Senescent Cell Secrets Reveal Why Lung Cancer Resists Immunotherapy

Lung adenocarcinoma, the most common subtype of non-small cell lung cancer, accounts for roughly 40 percent of all lung cancer diagnoses and claims more than 1.8 million lives each year worldwide. Although immune checkpoint inhibitors have revolutionized oncology by unleashing the body’s immune defenses against tumors, only 15 percent to 40 percent of patients achieve durable responses, and both intrinsic and acquired resistance remain major clinical roadblocks.

To unravel why certain tumors evade treatment so effectively, a research team led by scientists at Shandong Cancer Hospital investigated how senescent, or aging, cancer cells shape their surrounding environment. As cells undergo stress or DNA damage, they enter an irreversible arrest and start secreting a potent cocktail of inflammatory factors known as the senescence-associated secretory phenotype, or SASP. This secretory profile acts as a double-edged sword, capable of either summoning immune cells to clear out damaged tissue or driving chronic inflammation and immune evasion. The researchers published their findings in PeerJ.

How Cellular Senescence Divides Tumors Into Two Worlds

Using bioinformatic tools and machine learning algorithms on public transcriptomic datasets including The Cancer Genome Atlas, the authors mapped the expression patterns of 191 unique SASP-related genes. They applied non-negative matrix factorization, a mathematical technique that groups complex biological data into distinct patterns, to classify lung adenocarcinoma patients into two distinct molecular subtypes. Patients falling into Cluster 1 experienced significantly poorer overall survival and presented with a higher proportion of advanced-stage disease compared to those in Cluster 2.

Gene set enrichment analyses revealed that Cluster 1 tumors exhibited intense activation of cell cycle and cellular senescence pathways alongside a muted antitumor immune response. Computational estimations using immune deconvolution algorithms showed that Cluster 1 tumors harbored fewer infiltrating CD8+ positive T cells, natural killer cells, and dendritic cells, creating an immunosuppressive microenvironment that allows cancer cells to thrive unchecked.

Five Core Genes That Predict Survival Outcomes

To narrow down the most influential drivers within the SASP network, the team integrated three advanced machine learning methods, specifically least absolute shrinkage and selection operator regression, Random Forest, and extreme gradient boosting. This intersectional filtering isolated five core prognostic genes, namely ANGPTL4, CDKN2A, H2AX, MMP1, and CCL20. Transcriptomic and proteomic data confirmed that these five genes were significantly upregulated in tumor tissues compared to adjacent healthy lung tissue.

Building upon these core genes, the investigators calculated a composite metric known as the SASP Protein Regulator Score to stratify patients into high-risk and low-risk groups. Patients with high scores consistently demonstrated shorter overall survival times across multiple validation cohorts. Furthermore, time-dependent receiver operating characteristic curve analyses verified that the score possessed reliable prognostic performance for predicting one-year, three-year, and five-year survival rates.

Implications for Immunotherapy and Chemotherapy Resistance

The study also explored how the SASP score correlates with predicted responses to modern therapies. High-scoring patients tended to display features of immune exclusion, where suppressive cell types form a physical or chemical barrier blocking active T cells from infiltrating the tumor core. Consequently, high scores correlated with non-responder status and poorer survival outcomes in independent immunotherapy cohorts, including patients treated with anti-programmed death-ligand 1 antibodies.

Interestingly, drug-sensitivity prediction models suggested that high-scoring tumors might exhibit greater sensitivity to standard chemotherapeutic agents such as cisplatin, docetaxel, and paclitaxel, whereas low-scoring patients appeared more responsive to targeted therapies like erlotinib. The authors combined the SASP score with clinical parameters like tumor and node stages to construct a predictive nomogram, which demonstrated good calibration and net clinical benefit in estimating individual survival probabilities.

Study Limitations and Next Steps

Despite these promising correlations, the authors noted several important limitations. Because the study relied largely on retrospective data from public repositories, potential selection biases remain, and crucial clinical variables such as patient smoking history were occasionally missing. Furthermore, quantitative reverse transcription polymerase chain reaction and immunohistochemical stainings of cell lines and clinical specimens revealed heterogeneous expression patterns that did not always align perfectly with bulk transcriptomic trends.

The authors emphasize that predicted drug responses and immune cell infiltrations were computationally derived rather than directly measured through functional assays or multiplex tissue imaging. Therefore, prospective clinical trials and mechanistic laboratory studies are urgently needed to validate the clinical utility of the SASP score before it can be applied to guide personalized treatment decisions in patients with lung adenocarcinoma.

The research was published in PeerJ on January 1, 2026.