Ketogenic Diet and β‐Hydroxybutyrate Suppress Bladder Cancer Growth and Enhance Anti‐PD‐L1 Immunotherapy Efficacy Authors: Yuan Zeng, Gangli Gu, Lin Yang, Bohao Fan, Haochen Cui, Lei Yan, Zhao Liu Published: 2026 ABSTRACT ABSTRACTThe limited clinical efficacy of immune checkpoint inhibitors (ICIs) remains a major challenge in the treatment of bladder cancer (BCa). Here, we report that a ketogenic diet (KD) and its principal circulating metabolite, β‐hydroxybutyrate (β‐HB), suppress bladder tumor growth and enhance responses to PD‐L1 inhibitor. In a syngeneic MB49 model, KD reduced tumor growth compared with normal diet (ND). KD increased plasma β‐HB and was associated with higher intratumoral PD‐L1 expression. Exogenous β‐HB supplementation under ND recapitulated the antitumor effect of KD, further enhancing anti‐PD‐L1 efficacy. Mechanistically, β‐HB increased intracellular reactive oxygen species (ROS) and disrupted mitochondrial membrane potential in bladder cancer cells, leading to ATP depletion and enhanced apoptosis. Besides, β‐HB also upregulated PD‐L1 in vitro and in vivo through GPR109A‐JAK2‐STAT3 signaling axis. Immune profiling of treated tumors showed increased infiltration and effector function of CD8+ T cells and NK cells and decreased immunosuppressive populations, changes that were further amplified when combined with anti‐PD‐L1. Together, our results indicate that KD and β‐HB exert both tumor intrinsic and immune modulatory effects that sensitize bladder tumors to PD‐L1 inhibitor, supporting further evaluation of ketogenic interventions as adjuvants to immunotherapy in BCa. Abbreviations β‐hydroxybutyrate bladder cancer immune checkpoint inhibitors interferon‐γ ketogenic diet myeloid‐derived suppressor cells normal diet programmed cell death ligand 1 reactive oxygen species tumor necrosis factor‐α Introduction Bladder cancer is the tenth most common cancer worldwide, accounting for 3% of all cancer incidence, with approximately 573,000 newly diagnosed cases and 213,000 deaths globally each year, and the incidence rate in men is approximately four times that of women [1]. Although about 75% of bladder cancer patients have lesions confined to the mucosa (stage Ta or CIS) or submucosa (stage T1), the majority of patients recur or develop distant metastases, accompanied by a decline in survival [2, 3]. For limited or non‐muscle‐invasive bladder cancer, prompt surgical treatment has been the standard of care. In contrast, for muscle‐invasive bladder cancer and recurrence or distant metastasis due to it, systemic chemotherapy is relied upon [4]. Platinum‐based combination chemotherapy has been the chemotherapy of choice for bladder cancer in the past decades, and in recent years, the use of immune checkpoint inhibitors (ICIs) in cancer has provided new solutions for the treatment of metastatic bladder cancer. Atezolizumab, a human monoclonal immunoglobulin targeting programmed death ligand‐1 (PD‐L1), became the first immune checkpoint inhibitor approved by the U.S. Food and Drug Administration (FDA) for the treatment of advanced bladder cancer in 2016 [5]. However, only some patients benefit from immune checkpoint inhibitors, and a small number of patients experience rapid disease progression after immunotherapy due to primary and acquired resistance [6]. Therefore, the mechanism of immunotherapy resistance is also one of the issues worth exploring in the current treatment of bladder cancer. Metabolic pathways are receiving increasing attention as potential targets for cancer therapy. One of the hallmarks of all cancer cells is the aberrant metabolic phenotype‐Warburg effect, first described by Otto Warburg, characterized by a bioenergetic shift from oxidative phosphorylation to glycolysis to meet the increased demand for energy, biosynthesis, and regulation of gene expression in cancer cells, and therefore closely associated with cancer development, malignant growth, metastasis, and treatment resistance [7, 8, 9, 10, 11]. Based on this property of cancer cells, fasting, calorie restriction (CR) and carbohydrate‐restricted ketogenic diet (KD) can be used to achieve disruption of cancer glycolysis‐dependence by restricting glucose utilization, and may be used as potential anti‐tumor therapeutic strategies for slowing down cancer progression [12, 13, 14, 15, 16, 17, 18]. A ketogenic diet is low in carbohydrates and high in fat, resulting in elevated circulating ketone bodies such as acetoacetate (H3C–CO–CH2COO−), β‐hydroxybutyrate (H3C–CHOH–CH2–COO−) and volatile acetone (H3C–CO–CH3) [19, 20, 21]. Previous studies have demonstrated the therapeutic effect of KD in the treatment of convulsions and epilepsy, and the relationship between KD and cancer treatment has received increasing attention due to the ability to reduce glucose utilization by cancer cells while providing ketone bodies as bioenergetic fuel for normal cells [22, 23, 24]. Published studies have shown that the KD inhibits the growth of renal carcinoma and melanoma xenografts in mice and enhances the anticancer effects of immune checkpoint inhibitors [25, 26]. The anticancer effects of the KD are largely due to the lack of a major energy substrate for cancer cells, that is, a reduction in circulating glucose levels; besides, there is also emerging evidence that the physiological state of ketosis or elevated circulating ketone body levels can also exert anticancer effects [27, 28, 29]. In this study, we demonstrate that KD and its principal metabolite, β‐HB, suppress bladder tumor growth in an MB49 syngeneic model and inhibit proliferation, migration, and invasion of bladder cancer cells in vitro. Mechanistically, β‐HB induces ROS‐mediated mitochondrial dysfunction and apoptosis. Moreover, β‐HB upregulates PD‐L1 through the GPR109A‐JAK2–STAT3 signaling axis and remodels the tumor immune microenvironment, thereby enhancing the therapeutic efficacy of the anti‐PD‐L1 inhibitor. These findings support further investigation of ketogenic strategies or controlled β‐HB supplementation as adjuncts to immune checkpoint inhibitors. Materials and Methods Animals Female C57BL/6J and BALB/c nude mice (6‐week‐old) were obtained from Charles River Laboratories. All animal procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC), Qilu Hospital of Shandong