Pressurized intraperitoneal aerosol chemotherapy enhances cisplatin efficacy in colorectal cancer organoids.

Introduction
Peritoneal carcinomatosis (PC) may manifest either as primary peritoneal cancer or as a complication stemming from various types of malignancies such as ovarian, fallopian tube, colon, appendiceal, cholangiocarcinoma, and gastric cancers1. Peritoneal carcinomatosis in colorectal cancer is associated with the worst survival outcomes among metastatic sites, with a median overall survival of 16.3 months2. Although intraperitoneal chemotherapy (IPC) has proven to provide survival advantages, ineffective penetration of therapeutic agents into solid tumors remains the major obstacle. This is largely due to elevated intratumoral pressures that block the flow of therapeutic compounds into the tumor masses as compared to adjacent non-cancerous tissues3.
Pressurized Intraperitoneal Aerosol Chemotherapy (PIPAC) is a promising technique for treating peritoneal carcinomatosis, delivering chemotherapeutic agents directly to the peritoneal cavity in aerosolized form4. The mechanism behind PIPAC involves dispersing the medication in aerosol form under elevated intra-abdominal pressure, counteracting the high intratumoral pressure and enhancing drug delivery into tumor tissues. This technique has shown potential, with drug penetration depths reaching 500 to 600 μm and tissue concentrations up to 1.70 µmol/g, potentially surpassing those achieved with conventional liquid IPC5.
Although preliminary clinical results are promising, a subset of patients either does not respond to PIPAC treatment or experiences early disease recurrence following initial tumor reduction6. Mitomycin C and Oxaliplatin has been the primary drug of choice in PIPAC therapy, with the cytotoxic effects of combining multiple antitumor drugs with PIPAC remaining largely unexplored and typically inferred from liquid IPC clinical data7. Furthermore, due to its specific technical requirements, existing PIPAC technology is not adaptable for small animal models of peritoneal carcinomatosis, and a suitable large animal model approximating human dimension has yet to be developed8,9.
In this study, we aimed to investigate whether PIPAC could enhance the cytotoxicity of standard chemotherapeutic agents and induce transcriptomic changes that reflect therapeutic synergy in colorectal cancer. Using patient-derived organoid models, we demonstrate that PIPAC markedly increases the effectiveness of Cisplatin, and to a lesser extent Oxaliplatin and Paclitaxel, in reducing cell viability. Furthermore, transcriptomic profiling reveals that this enhanced cytotoxicity is accompanied by alterations in key oncogenic and cell cycle-related pathways. These findings support the hypothesis that the PIPAC delivery method confers both pharmacologic and molecular advantages, offering mechanistic insight into its potential clinical benefit for treating peritoneal carcinomatosis.

Results

Pressurized intra peritoneal aerosol chemotherapy (PIPAC) enhanced cytotoxic effect of cisplatin in SNU-4398S1-TO
Cisplatin, Oxaliplatin, and Paclitaxel were screened in conjunction with the increased temperature and pressure to evaluate their varying efficacies under the PIPAC condition (Fig. 1A and B). Viability reduction observed in drug-treated PIPAC conditions was compared against the PIPAC-only (DPBS) control to isolate the effect of the chemotherapeutic agent from the baseline mechanical or stress-related impact of aerosol delivery (Fig. 1B). Analysis of individual wells (well 1 to well 4) revealed no substantial variation in drug efficacy across different wells for each specific condition (Supplementary Table 1 A and 1B). Under atm conditions, Cisplatin was the least effective drug among the three drugs as indicated by a cell viability rate of approximately 83%. However, under PIPAC conditions, the average cell viability when treated with Cisplatin was significantly decreased to nearly 49%, reflecting a marked increase in drug efficacy. Similarly, both Oxaliplatin and Paclitaxel exhibited decreased cell viability under PIPAC conditions compared to atm conditions, signifying enhanced effectiveness, albeit to a lesser extent (Fig. 1B). In the context of PIPAC conditions, Cisplatin notably surpassed both Oxaliplatin and Paclitaxel in reducing cell viability, demonstrating greater efficacy against cancer cells in this environment.

