Intravesical VSVd51-GM-CSF virotherapy is superior to BCG in treating bladder cancer in preclinical and translational models

Abstract

Non-muscle-invasive bladder cancer (NMIBC) can progress to muscle-invasive disease, with transurethral resection followed by Bacillus Calmette-Guérin (BCG) immunotherapy reducing this risk. Effective immunotherapies for BCG-resistant NMIBC are lacking. This study directly compares the efficacy of the oncolytic vesicular stomatitis virus (VSVd51) encoding the granulocyte macrophage colony-stimulating factor (GM-CSF) transgene (VSVd51-GM-CSF) to BCG in preclinical and translational models of aggressive bladder cancer. VSVd51-GM-CSF and BCG were tested in mouse and human bladder cancer spheroids and in bladder cancer patient-derived organoids, to evaluate immunogenic cell death biomarkers, cytokine release, and immune cell activation. VSVd51-GM-CSF and BCG treatments were then administered to C57Bl/6 mice with MB49 or N-butyl-N-(4-hydroxybutyl)-nitrosamine(BBN)-induced bladder tumors via intravesical instillation. VSVd51-GM-CSF treatment induced a heightened release of immunogenic factors and cytokines, which then activated M1-like tumor-targeting monocytes. Mice treated with VSVd51-GM-CSF exhibited stronger tumor-infiltrating immune responses, longer survival, and reduced tumor volume compared to BCG-treated mice. Importantly, VSVd51-GM-CSF treatment extends survival in BCG-failed mice. This anti-tumor immunity was also observed in patient-derived organoids, suggesting clinical relevance. These translational findings suggest that VSVd51-GM-CSF has significant potential for early-phase clinical trials in NMIBC patients. As a promising viro-immunotherapy, it could provide an alternative for patients with BCG-resistant disease, marking an important step forward in bladder cancer immunotherapy.

Graphical abstract

Tai and colleagues compare a novel oncolytic virus to BCG, the standard treatment for bladder cancer. In animal models, the virus triggered stronger immune responses and better tumor control than BCG, even in BCG-resistant tumors, highlighting its potential as a next-generation therapy for difficult-to-treat bladder cancer.

Introduction

Bladder cancer (BC) ranks among the most prevalent malignancies globally, with non-muscle-invasive bladder cancer (NMIBC) representing a significant proportion of cases.¹ At diagnosis, 75% of cases are NMIBC, but many progress to muscle-invasive disease.^2,3 Standard treatment for high-risk NMIBC includes tumor resection and intravesical Bacillus Calmette-Guérin (BCG) immunotherapy, a 50-year-old technology, but recurrence and progression remain high (40%–60%), as many high-risk patients will not respond to BCG.^1,4,5,6,7 For these patients, radical cystectomy (bladder removal) is often required, with a 5-year survival rate of 50%.⁴ This life-altering surgery leads to major long-term morbidity. Patients who progress after cystectomy face a median survival of 12–15 months with salvage chemotherapy or 31.5 months with systemic immune checkpoint inhibitor immunotherapy once metastatic.⁸ BC has one of the highest lifetime costs due to the need for close surveillance and the high recurrence and progression rates, which necessitate multiple lines of treatment.^2,4,5,9 Emerging concerns over BCG shortages and failure rates have prompted exploration into alternative immunotherapies and targeted treatments to replace BCG, among which novel virotherapy has emerged as a promising therapeutic avenue.

Oncolytic viruses (OVs) offer a promising bladder-sparing treatment for BC by exploiting the cancer’s vulnerability to immune-mediated control.¹⁰ In the context of NMIBC, OVs are an ideal immunotherapy for BC because intravesical BCG therapy has been used for 50+ years in BC patients, providing feasibility for a direct bladder instillation approach. BC can also be exposed to high virus titers through instillation without off-target effects, and the papillary structure of BC increases the surface area for virus topical exposure. Importantly, unlike BCG, immune checkpoint inhibitors, gene therapy, and antibody drug conjugates, OVs are live, self-amplifying mobile pharmacies that actively spread through tumor tissues, deliver therapeutic payloads, and induce immunogenic cell death (ICD) following targeted tumor cell lysis.^10,11,12,13 Various OVs, including adenovirus, coxsackievirus, and vesicular stomatitis virus (VSV), have been studied in BC.¹⁰ VSV selectively targets interferon (IFN)-resistant cells, making it particularly effective against aggressive NMIBC.^11,12,13 An attenuated strain, VSVd51, with enhanced tumor specificity and safety due to a mutation in its matrix protein has been developed.¹⁴ VSVd51 circumvents challenges posed by pre-existing neutralizing antibodies (nAbs) and genotoxicity encountered by other OVs in preclinical and clinical settings.^14,15

