Abstract
BackgroundThe balance between immune-stimulatory and immune-suppressive mechanisms in the tumour microenvironment is associated with tumour rejection and can predict the efficacy of immune checkpoint-inhibition therapies.MethodsWe consider the observed differences between the transcriptional programmes associated with cancer types where the levels of immune infiltration predict a favourable prognosis versus those in which the immune infiltration predicts an unfavourable prognosis and defined a score named Mediators of Immune Response Against Cancer in soLid microEnvironments (MIRACLE). MIRACLE deconvolves T cell infiltration, from inhibitory mechanisms, such as TGFβ, EMT and PI3Kγ signatures.ResultsOur score outperforms current state-of-the-art immune signatures as a predictive marker of survival in TCGA (n = 9305, HR: 0.043, p value: 6.7 × 10−36). In a validation cohort (n = 7623), MIRACLE predicts better survival compared to other immune metrics (HR: 0.1985, p value: 2.73 × 10−38). MIRACLE also predicts response to checkpoint-inhibitor therapies (n = 333). The tumour-intrinsic factors inversely associated with the reported score such as EGFR, PRKAR1A and MAP3K1 are frequently associated with immune-suppressive phenotypes.ConclusionsThe association of cancer outcome with the level of infiltrating immune cells is mediated by the balance of activatory and suppressive factors. MIRACLE accounts for this balance and predicts favourable cancer outcomes.
Highlights
The balance between immune-stimulatory and immune-suppressive mechanisms in the tumour microenvironment is associated with tumour rejection and can predict the efficacy of immune checkpoint-inhibition therapies
Correlative studies in humans and experimental models suggest that checkpoint inhibitors are less effective in tumours characterised by a primary immune suppression, including the ones with low mutational load,[1,8] or dominated by the genomic dysregulations of oncogenic pathways leading to T cell exclusion such as WNT/beta-catenin,[9,10,11] mitogenactivated protein kinase[6,7,12] and transforming growth factor-β (TGFβ) pathways[13,14,15] The dichotomy between “immune active” and “immune silent” might be useful to explain a general phenomenon but do not reflect the high level of inter-patient heterogeneity and do not take into account the contribution of antagonist signals involved in primary immune suppression
We first identify cancer types from The Cancer Genome Atlas (TCGA) database in which a highly active immune phenotype is associated with favourable survival (HR < 1 with a p value < 0.1; Fig. S1) and cancer types in which this phenotype is associated with decreased survival (HR > 1 with a p value < 0.1; Fig. S1)
Summary
The balance between immune-stimulatory and immune-suppressive mechanisms in the tumour microenvironment is associated with tumour rejection and can predict the efficacy of immune checkpoint-inhibition therapies. The expression of such molecules as well as of other immuneregulatory markers (e.g., FOXP3, CTLA4 and PD-1) characterise a compensatory immune resistance, reflecting the presence of counter-regulatory mechanisms that follow, rather than precede,[4,5] the recognition of tumour antigens by T cells and the subsequent amplification of the inflammatory response.[6,7] Correlative studies in humans and experimental models suggest that checkpoint inhibitors are less effective in tumours characterised by a primary immune suppression ( called as “primary immune ignorance”), including the ones with low mutational load,[1,8] or dominated by the genomic dysregulations of oncogenic pathways leading to T cell exclusion such as WNT/beta-catenin,[9,10,11] mitogenactivated protein kinase[6,7,12] and transforming growth factor-β (TGFβ) pathways[13,14,15] The dichotomy between “immune active” (associated with the displaying of compensatory immune resistance) and “immune silent” (typified by the presence of a primary immune suppression) might be useful to explain a general phenomenon but do not reflect the high level of inter-patient heterogeneity and do not take into account the contribution of antagonist signals involved in primary immune suppression. One of the major limitations of the transcriptomic studies performed so far is the use of gene signatures or modules that capture only a dominant process or a group of processes tightly interconnected or correlated among each other
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