Pressurized intra peritoneal aerosol chemotherapy (PIPAC) improved cytotoxic effect of Cisplatin, oxaliplatin and paclitaxel in SNU-4398S4-TO
Molecular heterogeneity manifests various drug responses even within a single colon tumor mass. We assumed that the response to pressure of tumor cells varies in accordance with their molecular background as well. The SNU-4398S4-TO organoid line, originating from the same colorectal cancer (CRC) patient as SNU-4398S1-TO, was established from a distinct region of the same colon tumor mass. This patient was diagnosed with microsatellite instability-high cancer, which is associated with an increased risk of peritoneal seeding10. These organoids have been discussed in previous research11. Under PIPAC conditions, without the presence of drugs, SNU-4398S4-TO demonstrated an increased sensitivity, as evidenced by reduced cell viability (Fig. 2A and B, Supplementary Table 1B). When exposed to PIPAC conditions, all three drugs exhibited enhanced efficacy compared to atm conditions, as reflected by decreased cell viability. Notably, Cisplatin also showed the most substantial increase in efficacy under PIPAC conditions. While there are minor inconsistencies, Cisplatin under PIPAC conditions revealed a notable rise in variability, which may indicate a more diverse responses among different cell populations or under different experimental conditions. The findings suggest that the PIPAC condition amplifies the anti-cancer effectiveness of these drugs, particularly enhancing the efficacy of Cisplatin, as demonstrated by the significant reduction in cell viability.

Exposure to cytotoxic drugs under PIPAC condition dysregulated multiple biological pathways
We Subsequently analyzed the transcriptomic landscapes of colon organoids subjected to various drugs under Pressurized Intra Peritoneal Aerosol Chemotherapy (PIPAC) conditions. Principal component analysis (PCA) was conducted to evaluate transcriptomic shifts among treatment conditions. Rather than relying on formal differential gene expression (DEG) analysis, which was not feasible due to single-replicate sampling, we extracted representative genes with high loadings on PC1 and PC2 to identify those driving the observed separations. These genes were used for pathway enrichment analysis via single-sample GSEA (ssGSEA), highlighting key signaling pathways modulated by drug-PIPAC combinations. Specifically, PCA applied to the SNU-4398S1-TO line indicated that organoids treated with Paclitaxel exhibited unique mRNA expression profiles in contrast to those treated with other agents, as evidenced by principal component 1 (PC1), which explained 58.66% of the observed variance (Fig. 3A). Organoids treated with Paclitaxel under PIPAC conditions were closely aligned with those under atmospheric conditions, corroborating cell viability assays that suggested a negligible effect of PIPAC on Paclitaxel-treated samples (Fig. 3A). Principal component 2 (PC2), accounting for 15.61% of the total variance, effectively distinguished between samples treated with Cisplatin and Oxaliplatin, demonstrating the disparate effects of these drugs on colon organoids in PIPAC conditions through their separation on the PCA plot (Fig. 3A, Supplementary Table 2 A). Pathway analysis using genes consisting of PC2 revealed multiple biological pathways including amino acids metabolism as well as mRNA translation (Fig. 3B, Supplementary Table 2B).
Further investigation using the SNU-4398S4-TO line revealed transcriptomic profiles significantly different from those observed in the SNU-4398S1-TO line, with a predominant clustering influenced by the experimental conditions. Remarkably, SNU-4398S4-TO organoids were substantially affected by the PIPAC condition alone, even in the absence of pharmacological treatment, underscoring the heterogeneous response of colon tumor cells to enhanced pressure and temperature (Fig. 3C). Principle component 2 (PC2) better reflected the condition-induced expressional differences over principal component 1 (PC1). Pathway analysis, focusing on genes that fell within the top 17.4% of principal component 2 (PC2) unveiled several pathways altered by the PIPAC condition. Among these, pathways related to neutrophil degranulation and steroid metabolism were notably affected, highlighting the broad impact of PIPAC conditions on cellular processes (Fig. 3D, Supplementary Table 2D).
We further assessed the response of colon cells to each chemotherapeutic agent under Pressurized Intra Peritoneal Aerosol Chemotherapy (PIPAC) conditions. We mainly focused on expressional values of the SNU-4398S1-TO line since it better reflected the synergetic effect of drugs with PIPAC condition. To specifically focus on genes influenced by the synergistic effect of PIPAC combined with drug treatment, genes that changed more than twofold under PIPAC conditions compared to atmospheric (atm) conditions without any drug intervention were excluded from the following analysis (Supplementary Table 3 A).

