Mechanisms and Emerging Strategies to Overcome Immunotherapy Resistance in Cold Tumours of Colorectal Cancer

Introduction

Colorectal cancer (CRC) remains a leading cause of cancer-related mortality worldwide.¹ Globally, CRC accounts for nearly 1.9 million new cases and over 900,000 deaths annually,² with metastatic disease carrying a 5-year survival rate below 15%.^3,4 For decades, the therapeutic arsenal for metastatic CRC (mCRC) has been dominated by cytotoxic chemotherapy and biologic agents targeting angiogenesis and the epidermal growth factor receptor.^5,6 While these treatments have improved outcomes, responses are often transient and long-term survival remains limited.

The emergence of cancer immunotherapy,⁷ particularly immune checkpoint inhibitors (ICIs),⁸ has reshaped treatment for many malignancies.⁹ However, in CRC, ICIs have revealed a striking dichotomy. Their efficacy is almost exclusively confined to the small subset (≈4–5%) of patients with mismatch repair-deficient/microsatellite instability-high (dMMR/MSI-H) tumors.¹⁰ These “immunologically hot” tumors are sensitive to PD-1/PD-L1 blockade, demonstrating durable responses. In contrast, the vast majority of patients with microsatellite stable (MSS) or mismatch repair-proficient (pMMR) tumors exhibit primary resistance to single-agent ICIs, their “cold” immunological landscape posing a formidable challenge.^11,12

The biological basis of this dichotomy lies in distinct tumor microenvironment (TME) features. dMMR/MSI-H tumors have high mutational burden, abundant neoantigens, and dense CD8⁺ T-cell infiltration, whereas MSS/pMMR tumors are characterized by low mutational burden, T-cell exclusion, and dominance of immunosuppressive cells such as myeloid-derived suppressor cells and M2-polarized macrophages.^12,13 Consequently, ICIs are standard of care in dMMR/MSI-H mCRC, while no immunotherapy has yet been approved for MSS/pMMR disease.

Methods:This narrative review searched PubMed and ClinicalTrials.gov for articles published between January 2015 and December 2025 using terms including “colorectal cancer”, “immunotherapy”, “immune checkpoint inhibitors”, “MSS”, “pMMR”, “tumor microenvironment”, and “combination therapy”. We prioritized phase II/III trials, high-impact mechanistic studies, and recent systematic reviews. References from selected articles were hand-searched for additional relevant studies.

This review dissects the biological basis for this dichotomy, summarizes landmark clinical data in dMMR/MSI-H CRC, and critically evaluates strategies to overcome resistance in MSS CRC.

The dMMR/MSI-H Subset: A Paradigm for Immunotherapy Success

Biological Rationale: An Immunogenic Phenotype

The profound efficacy of ICIs in dMMR/MSI-H CRC is underpinned by its distinct tumor biology.¹⁰ The dMMR system, responsible for correcting DNA replication errors, is functionally impaired, leading to a hypermutator phenotype. This results in a very high tumor mutational burden (TMB),¹² with thousands of somatic mutations accumulating, particularly in coding microsatellite regions. These mutations frequently generate novel frameshift peptides, which serve as highly immunogenic neoantigens. These neoantigens are efficiently presented on major histocompatibility complex (MHC) molecules, leading to the priming and activation of a robust, tumor-specific T-cell response.¹⁴ Consequently, dMMR/MSI-H tumors are characterized by a pre-existing, T-cell-inflamed tumor microenvironment (TME), with dense infiltration of CD8+ cytotoxic T-cells¹⁵ and increased expression of immune checkpoint molecules like PD-1 and PD-L1. This adaptive immune resistance mechanism creates the perfect conditions for ICIs to act, by “releasing the brakes” on an already engaged anti-tumor immune response.¹⁶

Clinical Efficacy and Evolving Standards of Care

The translation of this biology into clinical practice has been transformative. Pivotal Phase II trials, such as KEYNOTE-164 and CheckMate-142,^17,18 demonstrated unprecedented activity of pembrolizumab and nivolumab (alone or combined with ipilimumab) in previously treated metastatic dMMR/MSI-H CRC, with objective response rates (ORR) of 40–55% and, critically, remarkably durable responses.^17,18 This led to regulatory approvals and established ICIs as the standard of care in the refractory setting.¹⁰

