Intratumoral Th2 predisposition combines with an increased Th1 functional phenotype in clinical response to intravesical BCG in bladder cancer

After decades of skepticism, immunotherapy is poised to become a mainstay of cancer treatment [1]. Currently, the most promising approach in activating therapeutic antitumor immunity is the blockade of immune checkpoints [2] as well as the use of genetically engineered T cells with chimeric antigen receptor (CAR) directed against tumor-associated antigens [3]. Treatment with Bacillus Calmette–Guérin (BCG) belongs to the most successful cancer immunotherapies, and in high-risk, non-muscle-invasive bladder cancer (NMIBC), it is the standard adjuvant treatment according to the European Association of Urology (EAU) guidelines [4]. Forty years after its introduction [5], the exact immune mechanism of BCG-induced antitumor activity is still not fully understood. Following BCG instillations, intravesical BCG–fibronectin complexes are internalized at the tumor resection site [6]. Antigen-presenting cells in the urothelium can phagocytose BCG and present BCG-derived antigens to CD4⁺ T cells. In vitro work by Brandau et al. has demonstrated that BCG activates natural killer (NK) cells in a monocyte-dependent manner [7]. It is well established that innate lymphocytes including NK cells not only participate in the early innate response but also promote and shape the subsequent adaptive response by triggering dendritic cell maturation [8] and are therefore essential for effective BCG immunotherapy [9, 10]. Different cytokines such as interleukin (IL)-1, IL-2, IL-6, IL-7, IL-8, IL-10, IL-12, tumor necrosis factor-(TNF)-α and interferon (IFN)-γ are released and can be detected in patients treated with BCG [11‐13]. Thus, BCG can induce the production of both Th1-type and Th2-type cytokines. This fact was confirmed in vitro showing that BCG stimulates cultured murine dendritic cells, which are able to induce both IL-12 and IL-10, resulting in a mixed, nontargeted Th1 and Th2 immune response [14]. However, a predominant Th1 cell-mediated immunity with an enhanced recognition of cancer cells through infiltrating effector cells into the bladder wall is required for subsequent BCG response [15]. IL-12- or IFN-γ-depleted animals were BCG-resistant with a poor cancer-specific survival [16], whereas therapeutic strategies administering BCG along with Th1 cytokines and concurrent blocking of Th2 cells may enhance BCG-induced IFN-γ production and BCG vaccine efficacy [17‐20]. Moreover, significant increases in urine concentrations of Th1-type cytokines during treatment were noticed in BCG responders [21, 22].

IFN-γ is an important stimulus for the enzyme GTP cyclohydrolase (GCH-I) in human monocyte-derived macrophages and dendritic cells, which induces neopterin production reflecting cellular immune activation [23, 24]. In parallel, IFN-γ activates the enzyme indoleamine 2,3-dioxygenase (IDO1), which converts tryptophan to kynurenine resulting in increased tryptophan breakdown, and elevated kynurenine-to-tryptophan ratio (KTR), [23]. Therefore, neopterin production and tryptophan breakdown are surrogate markers of IFN-γ production and thus of an ongoing Th1-type immune response. Currently, only a letter to the editor reported monitoring of neopterin in bladder cancer patients during intravesical BCG therapy [25]. Moreover, intravesical instillations of autologous IFN-γ-activated macrophages resulted in an increase in urinary neopterin [26].

It is well known that differentiation of type 1 and type 2 Th cells [27] as well as innate lymphoid cells [28] is controlled by the transcription factors T-bet and GATA3. Interestingly, a genome-wide analysis has revealed that T-bet is sufficient to induce GATA3 binding at Th1 specific sites, indicating its direct influence and responsibility for the redistribution of GATA3 in Th1 cells [29].

Recently, we confirmed a Th2 predisposition (GATA3>T-bet) of tumor-infiltrating immune cells in high-risk NMIBC patients with response to BCG [30]. The aim of the present follow-up study was to examine the relation between such a Th2 predisposition and the actual functional phenotype during treatment as a potential biomarker of BCG response.

