Introduction:

In 2023, 600,000 new cases of bladder cancer (BCa) were diagnosed worldwide and caused nearly 200,000 deaths. Most BCa patients present with non-muscle invasive urothelial cell carcinoma (NMIBC), for which the standard of care is transurethral resection followed by intravesical therapy. While therapies like mitomycin C, gemcitabine, and BCG induce remission in subsets of patients, they are associated with high recurrence and progression rates, adverse events, and global supply shortages (BCG). Thus many patients still undergo radical cystectomy (RC) as a salvage therapy. RC with urinary diversion is the standard of care for muscle-invasive BCa (MIBC) but is a life-changing and comorbid surgery and there is significant interest in bladder-sparing therapies for patients with organ-confined MIBC that are unfit or unwilling to undergo RC. Therefore, effective bladder-sparing treatments represent a major unmet clinical need in both high-risk NMIBC and MIBC.

Chimeric antigen receptor (CAR) T cell therapy, in which T cells are reprogrammed to express an artificial receptor to a known target, is an immunotherapeutic approach that is clinically effective in hematologic malignancies. Often described as ‘living drugs,’ CAR T cell therapies offer more specific and durable responses than systemic treatments and have the potential to induce long-lasting remissions. However, CAR T cell therapies have had limited success with solid tumors. The reasons for this are complex and multifactorial, including a paucity of antigens and failure of CAR T cells to traffic and persist in the tumor. Additionally, major systemic toxicities further limit the efficacy of CAR T cell therapies. A number of recent studies have reported locoregional cavitary delivery of CAR T cells to enhance trafficking into tumors and minimize systemic side effects. The barrier functions of the bladder normally limits systemic delivery of drugs to the bladder surface but also limits extravasation of intravesical contents systemically. This is a feature that is exploited in urologic practice to deliver high-doses of toxic therapeutics for enhanced efficacy and can potentially be used as a novel route of CAR T cell adoptive transfer for BCa.

We employed a simple antigen-discovery pipeline to identify MUC16 as a targetable antigen for BCa-specific CAR T cell therapy. MUC16 is a mucinous glycoprotein composed of several domains including a tandem repeat domain of >60 repeats of 156 amino acids. It displays pleiotropic functions including barrier, lubrication, cell adhesion, signaling functions which can promote metastasis, chemotherapeutic resistance, and tumorigenesis in various cancers. Concerning BCa, MUC16 expression correlates with tumor stage and grade, with published studies reporting detection rates of 30% in high-grade T1 tumors and up to 40% in MIBC. Further, MUC16 expression correlates with worse overall and recurrence-free survival, as well as chemotherapeutic resistance to both gemcitabine and cisplatin. To date, however, there are no pre-clinical or clinical studies that exploit MUC16 as a therapeutic target in BCa. We sought to investigate whether MUC16-targeted CAR T cells could be a therapy for BCa and whether the intravesical route of administration could be a viable mode of anti-tumoral adoptive CAR T cell transfer.

Methods:

Antigen Discovery Algorithm: To identify BCa-specific cell surface antigens, we used RNA-seq data from paired BCa tumors and tumor-adjacent normal bladder from The Cancer Genome Atlas (TCGA). DESeq2 was used to run differential expression analysis and extract genes in primary tumors versus normal bladder. Genes with log fold change (LFC)<2.5 or a p-value>0.01 were excluded. Cell surface encoding genes were identified using UniProt queries. Pan-normal tissue gene expression profiles were evaluated using RNA-seq data obtained through the Genotype-Tissue Expression Project (GTex). Validation data was sourced from Robertson et al. (GSE154261), Lindskrog et al. (EGAS00001004693), PURE01 (EGAC0001002276), de Jong et al. (https://github.com/CostelloLab/BRSpred), CCLE, and TCGA.

CAR Designs and CAR T Cell Production: Plasmids encoding CAR constructs in the SFG y-retroviral vector were used to transfect H29 cells. The retroviral supernatants were used to transduce GALV9 cells to generate stable viral packaging cell lines. Vectors were generated by restriction enzyme digest and Gibson Assembly. For the 3a5-28z CAR, the scFv for the anti-MUC16 3a5 monoclonal antibody was fused to a Myc tag, human CD28 transmembrane and intracellular domain, and human CD3z intracellular domain. For the MSLN-28z  CAR, a fragment human mesothelin (AA296-598) was similarly fused to the 28z backbone. Human T cells were isolated from healthy donors, activated with anti-CD3/CD28 beads in RPMI/10%FBS supplemented with 5ng/ml IL-7/IL-15, and subsequently transduced by spinoculation on retronectin-coated plates with retroviral supernatant from viral packaging cells. Transduction was confirmed at day 5 by expression of the Myc tag.

