Comprehensive genomic and immunological characterization of Chinese non-small cell lung cancer patients

Abstract

Deep understanding of the genomic and immunological differences between Chinese and Western lung cancer patients is of great importance for target therapy selection and development for Chinese patients. Here we report an extensive molecular and immune profiling study of 245 Chinese patients with non-small cell lung cancer. Tumor-infiltrating lymphocyte estimated using immune cell signatures is found to be significantly higher in adenocarcinoma (ADC, 72.5%) compared with squamous cell carcinoma (SQCC, 54.4%). The correlation of genomic alterations with immune signatures reveals that low immune infiltration was associated with EGFR mutations in ADC samples, PI3K and/or WNT pathway activation in SQCC. While KRAS mutations are found to be significantly associated with T cell infiltration in ADC samples. The SQCC patients with high antigen presentation machinery and cytotoxic T cell signature scores are found to have a prolonged overall survival time.

Fig. 1

Copy number variations in the CHOICE versus TCGA NSCLC data. a SQCC focal amplifications. b SQCC focal deletions. c ADC focal amplifications. d ADC focal deletions. The GISTIC 2.0 algorithm was used to identify significantly recurrent copy number gains and losses in ADC and SQCC population, respectively. The q values for amplification (a, c) and deletions (b, d) in the CHOICE study were plotted against the q values from the TCGA datasets. CNVs with q values <0.25 were significant. Representative genes in focal regions were used whenever possible. Red dots: significant focal events only in the CHOICE cohort (ADC); blue dots: significant focal events only in TCGA dataset (SQCC). Red dashed line: q value cutoffs. Source data are provided as a Source Data file

Fig. 2

Significantly mutated genes for ADC and SQCC in the CHOICE study. a Top 30 significantly mutated genes in ADC, samples were ordered based on their somatic non-synonymous mutation burden (top panel) and genes were ranked by mutation frequencies (left panel). Cancer stage, smoker status, and immune infiltration status are annotated in the bottom panel. b Top 30 significantly mutated genes in SQCC. Comparison of the mutation frequencies of significantly mutated gene between the CHOICE and TCGA study in ADC (c) and SQCC (d). Source data are provided as a Source Data file

Fig. 3

Tumor-infiltrating lymphocytes enrichment profiles. a Comparison of TILs enrichment profile in ADC (n = 131) versus SQCC (n = 114). b Comparison of TILs enrichment profile in smokers (n = 84) versus non-smokers (n = 47) in ADC. Percentage of patients with enriched immune cell signatures were calculated using ssGSEA (see Supplementary Methods). For each immune cell signature, enrichment is defined as q-value ≤0.1. Source data are provided as a Source Data file. a black bars indicate the percentage of patients having significant enrichment for the given immune cell type in ADC subtype, while gray bars represent the percentage in SQCC; b black bars represent the percentage of smokers, while gray bars represent the percentage of non-smokers. Immune cell signatures were classified to adaptive, innate and other

Fig. 4

Characterization of immune infiltration in the CHOICE population. a Heat map of normalized ssGSEA score for ADC, mutation status of TP53, KRAS, and EGFR are shown while CNV level of CASC5, FANCG, CDKN2A, CCND1, and SOX2 are shown. b Heat map of normalized ssGSEA for SQCC. Mutation status of TP53, NFE2L2, KEAP1, and EGFR are shown, while CNV profile of CDKN2A, CASC5, FANCG, CCND1, and SOX2 are shown. Tumor samples were classified into 3 immune status: HIGH, MIX, and LOW based on the signature score of 26 immune cell types. Samples were also labeled using 5 types of omics data. (1) Mutation burden for each sample (green). (2) Immune status (red, yellow, and blue for HIGH, MIX, and LOW). (3) Selected significantly mutated genes in each subtype (black for mutation and white for wild-type). (4) mRNA expression value for four immune marker genes (IFNG, PDL1, PD1, and CD8) (dark green indicates high expression and light green indicate low expression). (5) GISTIC 2 based CNV for the selective gene. Source data are provided as a Source Data file. Dark red color represents homozygous amplification, light red for heterozygous amplification, white for diploid, light blue for heterozygous deletion, and dark blue for homozygous deletion

Fig. 5

PI3K pathway activation versus T cell infiltration. a PTEN loss and PIK3CA amplification were enriched in immune LOW and MIX group SQCC (immune group red, yellow, and blue for HIGH, MIX, and LOW). b Comparison of PTEN and PIK3CA alteration with T cell signature. c Comparison of PTEN and PIK3CA alteration with CD8 IHC H-score. d PIK3CA expression versus T cell signature, amplification/high expression (red) was associated with low T cell signature. The top and bottom of the boxes are the lower and upper quartiles. The middle line in the box is median, and the whiskers are lowest and highest point within 1.5 times the interquartile range of the lower and upper quartile. e PTEN expression versus T cell signature, loss/low expression (blue) was associated with low T cell expression. Source data are provided as a Source Data file

Fig. 6

Immunohistochemistry staining of PD-L1 and CD8. a Membrane staining (3+) was seen in positive control cells (HDLM-2). b No staining was seen in negative control cells (PC3). c–f Representative images of membrane staining on tumor cells were shown with different signal intensities: c No staining. d Low/1+staining, e Medium/2+staining. f High/3+staining. Example of heterogeneous IHC staining: g Strong staining of PD-L1 from SQCC (h) Strong staining in ADC. i–l Heterogeneous expression of PD-L1 from the same tissue. i, k are the same tissue; two regions are selected (boxed). j Zoomed-in region of i. l Zoomed-in region of k. Correlation between PD-L1 and CD8. m–o Representative images of PD-L1/CD8 both high, both low and PD-L1 L/CD8 H. p Correlation between CD8 positivity and PD-L1 expression was observed in ADC (ANOVA p-value = 8.76e-13, F = 50.25, DF = 2). Source data are provided as a Source Data file

Fig. 7

Immune signatures associated with SQCC patient overall. a Immune signature, immune MIX and LOW groups were combined together and compared with the immune HIGH group. b APM signature. c Cytotoxic cell signature. d pDC signature. Source data are provided as a Source Data file