Treatment with Cisplatin under PIPAC conditions, relative to atm conditions, led to the alteration of several genes involved in GPCR ligand binding pathways, such as CCK and CX3CR1, as detailed in Fig. 4A and the accompanying Supplementary Table 3B and 3 C. Treatment with Oxaliplatin similarly affected multiple genes within the GPCR ligand binding pathway, in addition to negatively regulating TCF-dependent signaling through WNT ligand antagonism, as presented in Fig. 4B and the Supplementary Table 3 C and 3D. In contrast, treatment with Paclitaxel under PIPAC conditions did not result in significant pathway aberrations (Supplementary Table 3 F).
Three genes, CHGB (Fig. 4C), MPL (Fig. 4D), and YWHAZP4 (Fig. 4E) were found to be uniformly dysregulated across treatments with all three drugs in the SNU-4398S1-TO line. Notably, MPL, which is associated with the cause of thrombocytosis12, showed a significant reduction in mRNA expression when exposed to the combined PIPAC condition and drug treatment. This suggests the potential role of MPL gene expression in the response of colon cancer cells to cytotoxic drugs under conditions of elevated temperature and pressure. Previous research has documented that MPL overexpression can lead to resistance against ruxolitinib in myeloproliferative neoplasms13.
Further analysis was conducted on the mRNA expression of SNU-4398S4-TO (Supplementary Table 4 A-E). Of the three drugs, only Oxaliplatin, in combination with PIPAC conditions, led to significant dysregulation in pathways, including biological oxidations and glucuronidation (Supplementary Fig. 1 A, Supplementary Table 4 C and 4D). Additionally, the mRNA expression of CNTF was significantly altered when exposed to the drugs in conjunction with PIPAC conditions (Supplementary Fig. 1B). Interestingly, Cisplatin and Paclitaxel treatment resulted in increased expression of CNTF, while Oxaliplatin treatment reduced CNTF expression.

Exposure to cytotoxic drugs under PIPAC condition dysregulated multiple Tumor-related pathways
We conducted single-sample gene set enrichment analysis (ssGSEA) using the hallmark database to focus solely on cancer-related pathways. The clustering of enrichment scores for SNU-4398S1-TO samples indicated a significant impact on mRNA expressions by the combination of Cisplatin and PIPAC conditions. Oxaliplatin and Paclitaxel exhibited stronger influences than conditional differences, as evidenced by the grouping of samples treated with identical drugs. Consistent with previous results from principal component analysis (PCA), Cisplatin-treated samples closely clustered with PBS control, suggesting minimal genetic aberrations. Nonetheless, Cisplatin treatment demonstrated a potent cytotoxic effect under PIPAC conditions (Fig. 1B). Several pathways, including epithelial-mesenchymal transition, KRAS signaling down, and coagulation, were specifically down-regulated only under PIPAC conditions (highlighted with black rectangles). Genes involved in these pathways may contribute to the synergistic cytotoxic effect of Cisplatin under PIPAC conditions. Conversely, organoids treated with Oxaliplatin or Paclitaxel exhibited minimal differences between ATM and PIPAC conditions (Fig. 5A and Supplementary Table 5 A).
Enrichment scores using SNU-4398S4-TO samples indicated that PIPAC conditions exerted a stronger effect than drugs alone. However, Oxaliplatin-treated samples were grouped together, suggesting minimal expressional alterations due to conditional differences. PIPAC conditions enriched cell-cycle-related pathways such as G2M_CHECKPOINT, E2F_TARGETS, and MITOTIC_SPINDLE, while canonical tumor pathways such as TGF_BETA_SIGNALING, TNFA_SIGNALING_VIA_NFKB, and KRAS_SIGNALING were down-regulated (Fig. 5B and Supplementary Table 5B). This suggests that the overall cytotoxic effect of PIPAC in SNU-4398S4-TO resulted from accelerated uptake of cytotoxic drugs via up-regulation of cell cycle-related pathways and down-regulation of canonical oncogenic pathways.