Multiple VSVd51 variants with cytokine transgenes, including granulocyte macrophage colony-stimulating factor (GM-CSF), have shown enhanced oncolysis in preclinical models of melanoma.¹⁶ In human studies, oncolytic herpes simplex virus with GM-CSF was the first OV approved for end-stage melanoma.¹⁷ We recently engineered a VSVd51 variant with human GM-CSF for BC treatment. Both mouse and human VSVd51-GM-CSF variants increased oncolysis and ICD markers expression in BC cell lines compared to the parental virus. In the C57Bl/6-MB49 mouse model, VSVd51-GM-CSF improved tumor-infiltrating lymphocytes responses and survival. In patient-derived organoids (PDO), immune activation was enhanced compared to the parental virus.¹²

The observed response rates to novel immunotherapies and historical responses to BCG underscore that BC is a cancer susceptible to immune-mediated control.¹⁸ Perhaps, then, the ideal immunotherapy solution for patients high-risk NMIBC will require targeting, enhancing, and balancing anti-tumor immune functions. Virotherapy presents a rational and precision strategy to specifically target bladder tumor cells while minimizing systemic toxicity. Despite its theoretical advantages, the therapeutic efficacy of virotherapy in direct comparison to frontline BCG therapy and in the context of BCG failure remains to be investigated. In this report, we first delve into the immune mechanisms underlying both BCG therapy and virotherapy, using in vitro cellular models. Subsequently, we leverage two in vivo mouse models of BC and BC PDOs, evaluating the comparative efficacy and immune profiles of these therapies.

Results

VSVd51-GM-CSF treatment reduces cell viability and enhances ICD compared to BCG

There is strong evidence to suggest that the immune system influences the outcomes of both BCG and VSV immunotherapy.^6,18,19 We therefore compared the immunogenic potential of BCG and VSVd51-GM-CSF in inducing ICD in mouse and four human BC cell lines. VSVd51-GM-CSF induced significantly greater cell death than BCG (Figure 1A). Compared to BCG and controls, VSVd51-GM-CSF also significantly increased adenosine triphosphate (ATP) release (Figure 1B), ecto-calreticulin expression (Figure 1C), and the release of high-mobility group box 1 (HMGB1) and heat shock protein-90 (HSP90) (Figure 1D). BC cellular cytotoxicity induced by VSVd51-GM-CSF was also conducted using a range of multiplicity of infections (MOIs) (Figure S1). Together, the significantly higher levels of cellular death and presence of heightened danger-associated molecular patterns suggest a greater induction of ICD by VSVd51-GM-CSF.

Figure 1.

VSVd51-GM-CSF treatment reduces cell viability and enhances immunogenic cell death compared to BCG

(A) Cytotoxicity of BCG or VSVd51-GM-CSF was assessed in MB49 (mouse), T24 (human), TCCSUP (human), UM-UC-3 (human), and 5637 (human) BC cells at 48 h after infection at different MOI. (B) Luminometry measurement of ATP from cell-free supernatants, (C) measurement of cell surface calreticulin, and (D) western blot analysis of HMGB1 and HSP90 from cell-free supernatants of indicated BC cell lines infected with BCG (10 CFU) or VSVd51-GM-CSF (10 MOI) harvested 24 h after infection. All data are representative of at least three similar experiments where n = 3 for biological replicates, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, no significance.

Elevated M1-like macrophage markers on monocytes following exposure of human BC spheroids to VSVd51-GM-CSF

To test whether the enhanced ICD from BCG or VSVd51-GM-CSF infection translates to innate immune cell activation, we measured monocyte polarization in human BC cells. First, we grew all 4 human BC cell lines as 3D spheroids instead of in 2D monolayers to better mimic the physiology of the bladder urothelium. From BCG- or VSVd51-GM-CSF-infected spheroids, we collected conditioned media (CMs) and added them onto purified human CD14⁺ monocytes. Monocytes exposed to VSVd51-GM-CSF CM polarized toward an M1-like phenotype with higher CD80 and CD86 and lower PD-L1 expression, promoting anti-tumor responses,²⁰ compared to BCG and controls. In contrast, BCG CM induced more M2-like markers (CD163 and CD206), linked to tumor promotion²¹ (Figures 2A–2C).

Figure 2.

