Certain BCG-reactive responses are associated with bladder cancer prognosis

Abstract

A subset of bladder patients does not respond to BCG treatment effectively and the underlying reason behind this observation is currently unclear. CD4⁺ T cells are composed of various subsets that each expresses a distinctive set of cytokines and can potently shift the immune response toward various directions. In this study, we examined the CD4⁺ T-cell cytokine response in bladder cancer patients toward BCG stimulation. We found that bladder cancer patients presented a variety of responses toward BCG, with no uniform characteristics. Those patients with high IFN-γ and IL-21 expression in CD4⁺ T cells presented significantly better prognosis than patients with low cytokine secretion in CD4⁺ T cells. Tumor-infiltrating CD4⁺ T cells were significantly less potent in expressing IFN-γ, IL-4, and IL-17, and more potent in expressing IL-10 than circulating CD4⁺ T cells. In addition, we found no difference in CD80, CD86, or MHC II expression by macrophages from patients with different IFN-γ and IL-21 levels. However, the secretion of IL-12, a Th1-skewing cytokine, was released at significantly higher level by macrophages from patients with high IFN-γ or high IL-21 secretion. We also identified that modulating monocytes/macrophages by GM-CSF-mediated polarization resulted in significantly elevated expression of IFN-γ and IL-21 from CD4⁺ T cells. Overall, these results suggested that the specific types of responses mounted by CD4⁺ T cells were critical to the final outcome of bladder cancer patients and can be influenced by monocyte/macrophage polarization.

Introduction

Bladder cancer is one of the most common cancers worldwide [1]. In contrast with urinary tract infection, which occurs more often in women, bladder cancer is significantly more prevalent in men. The reason for this gender-specific difference is unclear, but it is possible that cellular immune responses toward bacterial antigens can play a protective role against the development of the cancer.

One of the treatment procedures of bladder cancer involves the induction of immune responses by intentional instillation of bacteria into the bladder lumen [2]. Intravesical instillation of BCG vaccine shortly after bladder tumor resection is performed for high-risk non-muscle invasive urothelial carcinoma and is shown to reduce the risk of tumor recurrence and slow disease progression [3, 4]. Like a natural urinary tract infection, BCG instillation induces the activation of the host immune system, starting with an infiltration of innate immune cells, including neutrophils, monocytes, and macrophages, and followed by an influx of adaptive T cells [5, 6]. An increase in proinflammatory cytokines is also observed to accompany the infiltration of immune cells. Neutrophils are capable of killing cancer cells directly due to their expression of apoptosis-inducing TRAIL molecule and other cytotoxic granule proteins, while monocytes and macrophages have the capacity to ingest bacterial and cellular debris and present antigens to incoming T cells [7, 8]. Interestingly, high macrophage numbers in tumor are correlated with increased risk of tumor recurrence [8]. T-cell infiltration, on the other hand, is associated with elevated protection against tumor recurrence [9]. T cells are broadly categorized into CD4⁺ Th cells and CD8⁺ cytotoxic T cells. In bladder cancer, markers of a robust Th1 responses in particular patients correlate with better patient prognosis [10]. IFNγ^−/− and IL-12^−/− mice are less resistant to tumor challenge, while IL-10^−/− mice demonstrate enhanced response to BCG therapy [11]. Consistent with these findings, polymorphisms in immune-associated genes are significantly overrepresented in bladder cancer patients who did not respond to BCG therapy [12, 13]. Together, these results suggest that proinflammatory T-cell responses induced by a microbial infection likely present antitumor immunity in bladder cancer patients.

In this study, we investigated the anti-bacterial response in bladder cancer patients undergoing curative resection.

Materials and methods

Participants

Thirty patients affected by bladder cancer were enrolled at Renmin Hospital of Wuhan University. No bladder cancer patient was treated with BCG regime at the time of sample collection, but all patients received BCG vaccine in childhood. Heparinized peripheral blood samples were collected before preparation for surgery. Tumor specimen was from the resected primary tumor. All patients were followed-up by phone, mail, or hospital visits for 36 months after the resection of the primary tumor.

PBMCs were purified from whole blood by centrifugation across a Ficoll (GE Healthcare) gradient for 30 min at 1500 rpm, and washed twice. TILs were collected by first mincing the tumor mass into small pieces. The minced cells were then filtered across a sterile strainer (Falcon). The resulting single tumor cells were centrifuged across a Ficoll gradient to isolate leukocytes.

BCG stimulation

For the stimulation of cells with BCG, PBMCs or tumor-infiltrating leukocytes were incubated with 1 µg vaccine per 10⁵ cells per 200 µL of sterile complete culture medium without antibiotics (RPMI 1640 supplemented with 15% FBS and 1 × l-glutamine; Thermo Fisher). The cells were incubated in a CO₂ incubator at 37 °C with 100% humidity for 72 h.

