Tumor γδ T-cell abundance is associated with favorable cancer treatment outcomes
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Purpose
Clinical response to immune checkpoint blockade (ICB) remains variable. We performed a reference-wide screen to identify immune-cell populations in the tumor microenvironment (TME) associated with benefit across treatment modalities and tumor types.
Experimental Design
We analyzed pretreatment bulk tumor RNA-seq to identify RNA-abundance patterns associated with favorable outcomes in (1) published ICB cohorts with response labels and (2) pretreatment TCGA cohorts with survival outcomes. To identify associated cell types, we compared favorable expression patterns to previously determined patterns for 157 cell types defined by the Human Primary Cell Atlas (HPCA), spanning immune, stromal, vascular, and epithelial lineages. Cell types were associated with favorable outcomes using Cox models (TCGA survival) and mixed-effects models of responder status that accounted for cohort, therapy class, tumor type (ICB cohorts). ICB analyses were restricted to pretreatment samples. Random-effects meta-analysis was used for the PRECOG cross-platform replication. After γδ T-cells emerged as the top-ranked population associated with ICB response, we adjusted γδ associations for CD8 + T-cell abundance, and evaluated γδ signals with TRUST4-based T-cell receptor gamma (TRG) and delta (TRD) gene reconstruction and single-cell RNA-seq. To test whether ICB-favorable cellular landscapes were shared with non-immunotherapy treatment response, we used TCGA RECIST-evaluable treatment records to derive immune-cell profiles associated with treatment response for chemotherapy, radiation, targeted therapy, and hormone therapy.
Results
The γδ T-cell program was the top-ranked cell-type program associated with ICB response and was also among the programs most strongly associated with favorable TCGA survival. Cell types associated with ICB response showed strong descriptive concordance with the chemotherapy-response profile derived from recorded TCGA outcomes (Pearson r of ranks = 0.92) and moderate concordance with the analogous radiation-associated profile (r = 0.58). Analyses of targeted and hormone therapies were limited by small sample sizes and showed weaker concordance. In pretreatment tumor biopsies, inferred γδ T-cell abundance was associated with ICB response and with survival, and these associations were not explained simply by CD8 + T-cell abundance. In TCGA, TRUST4-derived γδ T-cell receptor (TCR) abundance estimates correlated with signature-based γδ scores and stratified survival. In independent single-cell RNA-seq datasets, higher γδ TCR RNA abundance was also associated with a higher rate of ICB response.
Conclusion
A reference-wide cell-type screen identified γδ T-cells as the leading population associated with ICB response, and pretreatment γδ T-cell abundance was also associated with favorable outcomes in immunotherapy and non-immunotherapy settings. The evidence is associative and varies in strength by tumor context but supports prospective evaluation of γδ T-cell abundance as a component of pretreatment immune profiling alongside CD8, B-cell/TLS, stromal, and epithelial-state features.
Translational Relevance
Across cancers, clinical response to immune checkpoint blockade (ICB) is variable, with inconsistent performance from established biomarkers such as PD-L1 and tumor mutation burden. We asked which immune-cell transcriptional programs were associated with favorable outcomes across immunotherapy and non-immunotherapy treatment settings. Using pretreatment bulk tumor RNA-seq from published ICB cohorts and TCGA, we evaluated the inferred abundance of 157 subtypes in the tumor microenvironment (TME) with a common gene-level and sample-level framework. Pretreatment TMEs enriched in lymphoid cells or γδ T-cells were associated with ICB response and with favorable survival. The immune-cell landscape associated with favorable ICB outcomes was strongly concordant with chemotherapy response and moderately concordant with radiation response. The association between γδ abundance and favorable outcomes remained after accounting for CD8 + T-cell abundance. These observational findings support γδ T-cell abundance as a candidate for pretreatment immune profiling that warrants prospective evaluation.