The paradigm was further shifted by the Phase III KEYNOTE-177 trial,¹⁹ which compared pembrolizumab to standard chemotherapy as first-line treatment.²⁰ The study met its primary endpoint, demonstrating a significant improvement in median progression-free survival (16.5 vs. 8.2 months) with a more favorable toxicity profile.²¹ Based on these results, frontline pembrolizumab is now a standard first-line option, offering the potential for long-term disease control without the cumulative toxicities of chemotherapy.

More recently, the power of ICIs has been demonstrated in the neoadjuvant setting. Several landmark studies in locally advanced dMMR rectal cancer have reported a 100% clinical complete response rate²² to neoadjuvant dostarlimab or other ICIs, allowing patients to avoid radical surgery and preserve organ function.^22,23 This has ignited a new era of treatment for early-stage dMMR CRC.

The MSS/pMMR Majority: Confronting the “Cold” Tumor Challenge

It is important to note, however, that MSS CRC is not entirely homogeneous. Rare cases with POLE or POLD1 exonuclease domain mutations can exhibit ultra-high TMB and, despite being microsatellite stable, have shown responsiveness to immune checkpoint inhibitors in case series and small cohort studies. Similarly, a small fraction of MSS tumors with exceptionally high TMB through alternative mechanisms may also derive clinical benefit. Although these represent a minority of MSS cases, acknowledging this heterogeneity improves the scientific accuracy of the review.

The primary obstacle in CRC immunotherapy is the inherent resistance of MSS/pMMR tumors, which constitute over 95% of metastatic cases. The immunobiology of these tumors is characterized by multiple layers of immunosuppression¹² (Table 1).

Table 1 Comparison of Tumor Immune Microenvironment Characteristics Between dMMR/MSI-H and MSS/pMMR Colorectal Cancer

Mechanisms of Immunoresistance

The MSS phenotype is associated with low TMB and scarce neoantigens, resulting in inadequate T-cell priming.²⁴ Beyond this lack of immunogenicity, the TME is profoundly immunosuppressive. It is often infiltrated by myeloid-derived suppressor cells (MDSCs)²⁵ and M2-polarized tumor-associated macrophages (TAMs),^12,13 which secrete inhibitory cytokines (IL-10, TGF-β) that suppress T-cell function.^26,27 Regulatory T-cells (Tregs) are also abundant, contributing to local immune tolerance.^12,28

A key feature of many MSS CRCs is T-cell exclusion, where cytotoxic T-cells are physically prevented from infiltrating the tumor core. This “immune-excluded” phenotype is often driven by Wnt/β-catenin activation, which may inhibit recruitment of CD103⁺ dendritic cells needed for T-cell priming and trafficking.^29,30 Additionally, VEGF, commonly overexpressed in CRC, promotes abnormal vasculature that limits T-cell entry.

These resistance mechanisms often co-exist, creating a complex immunosuppressive landscape. Each mechanism represents a potential therapeutic vulnerability: T-cell exclusion may be targeted by pathway inhibitors or epigenetic modifiers; MDSC/M2 TAM dominance suggests a rationale for CSF-1R inhibitors or myeloid-targeting agents; and immunosuppressive cytokines like TGF-β may be neutralized by specific antibodies. This understanding underpins the rational combination strategies discussed below.

Established Immune Biomarkers in MSS CRC

Several immune-related biomarkers have prognostic value in CRC and may help guide immunotherapy patient selection.

Immunoscore^® quantifies CD3⁺ and CD8⁺ T-cell densities in the tumor core and invasive margin; it has been validated as a strong prognostic factor in stage I–III CRC and is being explored for predicting ICI response in MSS tumors.^31,32

Tumor-stroma ratio (TSR) assesses the proportion of stroma versus tumor cells; a stroma-rich phenotype correlates with immune exclusion, poor CD8⁺ infiltration, and worse prognosis in CRC.^33,34

CD8⁺ T-cell localization classifies tumors as inflamed, excluded, or desert. Most MSS CRCs show excluded or desert phenotypes, which are associated with primary ICI resistance.^35,36

Gut microbiome composition modulates systemic antitumor immunity. Specific bacteria (eg., Akkermansia) have been linked to improved ICI efficacy, and fecal profiling is emerging as a non-invasive biomarker in CRC.^37,38

Integrating these complementary biomarkers into future trials may enable biologically guided patient stratification (Figure 1).