Cell isolation

CD4⁺ T cells were isolated using EasySep Human CD4 T-cell Enrichment kit (Stemcell) from PBMCs or tumor-infiltrating leukocytes. Monocytes/macrophages were isolated using EasySep Human Monocyte Enrichment kit (Stemcell). All isolation processes were negative magnetic isolation following protocols from the manufacturer, resulting in untouched cells.

Luminex assay

Isolated CD4⁺ T cells or monocytes were incubated at 5 × 10⁴ cells per 250 µL complete culture medium at the bottom part of a 96-well transwell plate with 0.5 µm-pore membrane. The human cytokine capture beads (EMD Millipore) were added at the top. After 12 h incubation in a CO₂ incubator at 37 °C with 100% humidity, beads were aspirated and the cytokine expression was examined by Luminex.

Monocyte/macrophage polarization

Monocytes from PBMCs were incubated at 5 × 10⁵ cells per mL of complete culture medium supplemented with GM-CSF (1000 UI/mL; Miltenyi) for 5 days, in a CO₂ incubator at 37 °C with 100% humidity. To obtain unpolarized monocytes, no GM-CSF was added, while the rest of the procedures were kept the same.

Monocyte-T-cell coculture

10⁵ polarized or unpolarized monocytes/macrophages were incubated with 10⁵ CD4⁺ T cells in 200 µL of complete culture medium, in the presence of 2 µg BCG vaccine. After 72 h in a CO₂ incubator at 37 °C with 100% humidity, monocyte/macrophages and CD4⁺ T cells were isolated and each incubated with cytokine capture beads for Luminex assay as described above.

Statistical analyses

All statistical tests were done in Prism, and the specific tests used are indicated in the figure legend. P values less than 0.05 were required for reaching statistical significance.

Results

Patients with high IFN-γ and high IL-21 responses to BCG vaccination presented better prognosis

Bladder cancer patients donated peripheral blood samples at surgery and were tracked for 3 years after the resection of the primary tumor. We stimulated the PBMCs using the BCG vaccine. The CD4⁺ T cells were then isolated for analysis of secreted cytokines using multiplex Luminex assay. Data showed that the bladder cancer patients displayed highly variable cytokine responses across a wide range of concentrations (Fig. 1). The patients were then separated into High/Low IFN-γ, High/Low IL-4, High/Low IL-10, High/Low IL-17, and High/Low IL-21 groups based on their cytokine expression. The clinical outcome of the patients was then examined in all individuals (Fig. 2). High IFN-γ patients presented significantly better prognosis than low IFN-γ patients (Fig. 2a). No significant differences were observed between high IL-4 and low IL-4 patients (Fig. 2b), and high IL-10 and low IL-10 patients (Fig. 2c). High IL-17 patients appeared to present worse prognosis than low IL-17 patients, but the difference was borderline insignificant (Fig. 2d). High IL-21 patients presented better prognosis than low IL-21 patients (Fig. 2e).

Fig. 1.

Cytokine expression by CD4⁺ T cells following BCG stimulation. PBMCs from bladder cancer patients were incubated with 1 µg BCG vaccine per 10⁵ cells for 72 h. The CD4⁺ T cells were isolated after incubations and were incubated for an additional 12 h with cytokine capture beads. The cytokine expression in CD4⁺ T cells was examined by Luminex and expressed as pg/mL. Line indicates the median expression of each cytokine

Fig. 2.

Progression-free survival of bladder cancer patients with high or low expression of each cytokine. The bladder cancer cohort were separated into a High/Low IFN-γ, b High/Low IL-4, c High/Low IL-10, d High/Low IL-17, and e High/Low IL-21 based on the cytokine expression in Fig. 1. Statistical differences were examined by Log-rank test

Patients with higher IFN-γ expression in tumor-infiltrating CD4⁺ T cells presented better prognosis

Next, we investigated the cytokine expression by CD4⁺ T cells from tumor-infiltrating leukocytes, harvested from resected tumor. Compared to their counterparts in PBMCs, the tumor-infiltrating CD4⁺ T cells presented significantly lower IFN-γ, IL-4, and IL-17 secretion, and significantly higher IL-10 expression (Fig. 3). The patients were then separated into High/Low IFN-γ, High/Low IL-4, High/Low IL-10, High/Low IL-17, and High/Low IL-21 groups based on the cytokine expression from the tumor-infiltrating CD4⁺ T cells. Similar to the findings in PBMCs, high IFN-γ patients presented significantly better prognosis than low IFN-γ patients (Fig. 4a). No significant differences were observed between high IL-4 and low IL-4 (Fig. 4b), high IL-10 and low IL-10 patients (Fig. 4c), and high IL-17 and low IL-17 patients (Fig. 4d). High IL-21 patients also presented significantly better prognosis than and low IL-21 patients (Fig. 4e).