In vitro models: The following cell lines were used: 5637, T24, HT-1376, TCCSup, SV-HUC-1, UM-UC-3, UM-UC-7. MUC16 expression was evaluated by qRT-PCR and flow cytometric binding to anti-MUC16 mAbs and human mesothelin-Fc fusion proteins. For cytotoxicity assays, CAR T cells were co-cultured at varying E:T ratios with 1x10³ target luciferase-expressing tumor cells and killing quantified by luminescence. Supernatant cytokines were quantified by Luminex assay. For organoid experiments, patient-derived bladder tumor organoids (PDTO) were isolated from cystectomy specimens, stained with Cell Trace Far Red and seeded with media containing NucView 488 caspase-3 substrate (green) and CAR T cells at a 2:1 E:T ratio. Apoptotic PDTOs were identified as double-positive cells.

Animal model: Mice were housed under specific pathogen-free conditions and experiments were performed in accordance with IACUC-approved protocols. 6-8wk old female NSG mice from The Jackson Laboratory were used for xenograft experiments. For orthotopic tumor implants, mice were anesthetized and catheterized per urethra with a 24g angiocatheter, step-wise instilled with 0.01% poly-l-lysine (30 minutes), then 1x10⁶ HT-1376 cells (1h). Tumor bearing mice were randomized before experiments. For intravesical adoptive transfers, mice were instilled with 5x10⁶ CAR T cells (3h) via catheter. For intravenous adoptive transfers, 5x10⁶ CAR T cells were injected via tail-vein. At 7d, blood was sampled via retro-orbital access; blood for cellular analysis was collected into EDTA tubes. Serum was isolated from clotted blood and separated by centrifugation. Urine was obtained via urethral catheterization. Tumors were excised at pathologic or pre-specified endpoint of 60 days.

Results:

MUC16 is a Relevant Target for CAR T Cell Therapy in BCa

We used an antigen discovery pipeline using large-scale transcriptomics data from TCGA, leveraging the availability of paired normal bladder tissues. We identified differential gene transcripts and identified membrane-bound proteins not expressed on normal tissues (Figure 1A). The top candidate was MUC16 (Figure 1B), which is detected in 60% of BCa lines in the Cancer Cell Line Encyclopedia (Figure 1C), and in 20-40% of cases across the pathologic spectrum of BCa tumors from several datasets collectively representing 1,292 patients (Figure 1D). We also report that MUC16 is enriched in subsets of tumors with divergent differentiation or histologic subtypes (Figures 1E), and persists in NMIBC tumors that recur after BCG therapy (Figure 1F) and in MIBC tumors not responsive to neoadjuvant pembrolizumab (Figure 1G).

MUC16⁺ BCa Cells Can Be Targeted by Mesothelin-based CAR T cells

We validated MUC16 on multiple BCa cell lines by qRT-PCR and flow cytometric analysis (Figure 2A/B) and show mesothelin (MSLN), a differentiation antigen protein, binds to MUC16⁺ but not MUC16^- BCa cell lines (Figure 2A/B). This interaction is specific, as soluble CA-125, the cleaved form of MUC16, reduced surface binding of MSLN to MUC16⁺ HT-1376 cells in competitive binding assays (Figure 2B) and MSLN binding intensities correlate with normalized MUC16 transcripts in MUC16⁺ BCa cells (Figure 2B).

We next designed CD28-based second-generation MUC16-targeting CARs, including a novel CAR utilizing MSLN as the MUC16-binding moiety (Figure 2C). These CAR T cells lysed only MUC16⁺ cells, with MSLN-28z demonstrating consistently higher cytotoxicity and cytokine elaboration (Figure 2D/2E). Using orthogonal gain-of-function, loss-of-function, and protein cross-blocking approaches, MSLN-28z CAR T cells appear to require BCa cell-surface MUC16 expression, engagement, and intracellular signaling domains to exert cytotoxic functions, which is preserved in the presence of soluble CA-125 but can be inhibited through competitive binding by MSLN proteins. This specific cytotoxicity also extends to 3D patient-derived bladder tumor organoids, as the degree of organoid killing and T cell-derived cytokine production correspond with relative MUC16 expression levels.