Figure 1 Immunotherapy landscape and therapeutic framework in colorectal cancer (CRC). Immunotherapy efficacy in CRC is dictated by MMR/MSI status. dMMR/MSI-H “hot” tumors (≈4–5%) are ICI-sensitive; pMMR/MSS “cold” tumors (≈95%) are resistant. Strategies to overcome resistance include combination TME-remodeling approaches and next-generation immunotherapies.

Strategies to Sensitize MSS Colorectal Cancer to Immunotherapy

Overcoming the intrinsic resistance of MSS CRC requires a paradigm shift from empirical combination testing to mechanism-based strategies that target specific barriers within the immunosuppressive tumor microenvironment (TME). The heterogeneous nature of immune evasion in MSS tumors—ranging from T-cell exclusion and myeloid cell dominance to soluble immunosuppressive factors—demands a tailored approach that matches the dominant resistance mechanism with rationally selected interventions. This section systematically reviews emerging strategies organized by their primary mechanistic target, providing a framework for personalized immunotherapy in MSS CRC.

Targeting T-Cell Exclusion and Immune Desert Phenotypes

A subset of MSS CRCs exhibits an “immune-desert” or “immune-excluded” phenotype, characterized by the absence of T-cells within the tumor core despite their presence at the invasive margin or in surrounding stroma. This T-cell exclusion is often driven by oncogenic signaling pathways that actively prevent lymphocyte infiltration.

The Wnt/β-catenin pathway represents a prominent mechanism of immune exclusion. Constitutive activation of this pathway, frequently resulting from APC or CTNNB1 mutations, suppresses the production of CCL4, a chemokine essential for recruiting CD103+ dendritic cells. Without these dendritic cells, T-cell priming and subsequent trafficking into the tumor core are severely impaired.^29,39 Preclinical studies have demonstrated that pharmacological inhibition of Wnt/β-catenin signaling can restore chemokine expression, enhance dendritic cell recruitment, and sensitize tumors to immune checkpoint blockade. Several small molecule inhibitors targeting this pathway are currently in early-phase clinical development, though their tolerability and efficacy in combination with ICIs remain to be established.

Epigenetic modifiers represent another avenue for reversing immune exclusion. DNA methyltransferase inhibitors (eg., decitabine) and histone deacetylase inhibitors (eg., vorinostat) have been shown to upregulate the expression of endogenous retroviral elements, triggering a viral mimicry response that enhances tumor immunogenicity and promotes T-cell infiltration.⁴⁰ Early-phase trials combining epigenetic agents with ICIs in solid tumors have shown promising signals, though data specific to MSS CRC are still emerging.

For tumors with true immune-desert phenotypes—those lacking any evidence of a pre-existing T-cell response—therapeutic vaccines may be necessary to initiate de novo T-cell priming. Personalized neoantigen vaccines, generated by sequencing a patient’s tumor to identify mutation-derived immunogenic peptides, have demonstrated the ability to induce polyfunctional, tumor-specific T-cell responses in melanoma and other cancers.^24,41 In MSS CRC, early-phase studies are exploring neoantigen vaccines alone or in combination with ICIs, with the goal of creating an initial T-cell infiltrate that can then be unleashed by checkpoint blockade.⁴²

Remodeling the Immunosuppressive Myeloid Compartment

Myeloid cells dominate the immunosuppressive network in MSS CRC. MDSCs and M2-polarized TAMs accumulate abundantly and suppress T-cells via arginase, reactive oxygen species, and inhibitory cytokines.^12,13,25

CSF-1R targeting aims to deplete or reprogram TAMs. Preclinical studies suggest CSF-1R inhibitors can reduce TAM density, shift macrophages toward an M1-like phenotype, and enhance ICI efficacy.¹³ Clinical trials of CSF-1R inhibitors (eg., cabiralizumab, pexidartinib) combined with ICIs in advanced solid tumors, including CRC, have shown mixed results, possibly reflecting redundancy within the myeloid network.