University. Animals were maintained under specific pathogen‐free (SPF) conditions with ad libitum access to food and water and a 12 h light/12 h dark cycle. Animals were group‐housed in standard cages. The sex (female) and age of animals used are reported here. Details of randomization and group allocation are provided in Method Details. All in vivo experiments conform to the relevant institutional and national guidelines. Cell Lines Human bladder cancer cell lines T24 and 5637 were obtained from Procell Life Technology (CL‐0227 and CL‐0002, respectively). The murine bladder cancer cell line MB49 was obtained from Shanghai Zhong Qiao Xin Zhou Biotechnology (ZQ1050). Cells were cultured in RPMI1640 (Servicebio, G4531) supplemented with 10% fetal bovine serum (Umedium, 3022A), 100 U/mL penicillin and 100 μg/mL streptomycin (Servicebio, G4003) at 37°C in a humidified atmosphere containing 5% CO2. Animal Models and Tumor Establishment MB49 tumors were established by subcutaneous injection of 1 × 106 MB49 cells in 100 μL per mouse into the right flank of female C57BL/6J and BALB/c nude mice. Mice were randomized by body weight prior to tumor cell inoculation into experimental groups; the sample size per group is indicated in the figure legends. For the simple dietary intervention cohort, mice received normal diet (ND; F1515, Bio‐Serv) or ketogenic diet (KD; F3666, Bio‐Serv). Diet was started on day 4 after tumor implantation. Tumor size was measured every 2–3 days using digital calipers and tumor volume was calculated as follows: volume (mm3) = π/6 × (larger diameter) × (smaller diameter)2. Humane endpoint criteria were applied per the approved IACUC protocol and mice were humanely euthanized following the approved IACUC protocol by CO2 inhalation. Dietary and Pharmacologic Treatments For combination therapy experiments the following groups were used: ND + IgG isotype control (Bio‐XCell, Cat# BE0089), KD + IgG, ND + anti‐PD‐L1 (Bio‐XCell, Cat# BE0101), and KD + anti‐PD‐L1. Anti‐PD‐L1 antibody or IgG isotype control (200 μg in 100 μL PBS per mouse) was administered by intraperitoneal injection every 3 days starting on day 6 after tumor implantation for a total of three injections. For the β‐HB supplementation cohort, the KD group was replaced by a β‐HBi.p. group in which mice received ND but were administered β‐hydroxybutyric acid sodium salt (Millipore Sigma, Cat# 54965) intraperitoneally at 40 mg per mouse daily beginning on day 4 after tumor implantation. Treatment schedules and group sizes are indicated in the figure legends. In Vivo Bioluminescence Imaging Tumor growth was monitored using a DPM‐IVFM in vivo imaging system. Mice were injected i.p. with 200 μL D‐Luciferin Potassium Salt (Beyotime, ST196) 15 min prior to imaging. Imaging and image quantification parameters are reported in the figure legends. β‐Hydroxybutyrate (β‐HB) Measurement Plasma β‐HB concentration was measured using a β‐Hydroxybutyrate (Ketone Body) Colorimetric Assay Kit (Elabscience, E‐BC‐K785‐M) following the manufacturer's instructions. Blood was drawn prior to euthanasia to measure circulating β‐HB at the scheduled terminal sampling time. Plasma was prepared from whole blood collected into anticoagulant tubes, centrifuged, and stored at −80°C until assay. Absorbance of samples and standards was read at 450 nm, and concentrations were calculated from a standard curve. Cell Culture and In Vitro Treatments T24, 5637 and MB49 cells were maintained as described above. β‐hydroxybutyrate (Selleck, S1031), N‐acetylcysteine (Selleck, S1623), AG490 (Selleck, S1143) and mepenzolate bromide (MCE, HY‐17585) were prepared as stock solutions according to manufacturers' instructions and diluted to final concentrations in culture medium immediately prior to use. Where indicated, cells were treated with β‐HB at the concentrations and durations specified in the figure legends. All in vitro experiments were performed at least three independent times unless indicated otherwise. Cell Viability and Colony Formation Assays Cell viability was assessed using Cell Counting Kit‐8 (Beyotime, C0037). Cells were seeded at 2 × 103 cells/well in 96‐well plates, treated as indicated, and incubated with CCK‐8 reagent for 2 h at 37°C. Absorbance at 450 nm was measured on an Epoch2 microplate reader (Bio‐Tek). For colony formation assays, cells were seeded in 6‐well plates, treated as indicated, and cultured for 12 days. Colonies were fixed with 4% paraformaldehyde for 30 min, stained with crystal violet for 30 min, rinsed, and counted. EdU Incorporation Assay For DNA synthesis analysis, cells were treated as indicated, incubated with EdU for 2 h, fixed in 4% formaldehyde, permeabilized, and stained using the Apollo 643 reagent (EdU Cell Proliferation Kits, APExBIO K2243) per the manufacturer's instructions. Nuclei were counterstained with Hoechst 33342, and EdU‐positive cells were quantified by fluorescence microscopy. Wound Healing and Transwell Invasion Assays For wound healing assays, confluent monolayers in 6‐well plates were scratched using a pipette tip and washed to remove debris; medium was replaced with serum‐free medium and images were captured at 0 h and the indicated time point. Wound closure was quantified by measuring initial and final wound widths and calculating percent closure: [(initial − healed)/initial] × 100%. For Transwell invasion assays, 5 × 104 cells in serum‐free medium were seeded into the upper chamber of Matrigel‐coated Transwell inserts (Corning, Cat# 356234; 24‐well, Cat# 3422). Complete medium containing 10% FBS was added to the lower chamber as a chemoattractant. After 24 h, migrated cells on the lower membrane surface were fixed, stained with crystal violet, and counted under a microscope. Apoptosis Analysis by Flow Cytometry Apoptosis was measured by Annexin V‐FITC/PI staining (Beyotime, C1062) according to the manufacturer's protocol. Briefly, cells treated as indicated were harvested, washed, stained with Annexin V‐FITC and propidium iodide for 30 min at room temperature in the dark, and analyzed on a Beckman Coulter flow cytometer. The fractions of live, early apoptotic, and late apoptotic/necrotic cells were determined based on fluorescence signals. ATP Measurement Intracellular ATP levels were measured using an ATP assay kit (Beyotime, S0026). Cells (1 × 106) treated as indicated were lysed, centrifuged at 12,000× g for 5 min at 4°C, and luciferase intensity was measured on a multimode reader in luminance mode. ATP content was normalized to total protein concentration. Reactive Oxygen Species (ROS) Assays T24, 5637, and MB49 cells were planted in 6‐well plates for 24 h after different treatments. Intracellular ROS was measured with a ROS assay kit (Beyotime, S0033) according to the manufacturer's instructions and observed with a fluorescence microscope. Quantification was performed with ImageJ (Fiji) with background subtraction and consistent ROI thresholds across samples. Immunohistochemistry (IHC) and TUNEL Staining Tumor syngeneic allografts were fixed in 10% neutral buffered formalin, paraffin‐embedded, and sectioned. Sections were incubated with primary antibodies (anti‐PD‐L1 1:500; anti‐Ki‐67 1:1000) overnight at 4°C, followed by HRP‐conjugated secondary antibodies (1:250) for 40 min at 25°C. DAB chromogen was used for visualization and sections were counterstained with hematoxylin. Apoptotic cells were detected using a TUNEL kit (Servicebio, G1507). Images were captured on an Olympus BX53 microscope. Quantifications of staining were performed using ImageJ software. Immunofluorescence Cells were seeded on glass coverslips and treated as indicated. After fixation with 4% paraformaldehyde and permeabilization with 0.4% Triton X100, cells were blocked with 5% goat serum and incubated with primary antibodies overnight at 4°C. Secondary antibody (CoraLite 594‐conjugated goat anti‐mouse IgG, Proteintech SA00013‐3) was applied for 1 h at 37°C. Nuclei were counterstained with Hoechst 33342 and images acquired on an Olympus IX73 microscope. Immunofluorescence signal quantification was performed with ImageJ (Fiji) with background subtraction and consistent ROI thresholds across samples. Western Blot and Immunoprecipitation Whole‐cell lysates were prepared in RIPA buffer, protein concentrations determined, and samples separated by SDS‐PAGE followed by transfer to PVDF membranes. Membranes were blocked with 5% nonfat milk and incubated with primary antibodies overnight at 4°C, followed by HRP‐conjugated secondary antibodies for 1 h at room temperature. Signal was detected by enhanced chemiluminescence (Millipore). Densitometric quantification of bands was performed using ImageJ; protein expression was normalized to β‐actin and quantified from at least three independent experiments. RNA Extraction, qRT‐PCR and RNA‐Seq Total RNA was extracted from cell culture lysates using the SteadyPure Universal RNA Extraction Kit (Agbio, AG21017) and reverse transcribed using Evo M‐MLV RT Mix (Agbio, AG11728). qPCR was performed with SYBR Green Premix Pro Taq HS (Agbio, AG11701) on a CFX96 Real‐Time PCR System (Bio‐Rad). For transcriptome profiling, RNA‐seq libraries were prepared using VAHTS Universal V6 RNA‐seq Library Prep Kit and sequenced by Shanghai Biotechnology Corporation. Reads were processed and aligned to the reference genome; differential expression analysis was performed with edgeR (FDR < 0.05, fold‐change > 2). KEGG pathway analysis was performed using clusterProfiler, and Gene Set Enrichment Analysis (GSEA) was used for pathway enrichment. Tumor‐Infiltrating Lymphocyte Isolation and Flow Cytometry Tumors were collected after euthanasia, minced and digested in enzyme solution containing collagenase type II, hyaluronidase, and DNase I at 37°C for 40 min. Digests were mechanically dissociated, filtered through a 70 μm strainer, and subjected to low‐speed centrifugation to remove debris. The supernatant was centrifuged, RBCs lysed, and the pellet resuspended in 40% Percoll. After density gradient centrifugation, the lymphocyte‐enriched fraction was collected, washed, and kept on ice. For surface staining, 1 × 106 cells were incubated with fluorochrome‐conjugated antibodies for 30 min at 4°C in the dark, washed with PBS, and analyzed by flow cytometry. The corresponding gating strategies are shown in Figure S3. For intracellular cytokine detection, single‐cell suspensions were stimulated with PMA (1 ng/mL) and ionomycin (500 ng/mL); Brefeldin A (BioLegend, Cat# 420601) was added 1 h after stimulation. Surface staining was performed first, followed by fixation/permeabilization and intracellular staining for IFN‐γ and TNF‐α. Quantification and Statistical Analysis All statistical analyses were performed using GraphPad Prism 9. Data are presented as mean ± SD unless otherwise noted. For in vitro experiments, results are representative of at least three independent biological replicates. For in vivo experiments, exact sample sizes (n) for each experiment are indicated in the relevant figure legends and represent the number of animals per group. Statistical comparisons between two groups were performed using two‐tailed unpaired Student's t‐test; comparisons among multiple groups were made by one‐way ANOVA. p Values < 0.05 were considered statistically significant. Tests used, exact n, definition of center and dispersion measures (mean, SD), and any other statistical details are provided in the figure legends. Randomization was by body weight prior to tumor inoculation. Results Ketogenic Diet Inhibits Bladder Tumor Growth and Raises Circulating Level of β‐HB In Vivo To investigate the effect of ketogenic diet (KD) in bladder cancer growth, we established MB49 syngeneic allografts in C57BL/6J mice, comparing tumor progression between KD and normal diet (ND) groups (Figure 1A). Notably, body weight trajectories did not differ significantly between groups during the experiment (Figure 1B). Biochemical analysis confirmed effective metabolic adaptation to the dietary intervention, with KD‐fed mice demonstrating substantially elevated plasma β‐hydroxybutyrate (β‐HB) levels after 10‐day dietary conditioning (Figure 1C). This indicates successful induction of ketogenesis as an alternative energy substrate under glucose‐deprived physiological conditions. Tumor growth showed that KD markedly reduced tumor volumes and endpoint tumor weights relative to ND (Figure 1D–G). Following sacrifice, tumor tissues were harvested and subjected to immunohistochemical analysis. Comparative evaluation revealed that tumors from KD‐fed mice exhibited decreased expression of Ki‐67 compared to