Discussion
Our research aimed to evaluate the efficacy of the Pressurized Intra Peritoneal Aerosol Chemotherapy (PIPAC) technique combined with three distinct cytotoxic agents (Cisplatin, Oxaliplatin, and Paclitaxel) using patient-derived colorectal cancer organoid models. The study highlighted the potential of PIPAC to enhance the cytotoxic effects of these anticancer drugs by facilitating deeper drug penetration and affecting cell division cycles. Notably, Cisplatin showed markedly improved cytotoxicity when applied under PIPAC conditions, suggesting that the physical delivery environment may potentiate its therapeutic effect. While these findings imply a beneficial interaction, further studies are needed to determine whether a true pharmacological synergy exists.
Although mitomycin C is frequently utilized in hyperthermic intraperitoneal chemotherapy (HIPEC), particularly for appendiceal or colorectal peritoneal metastases14, its inclusion in PIPAC protocols remains limited due to pharmacokinetic concerns and drug-specific properties. Mitomycin C has a relatively long tissue half-life and limited volatility15, characteristics that may hinder its optimal aerosolization and distribution within the peritoneal cavity under PIPAC conditions. Furthermore, emerging clinical data suggest that mitomycin C may pose a higher risk of cumulative toxicity, especially when combined with systemic chemotherapy16, potentially limiting its safety profile in repetitive intraperitoneal treatments. In this context, our study focused on agents with more established aerosol performance profiles, Cisplatin, Oxaliplatin, and Paclitaxel, to better evaluate their mechanistic and cytotoxic effects in a controlled PIPAC setting using patient-derived organoid models. This approach allowed for clearer interpretation of transcriptomic responses and cytotoxic synergy, without confounding effects from mitomycin C’s unique pharmacodynamics. Also, the drug panel was selected to reflect current and emerging PIPAC strategies relevant to colorectal cancer, where oxaliplatin and paclitaxel are more frequently employed than agents like doxorubicin, which are more typical in gastric or ovarian cancer settings17.
Cisplatin and oxaliplatin are both platinum-based chemotherapeutic agents that induce cytotoxicity through DNA crosslinking and subsequent inhibition of DNA replication and transcription. However, their mechanisms diverge in the nature of the adducts formed and the cellular repair responses they activate. Cisplatin tends to form intrastrand crosslinks more efficiently, resulting in potent DNA damage and triggering apoptosis through p53-dependent pathways18. Oxaliplatin, on the other hand, induces bulky DNA adducts that interact differently with mismatch repair pathways and immune signaling19. Paclitaxel operates via a distinctly different mechanism by stabilizing microtubules and preventing their depolymerization, thus disrupting mitotic spindle formation and inducing mitotic arrest. This leads to apoptosis through the activation of stress response pathways20. PIPAC, as a pressurized aerosol delivery system, enhances drug-tissue contact time and promotes deeper drug penetration into the peritoneal and subperitoneal compartments. For cisplatin and oxaliplatin, the increased tissue perfusion and enhanced permeability may amplify DNA adduct formation in otherwise poorly accessible tumor regions. In the case of paclitaxel, which has a high molecular weight and limited diffusion, the aerosolized form may help circumvent pharmacokinetic limitations by maximizing local drug concentration. Collectively, these effects may potentiate the therapeutic efficacy of each drug, albeit through distinct molecular routes. Acknowledging these mechanistic differences is critical in interpreting the observed cytotoxic responses and in guiding future combinatorial strategies.
Oxaliplatin remains the predominant PIPAC agent for colorectal peritoneal metastases, demonstrating significant clinical efficacy with Peritoneal Regression Grading Scores indicating response rates of 42–79% and median overall survival ranging from 8 to 20.5 months21. However, recent discourse surrounding oxaliplatin’s efficacy in Hyperthermic intraperitoneal chemotherapy has prompted increased investigation into alternative agents22. Historically, cisplatin combined with doxorubicin served as a pioneering PIPAC regimen, exhibiting a favorable safety profile with notably reduced abdominal pain compared to oxaliplatin21. While paclitaxel’s application in colorectal peritoneal metastases remains nascent, emerging research explores both monotherapy and combination approaches23 though its primary clinical validation continues to be in ovarian and gastric malignancies. Despite oxaliplatin’s current prevalence, the optimal PIPAC agent for colorectal peritoneal metastases remains to be definitively established. Patient-derived organoid models present a promising avenue for personalized drug selection, potentially enabling tailored therapeutic approaches based on individual tumor characteristics.
The transcriptomic analysis revealed significant changes in gene expression patterns when colon organoids were treated under PIPAC conditions. Although DEG analysis is a standard tool for identifying statistically significant gene expression differences between conditions, our experimental design, consisting of single RNA-seq replicates per treatment group, limits its statistical application. Instead, our approach leveraged PCA-based gene loadings to highlight transcriptomic contributors to condition-specific variance. Notably, PCA identified the differentiation of effects between the drugs, with Paclitaxel-treated samples displaying unique mRNA expression profiles that suggest minimal impact by PIPAC conditions. Conversely, Cisplatin and Oxaliplatin treatments under PIPAC conditions led to marked transcriptomic alterations, underscoring the differential impacts of these drugs on colon organoids in the presence of PIPAC. The transcriptomic alterations observed under PIPAC conditions appear to reflect both drug-specific mechanisms and the physiological effects of pressurized aerosol delivery. Cisplatin and oxaliplatin exposure under PIPAC induced upregulation of DNA repair and cell cycle checkpoint pathways (e.g., G2/M arrest, p53 signaling), consistent with their roles as DNA-damaging agents. Paclitaxel-treated organoids under PIPAC showed enrichment of mitotic spindle assembly and apoptotic signaling, aligning with its mechanism of microtubule stabilization and mitotic disruption. These changes may reflect stress responses to the pressurized aerosol environment itself or secondary inflammatory and differentiation responses triggered by enhanced drug-tissue interaction.
