Colorectal cancer (CRC) is among the most common cancers worldwide, and when it spreads to distant organs, survival drops sharply. Brain metastases occur in roughly 1-2% of CRC patients, but when they do develop, they are devastating - median survival after diagnosis is only 3-7 months.
New therapies targeting the immune system, called immune checkpoint inhibitors, have shown promise in some CRC patients. However, these treatments don't work equally well in all settings. To improve outcomes, researchers need to understand how the immune environment inside brain tumors differs from that in the original colon or rectal tumor.
This study set out to map the immune landscape inside CRC brain metastases - focusing specifically on tumor-associated macrophages (TAMs), a type of immune cell that can either fight cancer or help it grow, and on PD-L1, a protein that acts as a brake on immune responses. Understanding this architecture is a critical step toward better therapies.
Macrophages are versatile immune cells that can take on very different roles depending on their environment. Researchers often describe this along an M1/M2 axis: M1 macrophages are inflammatory and can attack tumor cells, while M2 macrophages suppress immune responses and support tumor growth.
In most solid tumors, the macrophages that infiltrate the tumor - called tumor-associated macrophages (TAMs) - tend to be skewed toward the M2, pro-tumor side. This is associated with worse outcomes. Researchers use specific protein markers to identify these populations: CD86 marks M1-type macrophages and CD163 marks M2-type macrophages, while CD68 marks all macrophages regardless of type.
Crucially, macrophages also control immune checkpoints. TAMs are a major source of PD-L1, the protein that tells T cells (the immune system's attack cells) to stand down. In brain metastases, where the immune environment is already highly regulated, this macrophage-driven PD-L1 expression may be a key reason why immunotherapy often fails.
The study analyzed tumor tissue from 50 patients with CRC brain metastases treated at a single hospital between 2007 and 2022. Of these, 31 patients had matching samples from both their primary colon or rectal tumor and their brain metastasis, allowing direct comparisons between the two sites.
The researchers built tissue microarrays (TMAs) - compact collections of tiny tissue cores arranged on a single slide - allowing standardized analysis of all samples at once. Each tumor was sampled from three distinct regions: the tumor nests (clusters of cancer cells), the tumor stroma (the surrounding support tissue), and the invasive front (the edge where tumor meets normal tissue). This spatial approach is crucial because macrophage distribution varies enormously across these zones.
Slides were stained using immunohistochemistry (IHC) for CD68, CD86, CD163, and PD-L1. Each marker's intensity was scored using a semi-quantitative system called the immunoreactive score (IRS), and digital image analysis software helped ensure objective measurements across hundreds of tissue cores.
Macrophages were present throughout both primary CRC tumors and brain metastases, but their distribution was far from uniform. In both tumor types, macrophages were far more abundant in the stroma than in the tumor cell nests themselves. This compartment-specific pattern was statistically significant for all three macrophage markers.
Across all regions, CD163+ (M2-type) macrophages outnumbered CD86+ (M1-type) macrophages. This means the immune environment is broadly immunosuppressive - the type of macrophages that promote tumor survival dominate, while the anti-tumor macrophages are relatively scarce. This M2-skewed pattern was found in both primary tumors and brain metastases.
The key new finding was that brain metastases had significantly higher levels of CD163+ macrophages in the stroma compared to the matched primary tumors (p = 0.048). This enrichment was specific to the stromal compartment and was not accompanied by a general increase in all macrophage types - suggesting a targeted shift in the quality of the immune environment rather than simply more immune cells overall.
A particularly striking finding emerged when the researchers looked at how macrophage levels in the primary CRC tumor related to when patients developed brain metastases. High levels of CD68+ macrophages - the general macrophage marker - in the tumor nests and at the invasive front were both associated with developing brain metastases much sooner.
Specifically, patients with high CD68 expression at the invasive front had a median brain metastasis-free survival of only 9.5 months, compared to 37 months for patients with low CD68 expression. At the tumor nest level, the difference was 3 months versus 27 months. These are dramatic differences that point to a role for macrophages in driving early metastatic spread to the brain.