Activation of innate myeloid cells following exposure to VSVd51-GM-CSF in human BC spheroids

(A–C) Polarization and (D) depolarization of purified human monocytes in the presence of CM from human BC spheroids (T24, TCCSUP, UM-UC3, and 5637) infected with BCG or VSVd51-GM-CSF at 10 MOI and harvested 24 h post-infection. Controls include resting monocytes (M0); LPS (lipopolysaccharide)- and IFN-γ-treated monocytes (M1); and IL-4-, IL-10-, and transforming growth factor β-treated monocytes (M2). All data are representative of at least three similar experiments where n = 3 for biological replicates. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, no significance.

To further assess the mechanism of monocyte activation, we directly infected purified human monocytes with VSVd51-GM-CSF, recombinant GM-CSF, and UV-treated CM (virus) and observed a comparable upregulation of CD80 expression relative to CM-treated monocytes following VSVd51-GM-CSF treatment but diminished CD80 levels using recombinant GM-CSF or UV-treated CM (Figure S2). This finding suggests that VSVd51-GM-CSF can directly stimulate monocytes, in addition to activating them indirectly through tumor-derived soluble factors.

We also performed a monocyte depolarization assay where CD14⁺ human monocytes were first polarized to M2-like monocytes to more closely resemble tumor-infiltrating innate cells that are tumorigenic or immunosuppressive in the bladder tumor microenvironment (TME). Then we added CMs from the various conditions to these M2-like monocytes to assess whether immune-activating soluble factors from treated tumors can reverse this M2-like monocyte phenotype. From these cultures, we observed that CM from VSVd51-GM-CSF-infected spheroids was able to depolarize M2-like monocytes toward M1-like monocyte phenotypes compared to CM from BCG-infected spheroids (Figure 2D). These findings indicate that human BC cells infected with VSVd51-GM-CSF release a combination of immunostimulatory factors, including virus-derived signals and GM-CSF that collectively enhance innate immune activation.

VSVd51-hGM-CSF infection of human BC PDOs promotes ICD and autologous immune cell activation

To evaluate the clinical potential of VSVd51-GM-CSF, we tested its ability to induce an immunogenic signature in human BC patient tissues. Using 2 BC PDO lines established in our lab (BLC-38 and BLC-35), we treated them with BCG and VSVd51-GM-CSF or left them untreated. We also generated CM and cultured this with autologous (matched) peripheral blood mononuclear cells (PBMCs). In both PDO lines, we observed significant cell death in VSVd51-GM-CSF-infected PDOs (Figures 3A and 3B) and increased relative ATP release (Figure 3C), compared to BCG or controls. In cocultures with autologous PBMCs, we detected lower PD-L1⁺ exhaustion of dendritic cells (DCs) and higher natural killer (NK) cell CD107a degranulation and Perforin⁺/Granzyme B⁺ cytotoxicity markers following the addition of CM from VSVd51-GM-CSF-treated PDOs (Figures 3D and 3E). These translational results demonstrate that an ICD signature is present in BC PDOs following VSVd51-hGM-CSF infection and this phenotype has the capacity to activate autologous immune cells ex vivo.

Figure 3.

VSVd51-GM-CSF enhances immunogenic cell death and autologous immune cell activation in human BC patient-derived organoids

ICD and immune activation of BC PDOs from patient BLC-38 and BLC-35 were observed through (A) light microscopy and cytotoxicity quantified through (B) flow cytometry and (C) relative luminometry measurement of ATP. (D and E) Activation and functionality of autologous human CD14⁺ monocytes and CD56⁺/CD3⁻ NK cells in the presence of CM from infected human BC PDOs. Data are pooled from biological replicates, n = 3, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns, no significance. LPS, lipopolysaccharide.

VSVd51-GM-CSF enhances pro-inflammatory cytokine and chemokine secretion in bladder PDOs compared to BCG treatment

To investigate the role of cytokines and chemokines in the bladder TME, we analyzed the secretome of CM from BCG- and VSVd51-GM-CSF-infected PDOs using a panel of 48 inflammatory markers. VSVd51-GM-CSF treatment increased levels of GM-CSF, IFN-γ, interleukin (IL)-1α, IL-4, IL-8, IL-15, sCD40L, and tumor necrosis factor alpha compared to BCG and controls (Figures 4A and 4B). We also detected modest increases in CCL11/24/46, IL-12, and IL-13 in patient 38. Additionally, VSVd51-GM-CSF-treated PDOs showed higher levels of monocyte/macrophage-attracting chemokines CXCL9, CXCL10, MCP-1, MCP-3, and CCL5 (Figures 4A and 4B). These findings suggest that VSVd51-GM-CSF treatment profoundly reshapes the bladder TME by enhancing the secretion of pro-inflammatory cytokines and chemokines.