those in the ND group. Concurrently, we observed elevated TUNEL staining intensity in the KD cohort relative to ND controls (Figure 1H–J). Ketogenic diet inhibits bladder tumor growth and raises circulating level of β‐HB in vivo. (A) Schematic diagram of the experimental design. (B) Longitudinal monitoring of body weight in tumor‐bearing mice fed normal diet (ND) or ketogenic diet (KD) until experimental endpoint (n = 6). (C) β‐HB levels in plasma from C57BL/6J mice were determined using a β‐HB assay kit (n = 6). (D) Comparative analysis of excised tumor specimens at study termination. (E–G) Growth kinetics of MB49 bladder cancer in C57BL/6J mice. The minimal tumorigenic dose of MB49 bladder cancer cells was inoculated s.c. 4 days before starting the dietary intervention. Tumor size was monitored 5 times for 2 weeks (E), and the mean ± SD of tumor size (F) and weight (G) was calculated for both nutritional interventions. (H–J) Representative IHC images taken from syngeneic MB49 tumor tissues stained with the antibodies against Ki‐67 (H) and Tunel (I). All data are means ± SD; **p < 0.01, ****p < 0.0001, by unpaired Student's t‐test and one‐way ANOVA. We subsequently established the same MB49 tumors in immunodeficient BALB/c nude mice to dissect the immune‐dependent and independent effects of KD on bladder cancer growth. As shown in Figure S1A–G, KD significantly suppressed tumor growth in these immunodeficient animals, with reduced tumor volumes and weights compared to ND controls, demonstrating direct anti‐tumor effects independent of adaptive immunity. Collectively, these results indicate that the ketogenic diet significantly suppresses bladder tumor growth in vivo accompanied by reduced tumor cell proliferation and increased apoptosis. β‐HB Inhibits Bladder Cancer Cell Activity, Proliferation, Migration and Invasion In Vitro By elevating systemic β‐HB while suppressing glucose bioavailability [12, 15, 16], KD creates a metabolic environment unfavorable for tumor growth. Beyond metabolic substrate limitation, emerging evidence reveals direct tumor‐suppressive properties of β‐hydroxybutyrate. This prompted our investigation into its direct impact on bladder cancer cells in vitro. We next examined direct effects of β‐HB on bladder cancer cell lines; T24, 5637, and MB49 cells cultured in media supplemented with exogenous β‐HB exhibited significantly reduced viability compared to those grown in control media (Figure 2A). Colony formation was significantly inhibited by β‐HB across cell lines, with colonies counted at the endpoint (Figure 2B,C). EdU incorporation assays demonstrated a significant decrease in DNA synthesis following β‐HB exposure in all three cell lines (Figure 2D–F). Furthermore, we found that β‐HB markedly reduced the migration and invasion capacity of bladder cancer cells in vitro, as evidenced by both wound healing assays and Transwell invasion assays (Figure 2G–L). These results establish that β‐HB, the predominant ketone body generated during ketogenic states, exerts multimodal growth control in bladder cancer through direct cytotoxic activity and metastatic potential attenuation. β‐HB inhibits bladder cancer cells activity, proliferation, migration and invasion in vitro. (A) T24, 5637 and MB49 cells viability were decreased when grown in conditions with β‐HB (10 mM) supplementation for 24 h. (B, C) The effect of β‐HB on in vitro proliferation utilized colony‐formation assay. Numbers of colonies were counted at day 12. (D–F) T24 (D), 5637 (E) and MB49 (F) were cultured in control or β‐HB supplemented medium with 10 mM for 24 h and cell DNA synthesis was determined using an EdU assay. (G–J) Wound healing assay were performed following treatment with PBS or β‐HB (10 mM) in T24 (G), 5637 (H) and MB49 (I) cells. (K, L) The effect of β‐HB on in vitro invaded cell numbers was measured utilizing Transwell assay. Each experiment was performed at least three times, all data are mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001 ****p < 0.0001, by unpaired Student's t‐test and one‐way ANOVA. β‐HB Promotes Apoptosis in Bladder Cancer Cells by Increasing Their Reactive Oxygen Species Levels Previous studies have reported that β‐hydroxybutyrate can induce oxidative stress in smooth muscle cells, endometrial cells, cardiomyocytes, and hepatocytes, leading to cellular damage [30, 31, 32, 33]. While current evidence demonstrates β‐HB's capacity to induce lactate accumulation concomitant with reactive oxygen species (ROS) elevation specifically in glioma stem cell models, its broader oxidative stress‐modulating effects across cancer lineages require systematic investigation [20]. We assessed intracellular ROS after β‐HB treatment using the DCFH‐DA probe and fluorescence microscopy. A significant increase in ROS levels across bladder cancer cell lines treated with exogenous β‐HB was observed compared to controls (Figure 3A,B). Mitochondria serve as both the primary source of ROS generation and a key target of ROS‐mediated damage within cells. Excessive ROS can trigger the opening of the mitochondrial permeability transition pore (mPTP), resulting in mitochondrial swelling, rupture, cytochrome C release, and mitochondrial dysfunction accompanied by reduced ATP production, collectively driving cellular apoptosis [34]. Using TMRE fluorescence probes to quantify mitochondrial membrane potential (ΔΨm) before and after β‐HB treatment, we observed that β‐HB‐induced ROS accumulation caused a significant depolarization of ΔΨm (Figure 3C,D). Notably, intracellular ATP levels were substantially decreased in β‐HB‐treated cells (Figure 3E). Western blot analysis demonstrated significant activation of caspase‐3 in cells treated with β‐HB, confirming the occurrence of apoptosis (Figure 3F). Consistent with these changes, flow cytometry analysis revealed that apoptosis was significantly induced in T24, 5637, and MB49 cells exposed to β‐HB when compared to those under standard culture conditions, with the level of apoptosis showing a positive correlation with the concentration of β‐HB (Figure 3G–I). These findings suggest that β‐HB enhances ROS production in bladder cancer cell lines, which may play a critical role in its ability to attenuate cell proliferation and facilitate apoptosis. β‐HB promotes apoptosis of bladder