Inhibiting myeloid recruitment is an alternative. CCR2 and CCR5 on MDSC precursors mediate their trafficking into tumors. Small molecule inhibitors of these chemokine receptors are being evaluated to prevent MDSC accumulation and relieve T-cell suppression.²⁷

PI3Kγ regulates macrophage polarization. PI3Kγ signaling promotes an M2-like phenotype; preclinical data indicate that PI3Kγ inhibition can reprogram TAMs toward an M1-like, immunostimulatory state, enhancing CD8⁺ T-cell activation and ICI sensitivity.⁴³ Selective PI3Kγ inhibitors are entering clinical development as a strategy to “re-educate” rather than deplete myeloid cells.

Counteracting Soluble Immunosuppressive Factors

The MSS TME is rich in soluble factors that directly suppress effector immune cells and promote a tolerogenic milieu. Among these, transforming growth factor-beta (TGF-β) and vascular endothelial growth factor (VEGF) play particularly prominent roles.

TGF-β exerts pleiotropic immunosuppressive effects: it inhibits the proliferation and effector function of CD8+ T-cells and NK cells, promotes the differentiation of regulatory T-cells (Tregs), and drives the activation of cancer-associated fibroblasts (CAFs) that deposit extracellular matrix and create a physical barrier to T-cell infiltration.²⁶ Preclinical studies have demonstrated that TGF-β blockade can reverse these effects, enhancing T-cell infiltration and synergizing with ICIs. Several TGF-β receptor kinase inhibitors and ligand-trapping molecules are in clinical development. Notably, a phase Ib trial combining the TGF-β inhibitor galunisertib with nivolumab in advanced solid tumors showed acceptable safety and signals of clinical activity, though dedicated studies in MSS CRC are needed.

VEGF, in addition to its well-established role in promoting angiogenesis, is a potent immunosuppressive cytokine. VEGF inhibits dendritic cell maturation, promotes the expansion of MDSCs and Tregs, and contributes to the formation of abnormal, leaky vasculature that impedes effective T-cell trafficking.^44,45 These observations provide a strong rationale for combining anti-angiogenic agents with ICIs—a strategy that aims to achieve “vascular normalization,” improving both drug delivery and immune cell infiltration while directly counteracting VEGF-mediated immunosuppression.

The clinical exploration of this approach has yielded both promise and complexity. The phase Ib REGONIVO trial evaluated regorafenib (a multi-kinase inhibitor with anti-angiogenic activity) in combination with nivolumab in patients with advanced gastric and colorectal cancer. In the MSS CRC cohort, the combination achieved an objective response rate of 33%, a striking signal given the historical resistance of MSS CRC to immunotherapy.⁴⁶ However, subsequent larger studies have failed to consistently replicate this efficacy, suggesting that patient selection, dosing schedules, and the specific anti-angiogenic agent used may critically influence outcomes. Ongoing phase III trials, such as the LEAP-017 study evaluating lenvatinib plus pembrolizumab, will provide more definitive evidence.

The failure of the phase III IMblaze370 trial—which evaluated the MEK inhibitor cobimetinib combined with atezolizumab in previously treated MSS CRC—serves as an important cautionary tale.⁴⁷ Despite a strong preclinical rationale (MEK inhibition was thought to enhance antigen presentation and reduce MDSC accumulation), the combination failed to improve survival compared to regorafenib alone. This outcome underscores several critical lessons: the complex, sometimes counterproductive immunomodulatory effects of targeted agents; the importance of patient selection; and the need for robust predictive biomarkers to guide combination development.⁴⁸

Enhancing Effector Cell Function and Persistence

The ultimate goal of immunotherapy is to generate and maintain a robust, tumor-specific effector T-cell response. For tumors where such a response exists but is suppressed, ICIs alone may suffice. For MSS CRC, however, strategies that directly provide or enhance effector cells are increasingly being explored.

