixed model; multiplicity-adjusted post hoc comparisons versus control were obtained using Dunnett contrasts. The CD8⁺ Checkpoint Co-expression Score (CES; Day 3, within PATEC) was computed by z-scoring the %Gated frequencies of PD-1⁺LAG-3⁺TIM-3⁺, PD-1⁺TIGIT⁺LAG-3⁺, PD-1⁺TIGIT⁺TIM-3⁺ and PD-1⁺TIGIT⁺LAG-3⁺TIM-3⁺ and combining as CES = mean(z of the three triple positive subsets) + 2×z(quadruple positive). Tumor cell death at 72 h was summarized as Δlogit = logit(%dead_treated) – logit(%dead_control), and the contact-dependent component as Δ(contact) = Δlogit(Direct) − Δlogit(Insert). Associations with CES were evaluated by Gaussian LMMs with patient random intercepts (Δlogit ~ CES + (1|patient); Δ(contact) ~ CES + (1|patient)); fixed effect slopes (β) are reported in log odds per CES unit with Wald 95% confidence intervals, Satterthwaite P values (lmerTest), and Nakagawa R² (marginal; conditional). For Δ(contact), the random intercept variance was ≈ 0 (singular fit), hence R²_cond ≈ R²_marg; improvement over a null (intercept only) model was checked by likelihood ratio test and leave one PATEC out refits were used as a stability check. Synergy between STING agonist and RT was quantified using the Bliss independence model applied to control-normalized fractional killing. For each patient and experiment, the percentage of Zombie Violet™-positive tumor cells in untreated controls (d_ctrl) and in treated wells (d) was converted to fractions and expressed as additional killing on the surviving population, \(\:\varvec{E}=\frac{\varvec{d}-{\varvec{d}}_{\text{ctrl}}}{1-{\varvec{d}}_{\text{ctrl}}}\) , truncated to the interval [0,1]. Bliss expectation was calculated as \(\:({\varvec{E}}_{\text{bliss}}={\varvec{E}}_{\text{STING}}+{\varvec{E}}_{\text{RT}}-{\varvec{E}}_{\text{STING}}\times\:{\varvec{E}}_{\text{RT}})\) . The Bliss synergy index is defined as \(\:\varvec{\Delta\:}\varvec{E}={\varvec{E}}_{\text{obs}}-{\varvec{E}}_{\text{bliss}}\) ​, expressed as percentage points. For each patient, replicate level \(\:\varvec{\Delta\:}\varvec{E}\) values were summarized as the mean with 2,000-sample non-parametric bootstrap 95% confidence intervals; combinations were classified as synergistic when the entire confidence interval lay above + 5 percentage points, antagonistic when it lay below − 5 percentage points, and otherwise as neutral. Results Development and characterization of patient-derived autologous tumor-immune malignant effusion co-cultures (PATEC) We hypothesized that malignant effusions, which contain both tumor and immune cells within their native fluid microenvironment, could serve as a source for generating autologous co-cultures to study RT-IO responses. To test this, malignant pleural and peritoneal effusions (pleural = 23, ascites = 6) were prospectively collected and processed for biobanking and primary tumor cell culture (Fig. 1 A). Nine samples were excluded due to insufficient viable cell counts (< 100 × 10⁶ cells). Of the remaining 20 cultures initiated, six maintained continuous proliferation beyond three passages and were expanded as primary tumor cell cultures (PTCCs) for downstream functional co-culture assays (Fig. 1 B). Immunohistochemical and morphological validation of primary tumor cell cultures (PTCC) FCM analysis within the CD45⁻ compartment showed tumor cells expressed the epithelial marker EpCAM in five of six cases, consistent with their tumor entity, while one sample (EF4), derived from a small cell lung carcinoma (SCLC), lacked EpCAM expression, in line with its clinical histopathology report (Fig. 1 E). To confirm epithelial tumor origin, Immunohistochemical (IHC) staining of PTCCs was performed. PTCCs displayed strong cytoplasmic expression of pan cytokeratin (PanCK) and membranous expression of EpCAM, confirming epithelial origin (Fig. 1 D). Brightfield microscopy of PTCCs showed cohesive clusters of polygonal, adherent cells with characteristic epithelial morphology. Heterogeneous immune cell compositions of malignant effusions FCM demonstrated pronounced heterogeneity in the cellular composition of malignant effusions (EF1-EF6) (Fig. 1 F, Supplementary Fig. 2 ). Leukocytes (CD45⁺) were the most dominant population in most samples, making up over 90% of cells. The immune compartment showed patient-specific variation, with macrophages and neutrophils as the most abundant subsets in half of the patient samples, and variable proportions of CD4⁺ and CD8⁺ T cells, NK cells, NKT cells, dendritic cells, and B cells. This interpatient diversity reflects the individualized immune and tumor composition of malignant effusions, provides a genuine TIME and thus an optimal basis for patient-specific functional testing of immunomodulatory drugs. Collectively, these results establish a workflow to generate PTCCs that can be paired with autologous immune compartments, enabling downstream use of PATEC platform for functional RT-IO testing. With scalable patient-matched tumor and immune cells from one compartment, we next assessed therapeutic efficacy across immunomodulatory combinations to determine the most efficacious treatment regimen. A Schematic representation of experimental methodology. B CONSORT diagram for development of PTCC. C Cellular composition of malignant effusions (EF1-EF6, n = 6), showing proportions of CD45⁺ leukocytes and CD45⁻ non‑hematopoietic cells quantified by flow cytometry. D Immune subsets within the CD45⁺ compartment depicting relative abundances. E Immunohistochemical and morphological characterization of PTCC and PATEC Representative PTCC images show tumor cells positive for EpCAM and PanCK by IHC (brown chromogen) with hematoxylin counterstain (blue), and a bright‑field image illustrates a PATEC model. Combinatorial RT-IO screening in PATEC identifies STING agonism and RT as an effective therapeutic combination To functionally assess the capacity of the PATEC platform to capture immune mediated tumor cytotoxicity, we performed an initial combinatorial screen integrating innate and adaptive immunotherapies with RT (8 Gy) (Fig. 2 A, Supplementary Fig. 1A/B ). Quadruple RT-IO included two innate agonists targeting STING and TLR7/8 together with ICI PD-1, PD-L1, CTLA-4, and TIGIT. Tumor cell death was quantified at 72 h by FCM using Zombie Violet™ and Calcein AM. Representative histograms illustrate a marked shift towards Zombie Violet⁺ cells following treatment compared with RT alone (Fig. 2 B, left). Across four independent PATECs, quadruple RT-IO that included an innate agonist (STING agonist or TLR7/8 agonist) together with RT significantly increased tumor cell death compared with RT alone or untreated controls (Fig. 2 B, right; component wise linear mixed effects model with patient random intercept; Holm adjustment; p < 0.001). This demonstrates that PATEC can be used as an ex vivo human system capable of detecting immune mediated tumor killing under relevant combination regimens. Combinations with STING agonist contributed most strongly to tumor killing To delineate which innate immune modulatory components contributed to the observed response, we next compared STING agonist and TLR7/8-Ago, alone or in combination with RT, in representative PATECs (EF1 and EF3). Tumor cell death, expressed as fold change relative to matched controls, increased with either agonist, but the STING agonist + RT combination produced the most pronounced killing across replicates (one way ANOVA with Tukey multiple comparisons; p < 0.05 − 0.001; Fig. 2 C). STING agonist monotherapy also induced measurable cytotoxicity in the selected PATECs, whereas TLR7/8 agonism alone or with RT had more modest effects. Notably, RT potentiated STING agonist efficacy, suggesting cooperative mechanisms between DNA damage signaling and STING activation. Collectively, these findings indicate STING agonism in combination with RT is the leading combination in regard to tumor cell death in these select PATECs. We next investigated interpatient variability and potential synergy between STING agonism and RT across PATECs. A Experimental schematic. B Quadruple RT-IO combinations increase tumor cell death in PATEC. Representative histograms (left) depict tumor cell death measured with flow cytometry at 72 h culture after treatment. Boxplots (right) display the fold change in % of tumor cell death relative to the Control (cntrl; n = 4). Each boxplot summarizes all quadruple immunotherapy combinations containing the immunotherapeutic indicated on the x-axis, with RT (8 Gy) included in all. Colored dots represent independent biological replicates from PATEC. Median, IQR, whiskers 1.5×IQR. LMM; patient random intercept; Holm. C Tumor cell death in PATECs (EF1 and EF3, n = 2) after 72 h of RT, STING agonist, TLR7/8-Ago and their combinations. Representative flow cytometric density plots (left). Boxplots summarize tumor cell death, quantified as fold change relative to control (Δ Control, n = 2). Dots indicate biological replicate. One‑way ANOVA; Tukey multiple comparisons. *p < 0.05; **p < 0.01; ***p < 0.001 Patient-specific variability in response to STING agonist and RT in PATEC To assess patient-specific efficacy of STING agonism and RT, six PATECs (EF1-EF6) were treated with STING agonist, RT, or the combination for 72 h (Fig. 3 A). Across all patients, combination therapy showed the most pronounced tumor cell death, whereas the relative magnitude of response varied markedly between PATECs. EF2 and EF3 exhibited the most pronounced cytotoxicity (two to three-fold), while EF1 and EF6 displayed modest increases. Two-way ANOVA with Tukey’s multiple comparisons confirmed significant treatment effects in all PATECs (p < 0.05-0.0001) (Fig. 3 B). These results demonstrate that STING agonism, particularly in conjunction with RT, induces robust but patient-specific variable tumor cell killing in PATEC. Subset of PATECs exhibit synergistic STING-RT interactions Next, we examined whether combining STING agonism and RT results in greater-than-additive cytotoxicity. Bliss independence modelling identified synergy in a subset of PATECs (Fig. 3 C). EF2, EF3, and EF6 showed strong positive Bliss indices ( > + 5%) with non-overlapping 95% CIs from 2,000 bootstrap resamples, indicating synergy, whereas EF1, EF5, and EF4 fell within the neutrality band. The presence of both synergistic and neutral responses across patients underscores functional heterogeneity in innate immune and RT sensitivity within PATECs. STING agonist induced response depends on immune cells We compared primary tumor monocultures (TuC) with matched autologous tumor-immune co-cultures (PATEC) to assess whether the cytotoxic effects of STING activation are immune cell dependent. After 72 h of STING agonist treatment, tumor cell death increased significantly only in the presence of immune cells (paired Wilcoxon signed rank test; p < 0.01; Fig. 3 D), indicating that STING-mediated cytotoxicity is largely indirect and depends strongly on the immune cell compartment. STING expression in tumor cells varies across patients IHC staining of primary tumor cultures revealed variable cytoplasmic STING expression among patients (Fig. 3 E). EF1 and EF5 exhibited strong staining, whereas EF2 and EF3 were weak or negative. A descriptive comparison of mean optical density (OD) and corresponding fold change in tumor cell death after STING agonist exposure suggested that higher baseline STING expression may coincide with increased responsiveness in some samples (Fig. 3 D). However, although the limited cohort precludes definitive correlation. Together, these data demonstrate that STING agonism elicits variable but reproducible tumor cell killing across patients, that a subset of PATECs displays synergistic effects with RT, and that STING-dependent cytotoxicity is largely dependent on the presence of immune cells. Because immune cell participation appeared essential for STING efficacy, we next examined whether this killing required direct immune tumor contact (Fig. 4 ). A Tumor cell death in PATEC (EF1-EF6, n = 6) after 72 h of treatment (cntrl, STING agonist, RT, RT + STING agonist). B Bliss synergy index (ΔE, percentage points) for RT + STING agonist for PATEC (EF1-EF6), derived from control normalized fractional tumor cell death. Bars show mean ΔE with 95% bootstrap confidence intervals; dashed and dotted lines indicate ΔE = 0 and the ± 5% neutrality band. C Comparison of tumor cell death after 72 h of STING agonist treatment in primary tumor monocultures (TuC/PTCC) versus matched autologous co-cultures (TuCIC/PATEC) (n = 2 PATECs). D Heatmap of mean cytoplasmic STING expression (STINGMeanOD) and corresponding mean fold change in tumor cell death at 72 h after STING agonist across PTCC (n = 4). E Representative IHC images of formalin‑fixed paraffin‑embedded primary tumor cell cultures stained for STING (brown chromogen) and haematoxylin (blue). Treatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). A, C, D Fold change of % Zombie Violet™⁺ tumor cells relative to cntrl. A Bars represent mean ± s.e.m. of biological replicates. Two-way ANOVA; Tukey’s multiple comparisons. C Boxes show median and IQR; whiskers 1.5×IQR; dots are biological replicates. Paired Wilcoxon signed rank test. B Bars show mean ΔE with 95% bootstrap confidence intervals (2,000 resamples). D, E descriptive only; no additional statistical testing beyond that specified in the text and Methods. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 Direct immune tumor contact is critical for the cytotoxic efficacy of STING agonist-based RT-IO We next examined whether physical proximity between tumor and immune cells is necessary for the cytotoxic effects of STING agonist-based RT-IO. PATECs were treated for 72 h with STING agonist, RT, or the combination, either in direct co-culture or separated by a transwell insert (Fig. 4 A). Across five independent PATECs, RT + STING agonist elicited a marked increase in tumor cell death under direct contact conditions, whereas spatial separation abolished this response (two-way ANOVA with Tukey’s multiple comparisons; p < 0.05-0.0001). By contrast, the cytotoxic effect of RT monotherapy was preserved in the insert condition, indicating that RT acts primarily through tumor intrinsic processes (Fig. 4 C). Pooled analysis across PATECs confirmed significantly higher killing for both STING agonist alone and RT + STING agonist in direct contact compared with insert conditions (linear mixed effects model with patient random intercept; p < 0.01, p < 0.000; Fig. 4 D). Consistently, Bliss defined RT-STING synergy became neutral in all PATECs when repeated in the insert condition (Fig. 4 E). These findings demonstrate that the cytotoxic efficacy of STING agonist-based RT-IO is highly dependent on direct immune tumor contact, consistent with engagement of contact-dependent effector pathways such as immune synapse formation, directed degranulation, or death receptor mediated interactions. Radiotherapeutic efficacy is largely independent of immune cells in PATEC Next, we evaluated whether RT-induced killing also requires immune cells, we compared RT responses in tumor monocultures (TuC) versus autologous co-cultures (PATEC). RT alone significantly increased tumor cell death compared to control (p < 0.001), but responses were indistinguishable between TuC and PATEC conditions (linear mixed effects model with patient random intercept; Fig. 4 C). This indicates that, unlike STING agonist, RT elicits direct tumor cell toxicity that is largely immune cell independent. Immune cell contact alone does not promote spontaneous killing To determine whether autologous immune cells exert any basal cytotoxicity in the absence of therapeutic stimulation, we compared untreated tumor cell monocultures (TuC) with direct co-cultures (PATEC) and insert separated co-cultures (PATEC (well)). Baseline tumor cell death was indistinguishable across all three configurations (linear mixed effects model; ns), indicating that effusion-derived immune cells do not spontaneously kill tumor cells ex vivo , even when in direct physical contact (Fig. 4 B). In summary, these data show that immune cell activity in PATEC requires therapeutic augmentation and that the tumoricidal effects of STING agonist-based RT-IO derive from treatment induced, not constitutive, immune engagement. This provided the rationale to examine whether early immune activation and cytokine responses accompany the induction of cytotoxicity (Fig. 5 ). Treatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). A - D Fold change of % Zombie Violet™+ cells relative to cntrl. A Bars represent mean ± s.e.m. of 2–4 biological replicates (dots). Two-way ANOVA; Tukey’s multiple comparisons B - D Boxes show median and IQR; whiskers 1.5×IQR. LMM; patient random intercept; Tukey adjusted pairwise contrasts. ****p < 0.0001, ***p ≤ 0.001, **p < 0.01, *p < 0.05; ns, not significant. T cell activation in malignant effusion cultures following STING agonism and RT To assess early immune activation following treatment, malignant effusion cultures were exposed to STING agonist, RT, or combination for 24 h (Fig. 5 A). FCM analysis revealed distinct activation and degranulation profiles across T cell subsets. Expression of the activation marker CD69 was most pronounced after combined RT + STING agonist, exceeding both monotherapies (two-way ANOVA with Tukey; p < 0.05-0.0001; Fig. 5 B). In contrast, the degranulation marker CD107a most markedly increased after STING agonist monotherapy, with combination treatment showing moderate but consistent induction across patients’ malignant effusions. Both CD8⁺ and CD4⁺ T cells exhibited similar trends. Collectively, these data indicate that STING agonism rapidly promotes T cell degranulation, whereas its combination with RT primarily amplifies early activation, reflecting distinct but complementary mechanisms of effector engagement within the effusion microenvironment. Pro-inflammatory cytokine and IFN-I responses accompany T cell activation To explore whether cytokine release increases upon treatment, multiplex profiling of culture supernatants was performed after 24 h (n = 5). The combined RT + STING agonist treatment induced a broad pro-inflammatory cytokine response, marked by increased levels of TNF-α, IL-6, IP-10 (CXCL10), IFN-α, MIP-α (CCL3), and MIP-1β (CCL4) (two-way ANOVA with Tukey; p < 0.05 − 0.001; Fig. 5 d) (Fig. 5 C). In contrast, IFN-β, IL-8, MCP-1, RANTES, IFN-γ, and IL-10 showed no consistent modulation at this early time point ( Supplementary Fig. 3 ). This cytokine profile is characteristic of an early IFN-I-driven and chemokine rich milieu, consistent with STING pathway activation and recruitment of effector leukocytes. Overall, these results show that combined STING agonism and RT trigger rapid immune activation within malignant effusion cultures, marked by CD8⁺ and CD4⁺ T cell degranulation and a coordinated IFN-I cytokine response. This early immunostimulatory phase prompted us to investigate whether sustained checkpoint upregulation and functional engagement of CD8⁺ T cells correlate with tumor cell killing at later time points (Fig. 6 ). Treatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). B Values are normalized to each effusion’s matched cntrl. C Values are z-score normalized concentrations. B, C Boxes show median and IQR; whiskers 1.5×IQR; dots are means of biological replicates. Two-way ANOVA; Tukey’s multiple comparisons. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant. STING based RT-IO enhances checkpoint co-expression on CD8⁺ T cells Having established that STING agonism promotes early T cell activation and cytokine release (Fig. 5 ), we next examined whether prolonged stimulation elicits adaptive checkpoint engagement on cytotoxic T cells. After 72 h of treatment, PATEC cultures exposed to STING agonist + RT displayed broad upregulation of inhibitory and activation receptors on CD8⁺ T cells (Fig. 6 A). Single positive subsets (PD-1⁺, TIGIT⁺, LAG-3⁺) increased relative to matched controls, and combinatorial phenotypes, particularly PD-1⁺TIGIT⁺, PD-1⁺LAG-3⁺TIGIT⁺ and TIGIT⁺LAG-3⁺TIM-3⁺, were most enriched following RT + STING agonist (linear mixed effects models with Dunnett contrasts; p < 0.05 − 0.001). The simultaneous induction of multiple checkpoints under this regimen indicates a sustained antigenic stimulation program typical of recently activated effector cells rather than fixed exhaustion. Composite checkpoint expression correlates with tumor cell killing To quantify checkpoint co-expression, we derived a Checkpoint Co-expression Score (CES) that integrates z normalized frequencies of triple positive subsets weighted by quadruple positive enrichment. Across PATECs and treatments, higher CES correlated with greater tumor cell death at 72 h (Δlogit vs matched control; Fig. 6 B). In a Gaussian mixed effects model with patient as random intercept, the slope was β = 0.15 (95% CI 0.11–0.20), corresponding to an odds ratio (OR) of 1.16 per CES unit (p = 1.8 × 10⁻⁸; R²ₘarg = 0.37, R²cond = 0.41). Experiments from RT + STING agonist tended to be in the upper CES-cytotoxicity range, but the fitted relationship was consistent across all conditions, indicating that the degree of checkpoint co-expression quantitatively mirrors magnitude of cytotoxic efficacy irrespective of treatment type. CES is preferentially associated with the contact-dependent component of tumor cell killing Having established that STING agonist induced cytotoxicity contains a substantial contact-dependent component (Fig. 4 ), we next examined whether the CES cytotoxicity relationship was primarily attributable to this contact mediated fraction. To isolate this effect, we defined Δ(contact) as the difference in treatment induced cytotoxicity between direct co-culture and insert separated conditions: Δ(contact) = Δlogit(Direct) − Δlogit(Insert). Across PATECs, Δ(contact) increased with CES, with a mixed effects slope of β = 0.09 (95% CI 0.03–0.15), corresponding to a 10% increase in the odds of contact-dependent killing per CES unit (p = 0.0088; R²ₘₐ r g = 0.31; Fig. 6 C). Importantly, this association persisted after subtraction of contact independent cytotoxicity, indicating that CD8⁺ T cells with higher checkpoint expression preferentially arise under conditions of direct immune tumor interaction, consistent with an activation program linked to sustained antigen engagement. The distribution of checkpoint positive phenotypes is consistent with established CD8⁺ T cell differentiation trajectories. PD-1⁺TIGIT⁺ subsets are characteristic of recently stimulated, antigen-experienced effector populations that retain proliferative and cytotoxic potential, whereas the additional expression of LAG-3 or TIM-3 marks progressively constrained states associated with more sustained antigen exposure. At this early 72 h time point, the positive association between CES and tumor cell killing indicates that multi-checkpoint expression reflects an actively engaged effector program operating under inducible inhibitory feedback, rather than terminal dysfunction. To explore whether induction of individual checkpoints associated with benefit of having the corresponding checkpoint inhibitor in the PATEC, we used the quadruple RT-IO screen to correlate upregulation of PD-1⁺, PD-L1/PD-1⁺, TIGIT⁺ and CTLA-4⁺ on T cells after STING agonist + RT with the additional tumor cell death observed when Pembrolizumab, Atezolizumab, Tiragolumab or Ipilimumab are in culture. In CD8⁺ T cells, higher induction of PD-1, PD-L1/PD-1 and TIGIT tended to coincide with greater incremental killing, whereas analogous associations for CD4⁺ T cells were weaker and more variable; none reached statistical significance in this small cohort and these analyses were therefore considered exploratory ( Supplementary Fig. 4 ). These hypothesis-generating data indicate that multi-checkpoint expression on CD8⁺ T cells are associated with the magnitude of tumor cell killing, and that this association persists after isolating the contact-dependent component of cytotoxicity. Accordingly, higher CES values capture CD8⁺ states that coincide with conditions in which contact-mediated cytotoxicity constitutes a larger share of the overall response. A CD8⁺ checkpoint expression in PATEC after 72 h of treatment (cntrl, STING agonist, RT, RT + STING agonist) (n = 5 PATECs). Shows fold change of % CD8⁺ cells positive for the indicated checkpoints relative to matched cntrl. Boxes show median and IQR; whiskers 1.5×IQR; dots are individual experiments. LMM on logit-transformed data; patient random intercept; Dunnett contrasts versus cntrl. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant B Scatterplot of association between CD8⁺ Checkpoint Co-expression Score (CES) and overall tumor cell death at 72 h (n = 6 PATECs). CES plotted against Δlogit tumor cell death (change in logit-transformed % dead tumor cells vs cntrl). C Scatterplot of association between CES and the contact-dependent component of tumor cell killing at 72 h (n = 5 PATECs). Cell contact-dependent component is defined as Δ(contact) = Δlogit(Direct) − Δlogit(Insert). CES is plotted against Δ(contact). Treatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). B, C Each dot represents one experiment; colors indicate treatment. The black line shows the LMM regression fit and the grey band the 95% confidence interval. Discussion In this study, we developed PATEC as a human ex vivo platform to evaluate RT-IO regimens in an immunocompetent TIME. In a combinatorial screen, we found regimens combining RT with STING agonism were the most tumoricidal in this cohort. Furthermore, PATEC showed STING dependent cytotoxicity largely mediated by immune cells and strongly depends on direct immune tumor contact. In contrast, RT alone acts largely through tumor intrinsic mechanisms. STING based RT-IO is accompanied by effector T cell activation, an inflammatory/IFN-I associated cytokine milieu and upregulation of inhibitory checkpoints on CD8⁺ T cells. Together, these findings supporting the use of this platform to identify and characterize feasible RT-IO combinations. Within the current landscape of human ex vivo models, PATEC addresses some of the distinct limitations of the existing ones. While organoid immune co-cultures have been used to test ICI, their origin are solid tumors and their inherent immune infiltrate is sparse. They heavily rely on stem cell supporting growth factors, as well as typically requiring extraction, expansion and reintroduction of immune cells from PBMCs or TILs, rather than utilizing the native autologous TIME ( 28 , 29 ). Similarly, Tumor-on-a-chip platform mostly implement immune cell substitutes, require processing of scarce surgical or biopsy material and coincide with high costs due to microfabrication with a specialized infrastructure ( 30 – 34 ). Precision-cut tissue slices and PDTF platforms on the other hand preserve tissue architecture and stromal context. They are useful for short-term assessment of checkpoint and cytokine responses, but are still limited by restricted source material and a solid tissue interface that does not readily permit systematic separation or repeated perturbation of tumor and immune compartments ( 35 , 36 ). Malignant effusion-derived cultures published to date have focused on expanding epithelial cells for targeted or chemotherapeutic drug screening, isolating and stimulating T cells for cytotoxicity assays or testing CAR-T cell cytotoxicity using solely the pleural fluid ( 19 , 20 , 37 – 40 ). The configuration of the PATEC however uses the naturally enriched autologous tumor-associated immune cells in the fluid TIME retaining a plethora of myeloid and lymphoid repertoires of the malignant effusions. Cryopreserved aliquots enable experimental flexibility; spatial compartmentalization allow for dissection of the responsible cytotoxic effector and serial sampling offers the potential of longitudinal functional profiling during the course of therapy. The cellular baseline compositions of the cohorts’ effusions are consistent with previous descriptive characterizations. Across patients there was a pronounced interindividual heterogeneity, with immune compartments variably enriched for macrophages, neutrophils, CD4⁺ and CD8⁺ T cells, NK cells, and B cells, in line with findings from single cell sequencing and multiparametric studies showing diverse immune subsets and patient-specific effusion immune profiles ( 21 , 41 , 42 ). Although we did not specifically characterize exhaustion, regulatory or myeloid polarization states at baseline, the effusion-derived immune cells in PATEC did not exhibit spontaneous inherent cytotoxicity. This observation is concordant with mechanistic studies indicating the effusion TIME constitutes a maladaptive but potentially reversible effector compartment, in which CD8⁺ T cells and NK cells are chronically constrained by local factors yet can regain cytotoxic function upon appropriate stimulation ( 25 , 43 – 45 ). Within the PATEC system, combinatorial screening showed RT-IO regimens incorporating innate agonists were consistently among the most effective at inducing tumor cell death, with STING agonist containing combinations showing a modest but incrementally more consistent advantage over those including TLR7/8 agonists. These ex vivo findings are consistent with preclinical models demonstrating that ionizing radiation and STING agonists can be highly efficacious in combination ( 8 , 46 – 49 ). In contrast to these in vivo studies, the PATEC assay assesses only how RT and innate agonists modulate local effector activity and does not capture dendritic cell trafficking, priming in lymph node or abscopal responses, so the observed efficacy reflects direct tumor killing within the TIME rather than long-term systemic control ( 8 , 46 , 48 – 50 ). In this effector phase, Bliss independence analyses indicated that only a subset of PATECs exhibit a STING-RT synergistic effect, suggesting that patient-specific features of the TIME determine how effectively the combined actions of STING stimulation and irradiation are translated into cytotoxic effector activity. Such heterogeneity is consistent with preclinical data demonstrating that tumor intrinsic cGAS-STING pathway integrity, the mode of STING activation and its expression level critically modulate responsiveness to RT and to STING agonists ( 51 – 53 ). In this context, the modest and heterogeneous responses observed in early phase trials of intratumoral STING agonists, applied primarily as monotherapy or with PD-1 blockade, are plausibly attributable to variations in STING pathway and to the absence of rationally selected combinations and patient selection strategies ( 4 , 54 – 56 ). These data indicate that effector phase platforms such as PATEC could help functionally identify effective innate immunotherapy regimens and nominate patients for further in vivo evaluation. Comparing PATEC to matched tumor monocultures and transwell separated co-cultures, enabled us to distinguish tumor intrinsic from immune-mediated effects of STING based RT-IO. RT alone produced similar levels of cytotoxicity in primary tumor monocultures and in PATEC, retaining its efficacy when tumor and immune compartments were separated, suggesting that RT predominantly elicits DNA-damage-driven tumor-intrinsic cytotoxicity in this system ( 57 ). In contrast, the addition of the STING agonist increased tumor cell death only in the presence of immune cells. This benefit was almost completely lost under transwell separation, implicating that STING dependent cytotoxicity in PATEC is largely indirect and strongly dependent on close immune tumor apposition. This is consistent with canonical contact-dependent effector mechanisms, including formation of a cytotoxic immunological synapse, polarized/directed degranulation of lytic granules, and death receptor engagement or release of short-range soluble cytotoxic mediators by cytotoxic lymphocytes and activated myeloid cells ( 58 – 61 ). Untreated PATEC co‑cultures did not exhibit spontaneous tumor cell killing, even when tumor and immune cells were in close contact. Similarly human data shows that effusion‑resident NK and CD8⁺ T cells are chronically constrained by the effusion milieu yet can reacquire proliferative and cytotoxic function when removed from suppressive fluid or exposed to appropriate cytokine or antigenic stimulation ( 25 , 43 , 44 , 62 , 63 ). These converging lines of evidence support a model in which malignant effusions are thought to constitute an immunosuppressive effector niche that remains functionally reversible. In this context, these observations point to an interplay in which RT induces tumor-intrinsic DNA-damage and immunogenic antigen release, whereas STING agonism primarily boosts effusion resident cytotoxic lymphocyte and macrophage function. In PATEC, STING agonism and RT modulated T cell activation, checkpoint expression and secreted cytokine profile. After 24h, the STING agonist alone was sufficient to drive robust T cell degranulation. The combination of RT and STING agonist most strongly increased CD69 expression and induced a pro-inflammatory, IFN-I cytokine and chemokine profile marked by higher TNF-α, IL-6, CXCL10/IP-10, IFN-α, CCL3 and CCL4. This pattern is concordant with preclinical STING-RT studies in which IFN-I, TNF-α and CXCL9/10-rich milieus are induced that support effector activation, as well as with human co-culture studies that pharmacological STING activation can directly augment degranulation and cytotoxic function of T cells and NK cells ( 49 , 64 – 68 ). Consistent with sustained antigenic stimulation, STING agonist and RT induced the most pronounced significant upregulation of PD-1, TIGIT, LAG-3 and TIM-3 on CD8⁺ T cells, which was associated with cytotoxicity. Although high expression of PD-1, LAG-3, TIGIT and TIM-3 is often interpreted as an exhaustion signature, experimental and clinical data indicate that these receptors are induced sequentially along differentiation and chronic stimulation trajectories. They frequently mark tumor-experienced T cell populations whose functional state ranges from active to terminally exhausted depending on context and timing ( 69 , 70 ). In line with data showing that CD8⁺ T cells expressing multiple inhibitory receptors can remain polyfunctional mediating anti-tumor activity in patients, as well as humanized models and that gene signatures enriched for PD-1, LAG-3, TIM-3 and TIGIT expression are associated with clinical benefit from ICI, our data support a similar interpretation in PATEC ( 71 – 74 ). The positive association between CES and cytotoxicity in PATEC supports interpreting CES high CD8⁺ T cell states in this early window as engaged effector populations under inducible inhibitory feedback rather than terminally dysfunctional cells. The addition of PD-1, PD-L1, CTLA-4 or TIGIT blockade to STING or TLR7/8 based RT-IO in the quadruple therapy experiment did not produce a significant, consistent further increase in tumor cell death across PATECs. Exploratory analyses suggested that effusions with stronger induction of the corresponding checkpoint on CD8⁺ T cells under STING + RT tended to derive modest incremental killing from blockade of that pathway, particularly for TIGIT and PD-L1. These findings are hypothesis-generating but accord with reports that CD8⁺ T cells co-expressing multiple inhibitory receptors can remain functionally competent and enriched for checkpoint-responsive tumor reactivity. They are also consistent ex vivo “tumor avatar” studies in which PD-1/CTLA-4/TIGIT blockade predominantly modulates progenitor exhausted compartments, clonal composition and cytokine programs over several days rather than yielding significant gains in short-term cytotoxicity ( 71 – 78 ). Even though the PATEC model shows the potential for a screening platform application there are several limitations to consider. The successful generation of primary tumor cultures for PATECs was modest at 20% and was derived from pan-cancer, as well as diverse pretreatments. The inherent fact, that there is a finite amount of biological material obtainable from malignant effusions, restricts the complexity and number of follow up experiments conductible. The model’s fluid interface is a double-edged sword; while providing preferable culture conditions it lacks stromal architecture, vasculature, antigen priming environment and pharmacokinetics. Processes dependent on these, such as recruitment of leukocytes, clonal replacement, stromal immunosuppression or dose-limiting toxicities are not recapitulated. The functional behavior of the culture might also not reflect the in vivo response of the cells. Further work is required to establish the representability of culture regarding clinical response. Conclusion PATEC potentially provides a patient-derived ex vivo platform to functionally test new immunotherapeutic and RT-IO combinations in an autologous, immunocompetent TIME. In this proof-of-concept study combinations of RT with a STING agonist were the most tumoricidal regimens, with Bliss-defined synergy observed in a subset. STING-dependent cytotoxicity was largely immune-mediated and strongly dependent on direct immune tumor contact. STING-based RT-IO showed rapid T cell activation, an IFN-I-driven proinflammatory cytokine milieu and upregulated checkpoints on CD8⁺ T cells, which associated with tumor cell killing. These findings position malignant effusion-derived PATECs as an effector phase platform to rationally develop RT-IO regimens, dissect patient-specific synergy and generate functional hypotheses for immunotherapy-related biomarkers. Prospective integration of PATEC into precision oncology and early phase trials could help link patient-specific ex vivo responses to clinical outcome and support more rational selection of immunotherapeutic combinations for patients with advanced solid tumors. Abbreviations RT-IO radiotherapy-immunotherapy RT radiotherapy IO immunotherapy ICI immune checkpoint inhibition/inhibitor PATEC patient‑derived autologous tumor-immune effusion co‑culture PTCC primary tumor cell culture ME malignant effusion TIME tumor-immune microenvironment TME tumor microenvironment TuC tumor cell monoculture TuCIC tumor cell immune cell co‑culture (PATEC) STING stimulator of interferon genes STING‑Ago STING agonist TLR7/8 Toll‑like receptor 7/8 CTLA‑4 cytotoxic T‑lymphocyte associated protein 4 PD‑1 programmed cell death protein 1 PD‑L1 programmed death‑ligand 1 LAG‑3 lymphocyte activation gene 3 TIGIT T cell immunoreceptor with Ig and ITIM domains IFN‑I type I interferons IFN‑α interferon alpha IFN‑β interferon beta IFN‑γ interferon gamma TNF‑α tumor necrosis factor alpha IL interleukin IL‑1ra interleukin‑1 receptor antagonist IL‑6 interleukin 6 IL‑8 interleukin 8 IL‑10 interleukin 10 IP‑10 interferon‑gamma-induced protein 10 (CXCL10) MCP‑1 monocyte chemoattractant protein 1 (CCL2) MIP‑1α macrophage inflammatory protein 1 alpha (CCL3) MIP‑1β macrophage inflammatory protein 1 beta (CCL4) RANTES regulated on activation,normal T cell expressed and secreted (CCL5) PBMC peripheral blood mononuclear cell TIL tumor infiltrating lymphocyte NK cell natural killer cell NKT cell natural killer T cell IHC immunohistochemistry FCM flow cytometry FBS fetal bovine serum PBS phosphate buffered saline FC fold change CES checkpoint co‑expression score LMM linear mixed effects model ANOVA analysis of variance IQR interquartile range CI confidence interval OR odds ratio FDR false discovery rate DAB 3,3'-diaminobenzidine OD optical density Gy gray DMEM/F12 Dulbecco’s Modified Eagle Medium/Ham’s F‑12 nutrient mixture Declarations Ethics approval and consent to participate This study involved human participants and received approval from the Institutional Review Board of the Medical University of Vienna (No. 2042/2019). The study was conducted in strict accordance with Good Scientific Practice (GSP) guidelines of the Medical University of Vienna and the most recent Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion in the study. Consent for publication Not applicable. Availability of data and materials The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests. Funding The work was supported by internal funds of the Medical University of Vienna. There were no funding organizations that were involved in the study design, data acquisition, analysis, interpretation, manuscript preparation, or the decision to submit this work for publication. Authors’ contributions R.Z. conceived and designed the study, acquired the data, performed the analysis, interpreted the results, and drafted the manuscript. D.A., N.S., M.F., B.M., M.A.R.H. and A.T. acquired and analyzed data. M.B. and J.L. contributed to study conception and data interpretation. All authors critically reviewed the manuscript, as well as read and approved the final manuscript. Acknowledgements We thank Brigitte Wolf and Carolina Klicka for their day-to-day support in the laboratory and for coordinating the transfer of malignant effusion specimens from the hospital to the laboratory. We also thank Büsra Ehrnhofer for technical assistance with selected experiments. In addition, we are grateful to the members of our research group for constructive discussions and continued support throughout the project. References Haslam A, Prasad V. Estimation of the percentage of US patients with cancer who are eligible for and respond to checkpoint inhibitor immunotherapy drugs. JAMA Netw Open. 2019;2(5):e192535. Larkin J, Chiarion-Sileni V, Gonzalez R, Grob JJ, Cowey CL, Lao CD, et al. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N Engl J Med. 2015;373(1):23–34. Tawbi HA, Schadendorf D, Lipson EJ, Ascierto PA, Matamala L, Castillo Gutiérrez E, et al. Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma. N Engl J Med. 2022;386(1):24–34. Meric-Bernstam F, Sweis RF, Kasper S, Hamid O, Bhatia S, Dummer R, et al. Combination of the STING agonist MIW815 (ADU-S100) and PD-1 inhibitor spartalizumab in advanced/metastatic solid tumors or lymphomas: an open-label, multicenter, phase Ib study. Clin Cancer Res. 2023;29(1):110–21. Siu L, Brody J, Gupta S, Marabelle A, Jimeno A, Munster P, et al. Safety and clinical activity of intratumoral MEDI9197 alone and in combination with durvalumab and/or palliative radiation therapy in patients with advanced solid tumors. J Immunother Cancer. 2020;8(2):e001095. Cho BC, Abreu DR, Hussein M, Cobo M, Patel AJ, Secen N, et al. Tiragolumab plus atezolizumab versus placebo plus atezolizumab as a first-line treatment for PD-L1-selected non-small-cell lung cancer (CITYSCAPE): primary and follow-up analyses of a randomised, double-blind, phase 2 study. Lancet Oncol. 2022;23(6):781–92. Zhang Z, Liu X, Chen D, Yu J. Radiotherapy combined with immunotherapy: the dawn of cancer treatment. Signal Transduct Target Ther. 2022;7(1):258. Baird JR, Friedman D, Cottam B, Dubensky TW Jr, Kanne DB, Bambina S, et al. Radiotherapy combined with novel STING-targeting oligonucleotides results in regression of established tumors. Cancer Res. 2016;76(1):50–61. Stary V, Wolf B, Unterleuthner D, List J, Talic M, Laengle J, et al. Short-course radiotherapy promotes pro-inflammatory macrophages via extracellular vesicles in human rectal cancer. J Immunother Cancer. 2020;8(2):e000667. Spigel DR, Faivre-Finn C, Gray JE, Vicente D, Planchard D, Paz-Ares L, et al. Five-year survival outcomes from the PACIFIC trial: durvalumab after chemoradiotherapy in stage III non-small-cell lung cancer. J Clin Oncol. 2022;40(12):1301–11. Kelly RJ, Ajani JA, Kuzdzal J, Zander T, van Cutsem E, Piessen G, et al. Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer. N Engl J Med. 2021;384(13):1191–203. Petricevic B, Kabiljo J, Zirnbauer R, Walczak H, Laengle J, Bergmann M. Neoadjuvant immunotherapy in gastrointestinal cancers - the new standard of care? Semin Cancer Biol. 2022;86(Pt 2):834–50. Zirnbauer R, Hametner S, Bergler-Klein J, Kuehrer I, Kulu A, Ammon D et al. Incidence and immunopathology of myositis in rectal cancer patients treated with neoadjuvant immune checkpoint inhibitors and chemoradiotherapy: findings from the CHINOREC trial. MedComm (2020). 