cancer cells by increasing their reactive oxygen species levels. (A, B) ROS production was evaluated using fluorescence microscope in T24, 5637 and MB49 cells stained with DCFH‐DA after treatment with 10 mM β‐HB and quantitatively compared. (C, D) Mitochondrial membrane potential (mΔψ) in T24, 5637 and MB49 cells was determined with Tetramethylrhodamine ethyl ester (TMRE) probe after 10 mM β‐HB treatment and quantitatively compared. (E) The relative ATP level detected by multifunctional imaging enzyme labeling in T24, 5637 and MB49 cells after treatment with 10 mM β‐HB. (F) Expression levels of total caspase 3 and cleaved‐caspase 3 were determined by western blot analysis. (G–I) Apoptosis was determined using Annexin V‐FITC/PI staining followed by FACS analysis and the results were quantitatively compared in T24 (G), 5637 (H) and MB49 (I) cells. Each experiment was performed at least three times, all data were shown as mean ± SD; ****p < 0.0001, by unpaired Student's t‐test and one‐way ANOVA. β‐HB Regulates Immune Pathways in Bladder Cancer Cells Given the ability of β‐hydroxybutyrate to modulate oxidative stress levels in bladder cancer cells, we next aimed to investigate whether additional potential mechanisms are involved. RNA sequencing (RNA‐seq) was performed on the 5637 cell line after treatment with β‐hydroxybutyrate. The volcano plot illustrates the genes that were both up‐ and down‐regulated following β‐hydroxybutyrate treatment (Figure 4A). KEGG and GSEA of RNA‐seq data from β‐HB‐treated 5637 cells revealed statistically significant downregulation of cell‐cycle associated gene sets, including E2F targets and the G2M checkpoint (Figure 4B–D). Consistent with reduced transcriptional activity of cell‐cycle programs, EdU incorporation assays demonstrate that β‐HB reduces DNA synthesis in bladder cancer cells (Figure 2). Additionally, we examined the relationship between β‐hydroxybutyrate and various pathways associated with inflammation and immune response, encompassing signaling pathways such as IL6‐JAK‐STAT3, IL2‐STAT5, TNF‐α, and TGF‐β (Figure 4E–H). Our data suggest that β‐hydroxybutyrate is closely linked to cell cycle regulation and various immune pathways in bladder cancer cells. β‐HB regulates immune pathways in bladder cancer cells. (A) 5637 cells were treated with 10 mM β‐HB for 48 h. Cells were subjected to RNA‐seq analysis. (B) The KEGG enrichment analysis of RNA‐seq in β‐HB treated 5637 cells. (C–H) Gene Set Enrichment Analysis of G2M checkpoint pathway (C), E2F signaling pathway (D), IL2‐STAT5 signaling pathway (E), IL6‐JAK‐STAT3 signaling pathway (F), KRAS signaling pathway (G), and TGF‐β signaling pathway (H) in β‐HB treated 5637 cells. β‐HB Upregulates PD‐L1 in Bladder Cancer Cells PD‐L1 has been shown to play a crucial role in the metastatic progression of cancer. In recent years, immunotherapy has garnered increasing attention as a treatment option for bladder cancer, and our RNA sequencing results suggest that β‐HB may potentially influence the immune response of bladder cancer cells. Therefore, we further investigated the relationship between β‐HB and PD‐L1 in bladder cancer cells. We found that β‐HB treatment enhanced the mRNA transcription and protein expression of PD‐L1 in human‐derived bladder cancer cell lines T24 and 5637, as well as in the murine‐derived bladder cancer MB49 cell line, with the levels of upregulation exhibiting a concentration gradient correlation (Figure 5A–C). Cell immunofluorescence staining experiments also revealed a significant increase in the mean fluorescence intensity of PD‐L1 in cells treated with β‐HB, particularly on the cell membrane surface (Figure 5D–G). As expected, higher PD‐L1 membrane staining was also observed in KD group mice tumor samples (Figure 5H). As shown in our RNA‐seq results in Figure 4, β‐HB exerts a potential activating effect on the JAK2‐STAT3 pathway, which is a well‐established regulatory pathway for PD‐L1. To elucidate the regulatory mechanism of β‐HB on PD‐L1 expression, we assessed the activation levels of the JAK2‐STAT3 pathway using Western blot analysis. Consistent with our RNA sequencing findings, we observed that the phosphorylation levels of JAK2 and STAT3 were significantly upregulated (Figure 5I). Our findings suggest that β‐HB mediated activation of the JAK2‐STAT3 signaling pathway induces STAT3 phosphorylation, which subsequently functions as a transcription factor to enhance PD‐L1 gene transcription and protein expression, consistent with previous studies [35]. AG490 is a selective inhibitor of the JAK2‐STAT3 signaling pathway; treatment of BCa cells with AG490 resulted in a significant decrease in the expression of PD‐L1 after β‐HB treatment (Figure 5J,K). These results indicate that β‐HB regulates PD‐L1 expression through the activation of the STAT3 signaling pathway. β‐HB upregulates PD‐L1 in bladder cancer cells. (A) RT‐PCR analysis to show the CD274 gene's expression change in bladder cancer cells treated with β‐HB. β‐Actin served as an internal reference. (B, C) T24, 5637, and MB49 cells were treated with or without β‐HB (5, 10, and 20 mM) for 48 h. Western blot analysis of PD‐L1 protein expression. (D–G) Immunofluorescence staining showed PD‐L1 expression after treatment of β‐HB into T24 (D), 5637 (E), and MB49 (F) cells. (H) Immunohistochemical staining analysis of PD‐L1 expression in tumor cells in MB49 syngeneic models. (I) 5637 and T24 cells were treated with or without β‐HB (10 mM) for 48 h. Western blot analysis of JAK2‐STAT3 signaling pathway. (J, K) Western blot analysis of the effect of AG490 on PD‐L1 regulation of β‐HB. Each experiment was performed at least three times; all data are mean ± SD; **p < 0.01, ***p < 0.001, ****p < 0.0001, by unpaired Student's t‐test and one‐way ANOVA. β‐HB Exerts Anti‐Tumor Effects and Activates JAK2‐STAT3 Signaling at KD‐Induced Concentration In Vitro Given that systemic β‐HB levels in our in vivo KD model reached approximately 1.5 mM (Figures 1C and S1C), we conducted experiments at 1.5 mM β‐HB to validate whether the mechanistic findings observed at higher concentrations also occur at physiologically achievable levels. Our results revealed that β‐HB significantly inhibited bladder cancer cell viability and proliferation at 1.5 mM, although there was no significant