Fig. 3.

Cytokine expression by CD4⁺ T cells from tumor-infiltrating leukocytes. Tumor-infiltrating leukocytes were incubated with 1 µg BCG vaccine per 10⁵ cells for 72 h. The CD4⁺ T cells were isolated after incubations and were incubated for an additional 12 h with cytokine capture beads. The cytokine expression in CD4⁺ T cells was examined by Luminex and expressed as pg/mL. Line indicates the median expression of each cytokine. Statistical differences between circulating and tumor-infiltrating cells were examined by Mann–Whitney test

Fig. 4.

Progression-free survival of bladder cancer patients with high or low expression of each cytokine by tumor-infiltrating CD4⁺ T cells. The bladder cancer cohort were separated into a High/Low IFN-γ, b High/Low IL-4, c High/Low IL-10, d High/Low IL-17, and e High/Low IL-21 based on the cytokine expression in Fig. 3. Statistical differences were examined by Log-rank test

High IL-12 expression in tumor-infiltrating macrophages was associated with better IFN-γ and IL-21 responses in tumor-infiltrating CD4⁺ T cells

To elucidate the underlying mechanisms causing the varying cytokine expression in tumor-infiltrating CD4⁺ T cells, we assessed the phenotype and cytokine expression by tumor-resident macrophages. First, we examined the expression of surface molecules in tumor-infiltrating macrophages. After stimulation with the BCG vaccine, the tumor-infiltrating macrophages from high IFN-γ patients presented comparable CD80, CD86, and MHC II compared with tumor-infiltrating macrophages from low IFN-γ patients (Fig. 5a). Similarly, the tumor-infiltrating macrophages from high IL-21 patients presented similar levels of CD80, CD86, and MHC II compared with those from low IL-21 patients (Fig. 5b). The levels of cytokine secretion from the tumor-infiltrating macrophages after BCG stimulation were examined by Luminex. We found that in high IFN-γ patients, tumor-infiltrating macrophages presented significantly higher IL-12 expression (Fig. 5c). Similarly, high IL-21 patients also presented higher IL-12 expression by tumor-infiltrating macrophages (Fig. 5d).

Fig. 5.

Phenotype and cytokine expression of tumor-infiltrating macrophages in patients with high or low IFN-γ and IL-21 expression by tumor-infiltrating CD4⁺ T cells. Tumor-infiltrating macrophages were incubated with 1 µg BCG vaccine per 10⁵ cells for 72 h. The CD80, CD86, and MHC II expression by tumor-infiltrating macrophages from a High/Low IFN-γ patients and b High/Low IL-21 patients. The IL-12 and IL-10 secretion in the supernatant from c High/Low IFN-γ patients and d High/Low IL-21 patients were examined by Luminex. Mann–Whitney test

Circulating monocyte/macrophages can be primed into IFN-γ-promoting cells

Previously, studies have demonstrated that macrophages can be primed into proinflammatory or antiinflammatory macrophages via LPS, IFN-γ, and/or GM-CSF-mediated signal transduction [14, 15]. We examined the possibility of improving IFN-γ and IL-21 response by CD4⁺ T cells through polarizing macrophages toward the more proinflammatory subtypes using GM-CSF. In monocyte/macrophage culture with GM-CSF polarization, the IL-12 secretion was significantly higher than in monocyte/macrophage culture without GM-CSF polarization (Fig. 6a). We then incubated GM-CSF-polarized and unpolarized monocytes/macrophages with autologous CD4⁺ T cells, in the presence of BCG stimulation. Significantly higher IFN-γ and IL-21 secretion was found in CD4⁺ T cells incubated with GM-CSF-polarized monocytes/macrophages than in GM-CSF-unpolarized monocytes/macrophages (Fig. 6b, c).

Fig. 6.

a IL-12 expression by polarized or unpolarized monocytes/macrophages. Monocytes from PBMCs were incubated for 5 days without (blank) or with GM-CSF. The IL-12 secretion was then evaluated by Luminex assay. Wilcoxon test. b IFN-γ and c IL-21 expression by CD4⁺ T cells after incubation with autologous monocytes/macrophages. Wilcoxon test

Discussion

Although the precise mechanism for tumor recurrence is unknown, the current consensus suggests that recurrence occurs when the malignant tumor cells/tumor stem cells are not entirely removed from the body, allowing them to reestablish the tumor under permissible microenvironment [16]. It is expected that the immune system can detect some aberrant features of malignant tumor cells and, at least in part, mediate tumor rejection and elimination [17]. However, in many solid tumors, the immune system exhibit regulatory features with reduced capacity to mediate inflammation and cytotoxicity, contributing to tumor immune escape. Bladder cancer represents one of the cancers that present high recurrence rate following conventional chemotherapy and curative resection. Previous studies have demonstrated that BCG instillation is beneficial to the overall outcome of the bladder cancer patients, with potential to reduce the chance of recurrence. However, BCG is only effective in a subset of patients and the underlying reason behind this observation is currently unclear.