Immunity to pathogens including mycobacteria is mediated by type 1 effector Th cells (Th1 cells), which produce IFN-γ [27]. Whereas the T cell transcription factor T-bet (encoded by Tbx21) has a fundamental role in coordinating a type 1 immune response, GATA3 is considered to be the master regulator of the Th2 cell differentiation program characterized by the production of IL-4, IL-5 and IL-13. T-bet also promotes Th1-type responses by preventing GATA3-mediated Th2 cell development [41, 42]. Despite this established concept of mutually exclusive Th cell differentiation, there is also evidence that Th2-type cytokines such as IL-4 are important for memory T cell generation [43] and for the development of CD8⁺ T cell immunity against intracellular parasites [44].

A Th2 (GATA3>T-bet) predisposition has been reported in different cancer entities and metastatic lymph nodes and was found to be associated with cancer recurrence, progression and poor survival [45‐48]. This fact may support the hypothesis of an inverse correlation of GATA3 and T-bet expression, suggesting that the presence of Th2 immune cells limits functional Th1-type responses within the tumor microenvironment. In the present study, we observed a positive correlation between GATA3 and T-bet expression in tumor-infiltrating immune cells. In detail, when GATA3 expression was high, also T-bet was increased and vice versa. Our present findings together with results recently published by our study group [30] clearly indicate a Th2 predisposition of tumor-infiltrating immune cells in BCG responders. However, increased levels of GATA3 expression and GATA3/T-bet ratio were significantly associated with a prolonged, and not a poor recurrence-free survival [30]. Moreover, our current data also showed the tendency toward increased numbers of GATA3⁺ immune cells in BCG responders compared to BCG failure. These results are comparable with the trial by Nunez-Nateras et al. [49] analyzing the pretreatment immunologic tumor microenvironment within tumor-infiltrating lymphocytes. They also showed an increased pretherapeutic GATA3/T-bet ratio in responders in comparison with BCG failure.

Using real-time PCR to asses the gene expression of GATA3 and T-bet, a reduced Th2 phenotype (decreased GATA3 expression) correlated with disease aggressiveness (high-grade tumors and muscle-invasive status) and a poor cancer-specific survival in bladder cancer [50]. In line with this observation, we observed that the Th2 cytokine IL-4 can directly suppress the growth of tumor cells [51]. Whereas GATA3 expression remained mostly negative in prostate and renal cell carcinoma, up to 70% of invasive urothelial carcinomas were GATA3 positive in one study [52]. However, real-time PCR analysis confirmed higher expression of T-bet among BCG responders as compared to BCG failure (p = 0.02), whereas heavy smokers with low expression levels of GATA3 were poor responders to BCG (p = 0.01), [50].

Focusing on tumor cells, GATA3 has been shown to prevent bladder cancer progression and metastasis by inhibiting cell migration and invasion as well as epithelial-to-mesenchymal transition in vitro [53]. On the contrary, strong expression of GATA3 was associated with progression and poor cancer-specific survival in muscle-invasive bladder tumors [54]. Thus, the role of GATA3 has to be further validated and needs to be interpreted with caution. A possible explanation of these conflicting results may be the interaction of GATA3 with steroid hormone receptor signals influencing bladder cancer growth adversely by both stimulatory and inhibitory effects via androgen receptor (AR), estrogen receptor (ER)α and/or ERβ [55, 56]. For example, GATA3 expression correlated with a loss of ERβ as well as with AR/ERα overexpression in bladder cancer [54].