2025;6(7):e70275. Laengle J, Kuehrer I, Kulu A, Kabiljo J, Ammon D, Zirnbauer R, et al. Dual immune checkpoint inhibition plus neoadjuvant chemoradiotherapy in rectal cancer: a randomized clinical trial. JAMA Netw Open. 2025;8(8):e2527769. Tan AC, Bagley SJ, Wen PY, Lim M, Platten M, Colman H, et al. Systematic review of combinations of targeted or immunotherapy in advanced solid tumors. J Immunother Cancer. 2021;9(7):e002459. Yang J, Kang H, Lyu L, Xiong W, Hu Y. A target map of clinical combination therapies in oncology: an analysis of clinicaltrials.gov. Discov Oncol. 2023;14(1):151. Roscilli G, De Vitis C, Ferrara FF, Noto A, Cherubini E, Ricci A, et al. Human lung adenocarcinoma cell cultures derived from malignant pleural effusions as model system to predict patients chemosensitivity. J Transl Med. 2016;14:61. Kim SY, Lee JY, Kim DH, Joo HS, Yun MR, Jung D, et al. Patient-derived cells to guide targeted therapy for advanced lung adenocarcinoma. Sci Rep. 2019;9(1):19909. Wu CG, Chiovaro F, Curioni-Fontecedro A, Casanova R, Soltermann A. In vitro cell culture of patient derived malignant pleural and peritoneal effusions for personalised drug screening. J Transl Med. 2020;18:163. Wegmann R, Bankel L, Festl Y, Lau K, Lee S, Arnold F, et al. Molecular and functional landscape of malignant serous effusions for precision oncology. Nat Commun. 2024;15:8544. Huang ZY, Shao MM, Zhang JC, Yi FS, Du J, Zhou Q, et al. Single-cell analysis of diverse immune phenotypes in malignant pleural effusion. Nat Commun. 2021;12:6690. Chee J, Watson MW, Chopra A, Nguyen B, Cook AM, Creaney J, et al. Tumour associated lymphocytes in the pleural effusions of patients with mesothelioma express high levels of inhibitory receptors. BMC Res Notes. 2018;11:864. Wu MF, Lin CA, Yuan TH, Yeh HY, Su SF, Guo CL, et al. The M1/M2 spectrum and plasticity of malignant pleural effusion-macrophage in advanced lung cancer. Cancer Immunol Immunother. 2021;70(5):1435–50. Principe N, Kidman J, Lake RA, Lesterhuis WJ, Nowak AK, McDonnell AM, et al. Malignant pleural effusions-a window into local anti-tumor T cell immunity? Front Oncol. 2021;11:672747. Dhupar R, Okusanya OT, Eisenberg SH, Monaco SE, Ruffin AT, Liu D, et al. Characteristics of malignant pleural effusion resident CD8 + T cells from a heterogeneous collection of tumors. Int J Mol Sci. 2020;21(17):6178. Lv LL, Wang HB, Zhang YX, Zhai JW, Shen Y, Qu QX, et al. CD39 identifies a specific CD8 + T cell population in lung adenocarcinoma-related metastatic pleural effusion. BMC Immunol. 2023;24:53. Ge S, Zhao Y, Liang J, He Z, Li K, Zhang G, et al. Immune modulation in malignant pleural effusion: from microenvironment to therapeutic implications. Cancer Cell Int. 2024;24:105. Yuki K, Cheng N, Nakano M, Kuo CJ. Organoid models of tumor immunology. Trends Immunol. 2020;41(8):652–64. Wang J, Tao X, Zhu J, Dai Z, Du Y, Xie Y, et al. Tumor organoid-immune co-culture models: exploring a new perspective of tumor immunity. Cell Death Discov. 2025;11:195. Ao Z, Cai H, Wu Z, Hu L, Li X, Kaurich C, et al. Evaluation of cancer immunotherapy using mini-tumor chips. Theranostics. 2022;12(8):3628–36. Ayuso JM, Rehman S, Virumbrales-Munoz M, McMinn PH, Geiger P, Fitzgerald C, et al. Microfluidic tumor-on-a-chip model to evaluate the role of tumor environmental stress on NK cell exhaustion. Sci Adv. 2021;7(8):eabc2331. Chernyavska M, Masoudnia M, Valerius T, Verdurmen WPR. Organ-on-a-chip models for development of cancer immunotherapies. Cancer Immunol Immunother. 2023;72(12):3971–83. Liu H, Noguera-Ortega E, Dong X, Lee WD, Chang J, Aydin SA, et al. A tumor-on-a-chip for in vitro study of CAR-T cell immunotherapy in solid tumors. Nat Biotechnol. 2025. 10.1038/s41587-025-02845-z . Maulana TI, Teufel C, Cipriano M, Roosz J, Lazarevski L, van den Hil FE, et al. Breast cancer-on-chip for patient-specific efficacy and safety testing of CAR-T cells. Cell Stem Cell. 2024;31(7):989–e10029. Jagatia R, Doornebal EJ, Rastovic U, Harris N, Feyide M, Lyons AM, et al. Patient-derived precision cut tissue slices from primary liver cancer as a potential platform for preclinical drug testing. EBioMedicine. 2023;97:104826. Sullivan KM, Jiang X, Guha P, Lausted C, Carter JA, Hsu C, et al. Blockade of interleukin 10 potentiates antitumour immune function in human colorectal cancer liver metastases. Gut. 2023;72(2):325–37. Seo HY, Kim SC, Roh WL, Shin YK, Kim S, Kim DW, et al. Culture and multiomic analysis of lung cancer patient-derived pleural effusions revealed distinct druggable molecular types. Sci Rep. 2022;12:6345. Choi W, Kim YH, Woo SM, Yu Y, Lee MR, Lee WJ, et al. Establishment of patient-derived organoids using ascitic or pleural fluid from cancer patients. Cancer Res Treat. 2023;55(4):1077–86. Tano ZE, Kiesgen S, Chintala NK, Dozier J, Quach HT, Messinger J, et al. Ex vivo pleural effusion cultures to study chimeric antigen receptor T cell cytotoxicity in an immunocompetent environment. Cell Rep Methods. 2023;3(11):100622. Önder CE, Moustafa-Oglou M, Schröder SM, Hartkopf AD, Koch A, Seitz CM. Precision immunotherapy utilizing adapter CAR-T cells (AdCAR-T) in metastatic breast cancer leads to target specific lysis. Cancers (Basel). 2024;16(1):168. Huang XZ, Pang MJ, Li JY, Chen HY, Sun JX, Song YX, et al. Single-cell sequencing of ascites fluid illustrates heterogeneity and therapy-induced evolution during gastric cancer peritoneal metastasis. Nat Commun. 2023;14:822. Zheng X, Wang X, Cheng X, Liu Z, Yin Y, Li X, et al. Single-cell analyses implicate ascites in remodeling the ecosystems of primary and metastatic tumors in ovarian cancer. Nat Cancer. 2023;4(8):1138–56. Donnenberg VS, Luketich JD, Sultan I, Lister J, Bartlett DL, Ghosh S, et al. A maladaptive pleural environment suppresses preexisting anti-tumor activity of pleural infiltrating T cells. Front Immunol. 2023;14:1157697. Slattery K, Yao CH, Mylod E, Scanlan J, Scott B, Crowley JP, et al. Uptake of lipids from ascites drives NK cell metabolic dysfunction in ovarian cancer. Sci Immunol. 2025;10(107):eadr4795. Vacca P, Martini S, Mingari MC, Moretta L. NK cells from malignant pleural effusions are potent antitumor effectors: a clue for adoptive immunotherapy. Oncoimmunology. 2013;2(4):e23638. Deng L, Liang H, Xu M, Yang X, Burnette B, Arina A, et al. STING-dependent cytosolic DNA sensing promotes radiation-induced type I interferon-dependent antitumor immunity in immunogenic tumors. Immunity. 2014;41(5):843–52. Liu Y, Crowe WN, Wang L, Lu Y, Petty WJ, Habib AA, et al. An inhalable nanoparticulate STING agonist synergizes with radiotherapy to confer long-term control of lung metastases. Nat Commun. 2019;10:5108. Zhang P, Rashidi A, Zhao J, Silvers C, Wang H, Castro B, et al. STING agonist-loaded, CD47/PD-L1-targeting nanoparticles potentiate antitumor immunity and radiotherapy for glioblastoma. Nat Commun. 2023;14:1610. Zeng Q, Liu M, Wang Z, Zhou R, Ai K. Enhancing radiotherapy-induced anti-tumor immunity via nanoparticle-mediated STING agonist synergy. Mol Cancer. 2025;24:176. Luo M, Liu Z, Zhang X, Han C, Samandi LZ, Dong C, et al. Synergistic STING activation by PC7A nanovaccine and ionizing radiation improves cancer immunotherapy. J Control Release. 2019;300:154–60. Jagodinsky JC, Jin WJ, Bates AM, Hernandez R, Grudzinski JJ, Marsh IR, et al. Temporal analysis of type 1 interferon activation in tumor cells following external beam radiotherapy or targeted radionuclide therapy. Theranostics. 2021;11(13):6120–37. Drzyzga A, Czapla J, Matuszczak S, Łasut-Szyszka B, Cichoń T, Pilny E, et al. Differential response to local stimulator of interferon genes agonist administration in tumors with various stimulator of interferon genes statuses. Cancers (Basel). 2025;17(2):175. Hayman TJ, Baro M, MacNeil T, Phoomak C, Aung TN, Cui W, et al. STING enhances cell death through regulation of reactive oxygen species and DNA damage. Nat Commun. 2021;12:2327. Hines JB, Kacew AJ, Sweis RF. The development of STING agonists and emerging results as a cancer immunotherapy. Curr Oncol Rep. 2023;25(3):189–99. Luke JJ, Piha-Paul SA, Medina T, Verschraegen CF, Varterasian M, Brennan AM, et al. Phase I study of SYNB1891, an engineered Escherichia coli Nissle strain expressing STING agonist, with and without atezolizumab in advanced malignancies. Clin Cancer Res. 2023;29(13):2435–44. Meric-Bernstam F, Sweis RF, Hodi FS, Messersmith WA, Andtbacka RHI, Ingham M, et al. Phase I dose-escalation trial of MIW815 (ADU-S100), an intratumoral STING agonist, in patients with advanced/metastatic solid tumors or lymphomas. Clin Cancer Res. 2022;28(4):677–88. Sia J, Szmyd R, Hau E, Gee HE. Molecular mechanisms of radiation-induced cancer cell death: a primer. Front Cell Dev Biol. 2020;8:41. Deng A, Fan R, Hai Y, Zhuang J, Zhang B, Lu X, et al. A STING agonist prodrug reprograms tumor-associated macrophage to boost colorectal cancer immunotherapy. Theranostics. 2025;15(1):277–99. Nicolai CJ, Wolf N, Chang IC, Kirn G, Marcus A, Ndubaku CO, et al. NK cells mediate clearance of CD8 + T cell-resistant tumors in response to STING agonist. Sci Immunol. 2020;5(45):eaaz2738. Berger G, Knelson EH, Jimenez-Macias JL, Nowicki MO, Han S, Panagioti E, et al. STING activation promotes robust immune response and NK cell-mediated tumor regression in glioblastoma models. Proc Natl Acad Sci U S A. 2022;119(28):e2111003119. Ou L, Zhang A, Cheng Y, Chen Y. The cGAS-STING pathway: a promising immunotherapy target. Front Immunol. 2021;12:795048. Vacca P, Martini S, Garelli V, Passalacqua G, Moretta L, Mingari MC. NK cells from malignant pleural effusions are not anergic but produce cytokines and display strong antitumor activity on short-term IL-2 activation. Eur J Immunol. 2013;43(2):550–61. Hu CY, Zhang YH, Wang T, Chen L, Gong ZH, Wan YS, et al. Interleukin-2 reverses CD8(+) T cell exhaustion in clinical malignant pleural effusion of lung cancer. Clin Exp Immunol. 2016;186(1):106–14. Wang Q, Yu Y, Zhuang J, Liu R, Sun C. Demystifying the cGAS-STING pathway: precision regulation in the tumor immune microenvironment. Mol Cancer. 2025;24:178. Linder A, Nixdorf D, Kuhl N, Piseddu I, Xu TT, Holtermann AV, et al. STING activation improves T-cell-engaging immunotherapy for acute myeloid leukemia. Blood. 2025;145(19):2149–60. Knelson EH, Ivanova EV, Tarannum M, Campisi M, Lizotte PH, Booker MA, et al. Activation of tumor-cell STING primes NK-cell therapy. Cancer Immunol Res. 2022;10(8):947–61. Da Y, Liu Y, Hu Y, Liu W, Ma J, Lu N, et al. STING agonist cGAMP enhances anti-tumor activity of CAR-NK cells against pancreatic cancer. Oncoimmunology. 2022;11(1):2054105. Liu Y, Wang L, Song Q, Ali M, Crowe WN, Kucera GL, et al. Intrapleural nano-immunotherapy promotes innate and adaptive immune responses to enhance anti-PD-L1 therapy for malignant pleural effusion. Nat Nanotechnol. 2022;17(2):206–16. Cai L, Li Y, Tan J, Xu L, Li Y. Targeting LAG-3, TIM-3, and TIGIT for cancer immunotherapy. J Hematol Oncol. 2023;16:101. Joller N, Anderson AC, Kuchroo VK. LAG-3, TIM-3, and TIGIT: distinct functions in immune regulation. Immunity. 2024;57(2):206–22. Egelston CA, Avalos C, Tu TY, Simons DL, Jimenez G, Jung JY, et al. Human breast tumor-infiltrating CD8 + T cells retain polyfunctionality despite PD-1 expression. Nat Commun. 2018;9:4297. Mietz J, Kaulfuss M, Egli L, Opitz L, Münz C, Chijioke O. Human effector CD8 + T cells with an activated and exhausted-like phenotype control tumour growth in vivo in a humanized tumour model. EBioMedicine. 2024;106:105240. Ibáñez-Molero S, Veldman J, Simon Nieto J, Traets JJH, George A, Hoefakker K, et al. Tumour-reactive heterotypic CD8 T cell clusters from clinical samples. Nature. 2025. 10.1038/s41586-025-09754-w . Guégan JP, Peyraud F, Dadone-Montaudie B, Teyssonneau D, Palmieri LJ, Clot E, et al. Analysis of PD1, LAG3, TIGIT, and TIM3 expression in human lung adenocarcinoma reveals a 25-gene signature predicting immunotherapy response. Cell Rep Med. 2024;5(12):101831. Jenkins RW, Aref AR, Lizotte PH, Ivanova E, Stinson S, Zhou CW, et al. Ex vivo profiling of PD-1 blockade using organotypic tumor spheroids. Cancer Discov. 2018;8(2):196–215. Hulen TM, Friese C, Kristensen NP, Granhøj JS, Borch TH, Peeters MJW, et al. Ex vivo modulation of intact tumor fragments with anti-PD-1 and anti-CTLA-4 influences the expansion and specificity of tumor-infiltrating lymphocytes. Front Immunol. 2023;14:1180997. Eguren-Santamaría I, Rodríguez I, Herrero-Martin C, Fernández de Piérola E, Azpilikueta A, Sánchez-Gregorio S, et al. Short-term cultured tumor fragments to study immunotherapy combinations based on CD137 (4-1BB) agonism. Oncoimmunology. 2024;13(1):2373519. Thommen DS. Tumour avatars to model patients' responses to immunotherapy. Nat Rev Cancer. 2022;22(12):660. Additional Declarations No competing interests reported. Supplementary Files AdditionalFiles.docx Cite Share Download PDF Status: Published Journal Publication published 14 Apr, 2026 Read the published version in Journal of Experimental & Clinical Cancer Research → Version 1 posted Editorial decision: Revision requested 13 Feb, 2026 Reviews received at journal 13 Feb, 2026 Reviews received at journal 11 Feb, 2026 Reviewers agreed at journal 26 Jan, 2026 Reviewers agreed at journal 09 Jan, 2026 Reviewers invited by journal 07 Jan, 2026 Editor assigned by journal 06 Jan, 2026 Submission checks completed at journal 06 Jan, 2026 First submitted to journal 04 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8516163","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":571000612,"identity":"1a4eb872-0ffc-4eb3-8d22-37da7bbd1dc3","order_by":0,"name":"Rebecca Zirnbauer","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Rebecca","middleName":"","lastName":"Zirnbauer","suffix":""},{"id":571000620,"identity":"d3fb03cc-e2b7-4729-8565-8fee5d00b682","order_by":1,"name":"Daphni Ammon","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Daphni","middleName":"","lastName":"Ammon","suffix":""},{"id":571000622,"identity":"d87317d3-c6b1-48bb-b302-f029881916b0","order_by":2,"name":"Berta Mosleh","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Berta","middleName":"","lastName":"Mosleh","suffix":""},{"id":571000629,"identity":"3c222605-12bb-4175-a1bb-81b42700dcb2","order_by":3,"name":"Nora Speiser","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Nora","middleName":"","lastName":"Speiser","suffix":""},{"id":571000636,"identity":"eb2f0172-8378-402a-93f8-37c81f3d6482","order_by":4,"name":"Anna Theophil","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Theophil","suffix":""},{"id":571000637,"identity":"aa7a8ec0-c384-4f17-8707-a38577e7fbcf","order_by":5,"name":"Markus Fabits","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Markus","middleName":"","lastName":"Fabits","suffix":""},{"id":571000639,"identity":"1261e17a-b1fa-4cbc-9e64-b7e0a84060eb","order_by":6,"name":"Mir Ali Reza Hoda","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Mir","middleName":"Ali Reza","lastName":"Hoda","suffix":""},{"id":571000641,"identity":"568db998-c1ef-416c-89c3-8ea3cfdca4ba","order_by":7,"name":"Michael Bergmann","email":"data:image/png;base64,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","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":true,"prefix":"","firstName":"Michael","middleName":"","lastName":"Bergmann","suffix":""},{"id":571000651,"identity":"35758c04-f2d7-469e-a3a8-dc7b5da4381f","order_by":8,"name":"Johannes Laengle","email":"","orcid":"","institution":"Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Johannes","middleName":"","lastName":"Laengle","suffix":""}],"badges":[],"createdAt":"2026-01-05 02:23:48","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8516163/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8516163/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s13046-026-03707-5","type":"published","date":"2026-04-14T15:57:45+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":100011449,"identity":"ab18349f-005a-42d6-b851-9b82b00003cf","added_by":"auto","created_at":"2026-01-12 06:10:25","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16025319,"visible":true,"origin":"","legend":"","description":"","filename":"PATECManuscript2sub.docx","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/0deb39b7bdee980fd72b18eb.docx"},{"id":100011476,"identity":"a077f948-255d-4bbb-93f2-20d7a8ca7abc","added_by":"auto","created_at":"2026-01-12 06:10:26","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11130,"visible":true,"origin":"","legend":"","description":"","filename":"85087046075a49d9b6d2b61d2834de96.json","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/d03005c014cce6fb4bf47442.json"},{"id":100011491,"identity":"afa783d1-ca6a-4042-8e09-823e91f58285","added_by":"auto","created_at":"2026-01-12 06:10:27","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":713356,"visible":true,"origin":"","legend":"","description":"","filename":"AdditionalFiles.docx","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/4baa40d1e3d8721ee0306c80.docx"},{"id":100011484,"identity":"4ef0997f-f628-4c56-9f9a-2e0443371216","added_by":"auto","created_at":"2026-01-12 06:10:27","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":200128,"visible":true,"origin":"","legend":"","description":"","filename":"85087046075a49d9b6d2b61d2834de961enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/3119d58f4b94b0b30ff37966.xml"},{"id":100011482,"identity":"5c96f9d6-2bec-4aa2-b7a6-b3254f519474","added_by":"auto","created_at":"2026-01-12 06:10:26","extension":"emf","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10009016,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.emf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/f35a4e70beb960bd42d00ff3.emf"},{"id":100011516,"identity":"0734d17a-2cff-4c7b-85de-f93cb2619b8f","added_by":"auto","created_at":"2026-01-12 06:10:30","extension":"emf","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4720208,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.emf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/5b6853140a56314cbd17cfd1.emf"},{"id":100011496,"identity":"5a604401-f8ef-44e9-b242-c64dc0411d45","added_by":"auto","created_at":"2026-01-12 06:10:28","extension":"emf","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4537084,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.emf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/ffbdb3f61d7599fd9d7df8b9.emf"},{"id":100011498,"identity":"2ce0983e-a10c-4c16-a4e9-61cf05b971d9","added_by":"auto","created_at":"2026-01-12 06:10:28","extension":"emf","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1363168,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.emf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/806e4780f04f3ab78f18703b.emf"},{"id":100011479,"identity":"c877be93-2fbf-460c-93de-7d6c51cfb6fe","added_by":"auto","created_at":"2026-01-12 06:10:26","extension":"emf","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2637816,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.emf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/0653530b5a9d3c75e40f59db.emf"},{"id":100011503,"identity":"f4374e24-3cfa-4105-9bae-416f03bb1d9d","added_by":"auto","created_at":"2026-01-12 06:10:28","extension":"emf","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":905544,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.emf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/b8f576b2059d8ada6297df20.emf"},{"id":100361730,"identity":"cd19e48a-35de-43e7-9e39-f8c50ace5462","added_by":"auto","created_at":"2026-01-16 07:45:37","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":88020,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/03af7b663cd5d26bf7a89e20.png"},{"id":100011475,"identity":"688c483a-ed26-4fd5-84df-57ed1bb94451","added_by":"auto","created_at":"2026-01-12 06:10:26","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35685,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/cb8d39505fe7fafb362119b7.png"},{"id":100011520,"identity":"4302d665-e764-4cfb-9994-0b1f8faddd54","added_by":"auto","created_at":"2026-01-12 06:10:31","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49133,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/9cff22fbe591bbc62a801387.png"},{"id":100011490,"identity":"acc51372-61c2-40fb-8513-1bf33b837441","added_by":"auto","created_at":"2026-01-12 06:10:27","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22316,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/349da220c0a40d1392a0c4d3.png"},{"id":100011448,"identity":"59c2fcf6-0770-47b2-9290-cb5aff658005","added_by":"auto","created_at":"2026-01-12 06:10:24","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":26058,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/f5ebe17dff2bfa84e22b9212.png"},{"id":100011508,"identity":"12e5657e-d680-4bd1-afb7-de63dd4ed1f4","added_by":"auto","created_at":"2026-01-12 06:10:29","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18952,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/b822fdcf0bd8ca5ac48e5f8f.png"},{"id":100011505,"identity":"2581177f-bb5a-4569-bce6-9fe7a8f5fd70","added_by":"auto","created_at":"2026-01-12 06:10:29","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":198928,"visible":true,"origin":"","legend":"","description":"","filename":"85087046075a49d9b6d2b61d2834de961structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/6b30fa0742c9379d3cb32d6b.xml"},{"id":100011515,"identity":"c27eb768-4730-48e7-9782-66fe114fb19b","added_by":"auto","created_at":"2026-01-12 06:10:30","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":211504,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/1cc6511dda9496d7b3da98e6.html"},{"id":100011451,"identity":"15bddb5b-c7c3-43e0-ac0d-2bb8e0a60e8b","added_by":"auto","created_at":"2026-01-12 06:10:25","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":127440,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDevelopment of primary tumor cell cultures (PTCC) and patient-derived autologous tumor-immune effusion co-cultures (PATEC).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e Schematic representation of experimental methodology. \u003cstrong\u003eB\u003c/strong\u003e CONSORT diagram for development of PTCC.