Whole-genome characterization of large-cell lung carcinoma: A comparative analysis based on the histological classification

Abstract

Background: According to the 2015 World Health Organization classification, large cell neuroendocrine carcinoma (LCNEC) was isolated from Large-cell lung cancer (LCLC) tumors, which constitutes 2%-3% of non-small cell lung cancer (NSCLC). However, LCLC tumors are still fairly vaguely defined at the molecular level compared to other subgroups. Materials and Methods: In this study, whole-genome sequencing (WGS) was performed on 23 LCLC and 15 LCNEC tumor specimens. Meanwhile, data from the TCGA (586 LUADs and 511 LUSCs) and U Cologne (120 SCLCs) were analyzed and compared. Results: The most common driver mutations were found in TP53 (13/23, 57%), FAM135B (8/23, 35%) and FAT3 (7/23, 30%) in LCLC, while their counterparts in LCNEC were TP53 (13/15, 87%), LRP1B (6/15, 40%) and FAT1 (6/15, 40%). Notably, FAM135B mutations only occurred in LCLC (P = 0.013). Cosmic signature analysis revealed widespread defective DNA mismatch repair and tobacco-induced mutations in both LCLC and LCNEC. Additionally, LCNEC had a higher incidence of chromosomal copy number variations (CNVs) and structural variations (SVs) compared with LCLC, although the differences were not statistically significant. Particularly, chromothripsis SVs was significantly associated with CNVs. Furthermore, mutational landscape of different subtypes indicated differences between subtypes, and there seems to be more commonalty between our cohort and SCLC than with other subtypes. SMARCA4 mutations may be specific driver gene alteration in our cohort. Conclusion: Our results support that LCLC and LCNEC tumors follow distinct tumorigenic pathways. To our knowledge, this is the first genome-wide profiling comparison of LCLC and LCNEC.

Keywords: Large-cell lung cancer (LCLC); histological classification; large cell neuroendocrine carcinoma (LCNEC); non-small cell lung cancer (NSCLC); whole-genome sequencing.

FIGURE 1

Comparison of mutation landscape between LCLC and LCNEC. (A) A comparison of the mutational landscapes of LCLC and LCNEC is provided, along with the most frequently mutated driver genes. The top panel represents the TMB and the middle panel represents the matrix of frequently mutated genes. Columns represent samples, and clinicopathological characteristics of individual patients are presented below. Bar plots in the lower panel shows the contribution of six substitutions. (B) Forestplot shows the significant differences of driver genes between the two groups.

FIGURE 2

Mutational spectrum analysis for LCLC and LCNEC. (A) The pie chart shows the relative contribution of the six substitutions. (B) Relative contribution of 96 substitution subtypes SNV in each group. (C) Relative contributions of mutational signatures in each group.

FIGURE 3

Distinct CNA landscape of LCLC and LCNEC. (A) Overall copy number variation (CNV) profile of LCLC and LCNEC. Red represented amplification and blue represented deletion. (B) Comparison of the CNA burden between LCLC and LCNEC. (C) Somatic copy number alterations in each group. Deletions and amplifications are represented on the y-axis by blue or red bars, respectively. Each peak region (cytoband) is displayed together with its known or potential cancer-related genes.

FIGURE 4

Distribution of SVs in LCLC and LCNEC (A) Stacked bars show the genome-wide burden of each SV class (color) in each patient (x-axis). Lower panel: SVs resulting in copy-number gain or loss (B) Comparison of each SV burden between LCLC and LCNEC. (C) Pairwise associations between the numbers of SVs across patients. Color was determined by the magnitude of positive (blue) and negative (red) Spearman correlation coefficients, plotted only where q < 0.1. (D) Example of a chromothripsis event in chromosome 9 involving CN oscillations with interspersed loss of heterozygosity and templated insertions. Breakpoints corresponding to interchromosomal SVs are depicted as colored dots in the SV profile, whereas intrachromosomal SVs are represented with black dots and colored arcs.

FIGURE 5

Comparison of mutation landscape between this cohort and other three subtypes. (A)UpSetR plot shows the overlap of germline-regulated genes identified in the present study for the five lung cancer subtypes. (B) UpSetR plot shows the overlap of independent genomic loci that represent the genes shown in (A). (C) The TMB difference in the five lung cancer subtypes. (D)UpSetR plot shows the overlap of pathways from the Kyoto Encyclopedia of Genes and Genomes enriched with the germline-regulated genes.

FIGURE 6

Flowchart illustrating patient enrollment and analysis.