Antigen-specific T cell responses are a fundamental component of the anti-tumoral response and are key to help understand the intricate relationship between the tumor and host. Although the majority of targeted tumor antigens remain unknown, analysis of the intra-tumoral TCR repertoire provides a way to assess both the breadth and depth of the T cell immune response.
In our paper, which was developed as part of a collaborative project between the Quezada, Chain and Swanton laboratories, we examined the intratumoral TCR repertoire in patients with early stage NSCLC recruited to the lung TRACERx study. The lung TRACERx study provided us with a unique setting in which to systematically study genomic and immunological intratumoral heterogeneity in the context of NSCLC. We explored the spatial heterogeneity of individual TCRs within tumors and examined how this heterogeneity relates to genomic heterogeneity.
We have previously developed an experimental and computational pipeline for TCR repertoire sequencing. A key advantage of our TCR sequencing protocol is the ability to incorporate unique molecular identifiers (UMIs) into each cDNA molecule accounting for sequencing errors and PCR bias. Accordingly, the protocol enables us to achieve a high degree of quantitative precision. We utilized our pipeline to sequence TCRs from multi-region tumor samples, matched nontumor lung and peripheral blood from 72 patients enrolled into the lung TRACERx study.
We discovered a rich repertoire of TCRs within multi-region tumor samples and nontumor lung samples, comprising of thousands of different TCR sequences. The vast majority of the expanded intratumoral TCRs were differentially expressed in the tumor as compared to adjacent non-tumor lung. We observed a strong positive correlation between intratumoral TCR heterogeneity of expanded TCRs and spatial mutational heterogeneity. We utilized a robust statistical framework to categorize expanded intratumoral TCRs into ubiquitous TCRs (present in all tumor regions in a given patient), or regional TCRs (present in a subset of tumor regions).
We found that the number of ubiquitous and regional TCRs was correlated to the number of ubiquitous and regional nonsynonymous mutations respectively. TCR CDR3 clustering analysis showed that clusters with regional TCRs contained TCRs from fewer regions of the tumor than those with ubiquitous TCRs. We utilized flow cytometry and RNA sequencing methods to demonstrate that CD8+ tumor-infiltrating lymphocytes displaying ubiquitous TCRs exhibit a dysfunctional tissue-resident phenotype. Strikingly, we were able to detect a number of intratumoral ubiquitous TCRs in the blood of the patients at the time of primary tumor resection.
Taken together, the findings presented in our paper suggest that the expanded intratumoral TCR repertoire is driven by the intratumoral neoantigen landscape that may be sculpted by other immunological mechanisms such as focal HLA loss or antigen processing defects. Furthermore, our data suggest that a number of ubiquitously expressed TCRs within the tumor may be able to recognize antigens that are ubiquitously expressed throughout the tumor. Conversely, we propose that selected regional TCRs may recognize antigens expressed in a subset of tumor regions.
Our study highlights the dynamic complexity of the antitumoral immune response in NSCLC, across space and time. The finding that many of the ubiquitously expressed TCRs are detected in patients’ blood samples at the time of surgery is promising. The ability to track and isolate these cells from the peripheral blood rather than tumor tissue represents a noninvasive method for monitoring and accessing neoantigen-specific T cells for personalized immunotherapeutic strategies, ultimately to improve outcomes for patients with NSCLC.