Clonal expansion of tumor infiltrating lymphocytes (TILs) in the peripheral blood of metastatic melanoma patients is significantly associated with response to CTLA4 blockade-based immunotherapy

2541 Background: Patients with metastatic melanoma were treated on a clinical trial with tremelimumab and High Dose Interferon-Alfa (HDI) (Tarhini. J Clin Oncol. 2012). We previously reported that patients who achieved disease control and clinical response had significantly greater T-cell clonality (p = 0.0008) and T-cell fraction (p = 0.044) respectively in their pretreatment tumor biopsy samples (Tarhini. J Clin Oncol. 2017). In this study, we further characterize T-cell repertoire clonality and clonal expansion in the peripheral blood at different time points to evaluate the association between repertoire features and clinical response. Methods: Patients received tremelimumab 15 mg/kg I.V. every 12 weeks and HDI was given concurrently. Responses were assessed by RECIST as complete (CR) or partial (PR), stable disease (SD) or progression (PD). Peripheral blood mononuclear cells (PBMCs) from treated patients (N = 33) were obtained at baseline, day 29, and day 85 (following tremelimumab-HDI treatment); tumor samples at baseline were also obtained (N = 18). The T-cell receptor beta chain (TCRB) repertoire of PBMCs and tumor samples was immunosequenced using the immunoSEQ assay (Adaptive Biotechnologies), and repertoire clonality was assessed at baseline, day 29, and day 85. Differential abundance analysis was used to detect and quantify peripheral clonal expansion pre- versus post-treatment and identify the subset of peripheral clones also detected in the tumor repertoire. The Morisita Index of repertoire similarity was also calculated to compare global repertoire changes between pre- and post-treatment PBMC samples. Results: T-cell repertoire turnover, as measured by the Morisita Index, showed a trend towards responders (CR/PR) having greater turnover (lower Morisita Index) post-treatment than non-responders (SD/PD). Similarly, the total number of clones expanding in the peripheral repertoire varied over time within an individual (p = 0.034) but was not significantly affected by response to therapy (p = 0.275) or by on-treatment time point (p = 0.768). When the analysis was restricted to peripherally expanded clones that were also found in the tumor repertoire, responders had significantly more TILs expanded in the periphery at day 29 than non-responders (p = 0.036). Conclusions: Our analysis of the peripheral T-cell repertoire following treatment showed that detection of TILs in early peripheral clonal expansion correlates with response to therapy.