Figure 4.

VSVd51-GM-CSF enhances pro-inflammatory cytokine and chemokine secretion in bladder PDOs compared to BCG treatment

(A and B) ELISA analysis of 48 cytokines/chemokines from the cell-free supernatants of infected PDOs. PDOs were infected with BCG or VSVd51-GM-CSF at 10 MOI and harvested after 24 h. y axis breaks were used to allow visualization of differences among lower-expressed cytokines in the presence of extremely high GM-CSF levels due to transgene overexpression. Axis scaling was preserved across groups to allow fair comparison. Data are pooled from biological replicates, n = 3, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns, no significance.

VSVd51-GM-CSF treatment enhances immune cell activation compared to BCG in the C57Bl/6-MB49 mouse model

To assess whether the observed ICD and immune activation in our in vitro models translate to enhanced immune function in vivo, we compared intravesical VSVd51-GM-CSF and BCG treatment in C57Bl/6 mice with orthotopic MB49 tumors (Figure 5A). MB49 is one of the most-used murine BC cell lines and shares pivotal immunological and cell surface characteristics with aggressive human BC.^10,11,12,13 Treatment was initiated on day 2 after implantation following verification for tumor growth using small animal ultrasound (Figure S3). We conducted flow cytometric analysis on harvested and dissociated bladders to analyze tumor innate immune cells 3 days post-treatment. In VSVd51-GM-CSF-treated mice, we observed a significant increase in CD86⁺/F4-80⁺ M1 macrophages and a decrease in CD206⁺/F4-80⁺ M2 macrophages and granulocytic myeloid-derived suppressor cells (CD11b⁺/Ly6C^low/Ly6G^hi) compared to BCG and controls (Figure 5B). To measure adaptive immunity, we examined adaptive tumor immune cells at day 6 following treatment. In VSVd51-GM-CSF-treated mice, we detected augmented CD3⁺/CD8⁺ T cells and CD3^-/NK1.1⁺ NK cells frequencies compared to BCG treatment and controls (Figure 5C). Furthermore, tumor NK cells expressed higher levels of the CD107a degranulation marker compared to BCG treatment and controls (Figure 5C).

Figure 5.

Enhanced innate and adaptive immune cell activation following VSVd51-GM-CSF treatment in the syngeneic C57Bl/6-MB49 mouse model

(A) Timeline of the in vivo C57Bl/6-MB49 experiment. C57Bl/6 mice bladders were instilled with 1 × 10⁶ MB49 cells. Two days later, each group of mice received via intravesical instillation 50 μL of 5 × 10⁸ CFUs of BCG or 5 × 10⁸ PFU VSVd51-GM-CSF or vehicle for control groups. Innate immune cells (B) were assessed from dissociated bladder tumors at day 5 post-tumor implantation. Adaptive immune cells (C) were assessed from dissociated bladder tumors at day 7 post-tumor implantation. Immune cell suspensions from the dissociated bladders of mice following indicated treatments were stained with (B) myeloid markers (CD45⁺, CD11b⁺, CD86⁺, CD206⁺, F4/80⁺, Ly6C^low, and Ly6G^hi) and (C) NK cell markers (NK1.1⁺, CD3⁻, and CD107a⁺) and analyzed by flow cytometry. All flow cytometry data are representative of at least three similar experiments where n = 4–10 mice/treatment. A cohort of mice was monitored for Kaplan-Meier survival analysis (D) and tumor size progression (E). n = 8 mice/group. ∗p < 0.05; ∗∗p < 0.01; ns, no significance, log rank test. For Figure 5E, # represents BCG vs. VSVd51-GMCSG significance (#: p < 0.05; ##: p < 0.01); †† represent PBS vs. VSVd51-GM-CSF significance (††: p < 0.01).

To strengthen the link between immune modulation and therapeutic response, we performed a survival and tumor control study in the same MB49 orthotopic model. Although this tumor is highly aggressive and fast growing, reaching human endpoints around day 17 post-instillation, it offers a robust short-term window for assessing treatment efficacy. We observed that mice treated with VSVd51-GM-CSF demonstrated the longest survival (Figure 5D) and the most effective tumor control compared to BCG and untreated controls (Figure 5E). Taken together, these results suggest that VSVd51-GM-CSF treatment activates and expands both innate and adaptive tumor-infiltrating immune cells compared to BCG treatment.

