Front Immunol. 2026 ;17
1816086
Lisa Marie Grape,
Benjamin Hanke,
Belal Neyazi,
Ali Rashidi,
Vanessa Magdalena Swiatek,
Anna Schaufler,
Claudia Alexandra Dumitru,
Ulf Kahlert,
Roland Croner,
Christian Mawrin,
Sabine Franke,
I Erol Sandalcioglu,
Klaus-Peter Stein.
Background: Brain metastases from colorectal cancer (CRC) are associated with poor survival and limited treatment options. As immunomodulatory therapies gain relevance, a deeper understanding of the tumor microenvironment (TME) in this setting is needed. Tumor-associated macrophages (TAMs) are pivotal regulators of tumor immunity, yet their spatial organization, polarization, and relationship to PD-L1-mediated immune checkpoint regulation in CRC brain metastases remain poorly defined. We therefore characterized the compartment-specific architecture and functional orientation of TAMs in brain metastases and matched primary CRCs.
Methods: Immunohistochemical analyses of CD68 (pan-macrophages), CD86 (M1-associated), CD163 (M2-associated), and PD-L1 were performed on tissue microarrays from tumor specimens of 50 patients with CRC brain metastases, including 31 matched primary tumor-brain metastasis pairs. Compartment-specific TAM densities and PD-L1 expression were quantified to assess intra- and intertumoral heterogeneity and correlated with clinicopathological parameters and clinical outcomes.
Results: Both primary CRCs and brain metastases exhibited macrophage-rich TMEs characterized by stromal predominance and an M2-skewed polarization. Compared with matched primary tumors, brain metastases showed a significant stromal enrichment of CD163+ TAMs. Dexamethasone treatment was associated with reduced densities of CD86+ TAMs in brain metastases. PD-L1 expression was predominantly confined to immune cells and displayed marked intra- and intertumoral heterogeneity, with frequent discordance between matched primary tumors and brain metastases, including recurrent lesions. In primary CRCs, high densities of CD68+ TAMs at the invasive front were associated with shortened brain metastasis-free survival, whereas neither TAM infiltration nor PD-L1 expression correlated with overall survival.
Conclusion: CRC brain metastases exhibit a distinct, stroma-dominated and M2-polarized TME, consistent with site-specific enrichment of protumoral TAM phenotypes within the cerebral niche. This may reflect advanced disease biology, immunological adaptation to the brain microenvironment, or therapy- and selection-driven immune remodeling during metastatic progression. The association between dexamethasone treatment and reduced M1-associated TAM infiltration suggests therapy-related modulation of antitumoral immune activity, with potential implications for perioperative management. The heterogeneity and frequent discordance of PD-L1 expression highlight its dynamic regulation and support individualized assessment of metastatic lesions prior to immunotherapy. Collectively, these findings support site-specific immune profiling and identify TAMs as promising therapeutic targets within the TME of CRC brain metastases.
Keywords: PD-L1; TAMs; brain metastases; colorectal cancer; tumor microenvironment; tumor-associated macrophages