Adoptive cell therapy (ACT) represents one of the most powerful approaches to generate anti-tumor immunity. Tumor-infiltrating lymphocyte (TIL) therapy involves isolating T-cells from a patient’s resected tumor, expanding them ex vivo to large numbers, and reinfusing them following lymphodepleting chemotherapy.^49,50 This approach has achieved remarkable success in melanoma, with durable response rates exceeding 50%. In MSS CRC, early-phase TIL trials are now showing promising signals of activity, with objective responses observed in a subset of patients with liver metastases.⁵⁰ The major challenges include the technical complexity of TIL expansion and the variable presence of pre-existing tumor-reactive T-cells within individual tumors.⁵¹

Chimeric antigen receptor (CAR) T-cell therapy has transformed the treatment of hematologic malignancies but faces substantial hurdles in solid tumors like CRC.¹³ These challenges include: (1) a paucity of truly tumor-specific antigens that are uniformly expressed across all tumor cells; (2) on-target/off-tumor toxicity when targeting antigens shared with normal tissues; and (3) a hostile TME that rapidly inactivates infused CAR-T cells through immunosuppressive signals, metabolic restriction, and physical barriers.^8,52

Despite these obstacles, early-phase clinical trials are providing initial evidence that CAR-T cells can exert biological activity in CRC with an acceptable safety profile. In a Phase I dose-escalation study reported by Zhang C et al 10 heavily pretreated CRC patients received CAR-T cell therapy targeting a tumor-associated antigen across five dose levels (1×10⁵ to 1×10⁸ CAR+ cells/kg). The therapy was well-tolerated with no severe treatment-related adverse events attributable to CAR-T cells. Notably, among the 7 patients who had experienced progressive disease on all prior standard therapies, CAR-T infusion resulted in disease stabilization, with 2 patients maintaining stable disease for over 30 weeks and 2 others showing objective tumor reduction on PET/CT and MRI.⁵³ This emerging data, while preliminary, validates the potential of CAR-T therapy in CRC and provides proof-of-concept that adoptively transferred cells can traffic to tumors and exert anti-tumor effects even in the challenging MSS TME.

Strategies to enhance CAR-T cell efficacy in solid tumors include engineering “armored” CARs that constitutively secrete immunostimulatory cytokines (eg., IL-12, IL-18) to counteract the suppressive TME; designing logic-gated CARs that require recognition of two antigens for activation, thereby improving tumor specificity; and combining CAR-T cells with ICIs or TME-modulating agents to create a more permissive environment for their function.⁵²

Bispecific T-cell engagers (BiTEs) offer an alternative strategy to redirect T-cell cytotoxicity without the need for ex vivo cell manipulation. These antibody-based molecules simultaneously bind a tumor-associated antigen (eg., carcinoembryonic antigen, CEA) and the CD3 complex on T-cells, effectively forcing the formation of an immunological synapse and inducing T-cell activation and tumor cell killing.^54,55 The CEA-CD3 bispecific antibody cibisatamab has shown encouraging activity in early-phase trials, particularly when combined with atezolizumab, though cytokine release syndrome (CRS) remains a dose-limiting toxicity requiring careful management⁵⁵ (Table 2).

Table 2 Summary of Immunotherapy Combination Strategies and Key Clinical Trials for MSS Colorectal Cancer

Other Emerging Modalities

STING agonists activate the cGAS-STING pathway to induce type I interferons and T-cell priming. In CRC, activation of this pathway can enhance anti-tumor immunity, but its dual role in promoting intestinal inflammation underscores the need for precise delivery.⁵⁶ Preclinical studies have shown that combining a WIP1 inhibitor with a STING agonist synergistically enhances anti-tumor efficacy by amplifying IFNβ production to activate anti-tumor immune responses.⁵⁷ Dazostinag (TAK-676), a STING agonist, is being evaluated with pembrolizumab for colorectal cancer (iintune-1, NCT04420884).⁵⁸