\u003cstrong\u003eC \u003c/strong\u003eCellular composition of malignant effusions (EF1-EF6, n = 6), showing proportions of CD45⁺ leukocytes and CD45⁻ non‑hematopoietic cells quantified by flow cytometry. \u003cstrong\u003eD \u003c/strong\u003eImmune subsets within the CD45⁺ compartment depicting relative abundances. \u003cstrong\u003eE \u003c/strong\u003eImmunohistochemical and morphological characterization of PTCC and PATEC Representative PTCC images show tumor cells positive for EpCAM and PanCK by IHC (brown chromogen) with hematoxylin counterstain (blue), and a bright‑field image illustrates a PATEC model.\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/cd91b13868e59fce90e9937a.jpg"},{"id":100011528,"identity":"a3b6785d-9975-43cb-8aed-c6d6f5777488","added_by":"auto","created_at":"2026-01-12 06:10:34","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":98394,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCombinatorial RT-IO screening in PATEC identifies STING agonist and RT as effective therapeutic combination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA \u003c/strong\u003eExperimental schematic. \u003cstrong\u003eB\u003c/strong\u003eQuadruple RT-IO combinations increase tumor cell death in PATEC. Representative histograms (left) depict tumor cell death measured with flow cytometry at 72 h culture after treatment. Boxplots (right) display the fold change in % of tumor cell death relative to the Control (cntrl; n=4). Each boxplot summarizes all quadruple immunotherapy combinations containing the immunotherapeutic indicated on the x-axis, with RT (8 Gy) included in all. Colored dots represent independent biological replicates from PATEC. Median, IQR, whiskers 1.5×IQR. LMM; patient random intercept; Holm. \u003cstrong\u003eC\u003c/strong\u003e Tumor cell death in PATECs (EF1 and EF3, n = 2) after 72 h of RT, STING agonist, TLR7/8-Ago and their combinations. Representative flow cytometric density plots (left). Boxplots summarize tumor cell death, quantified as fold change relative to control (Δ Control, n= 2). Dots indicate biological replicate. One‑way ANOVA; Tukey multiple comparisons. *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/34cce3efb51f3fbe51a547ba.jpg"},{"id":100011506,"identity":"020343dd-b4f9-47bc-9a88-aa90ae2d0e71","added_by":"auto","created_at":"2026-01-12 06:10:29","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":97524,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePatient‑specific tumor cell killing and synergy of STING agonist and RT in PATEC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA \u003c/strong\u003eTumor cell death in PATEC (EF1-EF6, n = 6) after 72 h of treatment (cntrl, STING agonist, RT, RT + STING agonist).\u003cstrong\u003eB\u003c/strong\u003e Bliss synergy index (ΔE, percentage points) for RT + STING agonist for PATEC (EF1-EF6), derived from control normalized fractional tumor cell death. Bars show mean ΔE with 95% bootstrap confidence intervals; dashed and dotted lines indicate ΔE = 0 and the ±5% neutrality band.\u003cstrong\u003e C \u003c/strong\u003eComparison of tumor cell death after 72 h of STING agonist treatment in primary tumor monocultures (TuC/PTCC) versus matched autologous co-cultures (TuCIC/PATEC) (n = 2 PATECs). \u003cstrong\u003eD \u003c/strong\u003eHeatmap of mean cytoplasmic STING expression (STINGMeanOD) and corresponding mean fold change in tumor cell death at 72 h after STING agonist across PTCC (n = 4).\u003cstrong\u003e E \u003c/strong\u003eRepresentative IHC images of formalin‑fixed paraffin‑embedded primary tumor cell cultures stained for STING (brown chromogen) and haematoxylin (blue).\u003c/p\u003e\n\u003cp\u003eTreatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). \u003cstrong\u003eA, C, D \u003c/strong\u003eFold change of % Zombie Violet™⁺ tumor cells relative to cntrl. \u003cstrong\u003eA\u003c/strong\u003e Bars represent mean ± s.e.m. of biological replicates. Two-way ANOVA; Tukey’s multiple comparisons. \u003cstrong\u003eC\u003c/strong\u003e Boxes show median and IQR; whiskers 1.5×IQR; dots are biological replicates. Paired Wilcoxon signed rank test. \u003cstrong\u003eB\u003c/strong\u003e Bars show mean ΔE with 95% bootstrap confidence intervals (2,000 resamples). \u003cstrong\u003eD, E\u003c/strong\u003e descriptive only; no additional statistical testing beyond that specified in the text and Methods. *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001; ****p \u0026lt; 0.0001\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/070a5e5de8c58c6856a63690.jpg"},{"id":100011504,"identity":"fa8e5287-4a42-4e6a-8890-d456502e7bab","added_by":"auto","created_at":"2026-01-12 06:10:29","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":79507,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEffect of STING agonist and RT on cytotoxicity in PATEC depends on immune cell contact. A \u003c/strong\u003eTumor cell death in PATEC (n = 5) after 72 h direct co-culture with immune cells or well insert co-culture STING agonist, RT or RT + STING agonist. Comparisons with cntrl (black) and between treatments (red) are shown above bars. \u003cstrong\u003eB \u003c/strong\u003eTumor cell death in untreated TuC, direct PATEC and insert separated PATEC (well) at 72 h (n = 6); ns, not significant. \u003cstrong\u003eC \u003c/strong\u003eTumor cell death at 72 h in TuC and PATEC after RT or cntrl (n = 5). \u003cstrong\u003eD \u003c/strong\u003eDirect versus insert separated conditions for STING agonist and RT + STING agonist at 72h (n = 5). \u003cstrong\u003eE \u003c/strong\u003eBliss synergy index (ΔE, percentage points) for RT+STING agonist for insert separated condition, summarized per PATEC (EF1-EF5), summarized per PATEC (EF1-EF5) as mean ΔE with 95% bootstrap confidence intervals; dashed and dotted lines indicate ΔE = 0 and the ±5% neutrality band.\u003c/p\u003e\n\u003cp\u003eTreatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100).\u003cstrong\u003eA\u003c/strong\u003e-\u003cstrong\u003eD\u003c/strong\u003e Fold change of % Zombie Violet™+ cells relative to cntrl. \u003cstrong\u003eA\u003c/strong\u003eBars represent mean ± s.e.m. of 2-4 biological replicates (dots). Two-way ANOVA; Tukey’s multiple comparisons\u003cstrong\u003e B\u003c/strong\u003e-\u003cstrong\u003eD\u003c/strong\u003e Boxes show median and IQR; whiskers 1.5×IQR. LMM; patient random intercept; Tukey adjusted pairwise contrasts. ****p \u0026lt; 0.0001, ***p ≤ 0.001, **p \u0026lt; 0.01, *p \u0026lt; 0.05; ns, not significant.\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/4db21b5bf1c796a0a25c47ca.jpg"},{"id":100011494,"identity":"c69b3518-da41-43a7-a00f-972f5ad0e9f3","added_by":"auto","created_at":"2026-01-12 06:10:28","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":97860,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eT cell activation and cytokine induction in effusion cultures following STING agonism and RT. A \u003c/strong\u003eExperimental workflow. \u003cstrong\u003eB\u003c/strong\u003e CD8⁺ and CD4⁺ T cell activation (CD69) and degranulation (CD107a) across treatments (n = 9 effusions). Representative overlay histograms and density plots. \u003cstrong\u003eD \u003c/strong\u003eCytokine and chemokines in supernatants after 24 h of treatment (n=5).\u003c/p\u003e\n\u003cp\u003eTreatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). \u003cstrong\u003eB\u003c/strong\u003e Values are normalized to each effusion’s matched cntrl. \u003cstrong\u003eC\u003c/strong\u003eValues are z-score normalized concentrations. \u003cstrong\u003eB, C\u003c/strong\u003e Boxes show median and IQR; whiskers 1.5×IQR; dots are means of biological replicates. Two-way ANOVA; Tukey’s multiple comparisons. *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001; ****p \u0026lt; 0.0001; ns, not significant.\u003c/p\u003e","description":"","filename":"Picture5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/0bd30fcca6a8f7a10aa2e20d.jpg"},{"id":100011488,"identity":"416093aa-017d-4c31-8ef7-36d21c428b41","added_by":"auto","created_at":"2026-01-12 06:10:27","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":91036,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSTING agonist and RT induce CD8⁺ checkpoint upregulation and associates with immune cell contact-dependent cytotoxicity\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA \u003c/strong\u003eCD8⁺ checkpoint expression in PATEC after 72 h of treatment (cntrl, STING agonist, RT, RT + STING agonist) (n = 5 PATECs). Shows fold change of % CD8⁺ cells positive for the indicated checkpoints relative to matched cntrl. Boxes show median and IQR; whiskers 1.5×IQR; dots are individual experiments. LMM on logit-transformed data; patient random intercept; Dunnett contrasts versus cntrl. *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001; ns, not significant\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eB \u003c/strong\u003eScatterplot of association between CD8⁺Checkpoint Co-expression Score (CES) and overall tumor cell death at 72 h (n = 6 PATECs). CES plotted against Δlogit tumor cell death (change in logit-transformed % dead tumor cells vs cntrl).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eC\u003c/strong\u003eScatterplot of association between CES and the contact-dependent component of tumor cell killing at 72 h (n = 5 PATECs). \u0026nbsp;Cell contact-dependent component is defined as Δ(contact) = Δlogit(Direct) − Δlogit(Insert). CES is plotted against Δ(contact).\u003c/p\u003e\n\u003cp\u003eTreatments: STING agonist (10 µM ADU-S100), RT (8 Gy), RT + STING agonist (8 Gy + 10 µM ADU-S100). \u003cstrong\u003eB, C\u003c/strong\u003e Each dot represents one experiment; colors indicate treatment. The black line shows the LMM regression fit and the grey band the 95% confidence interval.\u003c/p\u003e","description":"","filename":"Picture6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/9bd70752a1033d92a3aba00b.jpg"},{"id":107351618,"identity":"db9da2a9-9d3d-4eba-9ecf-0f72d2197d2d","added_by":"auto","created_at":"2026-04-20 16:11:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1331938,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/c2e7a7ee-eb4c-46e4-8d20-a7cca64f5e69.pdf"},{"id":100011522,"identity":"ed6325f2-22f7-4cd4-9652-139495266a05","added_by":"auto","created_at":"2026-01-12 06:10:31","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":713356,"visible":true,"origin":"","legend":"","description":"","filename":"AdditionalFiles.docx","url":"https://assets-eu.researchsquare.com/files/rs-8516163/v1/62ab60fdfe178bb88cdb8af4.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Autologous tumor-immune effusion co-cultures enable ex vivo functional profiling of radiotherapy-immunotherapy combinations","fulltext":[{"header":"Background","content":"\u003cp\u003eAdvances in immunotherapy for solid cancers are perpetually evolving. Clinical use of immunomodulatory drugs to combat cancer has become more prominent over the last decade, with patients eligible for immune checkpoint inhibition (ICI) increased to 44% in 2018(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). However, only 12% of patients are estimated to respond to ICI and most indications involve monotherapy (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). This has driven the inclination of combining immunomodulatory therapies. Within a couple of years several trials administering dual immunotherapy, including antibodies programmed death 1 (PD-1), programmed death ligand 1 (PD-L1), cytotoxic T lymphocyte associated protein 4 (CTLA-4), lymphocyte activation gene 3 (LAG-3) or T cell immunoreceptor with Ig and ITIM domains (TIGIT), as well as combinations with innate agonists targeting stimulator of interferon genes (STING) or Toll like receptor 7/8 (TLR7/8), have been performed(\u003cspan additionalcitationids=\"CR3 CR4 CR5\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). Moreover, preclinical studies indicate that radiotherapy (RT) is a potent immunological adjuvant. Ionizing radiation induces immunogenic tumor cell death, enhances tumor antigen release and MHC class I expression, and activates cGAS-STING-mediated type I interferons (IFN-I) and inflammatory signaling within the tumor microenvironment (TME). These effects collectively promote recruitment and activation of effector lymphocytes and potentiate the efficacy of ICI and innate immune agonists (\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Clinical trials effectively combining RT with ICI in locally advanced solid tumors further support this rationale (\u003cspan additionalcitationids=\"CR11 CR12 CR13\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Notably, systematic analyses indicate that up to 72% of combinations lacked an explicit biological rationale and indicated most combinations follow an empirical rather than rational design (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOne rational approach to design effective radiotherapy-immunotherapy (RT-IO) combinations would be to identify them through functional testing in patient-derived systems. However, hardly any approaches have been implemented, most likely because testing immunotherapies, and ICIs in particular, requires scalable models that preserve a genuine, autologous tumor-immune microenvironment (TIME). Conventional immunocompetent mouse models and other standard preclinical models do not adequately recapitulate the complexity and interpatient heterogeneity of the human immune system in the context of the TIME. Precision-oriented platforms that use patient-derived material offer the potential to gain insights into the local interactions of the TIME and to design combination therapies with a translational rationale. Organotypic cultures derived from malignant effusions appear to offer a promising approach, allowing the expansion of effusion-derived tumor cells alongside accessible autologous immune cells within a clinically, relevant immunosuppressive TIME (\u003cspan additionalcitationids=\"CR18\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). Published effusion-derived cultures have predominantly expanded tumor compartments for chemosensitivity and targeted therapy screening while the autologous immune compartment is not preserved or not functionally investigated. Furthermore, they do not maintain a native effusion TIME for systematic testing of immunotherapeutic combinations in autologous co-cultures (\u003cspan additionalcitationids=\"CR18 CR19\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMalignant pleural and peritoneal effusions provide tumor cells, stromal fibroblasts and autologous enriched tumor-associated immune cells that are naturally adapted to the liquid environment, making them feasible for \u003cem\u003eex vivo\u003c/em\u003e experiments. T cells, B cells and myeloid cells have been characterized as being more similar to tumor-associated immune cells than peripheral blood mononuclear cells (PBMCs) in respect to their phenotypical and transcriptional states (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). Effusion-derived T cells display increased expression of inhibitory checkpoints and exhaustion markers on T cells are enriched, the transcriptional signatures for regulatory T cells are enhanced and macrophages have continuous phenotypes of M1/M2 with more prominent M2 polarization characteristics (\u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Functional and TCR sequencing studies further indicate that malignant effusions harbor clonally expanded, memory phenotype and CD39\u003csup\u003e+\u003c/sup\u003e CD8⁺ T cell subsets with features of tumor-reactive tumor-infiltrating lymphocytes, together with a cytokine and chemokine milieu enriched for Vascular Endothelial Growth Factor, Interleukin-6, Transforming Growth Factor-beta and C-C motif chemokine ligand 22 that recapitulates key immunosuppressive attributes of the solid TIME (\u003cspan additionalcitationids=\"CR25 CR26\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). Collectively, these phenotypic and functional similarities to the solid TIME support the use of malignant effusions as a platform to investigate therapeutics modulating the cancer-associated immune cell environment.\u003c/p\u003e \u003cp\u003eIn this study, we developed a patient-derived autologous tumor-immune malignant effusion co-culture (PATEC) and used it to assess the functional effects of RT-IO combinations. We systematically tested combinations of innate agonists, as well as ICI in PATEC and identified the STING agonist, particularly in combination with RT, as the most potent inducer of tumor cell death in a subset of patients. The STING agonist-mediated cytotoxicity is largely immune cell dependent and requires direct immune tumor contact. By condensing the distribution of multi checkpoint-positive CD8⁺ T cell phenotypes into a composite Checkpoint Co-expression Score (CES), we link these checkpoint co-expression states specifically to the magnitude of overall and contact-dependent tumor cell killing. Overall, these data provide proof‑of‑concept that malignant effusion-derived PATECs can serve as a patient‑specific \u003cem\u003eex vivo\u003c/em\u003e platform to functionally test and characterize RT-IO regimens within an autologous TIME.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eAcquisition and processing of malignant effusions\u003c/h2\u003e \u003cp\u003eMalignant effusions specimens (ascites and pleural) were prospectively collected from cancer patients undergoing drainage procedures performed by the Department of Thoracic Surgery and Department of General Surgery. Only pathological confirmed (Department of Pathology) malignant effusions were included. All patients provided written informed consent form (ICF). The study was approved by the local Ethics Committee (EC) of the Medical University of Vienna (No. 2042/2019). Samples were collected aseptically in sterile containers or fluid bags (1\u0026ndash;2 L) and transported to the laboratory. Processed within 3 h of collection under aseptic conditions in laminar flow cabinet samples were filtered through a 100 \u0026micro;m cell strainer (pluriStrainer\u0026reg;, pluriSelect) into 50 mL conical Falcon tubes (Sarstedt AG \u0026amp; Co.) removing fibrotic tissue and aggregates. After multiple washes with phosphate-buffered saline (PBS; Gibco) and multiple centrifugation steps at 400\u0026times;rcf for 10 min, erythrocytes were lysed using ACK lysis buffer (Gibco\u0026trade; A1049201). Cell counts were assessed using an automated Sysmex Cell Counter XN 350 (Sysmex). Specimens with fewer than 100\u0026times;10\u003csup\u003e6\u003c/sup\u003e viable cells were excluded from cultivation of primary tumor cell cultures. Cells were resuspended in CryoStor\u0026reg; CS10 medium (MKCL1984, C2874) in addition to 10 \u0026micro;M Y27632 dihydrochloride (MedChemExpress, HY 10071/CS 0131) and aliquoted into sterile 2 mL cryotubes (VWR), and gradually frozen overnight using CoolCell LX containers (BioCision) at -80\u0026deg;C before transfer to liquid nitrogen storage until further use.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eDevelopment and characterization of primary tumor cell cultures\u003c/h3\u003e\n\u003cp\u003ePrimary tumor cell cultures were developed by seeding CD45-depleted cells into standard adherent culture flasks (25 cm\u0026sup2;) using Advanced DMEM/F12 medium (Gibco, 12634 010) with 10% fetal bovine serum (FBS, Gibco), 1% Antibiotic Antimycotic solution (Sigma, A5955), and 10 \u0026micro;M Y 27632 dihydrochloride. Cryopreserved malignant effusion-derived cells were rapidly thawed in 100% FBS supplemented with 10 \u0026micro;M Y 27632, washed once with protocol medium and centrifuged at 400\u0026times;rcf for 5 min. Leukocyte depletion was performed using the EasySep\u0026trade; Human CD45 Depletion Kit (Stemcell Technologies, Cat#17898) according to the manufacturer\u0026rsquo;s protocol, using 5 mL polystyrene round bottom tubes (Stemcell Technologies, Cat#38007) and the EasySep magnetic cell isolation system. Tumor-enriched cells were cultured at 37\u0026deg;C in a humidified incubator with 5% CO₂, with medium changes every 2\u0026ndash;3 days. Successful primary tumor culture development was defined as growth and proliferation beyond three passages.\u003c/p\u003e\n\u003ch3\u003eImmunohistochemical (IHC) analysis of primary tumor cell cultures\u003c/h3\u003e\n\u003cp\u003ePrimary tumor cells seeded onto gelatin-coated slides were fixed in 2% formaldehyde for 20 min at room temperature after confluency. Antigen retrieval was performed using a citrate buffer in an autoclave at 121\u0026deg;C. Endogenous peroxidase was blocked (BLOXALL\u0026trade;, Vector Laboratories) for 10 min. Following blocking with 10% horse serum, slides were incubated for 60\u0026ndash;90 minutes at room temperature with primary antibodies diluted in 10% horse serum: EpCAM (clone HEA 125, MACS Miltenyi, 1:80), pan cytokeratin (clone C11, Cell Signaling, 1:500), and STING (clone D2P2F, Cell Signaling, 1:400). Detection was performed with ImmPRESS\u0026trade; Excel Polymer Reagent and visualized using DAB substrate (Vector Laboratories). Slides were counterstained with hematoxylin and mounted using Entellan\u0026trade; (Merck). Negative controls without primary antibody were included. Imaging was performed by bright-field microscopy using the Vectra Polaris\u0026trade;(Akoya Biosciences).\u003c/p\u003e\n\u003ch3\u003eFlow cytometry (FCM) analysis of cellular composition\u003c/h3\u003e\n\u003cp\u003eCells isolated from malignant effusions and primary tumor cultures were characterized using multiplex flow cytometry. Following thawing and washing, Fc receptor blockade was performed using human AB serum (diluted 1:1 with PBS containing 2% FBS) at 4\u0026deg;C. Cells were then stained with Zombie Violet\u0026trade; or Zombie yellow\u0026trade; viability dye (BioLegend) and incubated with fluorochrome conjugated monoclonal antibodies targeting CD45 (AF700), EpCAM (AF488), CD3 (FITC), CD4 (APC), CD8a (PE, AF700), CD56 (PE), CD14 (PE-Cy7), CD11c (PerCP-eFluor710), HLA-DR (APC-Cy7), CD20 (APC), CD163 (PerCP-eFluor710), CD66b (PE), CD11b (APC-Cy7), CD86 (BV650), CD107a (PE-Cy7), and CD69 (PE). For checkpoint receptor analysis, cells were stained with antibodies specific to CTLA-4 (PE-Dazzle), TIGIT (PerCP-eFluor710), LAG3 (PE-Cy7), PD-1 (BV510), PD-L1 (BV785), and TIM3 (APC-Cy7). Corresponding IgG isotype controls matched for each fluorochrome were included. Compensation was performed using UltraComp eBeads\u0026trade; (Invitrogen). Data were acquired on a CytoFLEX flow cytometer (Beckman Coulter) and analyzed using Kaluza Analysis 2.1 (Beckman Coulter).