Interestingly, once brain metastases were established, the TAM levels within those lesions did not predict overall survival. This suggests macrophages may primarily matter early in the disease process - helping cancer cells escape the primary tumor and seed the brain - rather than determining how patients fare once the brain tumor is present.
Dexamethasone is a steroid medication routinely given to patients with brain tumors or brain metastases to reduce dangerous brain swelling. It is often considered an essential supportive treatment. However, this study found an important potential downside.
Patients who had received dexamethasone before surgery showed significantly lower levels of CD86+ (M1-type, anti-tumor) macrophages in their brain metastases (p = 0.002). This suggests that steroids may be inadvertently suppressing the part of the immune response that can fight the tumor - the pro-inflammatory M1 macrophages.
This finding aligns with other research showing that dexamethasone use in brain tumors is associated with worse outcomes and potentially reduced benefit from immunotherapy. While steroids remain necessary to manage brain swelling, these results suggest that minimizing steroid exposure when possible - especially in patients being considered for immunotherapy - could be worth investigating in future clinical trials.
PD-L1 is a key biomarker in cancer immunotherapy. When tumors express PD-L1, it acts as a signal telling the immune system not to attack - and drugs that block PD-L1 (immune checkpoint inhibitors) can release that brake. The study found that PD-L1 expression was mostly found on immune cells rather than on the cancer cells themselves.
Comparing matched pairs of primary tumors and their corresponding brain metastases revealed frequent discordance - meaning PD-L1 status in the primary tumor often did not match that in the brain metastasis. In 11 of 30 matched pairs, the two sites had different PD-L1 status, with a tendency for brain metastases to acquire PD-L1 positivity that wasn't present in the original tumor.
This has important implications for treatment decisions. If doctors only test the primary tumor for PD-L1 and use that result to decide whether to give immunotherapy, they may get a misleading picture. The data suggest that testing the metastatic tissue directly - when feasible - would give more accurate information for guiding immunotherapy decisions in patients with CRC brain metastases.
The study's findings paint a coherent picture of why CRC brain metastases are so difficult to treat with immunotherapy. The brain creates a unique, immunosuppressive niche that strongly favors M2-type macrophages. This macrophage-driven suppression may be a key reason why checkpoint inhibitors - which work well in some CRC patients elsewhere in the body - have limited effectiveness in brain metastases.
The data point to macrophage-directed therapies as a potentially important complement to existing treatments. Strategies that either reduce M2 macrophage numbers, repolarize them toward an M1-type anti-tumor state, or block the signals they use to suppress the immune response could help overcome this resistance. Clinical trials exploring such approaches in brain metastasis patients are a logical next step.
The finding about dexamethasone is also clinically actionable: it adds to growing evidence that steroid-minimizing strategies in patients with brain metastases may preserve immune activity and improve the efficacy of immunotherapy. Combined with the PD-L1 discordance data, the overall message is that CRC brain metastases need to be understood and treated as biologically distinct from the primary tumor, not simply as copies of it in a different location.
This study provides the most detailed characterization to date of the immune microenvironment in CRC brain metastases. The core finding is that these tumors harbor a distinct, stroma-dominated, M2-polarized immune landscape that is qualitatively different from the primary CRC tumor - a difference driven by the unique biology of the brain as a metastatic site.
Macrophages emerge as central players on multiple fronts: they predict who will develop brain metastases early, they are modulated by corticosteroid treatment, and they are the primary source of PD-L1, the dominant immune checkpoint in this setting. This positions TAMs as a priority therapeutic target in this disease context.
The retrospective design and relatively small cohort size (50 patients) are acknowledged limitations. Future studies using spatial transcriptomics or multiplex imaging - technologies that can capture even more detail about the immune landscape - will be needed to fully validate these findings and translate them into new treatments. Larger prospective studies that test macrophage-targeted strategies alongside checkpoint inhibitors in CRC brain metastases are a clear priority.