Diminished bladder tumor burden and improved survival in the C57Bl/6-BBN model following VSVd51-GM-CSF treatment in frontline treatment against BCG and in second-line treatment of BCG-failed disease

To determine if the enhanced immune signatures in treated mice lead to better disease outcomes, we monitored survival and tumor volume using small animal ultrasound. Here, we used a second model, the C57Bl/6-BBN (N-butyl-N-(4-hydroxybutyl)-nitrosamine), a carcinogen-induced model for the spontaneous growth of aggressive BC. The C57Bl/6-BBN model recapitulates in vivo tobacco-induced BC development, which is the most common risk factor in industrialized countries where BC is highly prevalent. In addition, the BBN-induced model exhibits considerable tumor heterogeneity²² and has been used for studies on BCG response²³ (Figure 6A). In these mice, VSVd51-GM-CSF treatment reduced bladder tumor volume (Figures 6B, 6C, and S4A) and improved survival compared to BCG (Figure 6D). To mimic human BC patients who continue smoking (carcinogen exposure) while on treatment, we introduced a third treatment group where we maintained BBN exposure beyond week 12, while administering VSVd51-GM-CSF treatment. In this group, the survival benefit of VSV treatment was entirely lost (Figure 6D). While we are keen on replacing BCG with VSVd51-GM-CSF in frontline treatment for NMIBC, we acknowledge that the faster path to translation is to test the therapeutic potential of VSVd51-GM-CSF in the BCG-failed context. We, therefore, first treated BBN-tumor-bearing mice with BCG. This resulted in mice that responded to BCG (tumor volume reduction) and failed BCG (continued tumor growth despite 3 weeks of treatment). VSVd51-GM-CSF was then administered to BCG-failed mice, and their tumor growth and survival were compared to BCG responders. VSVd51-GM-CSF-treated mice that had failed BCG showed similar tumor control (Figures 6F and S4B) and survival when compared to BCG responders (Figure 6E). These in vivo findings suggest that VSVd51-GM-CSF is more effective than BCG in frontline therapy for aggressive BC and highlights the detrimental impact of continued carcinogen exposure on treatment efficacy and prognosis. Furthermore, these results demonstrate that VSVd51-GM-CSF can rescue BCG-failed disease and serve as a potential bladder-sparing treatment for NMIBC.

Figure 6.

Diminished bladder tumor burden and improved survival in the C57Bl/6-BBN model following VSVd51-GM-CSF treatment against BCG and in BCG-failed disease

(A) Timeline of the in vivo C57Bl/6-BBN experiment. Mice were fed on 0.5% BBN in their drinking water for 12 weeks to induce spontaneous BC. Once tumor volume reached 5–10 mm³ as measured by small animal ultrasound, mice received 5 × 10⁸ CFU of BCG (1 dose per week for 3 weeks) or 5 × 10⁸ PFUs of VSVd51-GM-CSF (3 doses every 2 days). All mice were monitored for tumor size progression and survival. (B) Representative bladder tumor volume image at weeks 3 and 7 following BCG or VSVd51-GM-CSF treatment by small animal ultrasound. (C) Longitudinal tumor volume assessment over time by small animal ultrasound in mice following indicated treatments. (D and E) Kaplan-Meier survival analysis of mice treated with indicated therapies. In (D), a cohort of mice was maintained on BBN. In (E), VSVd51-GM-CSF treatment started after mice failed BCG treatment. (F) Longitudinal tumor volume assessment of treated mice following BCG failure. n = 9–10 mice/group. ∗p < 0.05; ∗∗p < 0.01; ns, no significance, log rank test.

Discussion

This study offers preclinical evidence that VSVd51-GM-CSF is a promising alternative to BCG for treating aggressive NMIBC. While BCG is the standard of care for high-risk NMIBC, its limitations as a non-specific inflammatory mediator are evident, particularly in BCG-unresponsive cases.^9,24 Our findings show robust immune activation with VSVd51-GM-CSF and its superior ability to induce anti-tumor immunity. A key strength of our research is the use of 2 clinically relevant mouse models of aggressive BC and human PDOs, providing a comprehensive evaluation of the efficacy of VSVd51-GM-CSF. The increased release of ICD biomarkers by BC spheroids following VSVd51-GM-CSF infection corresponds with enhanced recruitment and activation of M1-like monocytes, crucial for anti-tumor immunity. Notably, monocyte activation was also observed following direct infection with VSVd51-GM-CSF, but not with recombinant GM-CSF or UV-inactivated CM, suggesting that both viral sensing and tumor-derived signals contribute to innate immune activation (Figure S2). This mechanistic insight highlights VSVd51-GM-CSF as a potent immune stimulator that elicits a strong anti-tumor response.