\u003c/p\u003e\n\u003ch3\u003ePatient-derived Autologous Tumor-immune Effusion Co-culture (PATEC) experiments\u003c/h3\u003e\n\u003cp\u003ePrimary tumor cells and matched immune cells derived from malignant effusions were co-cultured at a 1:5 ratio (5\u0026times;10⁴ tumor cells to 2.5\u0026times;10⁵ immune cells per well) in flat-bottom 24-well plates (Greiner Bio One). The cells were cultured in DMEM/F 12 medium (Gibco) in addition to 10% heat inactivated fetal bovine serum (FBS, Gibco) and 1% Antibiotic Antimycotic Solution (Sigma) in a humidified incubator at 37\u0026deg;C with 5% CO₂. After seeding tumor cells, they were incubated overnight prior to immune cell addition. Selected cultures were subjected to radiotherapy (8 Gy; YXLON Maxishot X-ray system, 200 kV, ~\u0026thinsp;1 Gy/min). Immunotherapeutic treatments included STING agonist ADU-S100 (10 \u0026micro;M, InvivoGen), TLR7/8 agonist Resiquimod (R848; 10 \u0026micro;g/mL, InvivoGen), and checkpoint inhibitors ipilimumab (anti CTLA-4; Yervoy, 10 \u0026micro;g/mL), atezolizumab (anti PD-L1; Tecentriq, 10 \u0026micro;g/mL), pembrolizumab (anti PD-1; Keytruda, 10 \u0026micro;g/mL), and tiragolumab (anti-TIGIT; eubio, 10 \u0026micro;g/mL), alone or in combinations as indicated. After 72 h, cells were detached using Accutase (Sigma) for flow cytometric analysis. Experiments were performed in duplicate.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSpatial separation of PATEC\u003c/h2\u003e \u003cp\u003eIn order to determine cell contact dependence, tumor cells and immune cells were cultured either in direct contact or spatially separated implementing ThinCert well inserts with 0.4 \u0026micro;m pore size (Greiner Bio One, 662640). The tumor cells were seeded at 5\u0026times;10\u003csup\u003e4\u003c/sup\u003e cells per well in 24-well plates, while immune cells were seeded into inserts. Immunotherapeutic treatment with the STING agonist (ADU-S100, 10 \u0026micro;M) and RT (8 Gy) was applied as described above. After 72 h incubation at 37\u0026deg;C and 5% CO₂, cells were harvested using Accutase, stained with fluorochrome-conjugated monoclonal antibodies, and analyzed by FCM.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eFCM analysis of treatment outcomes: tumor cell death and immune checkpoint receptor expression\u003c/h3\u003e\n\u003cp\u003eTumor cell death was quantified after 72 h incubation post-treatment by FCM using Zombie Violet\u0026trade; viability dye (BioLegend) and Calcein AM (BioLegend). Positive controls were generated through repeated freeze-thaw cycles. Viable (Calcein AM⁺/Zombie Violet⁻) and dead (Zombie Violet⁺) cell populations were quantified, and results were expressed as fold change relative to untreated controls. Additionally, immune checkpoint receptor expression was analyzed on CD8⁺ and CD4⁺ T cells.\u003c/p\u003e\n\u003ch3\u003eAssessment of T cell activation\u003c/h3\u003e\n\u003cp\u003eImmune cells isolated from malignant effusions were thawed and resuspended in assay medium (DMEM/F12, 10% FBS, antibiotic antimycotic solution). Cells were seeded at 1\u0026times;10\u003csup\u003e5\u003c/sup\u003e cells per well into 96-well U bottom plates (Greiner Bio One) and treated with STING agonist ADU-S100 (10 \u0026micro;M, InvivoGen) and/or RT (8 Gy; YXLON Maxishot, 200 kV X-ray, ~\u0026thinsp;1 Gy/min). Following 24 h of incubation at 37\u0026deg;C and 5% CO₂, cells were stained with fluorochrome-conjugated antibodies against CD3 FITC (BioLegend), CD4 APC (BioLegend), CD8 PE (BioLegend), CD45 eFluor450 (Invitrogen), CD69 PE (Invitrogen), and CD107a PECy7 (BioLegend). Compensation controls were prepared using UltraComp eBeads\u0026trade; (Invitrogen). Samples were acquired on a DxFLEX flow cytometer (Beckman Coulter), and expression levels of activation markers CD69 and CD107a on CD4⁺ and CD8⁺ T cells were analyzed using Kaluza software (Beckman Coulter).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMultiplex cytokine profiling of culture supernatants\u003c/h2\u003e \u003cp\u003eThe culture supernatants were harvested 24 h post-treatment and analyzed regarding cytokine secretion using the Luminex\u0026reg; Human Discovery Assay (13-Plex) LXSAHM-13 (Bio-Techne Ireland Limited) following the manufacturer's protocol. The cytokine panel included TNF-α, IL-6, IP-10, IFN-α, MIP-1α, MIP-1β, MCP-1, IL-1ra, IL-8, RANTES, IFN-β, IFN-γ, and IL-10. Data acquisition was performed on a Luminex\u0026reg; FlexMap 3D, and analyses were conducted using the xPONENT\u0026reg; 4.3 analysis software (Luminex\u0026reg;).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eStatistical analyses were performed in R (version 2025.05.1\u0026thinsp;+\u0026thinsp;513) using the lme4 and lmerTest packages for linear mixed‑effects models, emmeans or multcomp for multiple comparisons, and ggplot2 together with dplyr and tidyr for data handling and visualization. Treatment effects across multiple patients were evaluated using linear mixed effects models, adjusting for patient-specific random effects. Post hoc pairwise comparisons between treatment groups were conducted using Tukey\u0026rsquo;s multiple comparisons test. The cytokine concentrations were normalized using z-score transformation for comparative analyses. Additionally, intra-patient analyses were performed using one way ANOVA followed by Tukey\u0026rsquo;s post hoc test, and, where appropriate, by two-way ANOVA with factors Treatment and Culture condition followed by Tukey\u0026rsquo;s post hoc test. For checkpoint expression panels with multiple CD8⁺ subsets, family-wise multiplicity across the panel was controlled (Holm or Benjamini-Hochberg false discovery rate) in addition to the within-panel Dunnett adjustments. For cytokines, p-values were adjusted within the cytokine family using the Benjamini-Hochberg FDR. Statistical significance was defined as follows: *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, ****p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001.\u003c/p\u003e \u003cp\u003eCD8⁺ checkpoint panels and CES-killing analyses were conducted on the logit scale. Percent positive values were logit-transformed and analyzed using Gaussian linear mixed effects models with Condition as a fixed effect and patient as a random intercept; a random slope for Condition was included when supported by the split-plot design. For display, values are shown as fold change versus the patient matched control, while inference used the logit model. Omnibus p-values for Condition derive from the mixed model; multiplicity-adjusted post hoc comparisons versus control were obtained using Dunnett contrasts. The CD8⁺ Checkpoint Co-expression Score (CES; Day 3, within PATEC) was computed by z-scoring the %Gated frequencies of PD-1⁺LAG-3⁺TIM-3⁺, PD-1⁺TIGIT⁺LAG-3⁺, PD-1⁺TIGIT⁺TIM-3⁺ and PD-1⁺TIGIT⁺LAG-3⁺TIM-3⁺ and combining as CES\u0026thinsp;=\u0026thinsp;mean(z of the three triple positive subsets)\u0026thinsp;+\u0026thinsp;2\u0026times;z(quadruple positive). Tumor cell death at 72 h was summarized as Δlogit\u0026thinsp;=\u0026thinsp;logit(%dead_treated) \u0026ndash; logit(%dead_control), and the contact-dependent component as Δ(contact) = Δlogit(Direct) \u0026minus; Δlogit(Insert). Associations with CES were evaluated by Gaussian LMMs with patient random intercepts (Δlogit\u0026thinsp;~\u0026thinsp;CES + (1|patient); Δ(contact)\u0026thinsp;~\u0026thinsp;CES + (1|patient)); fixed effect slopes (β) are reported in log odds per CES unit with Wald 95% confidence intervals, Satterthwaite P values (lmerTest), and Nakagawa R\u0026sup2; (marginal; conditional). For Δ(contact), the random intercept variance was \u0026asymp;\u0026thinsp;0 (singular fit), hence R\u0026sup2;_cond\u0026thinsp;\u0026asymp;\u0026thinsp;R\u0026sup2;_marg; improvement over a null (intercept only) model was checked by likelihood ratio test and leave one PATEC out refits were used as a stability check.\u003c/p\u003e \u003cp\u003eSynergy between STING agonist and RT was quantified using the Bliss independence model applied to control-normalized fractional killing. For each patient and experiment, the percentage of Zombie Violet\u0026trade;-positive tumor cells in untreated controls (d_ctrl) and in treated wells (d) was converted to fractions and expressed as additional killing on the surviving population, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\varvec{E}=\\frac{\\varvec{d}-{\\varvec{d}}_{\\text{ctrl}}}{1-{\\varvec{d}}_{\\text{ctrl}}}\\)\u003c/span\u003e\u003c/span\u003e, truncated to the interval [0,1]. Bliss expectation was calculated as \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:({\\varvec{E}}_{\\text{bliss}}={\\varvec{E}}_{\\text{STING}}+{\\varvec{E}}_{\\text{RT}}-{\\varvec{E}}_{\\text{STING}}\\times\\:{\\varvec{E}}_{\\text{RT}})\\)\u003c/span\u003e\u003c/span\u003e. The Bliss synergy index is defined as \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\varvec{\\Delta\\:}\\varvec{E}={\\varvec{E}}_{\\text{obs}}-{\\varvec{E}}_{\\text{bliss}}\\)\u003c/span\u003e\u003c/span\u003e ​, expressed as percentage points. For each patient, replicate level \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\varvec{\\Delta\\:}\\varvec{E}\\)\u003c/span\u003e\u003c/span\u003e values were summarized as the mean with 2,000-sample non-parametric bootstrap 95% confidence intervals; combinations were classified as synergistic when the entire confidence interval lay above +\u0026thinsp;5 percentage points, antagonistic when it lay below \u0026minus;\u0026thinsp;5 percentage points, and otherwise as neutral.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eDevelopment and characterization of patient-derived autologous tumor-immune malignant effusion co-cultures (PATEC)\u003c/h2\u003e \u003cp\u003eWe hypothesized that malignant effusions, which contain both tumor and immune cells within their native fluid microenvironment, could serve as a source for generating autologous co-cultures to study RT-IO responses. To test this, malignant pleural and peritoneal effusions (pleural\u0026thinsp;=\u0026thinsp;23, ascites\u0026thinsp;=\u0026thinsp;6) were prospectively collected and processed for biobanking and primary tumor cell culture (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Nine samples were excluded due to insufficient viable cell counts (\u0026lt;\u0026thinsp;100 \u0026times; 10⁶ cells). Of the remaining 20 cultures initiated, six maintained continuous proliferation beyond three passages and were expanded as primary tumor cell cultures (PTCCs) for downstream functional co-culture assays (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eImmunohistochemical and morphological validation of primary tumor cell cultures (PTCC)\u003c/h2\u003e \u003cp\u003eFCM analysis within the CD45⁻ compartment showed tumor cells expressed the epithelial marker EpCAM in five of six cases, consistent with their tumor entity, while one sample (EF4), derived from a small cell lung carcinoma (SCLC), lacked EpCAM expression, in line with its clinical histopathology report (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). To confirm epithelial tumor origin, Immunohistochemical (IHC) staining of PTCCs was performed. PTCCs displayed strong cytoplasmic expression of pan cytokeratin (PanCK) and membranous expression of EpCAM, confirming epithelial origin (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). Brightfield microscopy of PTCCs showed cohesive clusters of polygonal, adherent cells with characteristic epithelial morphology.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eHeterogeneous immune cell compositions of malignant effusions\u003c/h2\u003e \u003cp\u003eFCM demonstrated pronounced heterogeneity in the cellular composition of malignant effusions (EF1-EF6) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF, \u003cb\u003eSupplementary Fig.\u0026nbsp;2\u003c/b\u003e). Leukocytes (CD45⁺) were the most dominant population in most samples, making up over 90% of cells. The immune compartment showed patient-specific variation, with macrophages and neutrophils as the most abundant subsets in half of the patient samples, and variable proportions of CD4⁺ and CD8⁺ T cells, NK cells, NKT cells, dendritic cells, and B cells. This interpatient diversity reflects the individualized immune and tumor composition of malignant effusions, provides a genuine TIME and thus an optimal basis for patient-specific functional testing of immunomodulatory drugs. Collectively, these results establish a workflow to generate PTCCs that can be paired with autologous immune compartments, enabling downstream use of PATEC platform for functional RT-IO testing. With scalable patient-matched tumor and immune cells from one compartment, we next assessed therapeutic efficacy across immunomodulatory combinations to determine the most efficacious treatment regimen.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eA\u003c/b\u003e Schematic representation of experimental methodology. \u003cb\u003eB\u003c/b\u003e CONSORT diagram for development of PTCC. \u003cb\u003eC\u003c/b\u003e Cellular composition of malignant effusions (EF1-EF6, n\u0026thinsp;=\u0026thinsp;6), showing proportions of CD45⁺ leukocytes and CD45⁻ non‑hematopoietic cells quantified by flow cytometry. \u003cb\u003eD\u003c/b\u003e Immune subsets within the CD45⁺ compartment depicting relative abundances. \u003cb\u003eE\u003c/b\u003e Immunohistochemical and morphological characterization of PTCC and PATEC Representative PTCC images show tumor cells positive for EpCAM and PanCK by IHC (brown chromogen) with hematoxylin counterstain (blue), and a bright‑field image illustrates a PATEC model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eCombinatorial RT-IO screening in PATEC identifies STING agonism and RT as an effective therapeutic combination\u003c/h2\u003e \u003cp\u003eTo functionally assess the capacity of the PATEC platform to capture immune mediated tumor cytotoxicity, we performed an initial combinatorial screen integrating innate and adaptive immunotherapies with RT (8 Gy) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cb\u003eSupplementary Fig.\u0026nbsp;1A/B\u003c/b\u003e). Quadruple RT-IO included two innate agonists targeting STING and TLR7/8 together with ICI PD-1, PD-L1, CTLA-4, and TIGIT. Tumor cell death was quantified at 72 h by FCM using Zombie Violet\u0026trade; and Calcein AM. Representative histograms illustrate a marked shift towards Zombie Violet⁺ cells following treatment compared with RT alone (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, left). Across four independent PATECs, quadruple RT-IO that included an innate agonist (STING agonist or TLR7/8 agonist) together with RT significantly increased tumor cell death compared with RT alone or untreated controls (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, right; component wise linear mixed effects model with patient random intercept; Holm adjustment; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). This demonstrates that PATEC can be used as an \u003cem\u003eex vivo\u003c/em\u003e human system capable of detecting immune mediated tumor killing under relevant combination regimens.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eCombinations with STING agonist contributed most strongly to tumor killing\u003c/h2\u003e \u003cp\u003eTo delineate which innate immune modulatory components contributed to the observed response, we next compared STING agonist and TLR7/8-Ago, alone or in combination with RT, in representative PATECs (EF1 and EF3). Tumor cell death, expressed as fold change relative to matched controls, increased with either agonist, but the STING agonist\u0026thinsp;+\u0026thinsp;RT combination produced the most pronounced killing across replicates (one way ANOVA with Tukey multiple comparisons; p\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u0026thinsp;\u0026minus;\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). STING agonist monotherapy also induced measurable cytotoxicity in the selected PATECs, whereas TLR7/8 agonism alone or with RT had more modest effects. Notably, RT potentiated STING agonist efficacy, suggesting cooperative mechanisms between DNA damage signaling and STING activation. Collectively, these findings indicate STING agonism in combination with RT is the leading combination in regard to tumor cell death in these select PATECs. We next investigated interpatient variability and potential synergy between STING agonism and RT across PATECs.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eA\u003c/b\u003e Experimental schematic. \u003cb\u003eB\u003c/b\u003e Quadruple RT-IO combinations increase tumor cell death in PATEC. Representative histograms (left) depict tumor cell death measured with flow cytometry at 72 h culture after treatment. Boxplots (right) display the fold change in % of tumor cell death relative to the Control (cntrl; n\u0026thinsp;=\u0026thinsp;4). Each boxplot summarizes all quadruple immunotherapy combinations containing the immunotherapeutic indicated on the x-axis, with RT (8 Gy) included in all. Colored dots represent independent biological replicates from PATEC. Median, IQR, whiskers 1.5\u0026times;IQR. LMM; patient random intercept; Holm. \u003cb\u003eC\u003c/b\u003e Tumor cell death in PATECs (EF1 and EF3, n\u0026thinsp;=\u0026thinsp;2) after 72 h of RT, STING agonist, TLR7/8-Ago and their combinations. Representative flow cytometric density plots (left). Boxplots summarize tumor cell death, quantified as fold change relative to control (Δ Control, n\u0026thinsp;=\u0026thinsp;2). Dots indicate biological replicate. One‑way ANOVA; Tukey multiple comparisons. *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003ePatient-specific variability in response to STING agonist and RT in PATEC\u003c/h2\u003e \u003cp\u003eTo assess patient-specific efficacy of STING agonism and RT, six PATECs (EF1-EF6) were treated with STING agonist, RT, or the combination for 72 h (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Across all patients, combination therapy showed the most pronounced tumor cell death, whereas the relative magnitude of response varied markedly between PATECs. EF2 and EF3 exhibited the most pronounced cytotoxicity (two to three-fold), while EF1 and EF6 displayed modest increases. Two-way ANOVA with Tukey\u0026rsquo;s multiple comparisons confirmed significant treatment effects in all PATECs (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05-0.0001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). These results demonstrate that STING agonism, particularly in conjunction with RT, induces robust but patient-specific variable tumor cell killing in PATEC.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eSubset of PATECs exhibit synergistic STING-RT interactions\u003c/h2\u003e \u003cp\u003eNext, we examined whether combining STING agonism and RT results in greater-than-additive cytotoxicity. Bliss independence modelling identified synergy in a subset of PATECs (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). EF2, EF3, and EF6 showed strong positive Bliss indices (\u0026thinsp;\u0026gt;\u0026thinsp;+\u0026thinsp;5%) with non-overlapping 95% CIs from 2,000 bootstrap resamples, indicating synergy, whereas EF1, EF5, and EF4 fell within the neutrality band. The presence of both synergistic and neutral responses across patients underscores functional heterogeneity in innate immune and RT sensitivity within PATECs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eSTING agonist induced response depends on immune cells\u003c/h2\u003e \u003cp\u003eWe compared primary tumor monocultures (TuC) with matched autologous tumor-immune co-cultures (PATEC) to assess whether the cytotoxic effects of STING activation are immune cell dependent. After 72 h of STING agonist treatment, tumor cell death increased significantly only in the presence of immune cells (paired Wilcoxon signed rank test; p\u0026thinsp;\u0026lt;\u0026thinsp;0.01; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD), indicating that STING-mediated cytotoxicity is largely indirect and depends strongly on the immune cell compartment.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eSTING expression in tumor cells varies across patients\u003c/h2\u003e \u003cp\u003eIHC staining of primary tumor cultures revealed variable cytoplasmic STING expression among patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE). EF1 and EF5 exhibited strong staining, whereas EF2 and EF3 were weak or negative. A descriptive comparison of mean optical density (OD) and corresponding fold change in tumor cell death after STING agonist exposure suggested that higher baseline STING expression may coincide with increased responsiveness in some samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). However, although the limited cohort precludes definitive correlation.\u003c/p\u003e \u003cp\u003eTogether, these data demonstrate that STING agonism elicits variable but reproducible tumor cell killing across patients, that a subset of PATECs displays synergistic effects with RT, and that STING-dependent cytotoxicity is largely dependent on the presence of immune cells. Because immune cell participation appeared essential for STING efficacy, we next examined whether this killing required direct immune tumor contact (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eA\u003c/b\u003e Tumor cell death in PATEC (EF1-EF6, n\u0026thinsp;=\u0026thinsp;6) after 72 h of treatment (cntrl, STING agonist, RT, RT\u0026thinsp;+\u0026thinsp;STING agonist). \u003cb\u003eB\u003c/b\u003e Bliss synergy index (ΔE, percentage points) for RT\u0026thinsp;+\u0026thinsp;STING agonist for PATEC (EF1-EF6), derived from control normalized fractional tumor cell death. Bars show mean ΔE with 95% bootstrap confidence intervals; dashed and dotted lines indicate ΔE\u0026thinsp;=\u0026thinsp;0 and the \u0026plusmn;\u0026thinsp;5% neutrality band. \u003cb\u003eC\u003c/b\u003e Comparison of tumor cell death after 72 h of STING agonist treatment in primary tumor monocultures (TuC/PTCC) versus matched autologous co-cultures (TuCIC/PATEC) (n\u0026thinsp;=\u0026thinsp;2 PATECs). \u003cb\u003eD\u003c/b\u003e Heatmap of mean cytoplasmic STING expression (STINGMeanOD) and corresponding mean fold change in tumor cell death at 72 h after STING agonist across PTCC (n\u0026thinsp;=\u0026thinsp;4). \u003cb\u003eE\u003c/b\u003e Representative IHC images of formalin‑fixed paraffin‑embedded primary tumor cell cultures stained for STING (brown chromogen) and haematoxylin (blue).