The Promise and the Problem Immune checkpoint inhibitors (ICIs) targeting PD-1/PD-L1 and CTLA-4 have transformed NSCLC treatment, but a significant proportion of patients show primary resistance (no initial response) or acquired resistance (response followed by relapse).
Scope of This Review This comprehensive review analyzes the biology of immune checkpoints, current approved inhibitors, mechanisms of resistance, and novel strategies to overcome those resistance mechanisms, including combination therapies, personalized vaccines, and microbiome-based approaches.
The Role of the Tumor Microenvironment The TME is central to both sensitivity and resistance. Immunosuppressive cells, cytokines, and metabolic factors within the TME create an environment where T cells are exhausted, excluded, or suppressed even in the presence of checkpoint blockade.
Emerging Tools AI and machine learning, novel biomarkers such as TCR clonality and T-cell inflamed gene signatures, and microbiome-modulating therapies represent the new frontier in predicting and enhancing ICI responses in NSCLC.
PD-1/PD-L1 Pathway PD-1 is expressed on activated T cells and inhibits their function when it binds to PD-L1 on tumor cells or immune cells in the TME. Tumor cells upregulate PD-L1 in response to inflammatory cytokines (IFN-gamma, TNF-alpha), effectively silencing effector T cells and allowing unchecked proliferation.
CTLA-4 Pathway CTLA-4 competes with the co-stimulatory receptor CD28 for binding to B7-1 and B7-2 on antigen-presenting cells, dampening T cell activation at an early stage. This checkpoint is exploited by regulatory T cells that naturally express CTLA-4 at high levels.
T Cell Exhaustion Chronic antigen exposure in tumors leads to progressive T cell exhaustion characterized by elevated inhibitory receptors (PD-1, CTLA-4, TIM3), reduced cytokine production, and impaired proliferation. While ICIs can partially reverse exhaustion, deeply exhausted T cells often fail to recover full function.
Metabolic Suppression in the TME Tumors accumulate adenosine and other metabolites that suppress T cell activation. Altered metabolic pathways compete with T cells for nutrients such as glucose and arginine, further impairing immune function within the tumor bed even when checkpoint blockade is active.
Primary Resistance Some patients never respond to ICIs because of pre-existing features such as low tumor mutational burden, absence of PD-L1 expression, lack of TILs, defects in antigen presentation machinery (beta-2-microglobulin loss, MHC-I downregulation), or constitutively active oncogenic pathways that suppress immune activation.
Acquired Resistance After an initial response, tumors can develop acquired resistance through antigen loss (mutations in target neoantigens), upregulation of alternative immune checkpoints (TIM-3, LAG-3, TIGIT), recruitment of additional immunosuppressive cells, and epigenetic silencing of immune recognition genes.
STK11/KEAP1 Mutations In NSCLC, STK11 and KEAP1 mutations are major drivers of primary ICI resistance by creating an immunosuppressive, T-cell-cold TME that is poorly responsive to PD-1/PD-L1 blockade regardless of PD-L1 expression levels.
Tumor Heterogeneity Intratumoral heterogeneity means that different regions of the same tumor may have different immune landscapes and resistance mechanisms. A responding clone may be eliminated while a resistant subclone expands, explaining the pattern of mixed or transient responses.
Dual Checkpoint Blockade Combining anti-PD-1 and anti-CTLA-4 (e.g., nivolumab plus ipilimumab) has shown improved outcomes in NSCLC, particularly in tumors with high TMB. This combination addresses both peripheral T cell priming (CTLA-4) and effector T cell suppression within the tumor (PD-1).
Targeting Novel Checkpoints TIM-3, LAG-3, and TIGIT are alternative immune checkpoints that are upregulated in ICI-resistant tumors. Clinical trials combining anti-PD-1 with anti-LAG-3 or anti-TIGIT are demonstrating promising results in expanding the population of patients who benefit from immunotherapy.
Chemotherapy Synergy Certain chemotherapy agents (paclitaxel, oxaliplatin, doxorubicin) induce immunogenic cell death, releasing damage-associated molecular patterns that stimulate dendritic cells and prime antitumor T cells. Combining these with ICIs creates a priming-and-releasing double action.
Epigenetic Modulators DNA methyltransferase inhibitors and HDAC inhibitors can reverse epigenetic silencing of immune recognition genes, upregulate MHC-I expression, and reduce immunosuppressive cell infiltration - all helping to convert cold tumors into ICI-responsive ones.
Personalized Neoantigen Vaccines Tumor-specific neoantigens - peptides arising from somatic mutations unique to each patient's cancer - can be identified via next-generation sequencing and used to create personalized vaccines. These vaccines prime the immune system against tumor-specific targets, potentially converting cold tumors into hot ones amenable to ICI therapy.
mRNA Vaccine Technology Advances in mRNA vaccine platforms (accelerated by COVID-19 vaccine development) have made personalized neoantigen vaccines more feasible and scalable. Early clinical trials combining meoantigen mRNA vaccines with PD-1 inhibitors in NSCLC and melanoma are showing encouraging results.
Gut Microbiome and Immunotherapy The gut microbiome profoundly influences systemic immune responses. Specific bacterial species (such as Akkermansia muciniphila and Bifidobacterium) are associated with better ICI responses, while dysbiotic microbiomes correlate with resistance and worse outcomes.
Microbiome Modulation Strategies Fecal microbiota transplantation (FMT) from ICI responders to non-responders has shown early promise in melanoma and is being explored in NSCLC. Probiotic supplementation and dietary interventions represent accessible microbiome-targeting approaches that could complement standard immunotherapy.
Limitations of PD-L1 and TMB PD-L1 is an imperfect biomarker - many PD-L1-positive patients do not respond and some PD-L1-negative patients do. TMB captures mutation quantity but not quality or the specific immune relevance of mutations, limiting its predictive precision.
TCR Clonality T-cell receptor (TCR) clonality - a measure of how focused or diverse the T cell repertoire within a tumor is - is emerging as a predictor of ICI response. Highly clonal TCR repertoires suggest pre-existing antigen-specific T cells that can be reactivated by checkpoint blockade.
T-Cell Inflamed Gene Signatures Gene signatures capturing the presence and activity of cytotoxic T cells within the tumor - sometimes called T-cell inflamed gene expression profiles - have been shown to predict ICI response across multiple cancer types including NSCLC.
AI and Machine Learning Applications Machine learning models can integrate multiple biomarkers (PD-L1, TMB, TCR clonality, gene expression, imaging features) to build composite predictive models that outperform any single biomarker. These models can analyze complex multi-dimensional datasets and identify patterns invisible to individual biomarker analysis.
Addressing the Non-Responder Majority Currently, the majority of NSCLC patients do not benefit significantly from ICIs. Future research must prioritize understanding resistance mechanisms in these patients and developing interventions that can convert non-responders to responders.
Multi-Omics Integration Combining genomics, transcriptomics, proteomics, and microbiomics data with AI-driven analysis will enable more comprehensive tumor immune profiling and identify which combinations of resistance mechanisms are present in individual patients.
Adaptive Trial Designs Novel adaptive clinical trial designs that can test multiple ICI combinations simultaneously, incorporating biomarker-based patient stratification, will accelerate discovery of effective regimens for specific resistance patterns.
Toxicity Management As combination immunotherapy becomes more complex, managing immune-related adverse events (irAEs) will become increasingly important. AI tools that predict which patients are at high risk of severe toxicity will be essential for safe clinical implementation.