Immune evolution from preneoplasia to invasive lung adenocarcinomas and underlying molecular features

Abstract

The mechanism by which anti-cancer immunity shapes early carcinogenesis of lung adenocarcinoma (ADC) is unknown. In this study, we characterize the immune contexture of invasive lung ADC and its precursors by transcriptomic immune profiling, T cell receptor (TCR) sequencing and multiplex immunofluorescence (mIF). Our results demonstrate that anti-tumor immunity evolved as a continuum from lung preneoplasia, to preinvasive ADC, minimally-invasive ADC and frankly invasive lung ADC with a gradually less effective and more intensively regulated immune response including down-regulation of immune-activation pathways, up-regulation of immunosuppressive pathways, lower infiltration of cytotoxic T cells (CTLs) and anti-tumor helper T cells (Th), higher infiltration of regulatory T cells (Tregs), decreased T cell clonality, and lower frequencies of top T cell clones in later-stages. Driver mutations, chromosomal copy number aberrations (CNAs) and aberrant DNA methylation may collectively impinge host immune responses and facilitate immune evasion, promoting the outgrowth of fit subclones in preneoplasia into dominant clones in invasive ADC.

Fig. 1. The immune landscape IPNs of different stages and associated genomic and epigenomic features.

Infiltration CD4+ T cells, CD8+ T cells inferred from immune gene expression using TIMER; regulatory T cells (Tregs, CD3+CD8–FOXP3+) and CD8+cytotoxic T lymphocytes (CTLs,CD3+CD8+granzyme B+) measured by multiplex immunofluorescence (mIF); T cell clonality and frequency of the top 100T cell clones by TCR sequencing are shown in upper panel. Genomic alterations from whole exome sequencing (WES) including EGFR/KRAS mutations, HLA LOH, copy number variation (CNV) burden, allelic imbalance (AI) burden, total number of mutations associated with predicted neoantigens, and total number of mutations associated with predicted neoantigens without promoter methylation; global methylation status using long interspersed transposable elements-1 (LINE-1) as a surrogate marker accessed by reduced representation bisulfite sequencing (RRBS) are shown in bottom panel. Data from 53 patients was used including AAH (typical adenomatous hyperplasia), AIS (adenocarcinoma in situ), MIA (minimally invasive adenocarcinoma), and ADC (invasive adenocarcinoma). (Source data is provided as a source data file).

Fig. 2. Immune landscape from preneoplasia to invasive lung adenocarcinoma.

a Significantly enriched functional pathways based on the 291 differentially expressed genes by Ingenuity Pathway Analysis (IPA®; Ingenuity Systems) software. Pathways with –log (p-value) > 10 (p-values are obtained from Fisher’s right-tailed exact test) and an absolute z-score > 0.5 are shown. Pathways that were predicted to be inhibited (negative Z scores) in later stages are in blue and pathways that were predicted to be activated (positive Z scores) in later stages are in orange. The heights of the bars indicate the significance of the enrichment (−log (p-value)) and the scales of the orange or blue colors represent the predicted directionality. Fractions of immune cells including CD4+ T cells (b), CD8+ T cells (c), and CD4/CD8 ratio (d) were estimated using TIMER based on the gene expression using nCounter PanCancer Immune Profiling Panel. Error bars indicate 95% confidence intervals and solid point represent mean value in each stage. The difference of cell fraction among different stages was evaluated using two-sided Kruskal–Wallis H test. Data from 38 patients was used including NL (normal lung tissue), AAH (atypical adenomatous hyperplasia), AIS (adenocarcinoma in situ), MIA (minimally invasive adenocarcinoma), and ADC (invasive adenocarcinoma). (Source data is provided as a source data file).

Fig. 3. T cell repertoire from preneoplasia to invasive lung adenocarcinoma.

a Distribution of T cell clones with frequency of top 1 (brown), top 2–10 (black), top 11–100 (orange), top 101–200 (purple), top 201–500 (green), top 501–1000 (red), and beyond 1000 (blue) in normal lung, AAH, AIS, MIA, and invasive ADC lesions. b T cell density, c T cell diversity, and d T cell clonality in normal lung, AAH, AIS, MIA, and invasive ADC. Error bars indicate 95% confidence intervals and solid point represent mean value in each stage. The difference of T cell matrix among different stages was evaluated using two-sided Kruskal–Wallis H test. Data from 51 patients was used including NL (normal lung tissue), AAH (atypical adenomatous hyperplasia), AIS (adenocarcinoma in situ), MIA (minimally invasive adenocarcinoma), and ADC (invasive adenocarcinoma). (Source data is provided as a source data file).

Fig. 4. The correlation between T cell clonality and T cell subtypes.

a Th1 (average value of marker genes IFNG, IL12A, IL12B) measured by gene expression from nCounter PanCancer Immune Profiling Panel. Fractions of CD8+ T cells (b) and CD4+ T cells (e) inferred from gene expression profiling using TIMER. c CD8+CTLs (CD3+CD8+granzyme B+), d ThCTLs (CD3+CD8–Granzyme B+), f CD3+CD8– T cells, and g Tregs (CD3+CD8–FoxP3+) measured by mIF. The correlation coefficient (rho) was assessed by two-tailed Spearman’s rank correlation test. Data from 35 patients was used including NL (normal lung tissue), AAH (atypical adenomatous hyperplasia), AIS (adenocarcinoma in situ), MIA (minimally invasive adenocarcinoma), and ADC (invasive adenocarcinoma). (Source data is provided as a source data file).

Fig. 5. Loss of heterozygosity in HLA (HLA LOH) in IPNs of different histologic stages and its association with chromosomal alterations.

a The proportion of AAH, AIS, MIA, and ADC lesions with evidence of HLA-LOH. Two-sided χ² test was used to assess the difference among different histologic stages. Comparison of AI burden (number of AI events) (b) and CNV burden (normalized as the percent of genes with CNV) (c) in lesions with (purple) and without (green) HLA-LOH. Two-sided Wilcoxon rank-sum test was used to assess the differences. Error bars indicate 95% confidence intervals and solid point represent mean value in each stage. Data from 35 patients was used including NL (normal lung tissue), AAH (atypical adenomatous hyperplasia), AIS (adenocarcinoma in situ), MIA (minimally invasive adenocarcinoma), ADC (invasive adenocarcinoma). (Source data is provided as a source data file).

Fig. 6. The potential impact of global methylation and immune infiltration.

The correlation between global methylation levels (using LINE-1 as a surrogate marker) with CD4+ T cells (a), CD4/CD8 ratio (b) inferred from immune gene expression by TIMER as well as Treg (CD3+CD8–FOXP3+) (c) and Treg(CD3+CD8−FOXP3+)/CD8(CD3+CD8+) ratio (d) measured by mIF. The correlation coefficient (rho) was assessed by two-tailed Spearman’s rank correlation test. Data from 24 patients was used including NL (normal lung tissue), AAH (atypical adenomatous hyperplasia), AIS (adenocarcinoma in situ), MIA (minimally invasive adenocarcinoma), and ADC (invasive adenocarcinoma). (Source data is provided as a source data file).