MSLN-28z CAR T Cells Delivered Intravesically Control Tumor Outgrowth

We assessed the therapeutic potential of MSLN-28z CAR T cells in vivo using HT-1376 cells orthotopically implanted the bladders of NOD-SCID gamma (NSG) mice. A single dose of 5x10⁶ MSLN-28z CAR T cells delivered intravesically significantly increased survival compared to control T cells or the equivalent doses delivered intravenously. Study endpoint analysis demonstrates that mice treated intravesically with MSLN-28z CAR T cells exhibited smaller tumor burdens with higher infiltrating human CD45⁺ lymphocytes. Notably, CAR T cells administered intravesically were not detectable in the peripheral blood of mice, suggesting their activity is localized to the local bladder environment. Last, MSLN-28z CAR T cells are well-tolerated in mice, as no acute infusion-related systemic toxicities are observed via either mode of adoptive transfer, including changes in body weight, relevant hematologic, liver, and renal clinical chemistries, or pathological changes in major organs after 14d of transfer.

Conclusion:

In this study, we leveraged a computational antigen-discovery pipeline to identify MUC16 as a target for BCa-directed CAR T cell therapies. This identification was based on stringent criteria, including tumor specificity and minimal pan-tissue expression. MUC16, and its soluble form CA-125, have been described in the BCa literature for prognostic purposes. Our study evaluated MUC16 expression across BCa patient cohorts spanning the spectrum of pathologic T stages, NMIBC tumors recurrent after BCG, MIBC tumors non-responsive to neoadjuvant pembrolizumab, and variant molecular subtypes. These findings suggest that MUC16-targeting CAR T cells may have use in treatment-refractory BCa, providing a therapeutic option for patients who have limited alternatives besides RC.

We showed that MUC16⁺ BCa cells can be bound by both antibodies and mesothelin proteins. We evaluated a novel CAR, MSLN-28z, which demonstrated significant cytotoxicity and cytokine responses across multiple MUC16⁺ BCa cell lines. We showed that MUC16 is both necessary and sufficient for MSLN-28z CAR T cells to induce cytolysis and cytokine production. Through cross-blocking experiments, we also showed soluble CA-125—which can be elevated in the sera and urine of BCa patients—does not impede the cytotoxicity or specificity of MSLN-28z CAR T cells, similar to prior studies evaluating CAR T cells targeting tandem-repeat domains of MUC16. However, soluble MSLN proteins—which are rarely expressed in BCa—can impede cytotoxicity in vitro in a dose-dependent fashion. This confirms the specificity of MSLN-28z CAR T cells for MSLN-binding sites on MUC16, but could pose a clinical limitation of systemic MUC16-targeted cell therapies in cases of elevated soluble mesothelin.

Our study demonstrates that MSLN-28z CAR T cells exhibit significant anti-tumor activity in an in vivo orthotopic model of BCa. The intravesical delivery of these CAR T cells was particularly effective, outperforming systemic adoptive transfers at equivalent doses; and this improved efficacy correlates with increased T cell infiltrates. We speculate this effect is similar to intratumoral inoculation, likely due to increased CAR T cells at the site of the tumor, which are otherwise diluted systemically with intravenous adoptive transfer. Intravesical adoptive transfer resulted in lower systemic engraftment in the peripheral blood and accompanying T cell-derived cytokines could only be detected in the urine. Taken together, these findings suggest that intravesical administration enhances anti-tumor efficacy and minimizes systemic exposures. These observations underscore the effect of localized treatment approaches in organ-confined BCa, where intravesical administration can maximize therapeutic impact while minimizing toxicities and is well-studied with other intravesical therapies, which demonstrate localized adverse events with negligible systemic toxicities. Our pre-clinical study extends these observations to T cell-based therapies.

Feasibility studies for an intravesical MSLN-28z CAR T cell clinical trial in MUC16⁺ BCa are currently underway at our institution, however given the enrichment of MUC16 in invasive disease and variant molecular subtypes, which have increased metastatic potential, additional studies evaluating MUC16-targeted CAR T cell therapies are also warranted for patients with advanced metastatic BCa.    

Funding: N/A