\u003c/p\u003e \u003cp\u003eTreatments: STING agonist (10 \u0026micro;M ADU-S100), RT (8 Gy), RT\u0026thinsp;+\u0026thinsp;STING agonist (8 Gy\u0026thinsp;+\u0026thinsp;10 \u0026micro;M ADU-S100). \u003cb\u003eA, C, D\u003c/b\u003e Fold change of % Zombie Violet\u0026trade;⁺ tumor cells relative to cntrl. \u003cb\u003eA\u003c/b\u003e Bars represent mean\u0026thinsp;\u0026plusmn;\u0026thinsp;s.e.m. of biological replicates. Two-way ANOVA; Tukey\u0026rsquo;s multiple comparisons. \u003cb\u003eC\u003c/b\u003e Boxes show median and IQR; whiskers 1.5\u0026times;IQR; dots are biological replicates. Paired Wilcoxon signed rank test. \u003cb\u003eB\u003c/b\u003e Bars show mean ΔE with 95% bootstrap confidence intervals (2,000 resamples). \u003cb\u003eD, E\u003c/b\u003e descriptive only; no additional statistical testing beyond that specified in the text and Methods. *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; ****p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eDirect immune tumor contact is critical for the cytotoxic efficacy of STING agonist-based RT-IO\u003c/h2\u003e \u003cp\u003eWe next examined whether physical proximity between tumor and immune cells is necessary for the cytotoxic effects of STING agonist-based RT-IO. PATECs were treated for 72 h with STING agonist, RT, or the combination, either in direct co-culture or separated by a transwell insert (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Across five independent PATECs, RT\u0026thinsp;+\u0026thinsp;STING agonist elicited a marked increase in tumor cell death under direct contact conditions, whereas spatial separation abolished this response (two-way ANOVA with Tukey\u0026rsquo;s multiple comparisons; p\u0026thinsp;\u0026lt;\u0026thinsp;0.05-0.0001). By contrast, the cytotoxic effect of RT monotherapy was preserved in the insert condition, indicating that RT acts primarily through tumor intrinsic processes (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Pooled analysis across PATECs confirmed significantly higher killing for both STING agonist alone and RT\u0026thinsp;+\u0026thinsp;STING agonist in direct contact compared with insert conditions (linear mixed effects model with patient random intercept; p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, p\u0026thinsp;\u0026lt;\u0026thinsp;0.000; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Consistently, Bliss defined RT-STING synergy became neutral in all PATECs when repeated in the insert condition (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE). These findings demonstrate that the cytotoxic efficacy of STING agonist-based RT-IO is highly dependent on direct immune tumor contact, consistent with engagement of contact-dependent effector pathways such as immune synapse formation, directed degranulation, or death receptor mediated interactions.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eRadiotherapeutic efficacy is largely independent of immune cells in PATEC\u003c/h2\u003e \u003cp\u003eNext, we evaluated whether RT-induced killing also requires immune cells, we compared RT responses in tumor monocultures (TuC) versus autologous co-cultures (PATEC). RT alone significantly increased tumor cell death compared to control (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), but responses were indistinguishable between TuC and PATEC conditions (linear mixed effects model with patient random intercept; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). This indicates that, unlike STING agonist, RT elicits direct tumor cell toxicity that is largely immune cell independent.\u003c/p\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eImmune cell contact alone does not promote spontaneous killing\u003c/h2\u003e \u003cp\u003eTo determine whether autologous immune cells exert any basal cytotoxicity in the absence of therapeutic stimulation, we compared untreated tumor cell monocultures (TuC) with direct co-cultures (PATEC) and insert separated co-cultures (PATEC (well)). Baseline tumor cell death was indistinguishable across all three configurations (linear mixed effects model; ns), indicating that effusion-derived immune cells do not spontaneously kill tumor cells \u003cem\u003eex vivo\u003c/em\u003e, even when in direct physical contact (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). In summary, these data show that immune cell activity in PATEC requires therapeutic augmentation and that the tumoricidal effects of STING agonist-based RT-IO derive from treatment induced, not constitutive, immune engagement. This provided the rationale to examine whether early immune activation and cytokine responses accompany the induction of cytotoxicity (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTreatments: STING agonist (10 \u0026micro;M ADU-S100), RT (8 Gy), RT\u0026thinsp;+\u0026thinsp;STING agonist (8 Gy\u0026thinsp;+\u0026thinsp;10 \u0026micro;M ADU-S100). \u003cb\u003eA\u003c/b\u003e-\u003cb\u003eD\u003c/b\u003e Fold change of % Zombie Violet\u0026trade;+ cells relative to cntrl. \u003cb\u003eA\u003c/b\u003e Bars represent mean\u0026thinsp;\u0026plusmn;\u0026thinsp;s.e.m. of 2\u0026ndash;4 biological replicates (dots). Two-way ANOVA; Tukey\u0026rsquo;s multiple comparisons \u003cb\u003eB\u003c/b\u003e-\u003cb\u003eD\u003c/b\u003e Boxes show median and IQR; whiskers 1.5\u0026times;IQR. LMM; patient random intercept; Tukey adjusted pairwise contrasts. ****p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001, ***p\u0026thinsp;\u0026le;\u0026thinsp;0.001, **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; ns, not significant.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003eT cell activation in malignant effusion cultures following STING agonism and RT\u003c/h2\u003e \u003cp\u003eTo assess early immune activation following treatment, malignant effusion cultures were exposed to STING agonist, RT, or combination for 24 h (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). FCM analysis revealed distinct activation and degranulation profiles across T cell subsets. Expression of the activation marker CD69 was most pronounced after combined RT\u0026thinsp;+\u0026thinsp;STING agonist, exceeding both monotherapies (two-way ANOVA with Tukey; p\u0026thinsp;\u0026lt;\u0026thinsp;0.05-0.0001; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). In contrast, the degranulation marker CD107a most markedly increased after STING agonist monotherapy, with combination treatment showing moderate but consistent induction across patients\u0026rsquo; malignant effusions. Both CD8⁺ and CD4⁺ T cells exhibited similar trends.\u003c/p\u003e \u003cp\u003eCollectively, these data indicate that STING agonism rapidly promotes T cell degranulation, whereas its combination with RT primarily amplifies early activation, reflecting distinct but complementary mechanisms of effector engagement within the effusion microenvironment.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e \u003ch2\u003ePro-inflammatory cytokine and IFN-I responses accompany T cell activation\u003c/h2\u003e \u003cp\u003eTo explore whether cytokine release increases upon treatment, multiplex profiling of culture supernatants was performed after 24 h (n\u0026thinsp;=\u0026thinsp;5). The combined RT\u0026thinsp;+\u0026thinsp;STING agonist treatment induced a broad pro-inflammatory cytokine response, marked by increased levels of TNF-α, IL-6, IP-10 (CXCL10), IFN-α, MIP-α (CCL3), and MIP-1β (CCL4) (two-way ANOVA with Tukey; p\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u0026thinsp;\u0026minus;\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). In contrast, IFN-β, IL-8, MCP-1, RANTES, IFN-γ, and IL-10 showed no consistent modulation at this early time point (\u003cb\u003eSupplementary Fig.\u0026nbsp;3\u003c/b\u003e). This cytokine profile is characteristic of an early IFN-I-driven and chemokine rich milieu, consistent with STING pathway activation and recruitment of effector leukocytes.\u003c/p\u003e \u003cp\u003eOverall, these results show that combined STING agonism and RT trigger rapid immune activation within malignant effusion cultures, marked by CD8⁺ and CD4⁺ T cell degranulation and a coordinated IFN-I cytokine response. This early immunostimulatory phase prompted us to investigate whether sustained checkpoint upregulation and functional engagement of CD8⁺ T cells correlate with tumor cell killing at later time points (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTreatments: STING agonist (10 \u0026micro;M ADU-S100), RT (8 Gy), RT\u0026thinsp;+\u0026thinsp;STING agonist (8 Gy\u0026thinsp;+\u0026thinsp;10 \u0026micro;M ADU-S100). \u003cb\u003eB\u003c/b\u003e Values are normalized to each effusion\u0026rsquo;s matched cntrl. \u003cb\u003eC\u003c/b\u003e Values are z-score normalized concentrations. \u003cb\u003eB, C\u003c/b\u003e Boxes show median and IQR; whiskers 1.5\u0026times;IQR; dots are means of biological replicates. Two-way ANOVA; Tukey\u0026rsquo;s multiple comparisons. *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; ****p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001; ns, not significant.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003eSTING based RT-IO enhances checkpoint co-expression on CD8⁺ T cells\u003c/h2\u003e \u003cp\u003eHaving established that STING agonism promotes early T cell activation and cytokine release (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), we next examined whether prolonged stimulation elicits adaptive checkpoint engagement on cytotoxic T cells. After 72 h of treatment, PATEC cultures exposed to STING agonist\u0026thinsp;+\u0026thinsp;RT displayed broad upregulation of inhibitory and activation receptors on CD8⁺ T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). Single positive subsets (PD-1⁺, TIGIT⁺, LAG-3⁺) increased relative to matched controls, and combinatorial phenotypes, particularly PD-1⁺TIGIT⁺, PD-1⁺LAG-3⁺TIGIT⁺ and TIGIT⁺LAG-3⁺TIM-3⁺, were most enriched following RT\u0026thinsp;+\u0026thinsp;STING agonist (linear mixed effects models with Dunnett contrasts; p\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u0026thinsp;\u0026minus;\u0026thinsp;0.001). The simultaneous induction of multiple checkpoints under this regimen indicates a sustained antigenic stimulation program typical of recently activated effector cells rather than fixed exhaustion.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003eComposite checkpoint expression correlates with tumor cell killing\u003c/h2\u003e \u003cp\u003eTo quantify checkpoint co-expression, we derived a Checkpoint Co-expression Score (CES) that integrates z normalized frequencies of triple positive subsets weighted by quadruple positive enrichment. Across PATECs and treatments, higher CES correlated with greater tumor cell death at 72 h (Δlogit vs matched control; Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). In a Gaussian mixed effects model with patient as random intercept, the slope was β\u0026thinsp;=\u0026thinsp;0.15 (95% CI 0.11\u0026ndash;0.20), corresponding to an odds ratio (OR) of 1.16 per CES unit (p\u0026thinsp;=\u0026thinsp;1.8 \u0026times; 10⁻⁸; R\u0026sup2;ₘarg\u0026thinsp;=\u0026thinsp;0.37, R\u0026sup2;cond\u0026thinsp;=\u0026thinsp;0.41). Experiments from RT\u0026thinsp;+\u0026thinsp;STING agonist tended to be in the upper CES-cytotoxicity range, but the fitted relationship was consistent across all conditions, indicating that the degree of checkpoint co-expression quantitatively mirrors magnitude of cytotoxic efficacy irrespective of treatment type.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCES is preferentially associated with the contact-dependent component of tumor cell killing\u003c/h3\u003e\n\u003cp\u003eHaving established that STING agonist induced cytotoxicity contains a substantial contact-dependent component (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), we next examined whether the CES cytotoxicity relationship was primarily attributable to this contact mediated fraction. To isolate this effect, we defined Δ(contact) as the difference in treatment induced cytotoxicity between direct co-culture and insert separated conditions: Δ(contact) = Δlogit(Direct) \u0026minus; Δlogit(Insert). Across PATECs, Δ(contact) increased with CES, with a mixed effects slope of β\u0026thinsp;=\u0026thinsp;0.09 (95% CI 0.03\u0026ndash;0.15), corresponding to a 10% increase in the odds of contact-dependent killing per CES unit (p\u0026thinsp;=\u0026thinsp;0.0088; R\u0026sup2;ₘₐ\u003csub\u003er\u003c/sub\u003eg\u0026thinsp;=\u0026thinsp;0.31; Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC). Importantly, this association persisted after subtraction of contact independent cytotoxicity, indicating that CD8⁺ T cells with higher checkpoint expression preferentially arise under conditions of direct immune tumor interaction, consistent with an activation program linked to sustained antigen engagement.\u003c/p\u003e \u003cp\u003eThe distribution of checkpoint positive phenotypes is consistent with established CD8⁺ T cell differentiation trajectories. PD-1⁺TIGIT⁺ subsets are characteristic of recently stimulated, antigen-experienced effector populations that retain proliferative and cytotoxic potential, whereas the additional expression of LAG-3 or TIM-3 marks progressively constrained states associated with more sustained antigen exposure. At this early 72 h time point, the positive association between CES and tumor cell killing indicates that multi-checkpoint expression reflects an actively engaged effector program operating under inducible inhibitory feedback, rather than terminal dysfunction.\u003c/p\u003e \u003cp\u003eTo explore whether induction of individual checkpoints associated with benefit of having the corresponding checkpoint inhibitor in the PATEC, we used the quadruple RT-IO screen to correlate upregulation of PD-1⁺, PD-L1/PD-1⁺, TIGIT⁺ and CTLA-4⁺ on T cells after STING agonist\u0026thinsp;+\u0026thinsp;RT with the additional tumor cell death observed when Pembrolizumab, Atezolizumab, Tiragolumab or Ipilimumab are in culture. In CD8⁺ T cells, higher induction of PD-1, PD-L1/PD-1 and TIGIT tended to coincide with greater incremental killing, whereas analogous associations for CD4⁺ T cells were weaker and more variable; none reached statistical significance in this small cohort and these analyses were therefore considered exploratory (\u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eThese hypothesis-generating data indicate that multi-checkpoint expression on CD8⁺ T cells are associated with the magnitude of tumor cell killing, and that this association persists after isolating the contact-dependent component of cytotoxicity. Accordingly, higher CES values capture CD8⁺ states that coincide with conditions in which contact-mediated cytotoxicity constitutes a larger share of the overall response.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eA\u003c/b\u003e CD8⁺ checkpoint expression in PATEC after 72 h of treatment (cntrl, STING agonist, RT, RT\u0026thinsp;+\u0026thinsp;STING agonist) (n\u0026thinsp;=\u0026thinsp;5 PATECs). Shows fold change of % CD8⁺ cells positive for the indicated checkpoints relative to matched cntrl. Boxes show median and IQR; whiskers 1.5\u0026times;IQR; dots are individual experiments. LMM on logit-transformed data; patient random intercept; Dunnett contrasts versus cntrl. *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; ns, not significant\u003c/p\u003e \u003cp\u003e \u003cb\u003eB\u003c/b\u003e Scatterplot of association between CD8⁺ Checkpoint Co-expression Score (CES) and overall tumor cell death at 72 h (n\u0026thinsp;=\u0026thinsp;6 PATECs). CES plotted against Δlogit tumor cell death (change in logit-transformed % dead tumor cells vs cntrl).\u003c/p\u003e \u003cp\u003e \u003cb\u003eC\u003c/b\u003e Scatterplot of association between CES and the contact-dependent component of tumor cell killing at 72 h (n\u0026thinsp;=\u0026thinsp;5 PATECs). Cell contact-dependent component is defined as Δ(contact) = Δlogit(Direct) \u0026minus; Δlogit(Insert). CES is plotted against Δ(contact).\u003c/p\u003e \u003cp\u003eTreatments: STING agonist (10 \u0026micro;M ADU-S100), RT (8 Gy), RT\u0026thinsp;+\u0026thinsp;STING agonist (8 Gy\u0026thinsp;+\u0026thinsp;10 \u0026micro;M ADU-S100). \u003cb\u003eB, C\u003c/b\u003e Each dot represents one experiment; colors indicate treatment. The black line shows the LMM regression fit and the grey band the 95% confidence interval.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we developed PATEC as a human \u003cem\u003eex vivo\u003c/em\u003e platform to evaluate RT-IO regimens in an immunocompetent TIME. In a combinatorial screen, we found regimens combining RT with STING agonism were the most tumoricidal in this cohort. Furthermore, PATEC showed STING dependent cytotoxicity largely mediated by immune cells and strongly depends on direct immune tumor contact. In contrast, RT alone acts largely through tumor intrinsic mechanisms. STING based RT-IO is accompanied by effector T cell activation, an inflammatory/IFN-I associated cytokine milieu and upregulation of inhibitory checkpoints on CD8⁺ T cells. Together, these findings supporting the use of this platform to identify and characterize feasible RT-IO combinations.\u003c/p\u003e \u003cp\u003eWithin the current landscape of human \u003cem\u003eex vivo\u003c/em\u003e models, PATEC addresses some of the distinct limitations of the existing ones. While organoid immune co-cultures have been used to test ICI, their origin are solid tumors and their inherent immune infiltrate is sparse. They heavily rely on stem cell supporting growth factors, as well as typically requiring extraction, expansion and reintroduction of immune cells from PBMCs or TILs, rather than utilizing the native autologous TIME (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). Similarly, Tumor-on-a-chip platform mostly implement immune cell substitutes, require processing of scarce surgical or biopsy material and coincide with high costs due to microfabrication with a specialized infrastructure (\u003cspan additionalcitationids=\"CR31 CR32 CR33\" citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). Precision-cut tissue slices and PDTF platforms on the other hand preserve tissue architecture and stromal context. They are useful for short-term assessment of checkpoint and cytokine responses, but are still limited by restricted source material and a solid tissue interface that does not readily permit systematic separation or repeated perturbation of tumor and immune compartments (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Malignant effusion-derived cultures published to date have focused on expanding epithelial cells for targeted or chemotherapeutic drug screening, isolating and stimulating T cells for cytotoxicity assays or testing CAR-T cell cytotoxicity using solely the pleural fluid (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan additionalcitationids=\"CR38 CR39\" citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e). The configuration of the PATEC however uses the naturally enriched autologous tumor-associated immune cells in the fluid TIME retaining a plethora of myeloid and lymphoid repertoires of the malignant effusions. Cryopreserved aliquots enable experimental flexibility; spatial compartmentalization allow for dissection of the responsible cytotoxic effector and serial sampling offers the potential of longitudinal functional profiling during the course of therapy.\u003c/p\u003e \u003cp\u003eThe cellular baseline compositions of the cohorts\u0026rsquo; effusions are consistent with previous descriptive characterizations. Across patients there was a pronounced interindividual heterogeneity, with immune compartments variably enriched for macrophages, neutrophils, CD4⁺ and CD8⁺ T cells, NK cells, and B cells, in line with findings from single cell sequencing and multiparametric studies showing diverse immune subsets and patient-specific effusion immune profiles (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). Although we did not specifically characterize exhaustion, regulatory or myeloid polarization states at baseline, the effusion-derived immune cells in PATEC did not exhibit spontaneous inherent cytotoxicity. This observation is concordant with mechanistic studies indicating the effusion TIME constitutes a maladaptive but potentially reversible effector compartment, in which CD8⁺ T cells and NK cells are chronically constrained by local factors yet can regain cytotoxic function upon appropriate stimulation (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan additionalcitationids=\"CR44\" citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWithin the PATEC system, combinatorial screening showed RT-IO regimens incorporating innate agonists were consistently among the most effective at inducing tumor cell death, with STING agonist containing combinations showing a modest but incrementally more consistent advantage over those including TLR7/8 agonists. These \u003cem\u003eex vivo\u003c/em\u003e findings are consistent with preclinical models demonstrating that ionizing radiation and STING agonists can be highly efficacious in combination (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan additionalcitationids=\"CR47 CR48\" citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). In contrast to these \u003cem\u003ein vivo\u003c/em\u003e studies, the PATEC assay assesses only how RT and innate agonists modulate local effector activity and does not capture dendritic cell trafficking, priming in lymph node or abscopal responses, so the observed efficacy reflects direct tumor killing within the TIME rather than long-term systemic control (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e, \u003cspan additionalcitationids=\"CR49\" citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). In this effector phase, Bliss independence analyses indicated that only a subset of PATECs exhibit a STING-RT synergistic effect, suggesting that patient-specific features of the TIME determine how effectively the combined actions of STING stimulation and irradiation are translated into cytotoxic effector activity. Such heterogeneity is consistent with preclinical data demonstrating that tumor intrinsic cGAS-STING pathway integrity, the mode of STING activation and its expression level critically modulate responsiveness to RT and to STING agonists (\u003cspan additionalcitationids=\"CR52\" citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e). In this context, the modest and heterogeneous responses observed in early phase trials of intratumoral STING agonists, applied primarily as monotherapy or with PD-1 blockade, are plausibly attributable to variations in STING pathway and to the absence of rationally selected combinations and patient selection strategies (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan additionalcitationids=\"CR55\" citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). These data indicate that effector phase platforms such as PATEC could help functionally identify effective innate immunotherapy regimens and nominate patients for further \u003cem\u003ein vivo\u003c/em\u003e evaluation.