Moreover, the in vivo MB49 and BBN results showing improved survival and tumor control in mice treated with VSVd51-GM-CSF in frontline against BCG and in BCG failure further support its potential therapeutic value. The observed activation of tumor-infiltrating effector cells underscores the virus’ ability to enhance localized anti-tumor immunity, which is crucial for controlling aggressive BC. This is especially significant when considering the translational relevance of our findings, as the virus-mediated immune response was also recapitulated in BC PDOs. Although PD-L1 can reflect late activation, the concurrent upregulation of CD80/CD86 and enhanced NK activity suggest functional activation of autologous DC rather than exhaustion. Nonetheless, further time course experiments are warranted, and these results may hold promise in improving anti-tumor immunity in clinical settings.

Despite the promising results, several questions remain. The long-term safety of VSVd51-GM-CSF treatment in humans needs further evaluation, particularly regarding the potential for viral dissemination or off-target effects. However, all of our in vivo administration of VSVd51-GM-CSF uses a direct bladder instillation approach. This prevents the virus from systemic infiltration, and no off-target effects or toxicity has been observed in our treated animals. Additionally, the impact of continued carcinogen exposure, as seen in mice that were maintained on BBN water, appears to compromise the efficacy of VSVd51-GM-CSF. These findings highlight the importance of considering patient lifestyle factors that could negatively influence treatment outcomes.²⁵

Looking forward, future studies should focus on the integration of VSVd51-GM-CSF into current BC treatment regimens, potentially in combination with immune checkpoint inhibitors, antibody drug conjugates, or other targeted therapies. The use of this OV in combination therapy could further enhance its anti-tumor activity, potentially providing a more comprehensive and durable response for patients with high-risk or recurrent BC.

Oncolytic adenovirus (oAd5) expressing GM-CSF—CG0070—is currently in phase 3 studies for high-risk BCG-failed NMIBC patients.¹¹ CG0070 is built on the Ad5 serotype that has infected the majority of the human population. Thus, the clinical dosing of Ad5 in patients could be hampered by pre-existing nAbs.^26,27 Baseline levels of Ad5 nAbs have been shown to suppress the immunogenicity of Ad5-based vaccines for HIV-1 and SARS-CoV-2 in clinical trials.^20,28 While pre-existing systemic nAbs are less relevant for intravesical delivery, local humoral responses may still arise after repeated instillations. In a phase 2 trial, 21 instillation doses of CG0070 at 1 × 10¹² plaque-forming units (PFUs) along with 34 doses of immune checkpoint inhibitor are administered to NMIBC patients to achieve a therapeutic effect.²⁹ This implies that an exceptionally high viral load is required to achieve saturation of nAbs. For future clinical application, the VSV rhabdovirus family is infrequently linked to human diseases, meaning that most populations worldwide are unlikely to have pre-existing nAb. This reduces the likelihood of a humoral response limiting its efficacy, a significant challenge faced by oncolytic adenoviruses in BC treatment. Finally, unlike DNA viruses (Ad), VSVd51’s replication occurs entirely within the cytoplasm of infected cells, without inducing genotoxicity. However, genotoxicity is primarily a concern for systemic OV delivery, and intravesical delivery mitigates many of these risks. In future preclinical and clinical studies, we expect to differentiate VSVd51-GM-CSF from CG0070 by demonstrating better in vivo efficacy when using the same or lower number of doses at the same or lower concentrations, resulting in diminished bladder tumor burden and improved survival.

In conclusion, our study demonstrates that VSVd51-GM-CSF represents a promising viro-immunotherapeutic with the potential to surpass BCG in both front- and second-line treatment of aggressive BC. The ability of VSVd51-GM-CSF to induce strong, multi-faceted immune responses makes it an attractive candidate for future clinical trials, with the ultimate goal of improving outcomes for patients with aggressive NMBIC or BCG-resistant NMIBC.

