# Comprehensive spatial and immune profiling of metastatic mismatch repair–deficient colorectal cancer reveals response to immunotherapy

## TL;DR

This study explores why some mismatch repair-deficient colorectal cancers respond to immunotherapy while others don't, using detailed spatial and immune profiling.

## Contribution

The study reveals how spatial immune architecture and macrophage polarization influence immunotherapy response in metastatic dMMR colorectal cancer.

## Key findings

- Responsive lesions showed enriched cytotoxic and interferon-γ signatures and greater CD8+ T-cell infiltration.
- Resistant lesions had immunosuppressive features like SPP1+ and CD163+ macrophages and TGF-β–mediated matrix remodeling.
- Spatial proximity between PD-L1+ tumor cells and M1-like macrophages was associated with treatment response.

## Abstract

Mismatch repair-deficient (dMMR) colorectal cancers (CRCs) exhibit variable clinical responses to immune checkpoint inhibitors (ICIs) despite their high immunogenicity.

To investigate the molecular and spatial determinants of this heterogeneity, we analyzed five metastatic sites (colon, liver, peritoneum, left ovary, and right ovary) from a dMMR CRC patient treated with pembrolizumab using bulk and single-cell RNA sequencing combined with multiplex immunohistochemistry.

Responsive lesions were characterized by enriched cytotoxic and interferon-γ signatures, greater CD8+ T-cell infiltration, and close spatial proximity between programmed death-ligand 1(PD-L1)+ tumor cells and M1-like macrophages. In contrast, resistant lesions demonstrated reduced effector cell presence and were enriched in immunosuppressive programs, including SPP1+ and CD163+ macrophages, noncanonical WNT5A/B signaling, and TGF-β–mediated matrix remodeling.

Collectively, these findings highlight the importance of spatial immune architecture and macrophage polarization in shaping ICI responses in dMMR CRC and underscore the need for spatial profiling to guide immunotherapy strategies in metastatic disease.

Graphical Abstract