\u003c/p\u003e \u003cp\u003eComparing PATEC to matched tumor monocultures and transwell separated co-cultures, enabled us to distinguish tumor intrinsic from immune-mediated effects of STING based RT-IO. RT alone produced similar levels of cytotoxicity in primary tumor monocultures and in PATEC, retaining its efficacy when tumor and immune compartments were separated, suggesting that RT predominantly elicits DNA-damage-driven tumor-intrinsic cytotoxicity in this system (\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). In contrast, the addition of the STING agonist increased tumor cell death only in the presence of immune cells. This benefit was almost completely lost under transwell separation, implicating that STING dependent cytotoxicity in PATEC is largely indirect and strongly dependent on close immune tumor apposition. This is consistent with canonical contact-dependent effector mechanisms, including formation of a cytotoxic immunological synapse, polarized/directed degranulation of lytic granules, and death receptor engagement or release of short-range soluble cytotoxic mediators by cytotoxic lymphocytes and activated myeloid cells (\u003cspan additionalcitationids=\"CR59 CR60\" citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). Untreated PATEC co‑cultures did not exhibit spontaneous tumor cell killing, even when tumor and immune cells were in close contact. Similarly human data shows that effusion‑resident NK and CD8⁺ T cells are chronically constrained by the effusion milieu yet can reacquire proliferative and cytotoxic function when removed from suppressive fluid or exposed to appropriate cytokine or antigenic stimulation (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e). These converging lines of evidence support a model in which malignant effusions are thought to constitute an immunosuppressive effector niche that remains functionally reversible. In this context, these observations point to an interplay in which RT induces tumor-intrinsic DNA-damage and immunogenic antigen release, whereas STING agonism primarily boosts effusion resident cytotoxic lymphocyte and macrophage function.\u003c/p\u003e \u003cp\u003eIn PATEC, STING agonism and RT modulated T cell activation, checkpoint expression and secreted cytokine profile. After 24h, the STING agonist alone was sufficient to drive robust T cell degranulation. The combination of RT and STING agonist most strongly increased CD69 expression and induced a pro-inflammatory, IFN-I cytokine and chemokine profile marked by higher TNF-α, IL-6, CXCL10/IP-10, IFN-α, CCL3 and CCL4. This pattern is concordant with preclinical STING-RT studies in which IFN-I, TNF-α and CXCL9/10-rich milieus are induced that support effector activation, as well as with human co-culture studies that pharmacological STING activation can directly augment degranulation and cytotoxic function of T cells and NK cells (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e, \u003cspan additionalcitationids=\"CR65 CR66 CR67\" citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e). Consistent with sustained antigenic stimulation, STING agonist and RT induced the most pronounced significant upregulation of PD-1, TIGIT, LAG-3 and TIM-3 on CD8⁺ T cells, which was associated with cytotoxicity. Although high expression of PD-1, LAG-3, TIGIT and TIM-3 is often interpreted as an exhaustion signature, experimental and clinical data indicate that these receptors are induced sequentially along differentiation and chronic stimulation trajectories. They frequently mark tumor-experienced T cell populations whose functional state ranges from active to terminally exhausted depending on context and timing (\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e, \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e). In line with data showing that CD8⁺ T cells expressing multiple inhibitory receptors can remain polyfunctional mediating anti-tumor activity in patients, as well as humanized models and that gene signatures enriched for PD-1, LAG-3, TIM-3 and TIGIT expression are associated with clinical benefit from ICI, our data support a similar interpretation in PATEC (\u003cspan additionalcitationids=\"CR72 CR73\" citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e). The positive association between CES and cytotoxicity in PATEC supports interpreting CES high CD8⁺ T cell states in this early window as engaged effector populations under inducible inhibitory feedback rather than terminally dysfunctional cells.\u003c/p\u003e \u003cp\u003eThe addition of PD-1, PD-L1, CTLA-4 or TIGIT blockade to STING or TLR7/8 based RT-IO in the quadruple therapy experiment did not produce a significant, consistent further increase in tumor cell death across PATECs. Exploratory analyses suggested that effusions with stronger induction of the corresponding checkpoint on CD8⁺ T cells under STING\u0026thinsp;+\u0026thinsp;RT tended to derive modest incremental killing from blockade of that pathway, particularly for TIGIT and PD-L1. These findings are hypothesis-generating but accord with reports that CD8⁺ T cells co-expressing multiple inhibitory receptors can remain functionally competent and enriched for checkpoint-responsive tumor reactivity. They are also consistent \u003cem\u003eex vivo\u003c/em\u003e \u0026ldquo;tumor avatar\u0026rdquo; studies in which PD-1/CTLA-4/TIGIT blockade predominantly modulates progenitor exhausted compartments, clonal composition and cytokine programs over several days rather than yielding significant gains in short-term cytotoxicity (\u003cspan additionalcitationids=\"CR72 CR73 CR74 CR75 CR76 CR77\" citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eEven though the PATEC model shows the potential for a screening platform application there are several limitations to consider. The successful generation of primary tumor cultures for PATECs was modest at 20% and was derived from pan-cancer, as well as diverse pretreatments. The inherent fact, that there is a finite amount of biological material obtainable from malignant effusions, restricts the complexity and number of follow up experiments conductible. The model\u0026rsquo;s fluid interface is a double-edged sword; while providing preferable culture conditions it lacks stromal architecture, vasculature, antigen priming environment and pharmacokinetics. Processes dependent on these, such as recruitment of leukocytes, clonal replacement, stromal immunosuppression or dose-limiting toxicities are not recapitulated. The functional behavior of the culture might also not reflect the \u003cem\u003ein vivo\u003c/em\u003e response of the cells. Further work is required to establish the representability of culture regarding clinical response.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003ePATEC potentially provides a patient-derived \u003cem\u003eex vivo\u003c/em\u003e platform to functionally test new immunotherapeutic and RT-IO combinations in an autologous, immunocompetent TIME. In this proof-of-concept study combinations of RT with a STING agonist were the most tumoricidal regimens, with Bliss-defined synergy observed in a subset. STING-dependent cytotoxicity was largely immune-mediated and strongly dependent on direct immune tumor contact. STING-based RT-IO showed rapid T cell activation, an IFN-I-driven proinflammatory cytokine milieu and upregulated checkpoints on CD8⁺ T cells, which associated with tumor cell killing. These findings position malignant effusion-derived PATECs as an effector phase platform to rationally develop RT-IO regimens, dissect patient-specific synergy and generate functional hypotheses for immunotherapy-related biomarkers. Prospective integration of PATEC into precision oncology and early phase trials could help link patient-specific \u003cem\u003eex vivo\u003c/em\u003e responses to clinical outcome and support more rational selection of immunotherapeutic combinations for patients with advanced solid tumors.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRT-IO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eradiotherapy-immunotherapy\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eradiotherapy\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eimmunotherapy\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eICI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eimmune checkpoint inhibition/inhibitor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePATEC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epatient‑derived autologous tumor-immune effusion co‑culture\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePTCC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eprimary tumor cell culture\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eME\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emalignant effusion\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTIME\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etumor-immune microenvironment\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTME\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etumor microenvironment\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTuC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etumor cell monoculture\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTuCIC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etumor cell immune cell co‑culture (PATEC)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSTING\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003estimulator of interferon genes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSTING‑Ago\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eSTING agonist\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTLR7/8\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eToll‑like receptor 7/8\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCTLA‑4\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecytotoxic T‑lymphocyte associated protein 4\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePD‑1\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eprogrammed cell death protein 1\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePD‑L1\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eprogrammed death‑ligand 1\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLAG‑3\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003elymphocyte activation gene 3\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTIGIT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eT cell immunoreceptor with Ig and ITIM domains\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIFN‑I\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etype I interferons\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIFN‑α\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterferon alpha\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIFN‑β\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterferon beta\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIFN‑γ\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterferon gamma\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTNF‑α\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etumor necrosis factor alpha\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIL\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterleukin\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIL‑1ra\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterleukin‑1 receptor antagonist\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIL‑6\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterleukin 6\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIL‑8\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterleukin 8\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIL‑10\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterleukin 10\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIP‑10\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterferon‑gamma-induced protein 10 (CXCL10)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMCP‑1\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emonocyte chemoattractant protein 1 (CCL2)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMIP‑1α\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emacrophage inflammatory protein 1 alpha (CCL3)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMIP‑1β\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emacrophage inflammatory protein 1 beta (CCL4)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRANTES\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eregulated on activation,normal T cell expressed and secreted (CCL5)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePBMC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eperipheral blood mononuclear cell\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTIL\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etumor infiltrating lymphocyte\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eNK cell\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003enatural killer cell\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eNKT cell\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003enatural killer T cell\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIHC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eimmunohistochemistry\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFCM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eflow cytometry\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFBS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003efetal bovine serum\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePBS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ephosphate buffered saline\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003efold change\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCES\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003echeckpoint co‑expression score\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLMM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003elinear mixed effects model\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eANOVA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eanalysis of variance\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIQR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterquartile range\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003econfidence interval\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eOR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eodds ratio\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFDR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003efalse discovery rate\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDAB\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003e3,3'-diaminobenzidine\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eOD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eoptical density\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGy\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egray\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDMEM/F12\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDulbecco\u0026rsquo;s Modified Eagle Medium/Ham\u0026rsquo;s F‑12 nutrient mixture\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study involved human participants and received approval from the Institutional Review Board of the Medical University of Vienna (No. 2042/2019). The study was conducted in strict accordance with Good Scientific Practice (GSP) guidelines of the Medical University of Vienna and the most recent Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe work was supported by internal funds of the Medical University of Vienna.\u0026nbsp;There were no funding organizations that were involved in the study design, data acquisition, analysis, interpretation, manuscript preparation, or the decision to submit this work for publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors’ contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eR.Z. conceived and designed the study, acquired the data, performed the analysis, interpreted the results, and drafted the manuscript. D.A., N.S., M.F., B.M., M.A.R.H. \u0026nbsp;and A.T. acquired and analyzed data. M.B. and J.L. contributed to study conception and data interpretation. All authors critically reviewed the manuscript, as well as read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Brigitte Wolf and Carolina Klicka for their day-to-day support in the laboratory and for coordinating the transfer of malignant effusion specimens from the hospital to the laboratory. We also thank Büsra Ehrnhofer for technical assistance with selected experiments. In addition, we are grateful to the members of our research group for constructive discussions and continued support throughout the project.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHaslam A, Prasad V. Estimation of the percentage of US patients with cancer who are eligible for and respond to checkpoint inhibitor immunotherapy drugs. JAMA Netw Open. 2019;2(5):e192535.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLarkin J, Chiarion-Sileni V, Gonzalez R, Grob JJ, Cowey CL, Lao CD, et al. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N Engl J Med. 2015;373(1):23\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTawbi HA, Schadendorf D, Lipson EJ, Ascierto PA, Matamala L, Castillo Guti\u0026eacute;rrez E, et al. Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma. N Engl J Med. 2022;386(1):24\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeric-Bernstam F, Sweis RF, Kasper S, Hamid O, Bhatia S, Dummer R, et al. Combination of the STING agonist MIW815 (ADU-S100) and PD-1 inhibitor spartalizumab in advanced/metastatic solid tumors or lymphomas: an open-label, multicenter, phase Ib study. Clin Cancer Res. 2023;29(1):110\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSiu L, Brody J, Gupta S, Marabelle A, Jimeno A, Munster P, et al. Safety and clinical activity of intratumoral MEDI9197 alone and in combination with durvalumab and/or palliative radiation therapy in patients with advanced solid tumors. J Immunother Cancer. 2020;8(2):e001095.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCho BC, Abreu DR, Hussein M, Cobo M, Patel AJ, Secen N, et al. Tiragolumab plus atezolizumab versus placebo plus atezolizumab as a first-line treatment for PD-L1-selected non-small-cell lung cancer (CITYSCAPE): primary and follow-up analyses of a randomised, double-blind, phase 2 study. Lancet Oncol. 2022;23(6):781\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Z, Liu X, Chen D, Yu J. Radiotherapy combined with immunotherapy: the dawn of cancer treatment. Signal Transduct Target Ther. 2022;7(1):258.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaird JR, Friedman D, Cottam B, Dubensky TW Jr, Kanne DB, Bambina S, et al. Radiotherapy combined with novel STING-targeting oligonucleotides results in regression of established tumors. Cancer Res. 2016;76(1):50\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStary V, Wolf B, Unterleuthner D, List J, Talic M, Laengle J, et al. Short-course radiotherapy promotes pro-inflammatory macrophages via extracellular vesicles in human rectal cancer. J Immunother Cancer. 2020;8(2):e000667.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpigel DR, Faivre-Finn C, Gray JE, Vicente D, Planchard D, Paz-Ares L, et al. Five-year survival outcomes from the PACIFIC trial: durvalumab after chemoradiotherapy in stage III non-small-cell lung cancer. J Clin Oncol. 2022;40(12):1301\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKelly RJ, Ajani JA, Kuzdzal J, Zander T, van Cutsem E, Piessen G, et al. Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer. N Engl J Med. 2021;384(13):1191\u0026ndash;203.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetricevic B, Kabiljo J, Zirnbauer R, Walczak H, Laengle J, Bergmann M. Neoadjuvant immunotherapy in gastrointestinal cancers - the new standard of care? Semin Cancer Biol. 2022;86(Pt 2):834\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZirnbauer R, Hametner S, Bergler-Klein J, Kuehrer I, Kulu A, Ammon D et al. Incidence and immunopathology of myositis in rectal cancer patients treated with neoadjuvant immune checkpoint inhibitors and chemoradiotherapy: findings from the CHINOREC trial. MedComm (2020). 2025;6(7):e70275.