Specifically, in Cisplatin-treated organoids exposed to PIPAC, we observed downregulation of hallmark oncogenic pathways such as epithelial-mesenchymal transition (EMT) and KRAS signaling, which are commonly associated with chemoresistance and tumor progression24. The suppression of these pathways may sensitize cells to DNA-damaging agents such as Cisplatin by limiting mesenchymal plasticity and reducing pro-survival signaling. Furthermore, pathway analysis revealed significant changes in GPCR ligand binding, including the downregulation of receptors such as CX3CR1, which has been implicated in tumor-immune interactions and metastasis25. Additionally, the repression of MPL, a gene associated with hematologic malignancies and treatment resistance, suggests that pressure- and temperature-induced modulation of signaling may augment drug efficacy. These pathway-level alterations collectively support the hypothesis that PIPAC enhances the therapeutic effect of Cisplatin not only through improved delivery but also by modifying the tumor’s molecular susceptibility. Although global mRNA expression patterns between atmospheric and PIPAC-delivered cisplatin were not markedly divergent, transcript-level differences in specific genes within critical pathways (e.g., pro-apoptotic regulators, EMT drivers, stress response factors) may underpin the enhanced cytotoxicity observed in functional assays. This supports the hypothesis that PIPAC may potentiate cisplatin’s efficacy not by triggering large-scale transcriptional reprogramming, but by amplifying localized effects on key cellular stress pathways. The PCA results, while showing clustering across drug types, also support the interpretation that condition-specific transcriptomic signatures exist and contribute to treatment response heterogeneity.
The transcriptomic divergence observed between S1 and S4 organoids under PIPAC-drug treatment further supports the presence of intra-tumoral heterogeneity. PCA analysis clearly separated these organoid lines, even under the same treatment conditions, indicating distinct global gene expression programs. For example, S1 organoids exhibited downregulation of EMT and KRAS signaling following PIPAC-Cisplatin exposure, whereas S4 organoids showed enrichment of DNA repair and apoptotic pathways. These differences suggest that the transcriptional landscape of each sub-region within a tumor may dictate differential pathway activation in response to therapy, ultimately contributing to variability in drug sensitivity. Incorporating transcriptomic heterogeneity into preclinical models, such as the S1/S4 paired organoids, enhances our ability to model patient-specific and region-specific treatment responses.
Despite the promising insights offered by this study, several limitations must be acknowledged. The small number of patient-derived colorectal cancer organoid lines used limits the statistical power and generalizability of our findings. While the observed cytotoxic and transcriptomic trends are consistent across the two models, inter-patient heterogeneity remains a critical factor in translational research. Future studies incorporating larger and more genetically diverse organoid cohorts will be essential to validate and refine the therapeutic implications of PIPAC-based combinatorial regimens. We acknowledge that viability was assessed at a single 72-hour time point, a commonly used and validated interval in colorectal cancer organoid models to capture peak cytotoxic responses. However, this approach may not fully capture delayed or long-term effects, and future studies incorporating extended time-course analyses would offer deeper insights into the kinetics and durability of treatment-induced cell death. It is noteworthy that PIPAC conditions alone resulted in moderate reductions in cell viability in both organoid lines. While this observation may raise concerns about non-specific cytotoxic effects, it is important to interpret these findings within context. Organoids represent a highly sensitive in vitro system that lacks protective stromal or immune components found in vivo. Moreover, pressure and temperature elevation are intrinsic to the PIPAC delivery mechanism, facilitating enhanced drug diffusion. Clinical studies have shown that such physical stressors are generally well-tolerated when properly regulated. Therefore, the observed PIPAC-only effects should be regarded as a conservative estimate of physical cytotoxicity rather than a contraindication for clinical application. The use of single RNA-seq replicates per condition also limits statistical generalization of our study. While the resulting transcriptomic analyses (e.g., PCA, pathway enrichment, ssGSEA) offer valuable insights into potential mechanisms of drug response under PIPAC conditions, they are presented as exploratory and do not include formal statistical testing. Future studies incorporating biological replicates will be essential to validate the gene- and pathway-level findings presented here. The absence of normal colon organoid controls also limits our ability to evaluate the baseline toxicity of PIPAC or PIPAC-delivered drugs in non-cancerous tissues. Future studies incorporating such controls will be necessary to further delineate therapeutic windows and potential off-target effects.
The findings of this study have significant implications for clinical practice, particularly in the treatment of peritoneal carcinomatosis. The enhanced efficacy of cytotoxic drugs under PIPAC conditions could lead to more effective treatment protocols, offering hope to patients with limited treatment options. Moreover, the study’s insights into the transcriptomic changes induced by PIPAC provide a valuable foundation for developing targeted therapeutic strategies that could improve patient outcomes in clinical settings.

