Combining Antiandrogens with Immunotherapy for Bladder Cancer Treatment

Take Home Message These preclinical studies in mice using the MBT-2 model of bladder cancer suggests that the combination of antiandrogens and immunotherapy for non–muscle-invasive bladder cancer may improve response rates.


Introduction
Bladder cancer (BCa) is the seventh most common cancer worldwide [1]. Men are three to four times more likely to be diagnosed than women [2], who often have more aggressive tumors at diagnosis [3]. At diagnosis, over 70% of tumors are classified as non-muscle-invasive bladder cancer (NMIBC). Intravesical bacillus Calmette-Guerin (BCG) is commonly used to decrease the risk of recurrence of highgrade NMIBC [4]. Nonetheless, approximately 35% of patients who receive BCG still have recurrences [5]. More recently, inhibition of programmed cell death protein-1 (PD-1) or programmed cell death-ligand 1 (PD-L1) has emerged as effective BCa immunotherapeutic treatment [6,7]. However, only 20-30% of patients have a significant response to inhibition of these immune checkpoints [8]. These incomplete response rates to both BCG and checkpoint inhibitors highlight the need for novel approaches to improve response rates.
The importance of sex and hormonal differences on BCa treatment response remains relatively unexplored. Large contemporary series do not suggest sex differences in response to BCG treatment reported in prior series [9,10]. Sociological, health system, and presentation differences nonetheless confound the assessment of biological treatment differences [11]. Existing studies suggest that hormonal differences may impact outcomes, with estrogens being potentially protective against BCa development but possibly supportive of BCa progression [12]. Androgens and the androgen receptor (AR) are more robustly implicated in bladder carcinogenesis and BCa progression [13]. Additionally, androgens are recognized to have immunosuppressive properties [14], which likely contribute to sex differences in various pathologies [15]. Recent studies have found AR-suppressive therapy associated with improved BCa outcomes, highlighting the potential for clinical impact [16].
In this study, we assessed the use of AR antagonism as an approach to improve the antitumor response induced by immunotherapy. We first demonstrate how the MBT-2 BCa murine model recapitulates human biology of NMIBC, including sex differences. We use the MBT-2 model to evaluate the AR antagonist enzalutamide in combination with BCG or anti-PD-1 immunotherapy.
Next, we compared differences in the immune composition tumors between sexes in NMIBC and MBT-2 tumors (Fig. 2). We observed sex differences in the ratio of CD45 + / CD45cells ratio ( Fig. 2A and 2B). This was also observed for TILs ( Fig. 2C and 2D) and the expression of the immune checkpoint markers LAG-3, CTLA-4, and PD-1 ( Fig. 2E and 2F). These observed differences between tumors from male versus female mice together recapitulate those observed in NMIBC patients and reported differences in human tumors in the literature [21]. We did not find these differences in the commonly used MB49 murine model of BCa (data not shown). Together, these data suggest that the MBT-2 murine model is relevant for the study of sex differences in the immune tumor microenvironment of BCa.

3.2.
Combination of enzalutamide and BCa immunotherapy synergizes to improve the therapeutic response in male MBT-2 mice Using different BCa cell lines, we next assessed whether AR antagonists inhibit proliferation in vitro and observed that AR antagonists decrease the proliferation of human and murine BCa cells, principally in lower-grade BCa cells ( Supplementary Fig. 1). These results also confirmed the choice of enzalutamide for use in in vivo studies.
Using the immunocompetent and syngeneic MBT-2 murine model, we first observed that treatment with anti-PD-1 mAb decreased tumor growth and resulted in a complete response in typically one out of six male mice, about half the number observed in female mice ( Fig. 3A-C). We also showed that the survival in response to PD-1 inhibition in female mice was superior to that observed in male mice ( Fig. 3B and C). To assess the effect of AR antagonism on survival, we treated male mice with enzalutamide. Treatment with enzalutamide alone resulted in tumor growth and survival similar to control (Fig. 3C). Subsequently, we assessed the combination of enzalutamide and anti-PD-1 mAb. We observed that the combination resulted in a significant increase in the survival of male mice as well as greater tumor growth inhibition compared with either monotherapy ( Fig. 3A-C). Notably, the combination of enzalutamide and anti-PD-1 mAb increased the proportion of complete responses. A rechallenge with reinoculation of MBT-2 cells was performed in the two available surviving mice, with no tumor growth observed indicating immune memory.
To confirm the therapeutic potential of enzalutamide to improve BCa immunotherapy, we further assessed the treatment with the combination of enzalutamide and BCG + poly(I:C) in the MBT-2 murine model. As expected, BCG + poly(I:C) treatment in male mice improved survival and decreased tumor growth compared with control, while no benefit was seen with enzalutamide alone (Fig. 3D-F). Similar to responses with anti-PD-1 mAb, BCG + poly(I:C) treatment was superior in female mice to that in male mice ( Fig. 3E and F). However, the combination of BCG + poly(I:C) and enzalutamide showed the highest tumor growth inhibition (Fig. 3D) and survival (Fig. 3F). Immune memory was also demonstrated, with no tumors observed growing in both mice treated with the combination of BCG + poly(I:C) and enzalutamide that were rechallenged. Importantly, a comparison between male and female mice highlights that the addition of enzalutamide improves the response rate for both BCG + poly(I:C) and PD-1 inhibition in male mice to resemble the superior results obtained in females (Fig. 3A-F). Differential gene expression on collected tumors shows an important difference between control and anti-PD-1 treatment among female mice but not among male mice ( Fig. 3G and H). Interestingly, female mice receiving anti-PD-1 showed greater differential gene expression changes than the male anti-PD-1 group (Fig. 3I). However, minimal changes were observed with enzalutamide or the combination of enzalutamide and anti-PD-1, suggesting that the observed synergy was not mechanistically related to gene expression changes. We therefore further investigated the immune composition of the same tumors using flow cytometry. We first observed that tumors from the male control group have significantly more TILs and CD8 + T cells, and significantly fewer CD4 + T cells (CD45 + CD3 + CD4 + cells) than female controls (Fig. 4A-C). Anti-PD-1 treatment significantly decreased TIL infiltration in tumors of female and male mice, but did not impact the proportion of CD4 + and CD8 + T cells in females, whereas in males, a significant increase of CD4 + T cells is observed (Fig. 4D-I). Notably, the therapeutic combination of enzalutamide and anti-PD-1   induces a decrease of TILs but an increase of the proportion of CD4 + T cells in tumors ( Fig. 4G and H).

