TCR Clonal Evolution in AML Patients in Morphologic Remission Treated with Anti-PD1 Antibody, Nivolumab

We are conducting a clinical trial titled "Randomized Phase II Study to Assess the Role of Nivolumab as Single Agent to Eliminate Minimal Residual Disease and Maintain Remission in Acute Myelogenous Leukemia (AML) Patients After Chemotherapy" (REMAIN trial) (NCT02275533).

A critical barrier in developing immunotherapies is the identification of predictive biomarkers of response to therapy. T lymphocytes play critical roles in response to immunotherapies but their clonality and temporal changes in the T cell repertoire during treatment have not been well investigated. Recent advances in deep sequencing technology make it possible to characterize the T cell receptor (TCR) repertoire generated following immunotherapy.

In this study, we characterized T cell repertoire in peripheral blood and/or bone marrow samples of three AML patients on the REMAIN trial before and after nivolumab treatment. Using Illumina MiSeq sequencer and total RNA from each sample, we conducted deep sequencing of TCR-α and -β chains, and calculated the diversity index (inverse Simpson's index) in their CDR3 sequences to assess overall clonality of T cells. We obtained total CDR3 clonotypes of 420,765 ± 155,449 (average ± standard deviation) for TCR-α and 410,786 ± 115,219 for TCR-β per each sample. Interestingly, we found that certain TCR-α and -β clonotypes were drastically enriched in the bone marrow samples after nivolumab treatment. Many of these enriched TCR clonotypes were minimal or undetectable before nivolumab treatment, indicating that nivolumab might induce expansion of anti-AML T cell subclones. Particularly, nivolumab treatment led to marked reduction of TCR diversity indexes in both peripheral blood and bone marrow samples of one AML patient, who had shown a clearance of minimal residual disease as detected by WT1 qRT-PCR.

Our results thus far indicate the feasibility of this type of comprehensive analysis of TCR repertoire in the context of immunotherapy for AML. Preliminary results suggest that such analysis may be utilized to predict response of immune checkpoint blockade, and could also be useful to identify high-affinity TCRs for adaptive T cell therapy approaches.