Methods

Colorectal cancer organoid cultures
SNU-4398S1-TO and SNU-4398S4-TO organoid lines were obtained from Korean Cell Line Bank (KCLB). The medium for human colorectal cancer (CRC) organoid cultures was replaced every three days. For passaging, the basement membrane extract (BME) dome was disrupted mechanically by pipetting using TrypLE Express solution (GIBCO, Cat# 12604-021), and the organoids were then transferred to a 15 mL conical tube. The BME dome was broken down through vigorous pipetting, after which the tube containing the organoid-BME mixture was incubated at 37 °C for approximately 5 to 10 min. Subsequently, the organoids were centrifuged at 1,000 rpm for 3 min, and the supernatant was discarded. After removing the BME, the cell pellet was resuspended in fresh BME, placed into a T-25 flask, and incubated at 37℃ for 10 min to allow the BME to solidify. Afterwards, 5 mL of HISC medium was added to cover the BME dome, and the cells were then cultured in an incubator set at 37℃ and 5% CO2. The HISC medium composition included 40% W/V basal culture medium, 50% W/V L-WRN conditioned medium, 1x B27 supplement (GIBCO, Cat# 17504-044), human EGF (50 ng/mL) (GIBCO, Cat# PHG0313), human FGF-10 (10 ng/mL) (Peprotech, Cat# 100 − 26), nicotinamide (10 mM) (Sigma-Aldrich, Cat# 72340), N-acetylcysteine (1.25 mM) (Sigma-Aldrich, Cat# A7250), A83-01 (500 nM) (Sigma-Aldrich, Cat# SML0788), SB202190 (3 µM) (Sigma-Aldrich, Cat# S7067), prostaglandin E2 (Sigma-Aldrich, Cat# P5640), and primocin (100 µg/mL) (GIBCO, Cat# ant-pm-1).

