EGFR mutation correlates with uninflamed phenotype and weak immunogenicity, causing impaired response to PD-1 blockade in non-small cell lung cancer

Abstract

Patients with EGFR mutations showed unfavorable response to programmed cell death-1 (PD-1) blockade immunotherapy in non-small cell lung cancer (NSCLC). Yet the underlying association between EGFR mutation and immune resistance remains largely unclear. We performed an integrated analysis of PD-ligand 1(PD-L1)/CD8 expression and mutation profile based on the repository database and resected early-stage NSCLC in Guangdong Lung Cancer Institute (GLCI). Meanwhile, 2 pool-analyses were set to clarify the correlation between EGFR mutation and PD-L1 expression, and the association of EGFR status with response to anti-PD-1/L1 therapy. Pool-analysis of 15 public studies suggested that patients with EGFR mutations had decreased PD-L1 expression (odds ratio: 1.79, 95% CI: 1.10-2.93; P = 0.02). Analysis of The Cancer Genome Atlas (TCGA) and the GCLI cohort confirmed the inverse correlation between EGFR mutation and PD-L1 expression. Furthermore, patients with EGFR mutation showed a lack of T-cell infiltration and shrinking proportion of PD-L1⁺/CD8⁺ TIL (P = 0.034). Importantly, patients with EGFR mutations, especially the sensitive subtype, showed a significantly decreased mutation burden, based on analysis of the discovery and validation sets. Finally, a pool-analysis of 4 randomized control trials confirmed that patients with EGFR mutation did not benefit from PD-1/L1 inhibitors (Hazard ratio [HR] = 1.09, P = 0.51) while patients with EGFR wild-type did (HR = 0.73, P < 0.00001). This study provided evidence of a correlation between EGFR mutations and an uninflamed tumor microenvironment with immunological tolerance and weak immunogenicity, which caused an inferior response to PD-1 blockade in NSCLCs.

Keywords: EGFR; mutation burden; non-small cell lung cancer; programmed cell death protein-1; uninflamed.

Figure 1.

Correlation between EGFR mutation and PD-L1 expression in patients with NSCLC (A) Forest plot of 15 studies evaluating the association between EGFR status and PD-L1 expression. Pooled odds ratio (OR) were computed using random-effects model. (B) Quantitative analysis of PD-L1 protein derived from reverse phase protein array based on EGFR status. (C) Quantitative analysis of PD-L1 mRNA expression from RNA-Seq profile of patients in Guangdong Lung Cancer Institute (GLCI) based on EGFR status. (D) Immunohistochemistry (IHC) analysis of PD-L1 protein expression based on EGFR status in a cohort of 255 resected NSCLC. mut, mutation; wt, wild-type

Figure 2.

EGFR mutant tumors show a lack of T-cell infiltration and low shrink proportion of PD-L1⁺/CD8⁺ TILs (A) Quantitative analysis of CD8A mRNA expression from RNA-Seq profiles of GLCI, based on EGFR status. (B) IHC analysis of CD8⁺ T lymphocytes infiltration, based on EGFR status in resected NSCLCs. (C and D) The correlation between EGFR status and tumor microenvironment immune phenotype classified based on PD-L1 and CD8A/CD8⁺ TIL expression, from the GLCI RNA-Seq data and IHC analysis. (E) Representative images of PD-L1 and CD8 immunostaining in resected NSCLC samples, based on EGFR status. dual-positive for PD-L1 and CD8A (PD-L1⁺/CD8A⁺) were defined as above-median mRNA expression for each gene. PD-L1⁺/CD8⁺ TIL was defined as PD-L1: TC2–3/IC2–3 and CD8 ≥ 25%. mut, mutation; wt, wild-type.

Figure 3.

EGFR mutation is correlated with decreased immunogenicity (A and B) Quantitative analysis of tumor mutational burden in EGFR mutation and wild-type tumors in the discovery set (TCGA and Broad cohorts). (C) Difference in tumor mutational burden based on EGFR status in the validation set (GLCI cohort). (D and E) Subgroup analysis of tumor mutational burdens between tumors with sensitive and resistant/unknown EGFR mutations in TCGA and Broad cohorts. mut, mutation; wt, wild-type; S-mut, sensitive mutation; R/UNK-mut, resistance/unknown sensitive mutation.

Figure 4.

Forest plots of hazard ratios (HRs) for (A) overall survival (OS) and (B) progression-free survival (PFS) from 4 randomized control trials, comparing second-line PD-1/L1 inhibitor with docetaxel in patients with EGFR mutations and wild-type non-small cell lung cancer. Pooled HRs were computed using the fixed-effects model.

Figure 5.

Anti-tumor activity and biomarker analysis of EGFR and PD-L1 in 3 patients with advanced non-small cell lung cancer treated with PD-1 inhibitors (A, B, and C).