Oncolytic viruses selectively replicate in malignant cells and mediate anti-tumor effects through direct oncolysis, immune activation, and modulation of tumor angiogenesis.⁵⁹ In KRAS-mutant MSS mCRC, the oncolytic reovirus pelareorep combined with FOLFIRI/bevacizumab achieved partial responses in 3 of 6 patients, with median PFS of 65.6 weeks and OS of 25.1 months.⁶⁰ IVX037, a bioselected oncolytic RNA picornavirus targeting CD55 and FcRn, has shown encouraging activity in MSS-CRC and is being evaluated in combination with sintilimab (anti-PD-1) in a Phase Ib trial.⁶¹

Antibody-drug conjugates (ADCs) are emerging as immuno-oncology agents that can induce immunogenic cell death and convert cold tumors into hot ones.⁶² In HER2-positive MSS CRC, trastuzumab deruxtecan (T-DXd) demonstrated an objective response rate of 45.3% in heavily pretreated patients in the DESTINY-CRC01 trial.⁶³ The DESTINY-CRC02 trial subsequently established 5.4 mg/kg as the preferred dose with a favorable benefit-risk profile.⁶⁴

Limitations and Future Challenges

Several important limitations must be honestly acknowledged. First, most combination strategies discussed in this review—including ICI plus anti-angiogenics (beyond the REGONIVO signal), ICI plus MEK inhibition (IMblaze370), and ICI plus chemotherapy—have failed to demonstrate consistent survival benefits in phase III trials.^46,47 Second, the majority of emerging modalities (CAR-T, TILs, bispecific antibodies, cancer vaccines) remain in early-phase development, with limited CRC-specific data and no mature survival outcomes.^50,53,55 Third, predictive biomarkers for patient stratification are still lacking; no validated biomarker panel is available for routine clinical use in MSS CRC.^35,65 Fourth, the inherent heterogeneity of MSS tumors—including rare but clinically relevant POLE/POLD1-mutated or high-TMB subsets—means that a “one-size-fits-all” strategy is unlikely to succeed.⁶⁶ Overcoming these barriers will require better preclinical models (eg., patient-derived organoids, immune-humanized mice), biomarker-driven basket trials, and international collaborative consortia.⁶⁷

Beyond tumor-intrinsic mechanisms, emerging evidence from molecular pathological epidemiology (MPE) indicates that long-term exposures—including genetic variants, dietary factors, lifestyle habits, and environmental agents—can shape tumor immune phenotypes over decades.^68,69 The prospective cohort incident-tumor biobank method (PCIBM), recently applied to large longitudinal studies such as the UK Biobank, enables integration of pre-diagnosis exposure data with detailed tumor molecular and immune profiling.⁷⁰ This approach holds promise for identifying modifiable risk factors that influence immune evasion patterns in CRC and for discovering novel biomarkers that predict immunotherapy response. Future studies applying PCIBM to MSS CRC may uncover opportunities for prevention and personalized treatment that extend beyond traditional therapeutic paradigms.

Through continued translational research and innovative trial designs that embed these principles, extending durable immunotherapy benefits to a broader population of CRC patients represents an active area of investigation with incremental progress.

Conclusion and Future Perspectives

Immunotherapy has substantially changed the treatment landscape for dMMR/MSI-H CRC, establishing a powerful new pillar of cancer care. However, it is critical to acknowledge that, despite extensive ongoing investigation, no immunotherapy approach has yet received regulatory approval for MSS/pMMR metastatic colorectal cancer.^46,47 The mission to extend these benefits to patients with MSS CRC therefore remains one of the most pressing challenges in oncology.

The path forward is not reliant on a single solution but on a mechanism-based, biomarker-driven approach that recognizes the heterogeneity of immune evasion within MSS tumors. We propose a three-pronged framework for future progress:

First, deeper biological insight is needed to deconvolute the heterogeneous immunosuppressive networks within the MSS TME at single-cell resolution.

Second, clinical trial design must evolve from empirical combinations to biology-guided strategies that match specific resistance mechanisms (eg., T-cell exclusion, myeloid dominance) with rationally selected interventions.

Third, the development of integrated biomarkers—combining T-cell-inflamed gene signatures, gut microbiome profiling, dynamic ctDNA monitoring, and emerging tools such as Immunoscore^®—is essential to identify the subset of MSS patients most likely to benefit from specific immunotherapy approaches.^31,67