generation of reactive oxygen species and mitochondrial damage (Figures 6A,B and S2A,B). β‐HB inhibits bladder cancer cells growth and activates JAK2‐STAT3 at KD‐induced concentration in vitro. (A) T24 and 5637 cells viability were decreased when grown in conditions with 1.5 mM β‐HB supplementation for 24 h. (B) T24 and 5637 were cultured in control or β‐HB supplemented medium with 1.5 mM and cell DNA synthesis was determined using an EdU assay. (C) Western blot analysis of the effect of GPR109A inhibitor on PD‐L1 regulation of different β‐HB concentrations. (D) Western blot analysis of the effect of GPR109A inhibitor on JAK2‐STAT3 pathway activation of different β‐HB concentrations. (E) Representative immunofluorescence images of STAT3 subcellular localization. Nuclei were stained with Hoechst 33342 (blue), and STAT3 was visualized in green. All data are mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001, by one‐way ANOVA. Crucially, Western blot analysis demonstrated that JAK2 and STAT3 phosphorylation were induced even at 1.5 mM β‐HB, albeit with lower magnitude compared to 10 mM treatments (Figure 6C,D). Immunofluorescence microscopy further confirmed STAT3 nuclear translocation at 1.5 mM β‐HB treatment and becoming more prominent at 10 mM (Figure 6E). This directly demonstrates that phosphorylated STAT3 translocates to the nucleus at physiologically relevant β‐HB concentrations and presumably engages PD‐L1 gene regulatory elements. These findings establish that the β‐HB‐induced PD‐L1 upregulation and JAK2‐STAT3 activation observed in our mechanistic studies are supraphysiological stress responses but rather represent genuine signaling events achievable through realistic dietary or supplemental ketogenic interventions. To identify the upstream receptor mediating β‐HB's effects, we investigated the role of GPR109A, a known β‐HB receptor on the cell surface, which has previously been reported to upregulate PD‐L1 and enhance anti‐PD‐1 therapy sensitivity via STAT3 activation [36]. Mepenzolate bromide, a selective GPR109A antagonist, was used to block GPR109A signaling [37]. Notably, GPR109A inhibition significantly reduced β‐HB‐induced JAK2 and STAT3 phosphorylation, as evidenced by Western blot analysis showing reduced p‐JAK2 and p‐STAT3 levels (Figure 6D). Correspondingly, GPR109A antagonism also attenuated the nuclear accumulation of STAT3 (Figure 6E) and suppressed β‐HB‐mediated PD‐L1 upregulation (Figure 6C). These findings strongly suggest that GPR109A plays a key role in mediating β‐HB's activation of the JAK2‐STAT3 pathway and subsequent PD‐L1 upregulation in bladder cancer cells. Ketogenic Diet and β‐HB Remodel the Tumor Microenvironment and Enhance the Efficacy of ICIs in Bladder Cancer Clinical studies indicate that tumor PD‐L1 expression can influence responses to anti‐PD‐1/PD‐L1 therapies. Patients with PD‐L1(+) tumors tend to exhibit higher response rates to immunotherapy and better overall survival rates [38, 39, 40, 41]. The tyrosine kinase inhibitor nintedanib has been shown to effectively enhance the response of colon cancer cells to anti‐PD‐L1 therapy by upregulating their PD‐L1 expression, thereby improving the therapeutic efficacy of ICIs and resulting in a favorable synergistic effect [42]. In melanoma, inhibition of the tricarboxylic acid cycle significantly increased the sensitivity to anti‐PD‐L1 immunotherapy through ATF3‐mediated PD‐L1 upregulation [43]. Based on these findings, combining drugs to enhance the level of PD‐L1 expression in cancer cells presents a promising strategy for increasing the sensitivity to immunotherapy. Given the upregulatory effect of KD and β‐HB on PD‐L1 in bladder cancer cells, we speculated whether the synergistic application of β‐HB with anti‐PD‐L1 could also alleviate potential immune resistance to some extent. Therefore, we tested whether KD or exogenous β‐HB supplementation (β‐HBi.p.) could enhance the efficacy of anti‐PD‐L1 therapy in vivo (Figure 7A). In MB49 syngeneic mice, KD and β‐HBi.p. each raised circulating β‐HB and β‐HBi.p. produced circulating β‐HB concentrations in the same range as KD group (Figure 7B). Exogenous β‐HB supplementation under normal diet significantly inhibited MB49 subcutaneous tumor growth compared with control group. The tumor growth inhibition by β‐HBi.p. closely matched that produced by KD. When combined with anti‐PD‐L1 therapy, β‐HBi.p. produced a significant enhancement of checkpoint blockade efficacy relative to anti‐PD‐L1 alone, similar to the enhancement observed with KD + anti‐PD‐L1. Representative IVIS images and endpoint tumor measurements are shown in Figure 7C–N. Ketogenic diet and β‐HB enhance the efficacy of ICIs in bladder cancer. (A) Experimental setting for combination of KD/β‐HBi.p. and ICIs in the MB49 model. (B) β‐HB levels in plasma from C57BL/6J mice. (C–H) KD ± αPD‐L1 experiment. (C) Representative in vivo bioluminescence images for the KD ± αPD‐L1 experiment. (D) Quantification of IVIS radiant efficiency for the groups shown in (C). (E) Comparative analysis of excised tumor specimens at study termination. (F) Tumor growth curves measured by caliper for each treatment group. (G) Endpoint tumor volumes for each group. (H) Endpoint tumor weights for each group. (I–N) β‐HBi.p. ± αPD‐L1 experiment. (I) Representative in vivo bioluminescence images for the β‐HBi.p. ± αPD‐L1 experiment. (J) Quantification of IVIS radiant efficiency for the groups shown in I. (K) Comparative analysis of excised tumor specimens at study termination. (L) Tumor growth curves measured by caliper for each treatment group. (M) Endpoint tumor volumes for each group. (N) Endpoint tumor weights for each group. Each experiment was performed at least three times, all data are mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, by one‐way ANOVA. To identify immune correlates, we analyzed tumor‐infiltrating leukocytes by flow cytometry. Compared with control group, tumors from β‐HBi.p. mice contained higher frequencies of immune effector populations, including CD8+ T cells and NK cells (Figure 8A,B). Conversely, immunosuppressive populations—CD11b+Gr‐1+ myeloid‐derived suppressor cells (MDSCs), a subset of CD11c+ dendritic cells, and CD4+FoxP3+ regulatory T cells—were reduced in β‐HBi.p. tumors relative to control group (Figure 