MGA Mutation as a Novel Biomarker for Immune Checkpoint Therapies in Non-Squamous Non-Small Cell Lung Cancer

Abstract

Background: Immune checkpoint inhibitors have changed the treatment landscape for advanced non-small cell lung cancer. However, only a small proportion of patients experience clinical benefit from ICIs. Thus, the discovery of predictive biomarkers is urgently warranted. Evidence have shown that genetic aberrations in cancer cells can modulate the tumor immune milieu. We therefore explored the association between oncogenic mutations and efficacy to ICIs in non-squamous NSCLC. Methods: We curated genomic and clinical data of 314 non-squamous NSCLC patients receiving ICIs from four independent studies for the discovery cohort. For external validation, 305 patients from an ICI-treated cohort and 1,027 patients from two non-ICI-treated cohorts were used. Relations between oncogenic mutations and outcomes of immunotherapy were examined. Multivariate Cox regression models were applied to adjust confounding factors. Further investigation on tumor antigenicity and antitumor immunity was performed in The Cancer Genome Atlas lung adenocarcinoma cohort. Results: A total of 82 oncogenes/tumor suppressor genes according to the Oncology Knowledge base database with a frequency greater than 3% were identified and investigated in the discovery cohort. Within these genes, MGA mutations were enriched in patients with durable clinical benefit (p = 0.001, false discovery rate q < 0.05). The objective response rate was also significantly higher in patients with MGA mutation (2.63-fold, p < 0.001, FDR q < 0.05). Longer progression-free survival was found in MGA-mutated patients (HR, 0.41; 95% CI, 0.23-0.73; p = 0.003), and the association remained significant after controlling for tumor mutational burden (TMB), programmed cell death ligand-1 expression, and treatment regimens. In the validation cohort, significant improvement in overall survival was found in patients harboring MGA mutation (HR, 0.39; 95% CI, 0.17-0.88; p = 0.02). Furthermore, the survival difference was not detected in non-ICI-treated cohorts. We also demonstrated that MGA mutation correlate with higher TMB, elevated neoantigen load and DNA damage repair deficiency. Gene set enrichment analysis revealed that gene sets regarding activated immune responses were enriched in MGA-mutated tumors. Conclusion: Our work provides evidence that MGA mutation can be used as a novel predictive biomarker for ICI response in non-squamous NSCLC and merits further clinical and preclinical validation.

Keywords: MGA; biomarker; immune checkpoint inhibitors; mutation; non-squamous non-small cell lung cancer.

FIGURE 1

Summary of mutational and clinical information of non-squamous NSCLC patients in the discovery cohort. Individual patients are represented in each column, organized by response category, with progression-free survival time in decreasing order. Categories of smoking status (never or ever) and treatment regimens (combination or monotherapy) are characterized. PD-L1 expression is stratified as 0% or greater than 1%. The occurrences of selected genes in each case are represented in the OncoPrint, with the percent frequency shown. *Mutational information unknown (not covered in the panel tested) are depicted in light gray on the OncoPrint. NSCLC, non small cell lung cancer; DCB, durable clinical benefit; NDB, no durable benefit; PD-(L)1, programmed cell death -(ligand)1; TMB, tumor mutation burden.

FIGURE 2

MGA mutation correlate with DCB, higher ORR and longer PFS in the discovery cohort of non-squamous NSCLC patients treated with ICIs (A) Enrichment of gene mutations in patients with DCB vs. NCB in the discovery cohort (two-tailed Fisher’s exact test, n = 113 patients with DCB, n = 181 patients with NCB). Red dashed line denotes BH FDR q = 0.05 (B) ORR were compared between subgroups according to mutational status of FAT1, TET1 and MGA. *p < 0.05, Fisher’s exact test, BH FDR q < 0.05 (C) Ratio of patients with CR/PR to patients with PD classified by FAT1, TET1 and MGA mutations. *p < 0.05, Fisher’s exact test, BH FDR q < 0.05 (D–E) Kaplan-Meier curves comparing PFS of patients with or without TET1 (D) and MGA (E) mutations in the discovery cohort. A two-sided log-rank test p < 0.05 is considered as a statistically significant difference. FDR, false discovery rate; BH, Benjamini-Hochberg method; DCB, durable clinical benefit; ORR, objective response rate; PFS, progression-free survival.

FIGURE 3

Validation of the predictive function of MGA mutation (A) Summary of mutational and clinical information of non-squamous NSCLC patients in the validation cohort. Individual patients are represented in each column, organized by response category, with progression-free survival time in decreasing order. Categories of smoking status (never or ever) and treatment regimens (combination or monotherapy) are characterized. The occurrences of selected genes in each case are represented in the OncoPrint, with the percent frequency shown. *Mutational information unknown (not covered in panel tested) are depicted in light gray on the OncoPrint (B–E) Kaplan-Meier curves comparing OS of patients with or without MGA mutations in the validation cohort (B), Non-ICI-treated cohort (C), TCGA-LUAD cohort (D) and TCGA-LUAD cohort with stage IV patients (E). Log-rank test was used in (B–E).

FIGURE 4

Association of MGA Mutation with tumor mutational burden, neoantigen load and DNA damage repair (DDR) deficiency in patients with non-squamous NSCLC (A) Comparison of tumor mutational burden between MGA-mutated and wild-type subgroups from ICI-treated NSCLC and the TCGA-LUAD cohorts (B) Comparison of neoantigen load between MGA-mutated and wild-type subgroups in the TCGA-LUAD cohort (C) Comparison of mutation amounts of DDR pathway genes between MGA-mutated and wild-type subgroups in the TCGA-LUAD cohort. Mann-Whitney U test was used to test the differences. ns: not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. BER, base excision repair; DR, direct damage reversal repair; FA, Fanconi anemia; HDR, homology-dependent recombination; MMR, mismatch repair; NER, nucleotide excision repair; NHEJ, non-homologous end joining; NP, nucleotide pool maintenance; TLS, translesion DNA synthesis.

FIGURE 5

Association of MGA Mutation with relative abundance of infiltrated immune cell by CIBERSORTx in the Chen cohort. Gene expression data were uploaded to the CIBERSORTx web portal (https://cibersortx.stanford.edu/), with batch correction performed and permutation number setting to1000 for significance analysis. *p < 0.05 (Mann-Whitney U test).

FIGURE 6

Gene set enrichment analysis (GSEA) was performed using the Hallmark gene sets (A) Inflammatory response pathway and TNFα-NFκB pathway were enriched in MGA mutated patients from the TCGA-LUAD cohort (B) IFN-α and IFN-γ pathways were enriched in advanced MGA mutated patients from stage IV TCGA-LUAD cohort (C) Inflammatory response pathway and TNFα-NFκB pathway were enriched in MGA mutated patients from stage IV TCGA-LUAD cohort (D) JAK-STAT pathways were enriched in MGA mutated patients with stage IV from the TCGA-LUAD cohort. TNFα, tumor necrosis factor-alpha; NFκB, nuclear factor kappa-B; IFN-α, interferon-alpha; IFN-γ, interferon-gamma; JAK-STAT, janus kinase-signal transducer and activator of transcription.