It is well established that Th1-type immunity is required for effective BCG-induced antitumor activity in bladder cancer. BCG efficacy may be increased by ex vivo tumor antigen-loading and dendritic cell activation as BCG stimulated dendritic cells to secrete TNF-α, which is responsible for phenotypic and functional changes [57]. Ponticiello et al. [58], for instance, noticed a significant increase in the CD4⁺ Th1 subsets during BCG therapy. Moreover, BCG resulted in an increase in Th1 cytokines and lower levels of Th2 cytokines during therapy in vitro and in humans [13, 21, 22, 59, 60]. Moreover, parenteral exposure to BCG before instillations triggered an accelerated T cell entry into the bladder in vitro, with an improved recurrence-free survival in patients [60]. Therefore, shifting the Th2 dominant immunologic landscape toward a Th1 response during therapy seems to be an important mechanism for adequate BCG response [61]. Another likely point of view is that a certain Th2 capacity is necessary for the generation of an adequate Th1 response. For instance, exposure to the Th2 cytokine IL-4 has been shown to result in an enhanced CD8⁺ T cell response to pathogens, improving a proinflammatory Th1 immune response [43]. However, the exact mechanism how IL-4 alters the frequency of CD8⁺ T cells in humans is still unclear. In mice, IL-4-producing NKT cells correlated with thymic innate memory CD8⁺ T cells [62]. In renal cell carcinoma, IL-4 and TNF-α synergistically induced apoptosis and cytokine production in vitro, promoting the recruitment of different immune effector cells [51]. Thus, simultaneous induction of both Th1 and Th2 response seems to be necessary for efficient systemic antitumor activity [63, 64]. As an example, a Th2-biased response to MAGE-6 epitopes prior to treatment shifted to a Th1-mediated response after therapy in two patients (renal cell carcinoma and melanoma) with complete therapeutic response [65]. Our present results confirm this view, since patients with Th2 predominant tumor-infiltrating immune cells displayed a Th1 functional phenotype during BCG induction.

IFN-γ-induced neopterin production and tryptophan breakdown are typical markers for Th1-activated cell-mediated immunity [23, 24]. So far, only a letter to the editor assessed changes in neopterin levels during BCG. Mack et al. [25] showed for the first time in 30 patients that BCG response was associated with a significant peak of serum and urinary neopterin after each BCG instillation as a sign of a Th1 cell-mediated immune activation. Our results are in line with these data as BCG responders also displayed higher levels of urinary and serum neopterin during intravesical BCG induction. In addition, we demonstrate for the first time enhanced tryptophan breakdown in the serum of BCG responders. Moreover, concentrations of neopterin correlated significantly with those for tryptophan breakdown at each time point confirming that IFN-γ induces simultaneously induces two different biochemical pathways: first, deprivation of tryptophan by IDO1; second, neopterin and reactive oxygen species (ROS) production by GCH-I [66]. Moreover, it is remarkable that neopterin and tryptophan metabolites can be detected in human serum as surrogate markers of a local IFN-γ response occurring in the bladder or the draining lymph nodes after BCG therapy.

As a result of a chronically activated immune system, also counter-regulatory and immunosuppressive mechanisms can be activated with decreased T cell responsiveness and development of immunodeficiency as a consequence of T_reg cell expansion [67, 68]. A positive correlation was found between high IDO1 expression in bone marrow-derived mesenchymal stem cells and elevated percentage of T_reg cells in acute myeloid leukemia [69]. Therefore, the proportion of CD4⁺CD25⁺CD39⁺ T_reg cells was analyzed in PMBCs at baseline, during and after BCG therapy. However, we noticed no significant correlations between serum levels of neopterin or tryptophan degradation and distribution of T_reg cells in PBMCs. Nevertheless, BCG responders showed the tendency toward increased T_reg cells compared to BCG failure, but without statistical significance.

The obvious limitation of this prospective pilot study is the relatively small sample size of 23 patients, which restricts statistical methods of interpretation. Therefore, further prospective and multi-institutional randomized trials with sufficient statistical power and long-term follow-up are required to validate these preliminary findings and to verify in detail the prognostic role of Th1-related inflammatory metabolites in the context of Th2-driving transcription factor overexpression.

In patients receiving intravesical BCG therapy, a general intratumoral Th2 predisposition at the level of transcription factors (GATA3>T-bet) was combined with an increased productive Th1-type immunity in responders compared to non-responders. A better understanding of Th1 regulation by Th2 components at the molecular level would be helpful in developing a more efficient and targeted cancer immunotherapy, particularly in bladder cancer patients with BCG failure.