8C–E). After short ex vivo stimulation, β‐HBi.p. tumors showed increased proportions of IFN‐γ+ and TNF‐α+ CD8+ T cells (Figure 8F,G). Combination treatment with anti‐PD‐L1 further amplified the increase in effector infiltration and cytokine production and further decreased suppressive subsets compared with antibody alone. In addition, ELISA results showed that the levels of IFN‐γ and TNF‐α increased in the β‐HBi.p. and combination treatment group (Figure 8H,I). β‐HB remodels the tumor microenvironment of bladder cancer. (A–E) Flow cytometry analysis of CD8+ T cells (A), NK cells (B), MDSCs (C), matured DCs (D) and Treg cells (E) in tumors after various treatments and quantification of flow cytometry analysis. (F–G) Flow cytometry analysis of IFN‐γ (F) and TNF‐α (G) expression levels in CD8+ T cells and quantification of flow cytometry analysis. (H, I) ELISA analysis of IFN‐γ (H) and TNF‐α (I) expression levels in CD8+ T cells after various treatments. Each experiment was performed at least three times, all data were shown as mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, by one‐way ANOVA. These findings indicate that KD and exogenous β‐HB supplementation each enhance anti‐PD‐L1 efficacy in the MB49 model and are associated with increased effector infiltration and function and decreased immunosuppressive cell populations in the tumor microenvironment. Discussion The ketogenic diet (KD) has attracted increasing attention as a potential adjunctive strategy in cancer therapy, supported by accumulating evidence from both preclinical and clinical studies. Preclinical research has demonstrated that the KD induces apoptosis in glioma cells and suppresses tumor growth in breast cancer models, providing proof‐of‐concept for its anti‐tumor potential [44, 45]. Clinical studies further corroborate these findings, with pilot studies indicating that KD combined with chemotherapy improves quality of life and reduces circulating tumor markers in patients with advanced cancers [46, 47, 48, 49]. Additionally, a randomized controlled trial demonstrated that KD can delay tumor progression to some extent [50]. The anticancer mechanisms of KD are multifaceted, encompassing glycolysis inhibition through glucose restriction, elevation of circulating ketone bodies, direct immunomodulation of the tumor microenvironment, and induction of cancer cell apoptosis. However, large‐scale randomized controlled trials remain needed to establish the clinical efficacy and safety profile of KD in cancer management. The ketogenic diet has been shown to exert anticancer effects in various malignancies including glioblastoma, colorectal cancer, and pancreatic cancer through mechanistically diverse pathways. These include energy restriction, modulation of the HOPX transcription factor, and participation in ferroptosis regulation [20, 51, 52]. However, the potential therapeutic application of KD in bladder cancer has not been previously characterized. Our study demonstrates that KD exerts significant tumor‐suppressive effects in a bladder cancer syngeneic model for the first time. Importantly, using immunodeficient BALB/c nude mice, we demonstrated that KD exerts anti‐tumor effects that are independent of adaptive immunity (Figure S1). This finding indicates that β‐HB possesses intrinsic anticancer properties directed at tumor cells themselves, rather than solely acting through immune mechanisms. This observation is consistent with our in vitro data showing that β‐HB directly inhibits bladder cancer cell proliferation, migration, and colony formation (Figure 2). Moreover, KD markedly elevated plasma β‐HB concentrations in mice and promoted apoptosis within tumor tissues. The Warburg effect, a cancer‐specific metabolic aberration characterized by preferential glycolysis under aerobic conditions [13, 20], establishes a therapeutic rationale for ketogenic diet. Supporting this concept, studies in colorectal and pancreatic adenocarcinoma models demonstrate dose‐dependent cytotoxic effects and proliferation arrest following β‐HB treatment [52, 53, 54]. Based on these findings, we examined the direct mechanisms by which β‐HB suppresses bladder cancer. In vitro experiments demonstrated that exogenous β‐HB treatment significantly decreased bladder cancer viability, impaired DNA synthesis and proliferative capacity, and attenuated cell migration and invasion (Figure 2). Mechanistically, β‐HB elevated intracellular ROS, which in turn disrupted mitochondrial membrane potential and depleted cellular ATP stores. ROS‐induced mitochondrial dysfunction coupled with energy depletion activates the apoptotic cascade through caspase‐3 activation (Figure 3). These findings establish that β‐HB exerts potent direct anti‐tumor effects through ROS‐mediated mitochondrial dysfunction and bioenergetic crisis. Immune checkpoint inhibitors (ICIs) have demonstrated significant therapeutic potential in the treatment of bladder cancer. However, the development of resistance remains a major clinical challenge [55]. A key mechanism underlying this resistance is the downregulation of PD‐L1 expression on tumor cells, which compromises the efficacy of ICIs by disrupting the PD‐L1/PD‐1 axis critical for modulating T‐cell activity. Evidence suggests that reduced PD‐L1 levels are associated with diminished treatment responses, as the loss of this immune checkpoint ligand enables tumor cells to evade immune surveillance. Furthermore, the complex tumor microenvironment, characterized by immune cell infiltration and the presence of immunosuppressive cytokines, may exacerbate PD‐L1 downregulation. Tumor heterogeneity and genetic alterations also contribute to variable PD‐L1 expression patterns, further complicating the therapeutic landscape [56, 57, 58]. Understanding the impact of PD‐L1 expression on ICIs therapeutic outcomes is essential for developing strategies to enhance ICIs efficacy and overcome resistance in bladder cancer. In this study, we revealed upregulated PD‐L1 expression in bladder tumor tissues from the KD group. Supporting this observation, in vitro experiments demonstrated that exogenous β‐HB stimulation directly activated both PD‐L1 mRNA transcription and protein expression in bladder cancer cells. RNA‐seq and Western blot analyses identified that this regulation was mediated