Abstract LB275: Transcriptomic reversal analysis yields cytokines and drugs mediating tumor microenvironmental reprogramming during cancer progression and therapy response

• Published in: Cancer research (Chicago, Ill.) Ill.), 2023-04, Vol.83 (8_Supplement), p.LB275-LB275
• Main Authors: Gopalan, Vishaka, Hannenhalli, Sridhar
• Format: Article
• Language: English
• ISSN: 1538-7445; 1538-7445
• DOI: 10.1158/1538-7445.AM2023-LB275

Recent computational work links therapy response to a drug based on the drug’s ability to reverse a tumor’s transcriptome towards a healthy state. This idea of reversal has been used to mine databases of cell-line transcriptional responses against drug libraries to prioritize anti-cancer drugs. However, there are two key shortcomings in these approaches: (1) though cytokines and their receptors are proposed as modulators of therapy response, there is no reversal-based method to prioritize cytokines as potential drugs or targets, and (2) responses of microenvironmental cell types to drugs, which dictate therapy response, have not been considered. We address these limitations by exploiting recent databases of cytokine transcriptional response and single-cell RNA-seq datasets of patient responses to cancer therapies. We first used a novel approach to derive drug response signatures from LINCS data that retained the coordinated nature of gene expression changes occurring during treatment. We then used these signatures to compute a transcriptional reversal score that ranks drugs by their ability to reverse TCGA RNA-seq profiles of a tumor towards its corresponding normal in GTEx. We found that FDA-approved drugs in prostate, lung, colorectal and breast adenocarcinomas have a significantly higher reversal potential than unapproved drugs, and that these drugs are more effective at in vitro cell killing amongst CTRP and GDSC cell viability measurements. Next, we extended our signature derivation and reversal computation approach to find cytokines in the CytoSig database that can reverse TCGA RNA-seq profiles. We found that the higher the reversal potential of a cytokine, the greater its association with better overall patient survival. In particular, our approach revealed the IL10 family and IL24 as potentially therapeutic cytokines based on their predicted ability to reverse cell states across tumor types and indeed found them to be associated with better overall and progression-free survival in the TCGA cohort. Finally, in two clinical cohorts where RNA-seq was carried out on breast cancer and multiple myeloma patients before and after therapy, cell types within the tumor microenvironment of responders showed a stronger reversal towards a healthy state compared to non-responders. Altogether, our work establishes the importance of transcriptional reversal in therapy response, particularly the role of microenvironmental transcriptional alterations. This allows us