Materials and methods

Cell lines and viruses

The mouse BC cell line MB49 was cultured in DMEM, while human BC cell lines 5637 in RPMI, UM-UC-3 and TCCSUP in DMEM, and T24 in modified McCoy’s 5A medium, all supplemented with 10% fetal bovine serum and 100 U/mL penicillin and 100 mg/mL streptomycin (complete media). Human cell lines were purchased from ATCC, and MB49 from MilliporeSigma; all were verified mycoplasma free and showed proper microscopic morphology. VSVd51-hGM-CSF was cloned from parental VSVd51,¹² and VSVd51-mGM-CSF was obtained from the Ottawa Hospital Research Institute (Ottawa, Canada).¹⁶ VSV-hGM-CSF was used on human cell lines and organoids, while VSVd51-mGM-CSF was used on the mouse cell line and in mouse models. All viruses were propagated on Vero cells, purified using OptiPrep methods, and titers determined by plaque assay. Viral cytotoxicity was assessed using the non-radioactive MTT assay (Promega) as previously described.¹²

Mice

Female and male C57Bl/6 mice (6–8 weeks old, 20–25g) were purchased from Jackson Laboratory (USA). Animals were housed in pathogen-free conditions at the Central Animal Facility of the Université de Sherbrooke with free access to food/water. Animals were euthanized by cervical dislocation under anesthesia. All studies were conducted in accordance with university guidelines and the Canadian Council on Animal Care, and protocols were approved by the Faculty Animal Care Committee.

C57Bl/6-MB49 syngeneic mouse model of BC

For orthotopic BC implantation, mice were anesthetized, and chemical lesions were induced by intravesical instillation of trypsin (Wisent) mixed 1:1 with DMEM. During this procedure, mice were maintained under anesthesia (3% induction, 1.5% maintenance isoflurane with 2% O₂).^12,30 Subsequently, 1 × 10⁶ MB49 BC cells were instilled for the immune cell analysis, and 5 × 10⁵ MB49 BC cells for the survival analysis. Mice were randomized by weight. Two days later, each group received two intravesical instillations of either 50 μL of 5 × 10⁸ PFUs of VSVd51-mGM-CSF¹² or 1 × 10⁸–7 × 10⁸ colony-forming units (CFUs) of BCG.^31,32,33 Screening of tumor-bearing mice prior to initiation of treatments was conducted by ultrasound to ensure equal tumor burdens. Bladder tumor growth was monitored bi-weekly using small animal ultrasound (Vevo 3100, VisualSonics).

C57Bl/6-BBN carcinogen-induced model of BC

Mice were given drinking water containing 0.5% BBN (TCI America) for 12 weeks to induce spontaneous BC.²³ Once tumor volume reached 5–10 mm³, mice received intravesical instillation of 50 μL of 5 × 10⁸ PFUs of VSVd51-GM-CSF (3 doses every 2 days in the bladder under anesthesia for 45 min) or 5 × 10⁶ to 4 × 10⁷ CFUs of BCG (3 doses once a week in the bladder under anesthesia for 2 h).²³ All mice were monitored for tumor progression and survival.

Mouse bladder tumor dissociation

Following euthanasia, bladders were immediately placed in complete RPMI and processed using the mouse tumor dissociation kit (Miltenyi Biotec). Briefly, bladders were minced (<2 mm³), treated with dissociation enzymes, and processed in the gentleMACS Octo dissociator using the mouse tumor dissociation kit (Miltenyi Biotec). Following dissociation, macroscopic fragments were removed with a 70 μm nylon cell strainer. The resulting single-cell suspensions were washed twice in complete RPMI before flow cytometry acquisition. Immune profiling was performed on bladder tissue by first gating on CD45⁺ cells after cell staining and fixation as described in the following.

Flow cytometry

For mouse splenocytes, cells were first incubated in ACK (Ammonium-Chloride-Potassium) lysis buffer for 5 min to lyse red blood cells. For spleen and tumor, 1 × 10⁶ cells were added to each well of a round-bottom 96-well plate. Dead cells were excluded using the LIVE/DEAD fixable near-IR dead cell stain kit (Invitrogen). After a 5-min Fc block at 4°C, cell surface antibodies were added. Cells were then fixed and permeabilized with the eBioscience Foxp3/transcription factor staining buffer set before intracellular staining. For assessment of NK and T cell functionality, cells were cultured with PMA (phorbol 12-myristate 13-acetate)/ionomycin (Sigma) for 4 h in the presence of brefeldin A (1:1,000) at 37°C. Then, cells were washed twice with PBS and stained as described. Samples were acquired on a CytoFLEX 20 (Beckman Coulter). Data were analyzed using FlowJo v.10 software (Table S1 for antibodies).