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLaengle J, Kuehrer I, Kulu A, Kabiljo J, Ammon D, Zirnbauer R, et al. Dual immune checkpoint inhibition plus neoadjuvant chemoradiotherapy in rectal cancer: a randomized clinical trial. JAMA Netw Open. 2025;8(8):e2527769.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTan AC, Bagley SJ, Wen PY, Lim M, Platten M, Colman H, et al. Systematic review of combinations of targeted or immunotherapy in advanced solid tumors. J Immunother Cancer. 2021;9(7):e002459.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang J, Kang H, Lyu L, Xiong W, Hu Y. A target map of clinical combination therapies in oncology: an analysis of clinicaltrials.gov. Discov Oncol. 2023;14(1):151.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoscilli G, De Vitis C, Ferrara FF, Noto A, Cherubini E, Ricci A, et al. Human lung adenocarcinoma cell cultures derived from malignant pleural effusions as model system to predict patients chemosensitivity. J Transl Med. 2016;14:61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim SY, Lee JY, Kim DH, Joo HS, Yun MR, Jung D, et al. Patient-derived cells to guide targeted therapy for advanced lung adenocarcinoma. Sci Rep. 2019;9(1):19909.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu CG, Chiovaro F, Curioni-Fontecedro A, Casanova R, Soltermann A. In vitro cell culture of patient derived malignant pleural and peritoneal effusions for personalised drug screening. J Transl Med. 2020;18:163.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWegmann R, Bankel L, Festl Y, Lau K, Lee S, Arnold F, et al. Molecular and functional landscape of malignant serous effusions for precision oncology. Nat Commun. 2024;15:8544.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang ZY, Shao MM, Zhang JC, Yi FS, Du J, Zhou Q, et al. Single-cell analysis of diverse immune phenotypes in malignant pleural effusion. Nat Commun. 2021;12:6690.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChee J, Watson MW, Chopra A, Nguyen B, Cook AM, Creaney J, et al. Tumour associated lymphocytes in the pleural effusions of patients with mesothelioma express high levels of inhibitory receptors. BMC Res Notes. 2018;11:864.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu MF, Lin CA, Yuan TH, Yeh HY, Su SF, Guo CL, et al. The M1/M2 spectrum and plasticity of malignant pleural effusion-macrophage in advanced lung cancer. Cancer Immunol Immunother. 2021;70(5):1435\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrincipe N, Kidman J, Lake RA, Lesterhuis WJ, Nowak AK, McDonnell AM, et al. Malignant pleural effusions-a window into local anti-tumor T cell immunity? Front Oncol. 2021;11:672747.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDhupar R, Okusanya OT, Eisenberg SH, Monaco SE, Ruffin AT, Liu D, et al. Characteristics of malignant pleural effusion resident CD8\u0026thinsp;+\u0026thinsp;T cells from a heterogeneous collection of tumors. Int J Mol Sci. 2020;21(17):6178.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLv LL, Wang HB, Zhang YX, Zhai JW, Shen Y, Qu QX, et al. CD39 identifies a specific CD8\u0026thinsp;+\u0026thinsp;T cell population in lung adenocarcinoma-related metastatic pleural effusion. BMC Immunol. 2023;24:53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGe S, Zhao Y, Liang J, He Z, Li K, Zhang G, et al. Immune modulation in malignant pleural effusion: from microenvironment to therapeutic implications. Cancer Cell Int. 2024;24:105.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuki K, Cheng N, Nakano M, Kuo CJ. Organoid models of tumor immunology. Trends Immunol. 2020;41(8):652\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang J, Tao X, Zhu J, Dai Z, Du Y, Xie Y, et al. Tumor organoid-immune co-culture models: exploring a new perspective of tumor immunity. Cell Death Discov. 2025;11:195.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAo Z, Cai H, Wu Z, Hu L, Li X, Kaurich C, et al. Evaluation of cancer immunotherapy using mini-tumor chips. Theranostics. 2022;12(8):3628\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAyuso JM, Rehman S, Virumbrales-Munoz M, McMinn PH, Geiger P, Fitzgerald C, et al. Microfluidic tumor-on-a-chip model to evaluate the role of tumor environmental stress on NK cell exhaustion. Sci Adv. 2021;7(8):eabc2331.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChernyavska M, Masoudnia M, Valerius T, Verdurmen WPR. Organ-on-a-chip models for development of cancer immunotherapies. Cancer Immunol Immunother. 2023;72(12):3971\u0026ndash;83.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu H, Noguera-Ortega E, Dong X, Lee WD, Chang J, Aydin SA, et al. A tumor-on-a-chip for in vitro study of CAR-T cell immunotherapy in solid tumors. Nat Biotechnol. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41587-025-02845-z\u003c/span\u003e\u003cspan address=\"10.1038/s41587-025-02845-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaulana TI, Teufel C, Cipriano M, Roosz J, Lazarevski L, van den Hil FE, et al. Breast cancer-on-chip for patient-specific efficacy and safety testing of CAR-T cells. Cell Stem Cell. 2024;31(7):989\u0026ndash;e10029.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJagatia R, Doornebal EJ, Rastovic U, Harris N, Feyide M, Lyons AM, et al. Patient-derived precision cut tissue slices from primary liver cancer as a potential platform for preclinical drug testing. EBioMedicine. 2023;97:104826.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSullivan KM, Jiang X, Guha P, Lausted C, Carter JA, Hsu C, et al. Blockade of interleukin 10 potentiates antitumour immune function in human colorectal cancer liver metastases. Gut. 2023;72(2):325\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeo HY, Kim SC, Roh WL, Shin YK, Kim S, Kim DW, et al. Culture and multiomic analysis of lung cancer patient-derived pleural effusions revealed distinct druggable molecular types. Sci Rep. 2022;12:6345.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoi W, Kim YH, Woo SM, Yu Y, Lee MR, Lee WJ, et al. Establishment of patient-derived organoids using ascitic or pleural fluid from cancer patients. Cancer Res Treat. 2023;55(4):1077\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTano ZE, Kiesgen S, Chintala NK, Dozier J, Quach HT, Messinger J, et al. Ex vivo pleural effusion cultures to study chimeric antigen receptor T cell cytotoxicity in an immunocompetent environment. Cell Rep Methods. 2023;3(11):100622.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e\u0026Ouml;nder CE, Moustafa-Oglou M, Schr\u0026ouml;der SM, Hartkopf AD, Koch A, Seitz CM. Precision immunotherapy utilizing adapter CAR-T cells (AdCAR-T) in metastatic breast cancer leads to target specific lysis. Cancers (Basel). 2024;16(1):168.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang XZ, Pang MJ, Li JY, Chen HY, Sun JX, Song YX, et al. Single-cell sequencing of ascites fluid illustrates heterogeneity and therapy-induced evolution during gastric cancer peritoneal metastasis. Nat Commun. 2023;14:822.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng X, Wang X, Cheng X, Liu Z, Yin Y, Li X, et al. Single-cell analyses implicate ascites in remodeling the ecosystems of primary and metastatic tumors in ovarian cancer. Nat Cancer. 2023;4(8):1138\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonnenberg VS, Luketich JD, Sultan I, Lister J, Bartlett DL, Ghosh S, et al. A maladaptive pleural environment suppresses preexisting anti-tumor activity of pleural infiltrating T cells. Front Immunol. 2023;14:1157697.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlattery K, Yao CH, Mylod E, Scanlan J, Scott B, Crowley JP, et al. Uptake of lipids from ascites drives NK cell metabolic dysfunction in ovarian cancer. Sci Immunol. 2025;10(107):eadr4795.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVacca P, Martini S, Mingari MC, Moretta L. NK cells from malignant pleural effusions are potent antitumor effectors: a clue for adoptive immunotherapy. Oncoimmunology. 2013;2(4):e23638.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeng L, Liang H, Xu M, Yang X, Burnette B, Arina A, et al. STING-dependent cytosolic DNA sensing promotes radiation-induced type I interferon-dependent antitumor immunity in immunogenic tumors. Immunity. 2014;41(5):843\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Crowe WN, Wang L, Lu Y, Petty WJ, Habib AA, et al. An inhalable nanoparticulate STING agonist synergizes with radiotherapy to confer long-term control of lung metastases. Nat Commun. 2019;10:5108.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang P, Rashidi A, Zhao J, Silvers C, Wang H, Castro B, et al. STING agonist-loaded, CD47/PD-L1-targeting nanoparticles potentiate antitumor immunity and radiotherapy for glioblastoma. Nat Commun. 2023;14:1610.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeng Q, Liu M, Wang Z, Zhou R, Ai K. Enhancing radiotherapy-induced anti-tumor immunity via nanoparticle-mediated STING agonist synergy. Mol Cancer. 2025;24:176.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo M, Liu Z, Zhang X, Han C, Samandi LZ, Dong C, et al. Synergistic STING activation by PC7A nanovaccine and ionizing radiation improves cancer immunotherapy. J Control Release. 2019;300:154\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJagodinsky JC, Jin WJ, Bates AM, Hernandez R, Grudzinski JJ, Marsh IR, et al. Temporal analysis of type 1 interferon activation in tumor cells following external beam radiotherapy or targeted radionuclide therapy. Theranostics. 2021;11(13):6120\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDrzyzga A, Czapla J, Matuszczak S, Łasut-Szyszka B, Cichoń T, Pilny E, et al. Differential response to local stimulator of interferon genes agonist administration in tumors with various stimulator of interferon genes statuses. Cancers (Basel). 2025;17(2):175.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHayman TJ, Baro M, MacNeil T, Phoomak C, Aung TN, Cui W, et al. STING enhances cell death through regulation of reactive oxygen species and DNA damage. Nat Commun. 2021;12:2327.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHines JB, Kacew AJ, Sweis RF. The development of STING agonists and emerging results as a cancer immunotherapy. Curr Oncol Rep. 2023;25(3):189\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuke JJ, Piha-Paul SA, Medina T, Verschraegen CF, Varterasian M, Brennan AM, et al. Phase I study of SYNB1891, an engineered Escherichia coli Nissle strain expressing STING agonist, with and without atezolizumab in advanced malignancies. Clin Cancer Res. 2023;29(13):2435\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeric-Bernstam F, Sweis RF, Hodi FS, Messersmith WA, Andtbacka RHI, Ingham M, et al. Phase I dose-escalation trial of MIW815 (ADU-S100), an intratumoral STING agonist, in patients with advanced/metastatic solid tumors or lymphomas. Clin Cancer Res. 2022;28(4):677\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSia J, Szmyd R, Hau E, Gee HE. Molecular mechanisms of radiation-induced cancer cell death: a primer. Front Cell Dev Biol. 2020;8:41.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeng A, Fan R, Hai Y, Zhuang J, Zhang B, Lu X, et al. A STING agonist prodrug reprograms tumor-associated macrophage to boost colorectal cancer immunotherapy. Theranostics. 2025;15(1):277\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNicolai CJ, Wolf N, Chang IC, Kirn G, Marcus A, Ndubaku CO, et al. NK cells mediate clearance of CD8\u0026thinsp;+\u0026thinsp;T cell-resistant tumors in response to STING agonist. Sci Immunol. 2020;5(45):eaaz2738.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerger G, Knelson EH, Jimenez-Macias JL, Nowicki MO, Han S, Panagioti E, et al. STING activation promotes robust immune response and NK cell-mediated tumor regression in glioblastoma models. Proc Natl Acad Sci U S A. 2022;119(28):e2111003119.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOu L, Zhang A, Cheng Y, Chen Y. The cGAS-STING pathway: a promising immunotherapy target. Front Immunol. 2021;12:795048.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVacca P, Martini S, Garelli V, Passalacqua G, Moretta L, Mingari MC. NK cells from malignant pleural effusions are not anergic but produce cytokines and display strong antitumor activity on short-term IL-2 activation. Eur J Immunol. 2013;43(2):550\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu CY, Zhang YH, Wang T, Chen L, Gong ZH, Wan YS, et al. Interleukin-2 reverses CD8(+) T cell exhaustion in clinical malignant pleural effusion of lung cancer. Clin Exp Immunol. 2016;186(1):106\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Q, Yu Y, Zhuang J, Liu R, Sun C. Demystifying the cGAS-STING pathway: precision regulation in the tumor immune microenvironment. Mol Cancer. 2025;24:178.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLinder A, Nixdorf D, Kuhl N, Piseddu I, Xu TT, Holtermann AV, et al. STING activation improves T-cell-engaging immunotherapy for acute myeloid leukemia. Blood. 2025;145(19):2149\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKnelson EH, Ivanova EV, Tarannum M, Campisi M, Lizotte PH, Booker MA, et al. Activation of tumor-cell STING primes NK-cell therapy. Cancer Immunol Res. 2022;10(8):947\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDa Y, Liu Y, Hu Y, Liu W, Ma J, Lu N, et al. STING agonist cGAMP enhances anti-tumor activity of CAR-NK cells against pancreatic cancer. Oncoimmunology. 2022;11(1):2054105.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Wang L, Song Q, Ali M, Crowe WN, Kucera GL, et al. Intrapleural nano-immunotherapy promotes innate and adaptive immune responses to enhance anti-PD-L1 therapy for malignant pleural effusion. Nat Nanotechnol. 2022;17(2):206\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCai L, Li Y, Tan J, Xu L, Li Y. Targeting LAG-3, TIM-3, and TIGIT for cancer immunotherapy. J Hematol Oncol. 2023;16:101.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoller N, Anderson AC, Kuchroo VK. LAG-3, TIM-3, and TIGIT: distinct functions in immune regulation. Immunity. 2024;57(2):206\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEgelston CA, Avalos C, Tu TY, Simons DL, Jimenez G, Jung JY, et al. Human breast tumor-infiltrating CD8\u0026thinsp;+\u0026thinsp;T cells retain polyfunctionality despite PD-1 expression. Nat Commun. 2018;9:4297.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMietz J, Kaulfuss M, Egli L, Opitz L, M\u0026uuml;nz C, Chijioke O. Human effector CD8\u0026thinsp;+\u0026thinsp;T cells with an activated and exhausted-like phenotype control tumour growth in vivo in a humanized tumour model. EBioMedicine. 2024;106:105240.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIb\u0026aacute;\u0026ntilde;ez-Molero S, Veldman J, Simon Nieto J, Traets JJH, George A, Hoefakker K, et al. Tumour-reactive heterotypic CD8 T cell clusters from clinical samples. Nature. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-025-09754-w\u003c/span\u003e\u003cspan address=\"10.1038/s41586-025-09754-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGu\u0026eacute;gan JP, Peyraud F, Dadone-Montaudie B, Teyssonneau D, Palmieri LJ, Clot E, et al. Analysis of PD1, LAG3, TIGIT, and TIM3 expression in human lung adenocarcinoma reveals a 25-gene signature predicting immunotherapy response. Cell Rep Med. 2024;5(12):101831.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJenkins RW, Aref AR, Lizotte PH, Ivanova E, Stinson S, Zhou CW, et al. Ex vivo profiling of PD-1 blockade using organotypic tumor spheroids. Cancer Discov. 2018;8(2):196\u0026ndash;215.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHulen TM, Friese C, Kristensen NP, Granh\u0026oslash;j JS, Borch TH, Peeters MJW, et al. Ex vivo modulation of intact tumor fragments with anti-PD-1 and anti-CTLA-4 influences the expansion and specificity of tumor-infiltrating lymphocytes. Front Immunol. 2023;14:1180997.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEguren-Santamar\u0026iacute;a I, Rodr\u0026iacute;guez I, Herrero-Martin C, Fern\u0026aacute;ndez de Pi\u0026eacute;rola E, Azpilikueta A, S\u0026aacute;nchez-Gregorio S, et al. Short-term cultured tumor fragments to study immunotherapy combinations based on CD137 (4-1BB) agonism. Oncoimmunology. 2024;13(1):2373519.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThommen DS. Tumour avatars to model patients' responses to immunotherapy. Nat Rev Cancer. 2022;22(12):660.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-experimental-and-clinical-cancer-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jecc","sideBox":"Learn more about [Journal of Experimental \u0026 Clinical Cancer Research](http://jeccr.biomedcentral.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/jecc/default.aspx","title":"Journal of Experimental \u0026 Clinical Cancer Research","twitterHandle":"@OncoBioMed","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Pleural Effusion, Malignant, Ascites, Neoplasms, Tumor Microenvironment, Coculture Techniques, Radiotherapy, Immunotherapy, Immune Checkpoint Inhibitors, Precision Medicine, PATEC","lastPublishedDoi":"10.21203/rs.3.rs-8516163/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8516163/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eCurrently, radiotherapy-immunotherapy (RT-IO) combinations provide limited and heterogeneous benefit in human solid cancers and are frequently selected empirically, partly because human models that preserve native, autologous tumor-immune interactions are lacking. We developed patient-derived autologous tumor-immune effusion co-cultures (PATEC) as an \u003cem\u003eex vivo\u003c/em\u003e platform to functionally evaluate RT-IO regimens within an immunocompetent tumor microenvironment.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eMalignant pleural and peritoneal effusions (n\u0026thinsp;=\u0026thinsp;29) from patients with metastatic solid cancers were processed for biobanking and primary tumor culture. Expandable tumor cultures were established in six effusions and recombined with matched autologous immune cells to generate PATEC. PATEC were treated with radiotherapy (RT), innate immune agonists (STING, TLR7/8) and immune checkpoint inhibitors (CTLA-4, PD-1, PD-L1, TIGIT) in combinatorial regimens. Tumor cell death, T-cell activation, cytokine secretion and CD8⁺ T-cell checkpoint expression was assessed using multiparametric flow cytometry and multiplex immunoassays. Contact dependence of cytotoxicity was evaluated by comparing tumor monocultures, direct co-cultures and transwell separated co-cultures.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eAcross conditions, regimens combining RT with a stimulator of interferon genes (STING) agonist were the most tumoricidal in PATEC, with marked interpatient variability and Bliss defined synergy in a subset of effusions (3/6). STING agonist-mediated cytotoxicity required immune cells and was attenuated by spatial separation of tumor and immune compartments, whereas RT alone produced similar cytotoxicity in monocultures and co-cultures, indicating a predominantly tumor-intrinsic effect. STING based RT-IO induced early T-cell activation and a type I interferon-rich cytokine milieu, followed by increased expression of multiple inhibitory checkpoints on CD8⁺ T cells. A composite CD8⁺ checkpoint co-expression score correlated with both overall and contact-dependent tumor cell death.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003ePATEC enables functional dissection of RT-IO combinations in a native effusion-derived tumor-immune microenvironment and shows that the additional tumor cell killing conferred by STING-based RT-IO depends on immune cells and direct tumor-immune contact and varies between patient samples. These findings support the use of PATEC as a functional \u003cem\u003eex vivo\u003c/em\u003e system for testing therapeutic combinations in a patient-specific setting.\u003c/p\u003e","manuscriptTitle":"Autologous tumor-immune effusion co-cultures enable ex vivo functional profiling of radiotherapy-immunotherapy combinations","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-12 06:09:10","doi":"10.21203/rs.3.rs-8516163/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-02-13T20:57:20+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-13T20:53:49+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-11T11:06:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"303119986729753016005040615445457529200","date":"2026-01-26T09:08:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"268472276079388764190804076773634077322","date":"2026-01-09T14:26:13+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-07T10:05:12+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-06T13:42:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-06T13:40:26+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Experimental \u0026 Clinical Cancer Research","date":"2026-01-05T02:08:22+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-experimental-and-clinical-cancer-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jecc","sideBox":"Learn more about [Journal of Experimental \u0026 Clinical Cancer Research](http://jeccr.biomedcentral.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/jecc/default.aspx","title":"Journal of Experimental \u0026 Clinical Cancer Research","twitterHandle":"@OncoBioMed","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"92fce165-0f71-4f2d-a7a5-0d48075dc93a","owner":[],"postedDate":"January 12th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-04-20T16:08:02+00:00","versionOfRecord":{"articleIdentity":"rs-8516163","link":"https://doi.org/10.1186/s13046-026-03707-5","journal":{"identity":"journal-of-experimental-and-clinical-cancer-research","isVorOnly":false,"title":"Journal of Experimental \u0026 Clinical Cancer Research"},"publishedOn":"2026-04-14 15:57:45","publishedOnDateReadable":"April 14th, 2026"},"versionCreatedAt":"2026-01-12 06:09:10","video":"","vorDoi":"10.1186/s13046-026-03707-5","vorDoiUrl":"https://doi.org/10.1186/s13046-026-03707-5","workflowStages":[]},"version":"v1","identity":"rs-8516163","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8516163","identity":"rs-8516163","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}