Drug treatment and PIPAC delivery
All experiments were conducted in duplicate. Organoids were enzymatically and mechanically dissociated into single cells by incubating and pipetting in TrypLE Express solution (GIBCO) for 5 to 10 min. Subsequently, the suspension was centrifuged at 1,500 rpm for 3 min. After the removal of the basement membrane extract (BME), the cellular pellet was resuspended in a 0.5:1.5 ratio of HISC medium to BME (GIBCO). A uniform BME mixture (100 µL, containing 20,000 cells/mL) was prepared and plated as domes into transparent 24-well plates (Corning, NY, USA) utilizing E1-ClipTip Electronic Pipettes (Thermo Fisher Scientific, MA, USA). The BME domes were allowed to solidify by inverting the culture plates and incubating them at 37 °C with 5% CO₂ for 15 min, which prevented cell settling and ensured uniform 3D structure formation within the matrix, before the addition of pre-warmed HISC medium (300 µL) to each well. After a period of 96 h, the procedure for Pressurized Intraperitoneal Aerosol Chemotherapy (PIPAC) was initiated.
The apparatus for the PIPAC model was maintained at a steady temperature of 37 °C within a water bath throughout the process. Prior to the PIPAC drug treatment, HISC medium was removed. The electronic pressure calibrator (LR-Cal XA1000, Kirchentellinsfurt, Germany) was set and maintained throughout the PIPAC process. Three drugs, Cisplatin, Oxaliplatin, and Paclitaxel (Selleckchem, TX, USA) were first transformed into aerosol form and then administered to the exposed tumor cells. We used a commercially available hermetically sealable plastic container with a 3.5 L capacity to simulate the abdominal cavity. Two trocars (12-mm and 5-mm, Kii®Balloon Blunt Tip System; Applied Medical, Rancho Santa Margarita, CA, USA) were positioned in the center of the container’s top cover. We inserted the aerosolizer nozzle (Capnopen®, Capnomed, Zimmern o.R., Germany) and a temperature/humidity monitoring probe (XA 1000; Lufft Messund Regeltechnik GmbH, Fellbach, Germany) through these trocars. The container was maintained at a constant temperature of 37˚C by placement in a water bath throughout the procedure. After sealing the container securely, we established a CO2 capnoperitoneum at 15mmHg (PneumoSure XL; Stryker Endoscopy, San Jose, United States) which was sustained for the duration of the PIPAC procedure. The drug administration was performed via a high-pressure line using an injection system (MEDRAD® Salient Compact Injection System; Bayer, Leverkusen, Germany) operating at 200 psi with a delivery rate of 0.5mL/s Culture plates containing organoids were vertically positioned against the chamber wall to avoid direct liquid contact with aerosol droplets while ensuring exposure to the aerosolized drug particles. A schematic of the entire setup is provided in Supplementary Fig. 2.
Three chemotherapeutic agents, cisplatin, oxaliplatin, and paclitaxel, were selected based on their clinical relevance to colorectal cancer and prior use in intraperitoneal chemotherapy protocols. Oxaliplatin is commonly employed in PIPAC regimens for peritoneal carcinomatosis of colorectal origin. Paclitaxel was chosen for its high molecular weight and favorable peritoneal retention, while cisplatin has been utilized in combination therapies for its potential to enhance cytotoxicity. Although doxorubicin is widely used in PIPAC for other cancer types, it is less commonly applied in colorectal cancer-specific protocols, which guided our drug selection26. To evaluate the cytotoxic effect of the PIPAC test atmosphere on organoid cultures, we established three distinct treatment groups:
(1) Atmospheric control (atm): Organoids were incubated at room temperature (~ 22–24 °C) under atmospheric pressure (1 atm) without any aerosol exposure.
(2) PIPAC-PBS group: Organoids were exposed to a 5-minute aerosol phase using 50 mL sterile PBS, followed by a 15-minute exposure phase under a CO₂ capnoperitoneum of 15 mmHg at 37 °C in a closed chamber.
(3) Drug-PIPAC groups: Organoids were exposed to aerosolized chemotherapeutic agents (cisplatin, oxaliplatin, or paclitaxel) using the same PIPAC parameters as above.
This experimental design enabled differentiation between the effects of the PIPAC environment itself and those attributable to the chemotherapeutic agents under PIPAC conditions.
Doses of Cisplatin, Oxaliplatin, Paclitaxel were 5µM, 100µM, and 1µM respectively diluted in DPBS at room temperature. The concentrations of Cisplatin (5 µM), Oxaliplatin (100 µM), and Paclitaxel (1 µM) were selected based on previously published organoid-based drug screening studies11 and clinical pharmacokinetic data derived from intraperitoneal chemotherapy protocols. These concentrations were shown to induce quantifiable but submaximal cytotoxicity, facilitating comparison of drug efficacy under different delivery conditions. HISC (200 µL) was added 30 min after drug treatment. After 72 h, ATP detection assay followed. For atmospheric (atm) control conditions, the same concentrations of Cisplatin, Oxaliplatin, and Paclitaxel were prepared in liquid form and added directly to the culture wells using pipette-based manual administration. These treatments were conducted under standard atmospheric pressure and temperature (37 °C, 5% CO₂) without exposure to aerosolized delivery or elevated pressure. This setup was designed to reflect conventional in vitro drug application, allowing for direct comparison with the PIPAC-treated organoids.

