Percutaneous BCG enhances innate effector antitumor cytotoxicity during treatment of bladder cancer: a translational clinical trial

Abstract

Background: Intravesical bacillus Calmette-Guérin (BCG) is the gold standard immunologic agent for treating patients with high-grade non-muscle invasive bladder cancer (NMIBC). Nevertheless, relapse rates remain high and BCG unresponsive NMIBC often requires bladder removal. Preclinical data suggest that priming with percutaneous BCG vaccine could improve response to intravesical BCG. Methods: A single-arm trial (NCT02326168) was performed to study the safety, immunogenicity, and preliminary efficacy of priming. Percutaneous BCG was given 21 days prior to intravesical BCG instillation in patients (n = 13) with high-risk NMIBC. Immune responses were monitored and compared to a sequentially enrolled cohort of nine control patients receiving only intravesical BCG. The effect of BCG on natural killer (NK) and γδ T cell in vitro cytotoxicity was tested. γδ T cell subsets were determined by T cell receptor gene expression with NanoString. Results: Priming was well tolerated and caused no grade ≥3 adverse events. The 3-month disease-free rate for prime patients was 85% (target goal ≥ 75%). Priming boosted BCG-specific immunity at 3 months and increased the activation status of in vitro expanded circulating NK and γδ T cells and their cytotoxicity against bladder cancer cells through receptor NKG2D. BCG enhanced the cytotoxicity of NK and γδ T cells against K562, RT4, and UM-UC6 but not against T24, UM-UC-3, or UM-UC-14 cells. Infiltrating γδ T cell subsets identified in the bladder includes γ9δ2 and γ8δ2. Conclusions: BCG priming is safe and tolerable. Poor sensitivity to NK and γδ T cell cytotoxicity by some bladder tumors represents a potential BCG-resistance mechanism.

Keywords: BCG; Priming; bladder cancer; clinical trial; vaccine.

Figures

Figure 1.

Trial Schema and results of PPD reading in prime patients. (a) Eligible patients were BCG naïve, PPD negative, and had high-grade non-muscle invasive bladder cancer (NMIBC). BCG vaccination was administered percutaneously and 21 days later, patients were given standard-of-care intravesical BCG instillation. (b) Examples of PPD response. The “ballpoint pen” technique was used to measure response at 48–72 after PPD placement. Shown are the examples of a patient with negative PPD response (induration

Figure 2.

BCG-induced urinary cytokines during intravesical BCG therapy. Urine supernatant collected from prime patients 4–6 h after BCG instillation was measured for cytokine level by Luminex Multi-plex assay from all samples available. Cytokine concentrations were corrected by individual urinary creatinine concentrations and normalized to 100 mg/dl creatinine for data deduction and plotting as mean ± SEM. Urinary cytokines compared at (a) week 1 and 3 for control versus prime patients and at (b) weeks 1–6 for PPD conversion versus PPD non-conversion patients by 3 months. Filled symbols, prime patients, n = 13; open symbols, control patients, n = 5. Comparison between control and prime groups was performed using Mann–Whitney test. Filled circles, PPD+, 3-month PPD converted patients, n = 7; open circles, PPD-, PPD non-converted patients by 3 months, n = 6. Area under curves (AUCs) were compared between groups using Mann–Whitney test. * p < .05, **, Pp< .01, NS, not significant.

Figure 3.

Priming boosts BCG-specific immunity and the simultaneous proliferation and effector function of memory T cells in circulation. ELISPOT assay in response to live BCG was conducted from PBMCs collected from patients at baseline (prior to BCG induction), 3 and 6 months. Response from all available samples is compared for: (a) prime (filled circles, n = 13) versus control patients (open circle, n = 7), or (b) prime patients who PPD converted by 3 months (filled diamonds, n = 7) versus who did not PPD convert (open diamonds, n = 6). SFC, spot-forming cells represent the absolute number of BCG-specific IFN-γ-producing PBMCs. (c) 1 × 106 patient PBMCs labeled with CFSE were cultured with the presence of live BCG for 7 days, then incubated with the cocktail of PMA, ionomycin and Golgi blocker for 5 h, prior to cell-surface staining for CD4, CD8, and γδ TCR markers, followed by intracellular staining for IFN-γ. Absolute numbers (AN) of proliferated memory T cell subsets (defined as both CFSElo and IFN-γ+) were calculated by multiplying total viable cells by the percentage of T cell subset acquired from flow cytometry. Shown are AN of proliferated memory CD4+, CD8+, and γδ T cells from circulation in response to live BCG at 3 or 6 months over baseline for PBMCs from prime (filled circles, n = 13) and control patients (open circles, n = 7). Wilcoxon matched-pairs signed rank test was performed for each group at either 3-month or 6-month in comparison with baseline. *, Pp< .05, **, p < .01, NS, not significant.

Figure 4.

BCG increases the cytotoxicity of innate effector cells expanded in vitro. PBMCs from a patient with BC were cultured with IPP and human IL-2 for 14 days with or without the presence of live BCG (added on day 0 and day 7 of culture at MOI = 0.1) to expand innate effector cells. Then, these cells (effector) were co-cultured with CFSE-labeled human leukemia K562 cells or human bladder cancer cells (target) for 5 h at effector: target (E: T) ratio 0: 1 and up to 25: 1, followed by staining of fixable viability dye to analyze percentage cell death of target cells on flow cytometer. Results plotted as mean ± SEM based on assay replicates and groups (BCG vs. media only) were compared with two-way ANOVA, **, p < .001, *, Pp< .01, NS, not significant.

Figure 5.

BCG priming improves BC cytotoxicity of innate effector cells expanded in vitro. PBMCs were cultured in vitro with IPP and IL-2 to expand innate effector cells before testing cytotoxicity as in Figure 4. In parallel, percentage and expression of activation markers of γδ T cells or NK cells were measured by flow cytometry. Shown are (a) percentage of CD56-expressing population among γδ T and NK cells after in vitro expansion, as well as (b) percentage of cell death in target cells after cytotoxic killing by effector cells generated in vitro from PBMCs of control or prime patients at baseline versus 3 months post BCG treatment for patients with adequate samples remaining for analysis (prime, n = 7; control, n = 5). Mean ± SEM. Wilcoxon matched-pairs signed rank test; *, p < .05; NS, not significant. Bonferroni corrections were applied to p value if multiple comparisons were used. (c) Innate effector cells from prime patients were pre-incubated with anti-γδTCR or anti-NKG2D blocking antibody for 45 min prior to cytotoxicity assay against RT4 cells. Shown are representative results from three experiments (prime patient 8 and 11, 3-month PBMCs). Mean ± SEM, unpaired t-test, **, p < .01, ****, p < .0001. (d) Cytotoxicity against RT4 was compared between the unsorted mixture of innate effector cells versus sorted γδ T cells or NK cells. Ratio of sorted γδ T and NK cells versus target RT4 cells are similar to those present in the unsorted mixture. Shown are the representative result from six experiments. two-way ANOVA, *, p < .05.