Uncovering Cellular Interactome Drivers of Immune Checkpoint Inhibitor Response in Advanced Melanoma Patients
Open Access
- 8 September 2025
- journal article
- research article
- Published by Springer Nature in Cellular and Molecular Bioengineering
- Vol. 18 (5) , 519-541
Abstract
Purpose: Despite the success of immune checkpoint inhibitors (ICIs) that target immunosuppressive interactions, treatment resistance remains a major clinical challenge. The tumor microenvironment is comprised of tumor, immune, and stromal cell types that communicate through secreted and cell surface proteins. This can be represented by a weighted, directed network where pairs of cell types communicate via multiple ligand-receptor interactions with varying strengths. Identifying interaction network motifs that are linked with outcome or evolve pre- to post-ICI presents a rational framework to identify combination therapeutic targets. Methods: Interaction inference was performed on publicly available single-cell RNA-sequencing data from melanoma patients. The constructed patient-specific networks were input to multivariate statistical learning approaches to identify network motifs that predicted response pre-treatment and that shifted pre- to post-treatment. Relevance of interactions was validated by (1) differential expression of related pathways in single cell RNA sequencing (scRNA-seq) data, (2) survival associations in an independent bulk RNA-seq dataset, and (3) repeated analyses of scRNA-seq data in a second cohort. Results: Immune-immune interactions with roles in T cell activation, chemotaxis, and adhesion were upregulated in patients who respond to therapy pre-treatment. Related pathways were perturbed in involved immune cells and expression of these genes was associated with improved survival. The interactome also distinguished pre- and post-treatment biopsies with high accuracy despite no significant differences in individual interactions. Analysis in the validation dataset with mixed responses pre-treatment recapitulated results from the discovery analyses. Conclusion: Unbiased analysis of interaction networks and their evolution is a powerful framework to guide prognostic indicators and novel combination targets to improve patient outcomes.Keywords
Funding Information
- National Institute of General Medical Sciences (T32-GM145443)
- Division of Microbiology and Infectious Diseases (T32-AI007496)
This publication has 140 references indexed in Scilit:
- Prediction of patient‐specific quality assurance for volumetric modulated arc therapy using radiomics‐based machine learning with dose distributionJournal of Applied Clinical Medical Physics, 2023
- Improving performance of medical image alignment through super-resolutionBiomedical Engineering Letters, 2023
- Use of First-Line Immune Checkpoint Inhibitors and Association With Overall Survival Among Patients With Metastatic Melanoma in the Anti–PD-1 EraJAMA Network Open, 2022
- The F11 Receptor (F11R)/Junctional Adhesion Molecule-A (JAM-A) (F11R/JAM-A) in cancer progressionMolecular and Cellular Biochemistry, 2021
- SNR‐enhanced diffusion MRI with structure‐preserving low‐rank denoising in reproducing kernel Hilbert spacesMagnetic Resonance in Medicine, 2021
- Tissue‐resident mucosal‐associated invariant T (MAIT) cells in the human kidney represent a functionally distinct subsetEuropean Journal of Immunology, 2020
- Lubricin/Proteoglycan 4 Binding to CD44 Receptor: A Mechanism of the Suppression of Proinflammatory Cytokine–Induced Synoviocyte Proliferation by LubricinArthritis & Rheumatology, 2015
- The immunoregulatory role of type I and type II NKT cells in cancer and other diseasesCancer Immunology, Immunotherapy, 2014
- THE LASSO METHOD FOR VARIABLE SELECTION IN THE COX MODELStatistics in Medicine, 1997
- Signaling by vascular cell adhesion molecule‐1 (VCAM‐1) through VLA‐4 promotes CD3‐dependent T cell proliferationEuropean Journal of Immunology, 1991