Nonmuscle invasive urothelial cancer— Bacillus Calmette–Guérin instillation or checkpoint inhibitor immunotherapy?

Adjuvant BCG therapy with maintenance for nonmuscle invasive high-grade bladder cancer (NIMBC) by BCG instillation therapy has been shown to reduce recurrence and progression to muscle-invasive disease [1]. Intravesical instillation of BCG is performed after complete transurethral resection of the tumor using a 6-weekly induction schedule and followed by maintenance therapy with the SWOG schedule being the most recognized [2]. Patients failing BCG can be categorized into three groups: BCG refractory, early and late BCG relapsing and BCG intolerant patients. BCG refractory patients present with persistent high-grade disease at 6 months after the start of induction therapy or show progression by grade or stage 3 months after the start of induction therapy. Early relapse is defined by tumor recurrence within 6 months of last BCG therapy. Both BCG refractory and early relapsing patients are termed “BCG unresponsive”. Late BCG relapse is seen as recurrence after a disease-free state of at least 6 months or later after last BCG exposure and these patients have a better prognosis than BCG unresponsive patients [3, 4]. Currently ongoing CKI trials mainly focus on BCG naïve or unresponsive patients aiming at providing an alternative or enhancement through synergy to BCG therapy or an alternative to salvage radical cystectomy in the case of BCG unresponsiveness.

In this review, we want to discuss the effects of BCG and CKI therapy on the immune system for NMIBC therapy and highlight the most important clinical trials involving CKIs in the setting of NMIBC.

BCG is derived from a strain of attenuated Mycobacterium bovis and is thought to attach to bladder urothelium after instillation [5]. Attachment and uptake of the bacterium are a key step in innate immune system activation and cytokine expression, which subsequently attracts more innate immune cells to the bladder. The innate immune system then triggers a strong T helper 1 (T_H1) immune response involving the production of T_H1 cytokines. These T_H1 cytokines activate macrophages and CD8+ killer cells, which are able to eliminate infected cells [6‐8]. A T_H2 helper cell response, however, has been correlated with BCG failure [9]. In order to enhance the effect of BCG, several trials have been testing the co-administration of the bacterium together with immunostimulating agents, such as interferon alpha (INFα), a cytokine that has been shown to induce bladder cancer apoptosis [10]. However, BCG plus INFα was not proven to be superior to BCG therapy alone, and agents that could boost the reaction of immune system towards BCG are currently a major area of research in BCG therapy optimization [11‐13].

Activation of cytotoxic T‑cell activity is, besides cytokine regulation, also modulated via specific receptors on T‑cells or ligands binding to these receptors present on interacting cells [14]. PD-L1 is a ligand on malignant cells that can down-regulate T‑cell activity by binding to PD‑1 on T‑cells [15, 16]. CTLA4 is a receptor on T‑cells that transmits inhibitory signals when regulatory or antigen-presenting cells bind to it [17]. Monoclonal antibodies such as pembrolizumab, atezolizumab, or ipililumab targeting PD‑1, PD-L1, or CTLA4 block transmission of inhibitory signals and are thus able to enhance T‑cell activation [18]. In late-stage metastatic bladder cancer, CKIs have been shown to result in sustainable responses in approximately 20–30% of patients [19, 20]. Regarding side effects, CKIs are generally well tolerated with immune-related adverse events occurring in less than 17% of patients. Side effects are usually self-limiting, and only some patients need short-course immunosuppressive drugs or TNFα-receptor antagonists when glucocorticoids fail [21]. Response to CKIs depends on multiple factors involving molecular characteristics of the tumor and interaction with the immune system, but expression of PD-1/PD-L1, tumor mutational burden and tumor immune infiltration seem to play a role for adequate response [22]. In bladder cancer, NMIBC tumors have been reported to express lower levels of PD‑1 than MIBC, but BCG infection can induce PD-L1 expression in regulatory T‑cells [23]. Furthermore, PD-L1 is enhanced on tumor tissue after BCG treatment in BCG resistant patients, making combination or sequential CKI therapy a promising option [24]. Also, a subset of high-grade NMIBC harbor mutations in DNA damaged genes which are known to be associated with a higher mutational load potentially resulting in better response rates to CKIs [25].

Due to the observed success rates in metastatic bladder cancer, CKI trials have recently also been initiated in the adjuvant setting for NMIBC. Table 1 gives an overview of currently ongoing, selected phase II and III trials. We especially want to highlight recent results for the single-arm KEYNOTE-57 phase II trial evaluating pembrolizumab for NMIBC patients unresponsive to BCG. The latest trial update reported a 3-month complete response rate (CRR) of 40%, and a 53% maintained complete response for more than 9 months. Importantly, no progression to MIBC or metastatic disease was observed [26]. A phase III trial (KEYNOTE-676, NCT03711032) for pembrolizumab + BCG vs BCG monotherapy in patients having received at least one course of BCG induction therapy has recently been started, and pembrolizumab as single adjuvant therapy for high risk T1 NMIBC tumors is also being tested (NCT03504163). For all other CKIs, no trial results have been reported yet, but phase II trials for atezolizumab (NCT02844816), durvalumab (NCT02901548, NCT03759496), nivolumab (CheckMate 9UT, NCT03519256) and avelumab (PREVERT, NCT03950362) in the BCG unresponsive setting are ongoing. Noteworthy, phase III trials evaluating atezolizumab (ALBAN, NCT03799835) and durvalumab (POTOMAC, NCT03528694) are already actively recruiting.

A multitude of clinical trials assessing CKI therapy in BCG unresponsive patients as well as first-line combination therapy with BCG are being conducted, and—depending on outcomes of these trials—addition of CKI to BCG may become a standard option in the future, thereby, reducing the need for radical cystectomy. However, it is still unclear which patients respond to CKIs and which do not. From mechanistic studies in patients receiving CKIs, we know that tumors with sufficient PD-L1 expression, high in mutational burden/neoantigen load, and inflamed tumors are the most likely to respond to CKIs [27]. Current trials will need to evaluate whether combination with BCG can synergistically enhance CKI efficacy and lead to stronger immune responses without increasing adverse events, or whether only a subset of patients with adequate molecular profiles will profit from CKIs as therapy for high-risk NMIBC. In conclusion, there is no doubt that we are discovering dimensions to the therapy of NMIBC never seen before, and identifying the best therapy for each tumor in each patient at the right time may finally usher in an age of tailored immunotherapy.