through β‐HB induced activation of the JAK2‐STAT3 signaling pathway (Figure 5). A key consideration in interpreting our findings is the relationship between β‐HB concentrations used in in vitro experiments and those achieved physiologically. Our in vitro experiments employed 10 mM β‐HB, which notably exceeds the circulating levels (~1.5 mM) observed in KD‐fed mice. We selected this concentration based on established precedent in the cancer metabolism literature, where investigators typically examine 5–20 mM β‐HB to characterize its tumor suppression effects [52, 53, 54]. Critically, we found that 1.5 mM β‐HB, which approximates the plasma level achieved in our in vivo KD model, significantly inhibited bladder cancer cell viability and DNA synthesis, and notably activated JAK2‐STAT3 signaling with attendant STAT3 nuclear translocation. GPR109A antagonism substantially attenuated β‐HB‐induced JAK2‐STAT3 activation and PD‐L1 upregulation, implying that the GPR109A receptor plays a critical role in this signaling cascade (Figure 6). These findings establish that the biological effects observed at higher concentrations are not merely supraphysiological stress responses, but rather represent genuine dose‐dependent signaling events occurring at realistic in vivo concentrations. As a hollow visceral organ, the bladder is distinctly amenable to regional therapeutic administration. Intravesical chemotherapy and intravesical BCG immunotherapy represent well‐established approaches for bladder cancer treatment [59]. Theoretically, intravesical delivery of β‐HB or β‐HB‐generating substrates could achieve local tumor concentrations substantially exceeding systemic blood levels, potentially reaching the 5–20 mM range that demonstrates pronounced anti‐tumor effects in our experiments. Our data thus provide an important mechanistic foundation and preliminary rationale for future exploration of intravesical β‐HB‐based therapeutic approaches, though direct testing of this concept remains to be undertaken. PD‐L1 is widely recognized as a key mediator of tumor immune evasion, thereby promoting cancer progression and metastasis. Clinically, elevated PD‐L1 expression correlates strongly with enhanced tumor aggressiveness, metastatic potential, and reduced patient survival rates. However, there are also studies indicating that the overexpression of PD‐L1 on the surface of tumor cells was considered as a positive marker of response to ICIs treatment, PD‐L1 positivity in tumors is associated with improved objective response rates and prolonged progression‐free survival [60, 61]. Our in vivo results demonstrate that both KD and exogenous β‐HB supplementation (β‐HBi.p.) suppress MB49 tumor growth and, importantly, enhance the antitumor activity of PD‐L1 blockade. These treatment arms produced comparable increases in circulating β‐HB and similar magnitudes of tumor control, indicating that the principal circulating ketone can recapitulate many KD‐associated effects in this model. Analysis of tumor infiltrates revealed a consistent pattern of immune microenvironment remodeling after β‐HBi.p.: increased frequencies of immune effector populations, elevated effector cytokine production and reduced representation of several immunosuppressive subsets. Combination treatment with anti‐PD‐L1 further amplified effector's infiltration and function, and further diminished suppressive populations (Figure 8). Collectively, these data show that direct β‐HB exposure creates a more immunologically active, targetable tumor microenvironment that likely contributes to improved checkpoint blockade efficacy. While the present study provides multiple lines of evidence that KD/β‐HB favorably reprogram the tumor immune microenvironment and enhance PD‐L1 mediated immunotherapy response, several important limitations remain. First, while we validated KD and β‐HB effects in an independent MB49 cohort, additional validation in patient‐derived material will be important to establish generalizability. Second, although we identified GPR109A's involvement, the detailed mechanisms linking β‐HB receptor engagement to JAK2 activation remain to be elucidated. Third, although we measured β‐HB, we did not systematically evaluate the roles of other ketone species or of altered glucose/insulin signaling downstream of KD; additional metabolomic and flux analyses would clarify the full metabolic context. Future research should prioritize clinical translation of the ketogenic diet and β‐HB in bladder cancer, addressing interpatient heterogeneity in dietary responses and potential combinatorial efficacy with existing therapies. In summary, our study delineates that the ketogenic diet and its mediator β‐HB exert tumor suppression effect in bladder cancer through both direct tumor inhibition and immune modulation (Figure 9). These findings support further mechanistic studies and validation toward translational evaluation of ketogenic approaches as immunotherapy adjuvants. Schematic diagram of this study. Author Contributions Yuan Zeng: conceptualization, writing – review and editing, writing – original draft, formal analysis, data curation. Gangli Gu: conceptualization, supervision, funding acquisition. Lin Yang: formal analysis, data curation. Bohao Fan: methodology. Haochen Cui: methodology. Lei Yan: methodology. Zhao Liu: writing – review and editing, supervision, funding acquisition. Funding The authors declare financial support was received for the research, authorship, and publication of this article. This study was financially supported by the Natural Science Foundation of Shandong Province (nos. ZR2025MS1312, ZR2025MS1429, ZR2020QH068, and ZR2021MH049). Disclosure None of the authors of this manuscript are a current Editor or Editorial Board Member of Cancer Science. Ethics Statement Informed Consent: N/A. Registry and the Registration No. of the study/trial: N/A. Animal Studies: All animal procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC), Qilu Hospital of Shandong University (no. DWLL‐202500043). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1: Ketogenic diet inhibits bladder tumor growth in BALB/c nude mice. Figure S2: Effects of physiological β‐HB concentration on ROS production and mitochondrial membrane potential. Figure S3: Gated strategies for flow cytometry.