The landscape of cancer rewired GPCR signaling axes

Abstract

We explored the dysregulation of GPCR ligand signaling systems in cancer transcriptomics datasets. We derived a network of interacting ligands and biosynthetic enzymes from public databases, that we combined with cognate GPCRs and downstream effectors to quantify GPCR signaling pathways. We found multiple GPCRs differentially regulated together with their ligands across cancers, suggesting a widespread perturbation of these signaling axes in specific cancer molecular subtypes. We showed that biosynthetic pathway enrichment from enzyme expression recapitulates pathway activity signatures from metabolomics datasets, therefore providing valuable surrogate information for receptor-organic ligand systems. The expression of several GPCRs signaling components is significantly associated with either lower or higher patient survival in a cancer subtype specific fashion. In particular, the expression of both receptor-ligand (or biosynthetic enzyme) interaction partners improves the stratification of patients based on survival, suggesting a potential synergistic role for the activation of certain receptor axes in modulating cancer phenotypes. Strikingly, we have found that receptors from these actionable axes are the targets of several drugs displaying anti-growth effects in large, drug repurposing screens in cancer cell lines. This study provides a comprehensive map to exploit GPCR signaling axes as actionable targets for personalized cancer treatments. We have made the results generated in this study freely available for further exploration to the community through a webapp (gpcrcanceraxes.bioinfolab.sns.it).

Learning and actioning general principles of cancer cell drug sensitivity

“…For instance, two Adenosine Receptors antagonists, i.e. CGS-15943 and MRS-1220, which have been recently proposed as effective therapies against multiple cancers (Arora et al, 2023; Corsello, Spangler, et al, 2020), are indeed predicted for multiple samples of breast (BRCA), liver (LIHC), prostate (PRAD) as well as gastric (STAD) cancers (Figure 6E).…”

“…Considering the high anti-cancer potential of many non-oncological drugs against certain targets (e.g. GPCR drugs) (Arora et al, 2023; Wu et al, 2023), it will be interesting to explore in the future alternative ML frameworks to model the sensitivity of those drugs for which we obtained low performances due to high specificity and low variability of the response.…”

“…For instance, two Adenosine Receptors antagonists, i.e. CGS-15943 and MRS-1220, which have been recently proposed as effective therapies against multiple cancers (Arora et al, 2023;, are indeed predicted for multiple samples of breast (BRCA), liver (LIHC), prostate (PRAD) as well as gastric (STAD) cancers (Figure 6E). Therefore, we demonstrated that the CellHit models can reliably predict anti-cancer therapies based on patients' transcriptomics signatures.…”