Using DNA sequencing data to quantify T cell fraction and therapy response

Abstract

The immune microenvironment influences tumour evolution and can be both prognostic and predict response to immunotherapy 1,2 . However, measuring tumour infiltrating lymphocytes (TILs) is restricted by lack of appropriate data. Whole exome sequencing (WES) of DNA is frequently performed to calculate tumour mutational burden and identify actionable mutations.Here we develop a method for T cell fraction estimation from WES samples, utilising a signal from T cell receptor excision circle (TRECs) loss during VDJ recombination of the T cell receptor alpha (TCRA) gene. This score significantly correlates with orthogonal TIL estimates and is agnostic to sample type. Blood TCRA T cell fraction is higher in females and correlates with both tumour immune infiltrate and presence of bacterial sequencing reads. Tumour TCRA T cell fraction is prognostic in lung adenocarcinoma and using a meta-analysis of immunotherapy-treated tumours, we show that this score predicts immunotherapy response, providing value beyond tumour mutational burden. Applying this score to a multi-sample pancancer cohort revealed high diversity in immune infiltration within tumours. Subclonal loss of 12q24.31-32, encompassing SPPL3, was associated with reduced TCRA T cell fraction. Our method, T cell ExTRECT (T cell Exome TREC Tool), quantifies the T cell infiltrate of WES samples.Here we propose a method for the estimation of the T cell fraction present in a WES sample. This method utilises a somatic copy number-based signal from VDJ recombination and the loss of TRECs. We explore the underlying features which predict T cell infiltrate in tumours and blood and evaluate determinants of immune heterogeneity within tumours. Finally, we demonstrate that our estimated T cell fraction can be used as a predictor of clinical response to CPI therapy. Results Inferring T cell fraction from WES dataT cell diversity, which is required for immune system recognition of foreign antigens, is a product of VDJ recombination, where segments within the T cell receptor genes recombine.The alpha chain of the T cell receptor is encoded by the TCRA gene (also known as TRA) and the result of VDJ recombination is the excision of unselected gene segments from TCRA as TRECs, with the T cell undergoing a deletion event within TCRA.Tools to infer cancer somatic copy number alteration (SCNA) [6][7][8][9] rely on the read depth ratio (RDR), reflecting the log of the ratio of reads between the tumour sample and its matched control (e.g. buffy coat in a centrifuged blood sample). Deviation in the RDR from zero is assumed to reflect a tumour SCNA. However, within TCRA this assumption does not hold; a deviance in the RDR may reflect T cell specific deletion events and SCNA tools may thus erroneously infer tumour SCNA. Indeed, in the TRACERx100 cohort multiple SCNA within TCRA were inferred in 165/327 tumour regions (Extended Data Fig. 1a). The RDR deviated the most within segments frequently included within TRECs (Extended Data Fig. 1b-c). This suggests that most detected SCNAs within TCRA...

Biological and Molecular Factors Predicting Response to Adoptive Cell Therapies in Cancer

“…These DNA methylation profiles that are able to predict clinical response to immunotherapy have also been extended to melanoma patients ( 5 ). In addition, recent development of sophisticated computational methods allows the inference of T-cell infiltration into tumor samples from whole exome sequencing data, which can predict the response to immune checkpoint inhibitors independently of tumor mutational burden ( 6 ). Finally, the advancement of single-cell technologies also holds great potential to refine the search for predictors of treatment response, personalize patients’ management, and increase the responders to nonresponders’ ratio ( 7-9 ).…”

Mutation burden-orthogonal tumor genomic subtypes delineate responses to immune checkpoint therapy

“…Furthermore, the results were similar when using the cohort’s optimal TMB cutoff determined by the ROC curve and the Youden index or using log(10)-transformed TMB as a continuous variable (Figure S15). Even when the recently reported score for estimating T-cell infiltration in tumors from WES data 21 was added as a covariate, there remained a significant correlation between irGS and ICI response (Figure S16). Genome subtyping and ICI response analysis by GS-PRACTICE for each of the 13 individual ICI studies comprising the combined 973 patients are described in Figure S17 and Table S4.…”

Quantity of Immune Cells Predict Response to Immunotherapy in Cancer

“…then TILs, using morphology, evaluated according to established guidelines, or with immunohistochemistry for CD8, or even by DNAsequencing [10], may one day be used in daily practice in a rational and biological plausible manner.…”