ICD assays

Calreticulin, HMGB1, and ATP measurements were conducted as previously described.³⁴ Briefly, calreticulin exposure was measured by flow cytometry; HMGB1 and HSP90 protein and ATP levels were measured by western blot and luminometry, respectively (Table S1 for all antibodies).^12,34

Human monocyte de/polarization assays

For spheroid cultures, CMs were obtained by resuspending 2.5 × 10⁴ cells in 20 μL of Matrigel (Corning) per well of a 48-well (Thermo Fisher Scientific) plate for 6 days followed by infection with BCG and VSVd51-GM-CSF at indicated CFU or MOI. Human monocytes were isolated from peripheral blood (human CD14⁺ isolation kit, STEMCELL). 1 × 10⁵ monocytes were seeded in 24-well plates in complete RPMI and incubated overnight at 37°C and 5% CO₂. 24 h later, the monocyte media were replaced with the CM of infected human spheroids. For monocyte depolarization assays, CD14⁺ monocytes were first polarized to M2-like monocytes for 24 h. Controls, including cytokines used to depolarize monocytes, were used as previously published.^12,34

CM cytokine analysis

Cell-free supernatants from control, BCG-, or VSVd51-GM-CSF-treated PDOs were analyzed for cytokine content using a human cytokine/chemokine panel 48-plex discovery assay (Eve Technologies).

Human BC PDOs

Our human protocol number is: REB (research ethics board)#: 2018–2465, provided by CIUSSS (Centre intégré universitaire de santé et de services sociaux de l'Estrie) de l’Estrie - CHUS (centre hospitalier universitaire de Sherbrooke). Patients or the public were not involved in the design, conduct, reporting, or dissemination plans of our research. BCL035 and BLC038 tissue from NMIBC patients were collected after resection and placed in complete DMEM. Tumors were dissociated using the human tumor dissociation kit with the gentleMACS Octo dissociator. Cells were viably frozen down or freshly used for downstream experiments. Dissociated cells (1 × 10⁵) were centrifuged, resuspended in 20 μL Matrigel, and plated in a 48-well plate. After Matrigel solidification, human BC PDO media were added (Table S1). VSVd51-GM-CSF or BCG infections were performed as in ICD assays, and bioimaging was performed using an automated fluorescence microscope (Celldiscoverer 7, Zeiss).

Ex vivo human immune cell activation

Monocytes were obtained by CD14⁺ selection (STEMCELL) from frozen human PBMCs. Sorted cells were incubated for 6 days with 500 U/mL of recombinant human IL-4 and 50 ng/mL of recombinant human GM-CSF (Bio Basic).³⁵ For monocyte:PBMC coculture assays, matched PBMCs were incubated for 24 h with 100 U/mL of recombinant human IL-2 (PeproTech). Then, generated immature DCs and lymphoid cells were incubated for an additional 24 h with CM from infected autologous PDOs and acquired for CD80/86 and HLA (human leukocyte antigen)-DR on CD14⁺ DCs; T and NK cell CD107a degranulation assessment was conducted following coculture (Table S1).

Statistical analysis

All analyses were conducted using Prism 10 (GraphPad). Unpaired two-tailed t tests were used for comparing cell cultures and mice experiments that included 2 treatment groups, whereas the Kruskal-Wallis test was used for more than 2 treatment groups. Survival studies were assessed using Kaplan-Meier curves and analyzed by log rank testing. p < 0.05 was considered statistically significant.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Acknowledgments

The authors would like to thank Dr. Leonid Volkov for his flow cytometry expertise. The authors would like to thank the following funding organizations: FRQS Sr Salary Award (L.-H.T.) provided salary for the principal investigator, CIHR Project Grant and CRMUS Chair (L.-H.T.) provided operating funds for this study, and CRCHUS (centre de recherche du centre hospitalier universitaire de Sherbrooke) Postdoctoral Fellowship (L.D.) provided the fellowship for the first author.

Author contributions

L.D., K.S.N., H.G., L.J.C., J.L.L., and G.S.-C. executed experiments and read and approved the manuscript; L.D., H.G., J.L.L., and P.O.R. contributed to writing and critically revised the manuscript; N.E.N. assisted with tissue processing; P.O.R. and L.-H.T. conceived and designed the experiments, were a major contributor in writing the manuscript, and supervised the study. All authors have read and approved the manuscript.

Declaration of interests

The funding bodies did not play a role in the design of the study; collection, analysis, and interpretation of data; and writing the manuscript. L.-H.T. is the inventor of a submitted patent on recombinant VSV for bladder tumors.