Discussion
While differences between sexes in BCa are well known, application of these differences to treatment strategies has yet to be made. Our results suggest the potential to use AR antagonists in combination with immunotherapies in males to improve treatment response rates. We also demonstrate key aspects of the MBT-2 mouse model, highlighting how it is comparable with the immune composition of NMIBC tumors from patients and sex differences reported in the literature.
Our results provide particular support to combine antiandrogens with BCG treatment for NMIBC. We demonstrate notable similarities from the MBT-2 model to NMIBC in patients, which importantly include sex differences in the immune infiltrate. Clinically, there is limited evidence for AR antagonists for BCa, although accumulating data available support efficacy for 5-alpha reductase inhibitors [22]. Nonetheless, given the dose-response rates observed in vitro, it remains plausible that the more potent AR antagonism may induce greater synergy, as seen in our murine experiments.
Our in vivo results demonstrating synergy with AR antagonism and BCG or anti-PD-1 mAb suggest a broader mechanism not specific to either immunotherapy. Initial clinical results combining pembrolizumab and enzalutamide in prostate cancer suggest similar synergy [23], with a recent study suggesting that androgens play a role in suppressing T-cell function [24] important for response to immune checkpoint inhibitors. As the composition of the immune cell infiltrate in the BCa microenvironment is critical for the response to both immune checkpoint inhibitors and BCG, differences at these levels may explain the synergy, consistent with our flow cytometry and RNA-sequencing studies. Indeed, the accumulation of immunosuppressive innate immune cells portends a worse prognosis and is also associated with a poorer response to BCG [25][26][27]. Consistent with this literature, we found that the combination of enzalutamide and anti-PD-1 treatment increased proinflammatory TIDC1 and decreased anti-inflammatory TIDC2 and MDSCs in male MBT-2 tumors. The low AR expression in MBT-2 cells (data not shown) also supports an indirect mechanism acting on the immune microenvironment and helps explain why enzalutamide alone had no benefit. Nonetheless, our tumor analyses were performed on tumors remaining after treatment; it is possible that the differences observed between control and treated female mice reflect an increased immunoregulatory response due to treatment selection. For the male mice, the less effective immune response to tumors and the fewer differentially expressed genes may represent tumors that resemble more MBT-2 tumors, which have less selection pressure from the immune system. In this case, the changes we observed with the combination of enzalutamide and anti-PD-1 immunotherapy may reflect greater selection pressure on tumors. In either case, changes to the immune composition are much more pronounced than gene expression changes, reflecting the importance of understanding changes in the immune tumor microenvironment to improve immunotherapy response rates.
Our study has several limitations. Owing to sample availability, RNA sequencing included only two tumors in the male combination group. The collection of tumors at maximal tumor size may also obscure changes induced at the transcriptional level. Another limitation is the low number of mice per group for each animal study. Finally, additional mechanistic possibilities including epigenetic or unmeasured immune cell populations not evaluated may provide further insights.

Conclusions
In summary, our preclinical studies suggest that AR antagonism may synergize to improve BCa immunotherapy response rates in men through modulation of the immune cell composition of tumors. Translation of our results to patients is facilitated by the large urological experience with AR antagonists, with a phase II trial of bicalutamide in men receiving BCG underway (NCT05327647).
Author contributions: Paul Toren had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.