ATP detection assay
Following a 72-hour period of drug exposure, 10 µl of 3D Reagent (Promega #G968B) was added to each organoid well. This was succeeded by five minutes of vigorous agitation. Subsequently, after a 30-minute incubation at ambient temperature and an additional minute of agitation, the mixture was transferred to white-walled 96-well plates (Corning, Cat# 3903). Luminescence measurements were performed using a Luminoskan Ascent (Thermo Scientific) with an integration time of 1000 milliseconds. The results were standardized against the control vehicle, graphed, and the EC50 values were determined utilizing Prism 5 software. In the context of high-throughput drug screening, DMSO served as the control. The values were then adjusted based on the control vehicle. The cytotoxicity endpoint was assessed at 72 h post-treatment based on prior time-course optimization, which indicated that maximal ATP depletion and cell death plateau occur between 72 and 96 h in colorectal cancer organoids. Although earlier time points were not analyzed in this study, the 72-hour window is commonly used and provides sufficient resolution to compare relative drug efficacies across conditions17.

Analysis of RNA sequencing
Total RNA was extracted 72 h after treatment across all conditions using the RNeasy Mini Kit (Qiagen, Hilden, Germany), following the manufacturer’s protocol. This timepoint was selected to match the viability assessment window and capture the downstream transcriptional responses. RNA-seq analysis was performed once per condition due to material limitations inherent in patient-derived organoid systems. Paired end sequencing reads of cDNA libraries (101 bp) generated from a NovaSeq6000 instrument were verified its sequence quality with FastQC v 0.11.7. For data preprocessing, Trimmomatic v0.38 was employed to remove low-quality bases and adapter sequences from the reads. Subsequently, the processed reads were aligned to the human genome (UCSC hg19) utilizing HISAT v2.1.0. StringTie v1.3.4d was then used for transcript assembly, which included known, novel, and alternatively spliced transcripts. The expression levels of transcripts and genes were quantified in terms of read count per sample.
To investigate transcriptomic alterations, we applied principal component analysis (PCA) to reduce data dimensionality and uncover key patterns in gene expression variability. Sample clustering and potential batch effects were assessed by visualizing the principal components using the FactoMineR and factoextra packages in R. Read counts were used as input for PCA, and the top principal components were visualized using the ggplot2 package. Genes with the highest absolute loadings in PC1 and PC2 were identified and used for exploratory pathway interpretation. Additionally, single-sample gene set enrichment analysis (ssGSEA) was performed to quantify pathway activity within individual samples. Gene sets for enrichment analysis were obtained from the Molecular Signatures Database (MSigDB v2023.1), including the hallmark gene sets. Pathway enrichment was analyzed using the ReactomePA package in R, and statistical significance was defined as p < 0.05. No batch correction or sample scaling was applied due to the uniform experimental setup.

Statistical analysis
Cell viability assays were conducted in replicate, and the mean values were compared using two-tailed paired t-tests to assess differences in drug response between atmospheric and PIPAC conditions within the same organoid line. Statistical significance thresholds were defined as follows: p < 0.05 (*), p < 0.005 (**), and p > 0.05 (ns).

Supplementary Information
Below is the link to the electronic supplementary material.