The Inflammation-Cancer Connection Chronic inflammation is increasingly recognized as a driver of cancer development. This study investigated whether specific inflammatory signaling proteins - cytokines and immune mediators - directly increase or decrease the risk of developing non-small cell lung cancer (NSCLC), the most common form of lung cancer.
Multi-Dimensional Approach Rather than relying on a single dataset or method, the researchers combined four complementary analytical approaches: Mendelian Randomization (a genetic epidemiology technique), transcriptomics (gene expression analysis), network pharmacology (drug-target network analysis), and machine learning classification.
Why Mendelian Randomization? Standard observational studies cannot prove that inflammation causes cancer - the association might be reversed (cancer causing inflammation) or explained by confounding factors. Mendelian Randomization uses naturally occurring genetic variants (SNPs) as instruments to test causal relationships, mimicking the logic of a randomized controlled trial.
Study Scope The analysis examined 91 inflammatory mediators spanning cytokines, chemokines, and growth factors. The goal was to identify which specific inflammatory proteins causally drive NSCLC risk, which protect against it, and whether these proteins could serve as machine learning prediction features.
TGFB1 Increases NSCLC Risk Transforming Growth Factor Beta 1 (TGFB1) was identified as a causal risk factor for NSCLC with an odds ratio of 1.173, meaning each genetically predicted unit increase in TGFB1 is associated with a 17.3% higher odds of developing NSCLC. TGFB1 promotes immune suppression and epithelial-to-mesenchymal transition - both well-known cancer-promoting mechanisms.
CCL11 Also Elevates Risk CCL11 (eotaxin), a chemokine that recruits eosinophils, had an odds ratio of 1.192, making it the stronger risk factor in this analysis. Elevated CCL11 may reshape the tumor microenvironment toward an immune-tolerant state that allows cancer cells to escape surveillance.
CD40 Is Protective CD40, a co-stimulatory immune receptor important for T cell activation, showed a protective effect with odds ratio of 0.857 - each unit increase in CD40 signaling was associated with a 14.3% lower NSCLC odds. This is consistent with CD40's known role in activating anti-tumor immune responses.
CCL4 Also Protective CCL4 (MIP-1 beta), a chemokine that recruits NK cells and cytotoxic T cells, had an odds ratio of 0.896. Higher CCL4 levels may increase immune surveillance and tumor cell killing, providing protection against NSCLC development.
CDK2 and CDK3 Suppress Regulatory T Cells Transcriptomic analysis of NSCLC tumor samples revealed that CDK2 and CDK3 (cyclin-dependent kinases involved in cell cycle regulation) are negatively correlated with the infiltration of regulatory T cells (Tregs) in the tumor microenvironment. Tregs suppress anti-tumor immunity, so less Treg infiltration typically correlates with better immune control of tumors.
CDK5 Promotes Tregs In contrast, CDK5 - a CDK family member with additional roles in neuronal function and inflammation - was positively correlated with Treg infiltration. This distinction within the CDK family highlights that not all CDKs have the same immunological effects.
Network Pharmacology Targets Network pharmacology analysis identified potential drug targets at the intersection of the causal inflammatory pathways and the transcriptomic signature. These include molecules where existing or investigational drugs could theoretically modulate the inflammatory environment to reduce NSCLC risk or improve treatment response.
Immune Landscape Complexity The combined analysis reveals that the NSCLC tumor microenvironment is shaped by a complex balance of pro-inflammatory signals (like TGFB1) that create immune suppression and protective signals (like CD40) that enable immune surveillance - understanding this balance is key to designing effective immunotherapy combinations.
ML Model Performance Machine learning classifiers built using the identified inflammatory biomarkers achieved AUC values of 0.723 to 0.763 across different model types. While not exceptional, these values demonstrate that inflammatory protein signatures alone carry meaningful predictive signal for NSCLC classification.
Feature Importance The Mendelian Randomization results helped prioritize which inflammatory features to include as model inputs. Features with established causal relationships (like TGFB1 and CCL11) were more informative than non-causal associations, demonstrating the value of biological prior knowledge in feature selection.
Integration Value The multi-dimensional framework - combining causal genetics, transcriptomics, and machine learning - produced a more biologically grounded predictive model than any single approach alone. Causal features are less likely to be confounders, improving model generalizability.
Potential Clinical Application If validated in larger cohorts, a blood-based inflammatory protein panel (measuring TGFB1, CCL11, CD40, CCL4, and related markers) could potentially serve as a risk stratification tool - identifying which high-risk individuals (smokers, occupational exposures) have the highest near-term cancer risk and should receive priority LDCT screening.
GWAS Data Constraints Mendelian Randomization relies on genome-wide association study (GWAS) summary statistics, which may have population-specific biases. Most GWAS data come from European ancestry populations, limiting generalizability to other ethnic groups where NSCLC incidence and etiology may differ.
Moderate ML Performance AUC values of 0.72 to 0.76, while statistically meaningful, are not strong enough for clinical decision-making without further improvements. Combining inflammatory markers with imaging biomarkers, clinical risk factors, and genomic data could substantially boost performance.
Mechanistic Gaps The study identifies causal associations but does not fully resolve the mechanisms by which TGFB1 and CCL11 promote NSCLC. Cell-based and animal model experiments are needed to confirm these causal pathways mechanistically.
Future Directions Researchers should validate the identified inflammatory risk signatures in prospective NSCLC cohorts, investigate whether CD40 agonists or TGFB1 inhibitors can reduce NSCLC incidence in high-risk populations, and explore how CDK inhibitors (already used as cancer drugs) interact with the immune microenvironment in lung cancer.