High atezolizumab antidrug antibody levels are associated with unfavorable clinical outcomes and diminished T cell responses following atezolizumab and bevacizumab treatment in advanced hepatocellular carcinoma.

Abstract

4105 Background: Systemic administration of humanized monoclonal antibodies can be immunogenic and trigger unwanted anti-drug antibody (ADA) responses. In the IMbrave 150 study, 29.6% of advanced hepatocellular carcinoma (HCC) patients developed atezolizumab ADAs after atezolizumab and bevacizumab (atezo/bev) treatment. ADAs could impair the action of the therapeutic antibody by reduction of serum concentration or neutralization. We determined the clinical and immunological implications of high ADA levels in advanced HCC patients after atezo/bev treatment. Methods: Advanced HCC patients (n = 132) treated with first-line atezo/bev were prospectively enrolled (discovery cohort: 50 from an institute; validation cohort: 82 from four institutes). Serum levels of atezolizumab ADA at baseline and three weeks (C2D1) and atezolizumab concentrations at C2D1 were measured by competitive ELISA. The effects of ADA on T cell immunity were examined by multiplex flow cytometry. Results: Strong ADA (≥ 1000 ng/ml) responses at C2D1 were observed in 17.4% of advanced HCC patients. ADA elevation after atezo/bev at C2D1 was evident in non-responders but not significant in responders. In the discovery cohort, patients with high ADA at C2D1 showed a decreased response rate (ADA-high: 11% and ADA-low: 34%) and shorter progression-free survival (PFS) and overall survival (OS) with atezo/bev compared to those with low ADA levels ( P = 0.004 for PFS; P = 0.009 for OS). In the validation cohort, patients with high ADA at C2D1 showed reduced response rate than those with low ADA (ADA-high: 7% and ADA-low: 29%). PFS and OS were worse in ADA-high group than in ADA-low group ( P = 0.001 for PFS; P < 0.001 for OS). In multivariate Cox regression analysis, the clinical significance of high ADA levels was independently associated with shorter PFS and OS after adjustment for age, sex, ECOG performance status, Child-Pugh score, AFP, macroscopic vascular invasion, extrahepatic spread, and neutrophil-to-lymphocyte ratio (PFS: HR 2.27, P = 0.006; OS: HR 3.04, P = 0.006). The atezolizumab serum concentrations were 29.8% lower in patients with high ADA levels than in ADA-negative patients. Atezolizumab concentration at C2D1 was inversely correlated with ADA levels. Moreover, patients with high ADA lacked CD8+ T cell proliferative responses to atezo/bev treatment. Furthermore, patients with high ADA had decreased secretion of effector cytokines such as IFN-γ and TNF-α from CD8+ T cells compared to those with low ADA. Conclusions: Highly elevated ADA at C2D1 is associated with unfavorable clinical outcomes in advanced HCC patients treated with atezo/bev. High ADA levels were associated with reduced atezolizumab exposure and could limit the drug’s anti-cancer efficacy.