CD4⁺ T cells are composed of various subsets, and each subset expresses a distinctive set of cytokines. By cytokine expression and other stimulatory/inhibitory molecules, different CD4⁺ T-cell subsets can potently shift the immune response toward various directions. In this study, we examined the CD4⁺ T-cell cytokine response following BCG stimulation. We found that patients with high secretion of IFN-γ and IL-21 in CD4⁺ T cells presented significantly better prognosis than patients with low secretion of IFN-γ and IL-21 in CD4⁺ T cells. This effect was identified in both the circulating CD4⁺ T cells and tumor-infiltrating CD4⁺ T cells. No significant association was discovered between CD4⁺ T-cell-mediated IL-4, IL-17, or IL-10 secretion and patient prognosis, although in some cases, the statistical difference was borderline insignificant. We also observed that tumor-infiltrating CD4⁺ T cells were significantly less potent in expressing IFN-γ, IL-4, and IL-17, while more potent in expressing IL-10. We attempted to investigate the difference in CD4⁺ T-cell-mediated cytokine secretion by examining the autologous macrophages, which are known to modulate T-cell responses by expression of costimulatory/inhibitory molecules during antigen presentation and influence the surrounding environment by release of soluble factors. We found no difference in CD80, CD86, or MHC II expression by macrophages from patients with high or low IFN-γ expression and patients with high or low IL-21 expression. However, the secretion of IL-12, a Th1-skewing cytokine [18], was released at significantly higher level by macrophages from patients with high IFN-γ or high IL-21 secretion. We also found that modulating monocytes/macrophages by GM-CSF-mediated polarization resulted in significantly elevated expression of IFN-γ and IL-21 from CD4⁺ T cells. Overall, these results suggested that the specific types of responses mounted by CD4⁺ T cells were critical to the final outcome of bladder cancer patients.

IFN-γ and IL-21 have been previously described to present antitumor functions [19, 20]. IFN-γ has been shown to exert anti-angiogenesis and pro-apoptosis effects to tumor cells and could exert divergent proinflammatory functions through the activation of STAT1 signaling pathway [21]. IL-21 has been shown to promote the expansion of cytotoxic CD8⁺ T cells, NK cells, and follicular helper T cells, and the differentiation of B cells into plasmablasts [22, 23]. The association between patient prognosis and higher IFN-γ and IL-21 expression by CD4⁺ T cells suggests that these cytokines are involved in the antitumor responses in bladder cancer patients. Further functional analyses are required. The lack of an association between IL-10 expression and disease progression was surprising to some extent, because serum IL-10 levels were associated with worse outcomes in a variety of solid and mobile-phase cancers [24]. Given that IL-10 presents pleiotropic roles, it is possible that the location and context of IL-10 expression are as important as its concentration. Furthermore, accumulating evidence from more recent studies suggests that intratumoral IL-10 could reduce the apoptosis and enhance the proliferation of activated CD8⁺ T cells [25, 26]. Perhaps, the proinflammatory properties of IL-10 counteracted some of its suppressive functions in bladder cancer patients.

At the time of sample collection, no subject of our patient cohort had undergone BCG instillation, but all has received BCG vaccination in childhood. Hence, it is difficult to differentiate whether the BCG stimulation reactivated pre-existing CD4⁺ T-cell memory responses, or stimulated new responses. The finding that stronger IFN-γ and IL-21 responses toward BCG from CD4⁺ T cells, even before BCG instillation for bladder cancer patients, could significantly predict better prognosis possibly suggests that pre-existing immunity toward BCG could exert beneficial effects. Alternatively, the patients could have presented de novo responses toward BCG stimulation via the danger signal receptors expressed by monocytes/macrophages, T cells, and other immune cells, which is a mutually non-exclusive possibility. Further research is required to elucidate the underlying mechanisms behind our current observations.

In conclusion, our study suggested that the specific types of responses mounted by CD4⁺ T cells were critical to the final outcome of bladder cancer patients, which could be affected by monocyte/macrophage. These data shed lights on understanding the mechanism of BCG treatment in bladder cancer.

Author contributions

GS, TT, and YX conceived and designed the study. GS, HQ, and YX conducted the experiments. GS, TT, and YX analyzed the data. GS and TT wrote the paper.

Funding

None.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval and ethical standards

This study was reviewed and approved by the Internal Review Board of the Renmin Hospital of Wuhan University (No. RHWU10536).

Informed consent

Informed consents were obtained from all participants in the study.