Development of a genome atlas for discriminating benign, preinvasive, and invasive lung nodules

Liang Peng, Minhua Peng, Jinsheng Tao, Bo Wang, Jinwang Wei, Lixuan Lin, Bo Cheng, Shan Xiong, Jianfu Li, Caichen Li, Ziwen Yu, Chunyan Li, Jun Wang, Hui Li, Zhiwei Chen, Jian‐Bing Fan, Wenhua Liang, Jianxing HeMedComm 2024 · DOI

TL;DR

This study develops a comprehensive "genome atlas" by integrating multi-omics data (genomics, transcriptomics, epigenomics) from over 400 lung nodules. This atlas reveals key molecular drivers like EGFR and TP53 mutations and distinct molecular subtypes, offering new insights into lung cancer progression and potential biomarkers for early detection and targeted therapies.

What Problem Does This Paper Address?

Lung cancer screening using low-dose computed tomography (LDCT) is effective but often leads to misdiagnosis and unnecessary procedures due to the difficulty in distinguishing benign nodules from preinvasive or invasive ones. Adenocarcinoma, the most common type, progresses through several stages (AAH, AIS, MIA, IA). A deeper understanding of the molecular mechanisms driving this progression is crucial for developing accurate diagnostic tools and effective treatments. Current research often focuses on single omics data, leaving a gap in a holistic, multi-omics view of early-stage lung nodule development.

Methods

The researchers employed a multi-omics approach, integrating:

  • Whole Exome Sequencing (WES): To identify somatic mutations in genes across the genome.
  • RNA-Sequencing (RNA-Seq): To analyze gene expression profiles and understand cellular functions and immune microenvironment.
  • DNA Methylation Sequencing: To investigate epigenetic alterations that can influence gene activity.

These datasets were analyzed using various bioinformatics tools for:

  • Somatic Variant Identification: Calling single nucleotide variants (SNVs) and short insertions/deletions (indels).
  • Tumor Mutation Burden (TMB) Calculation: Quantifying the number of somatic mutations.
  • Transcriptomic Subgrouping: Unsupervised hierarchical clustering of gene expression data to identify distinct molecular subtypes.
  • Immune Cell Infiltration Analysis: Using ssGSEA to score immune cell populations.
  • DNA Methylation Clustering: Hierarchical clustering of methylation data to identify epigenetic subgroups.
  • Integrated Analysis: Correlating DNA methylation and gene expression data to find inverse relationships.
  • Pathway Enrichment Analysis: Using KEGG and MSigDB to understand the biological functions associated with identified molecular alterations.

Datasets

The study utilized 432 archived Formalin-fixed paraffin-embedded (FFPE) lung nodule tissue samples from Chinese patients. After quality control, 422 samples were used for WES, 228 for RNA-Seq, and 83 for DNA methylation sequencing. These samples were categorized into three main groups:

  • Benign: (n=124)
  • Preinvasive: Atypical adenomatous hyperplasia (AAH) + Adenocarcinoma in situ (AIS) (n=32)
  • Invasive: Minimally invasive adenocarcinoma (MIA) + Invasive adenocarcinoma (IA) (n=266)

The study did not introduce a new benchmark dataset but rather compiled and analyzed a unique multi-omics dataset from a specific patient population.

Key Insights and Ideas

  • EGFR and TP53 as Key Drivers: These mutations are strongly associated with the progression from preinvasive to invasive lung nodules, with EGFR mutations becoming increasingly prevalent as the disease advances.
  • EGFR Pathway Activation: Both EGFR mutations and hypomethylation of EGF/EGFR genes contribute to activating the EGFR pathway, fueling cancer growth.
  • Distinct Molecular Subtypes: Transcriptomic analysis revealed four subgroups (EGFR-driven, EGFR/TP53 co-mutation, inflamed, and stem-like), each with unique molecular features and immune microenvironments, suggesting different underlying biological processes.
  • Estrogen/Androgen Signaling Link: These pathways are associated with the EGFR-driven subtype, opening avenues for combination therapies (e.g., TKIs with antiestrogens).
  • Preinvasive Nodules and Stem Cells: Preinvasive nodules show enrichment in stem cell pathways, which might play a role in hindering invasion initially.
  • Immune Microenvironment Differences: The "inflamed" and "stem-like" groups show higher immune cell infiltration, while EGFR-driven and EGFR/TP53 co-mutation groups are "cold" with less infiltration, impacting potential immunotherapy responses.
  • Epigenetic Regulation is Crucial: DNA methylation patterns change significantly across nodule progression, with specific genes showing inverse correlations between methylation and expression, highlighting epigenetics as a key player in lung cancer development.

Key Results

  • EGFR Mutation Frequency: Increased significantly from benign (3%) to preinvasive (16%), MIA (35%), and IA (73%).
  • TP53 Mutation Frequency: Significantly higher in IA (35.43%) compared to MIA (2.2%).
  • Four Transcriptomic Subgroups: Identified with distinct molecular characteristics and immune profiles.
  • DNA Methylation Clusters: Four groups (M1-M4) identified, with M2 and M4 showing distinct patterns related to nodule progression.
  • Inverse Methylation-Gene Expression: Identified fourteen genes with significant inverse correlations, including BDNF, CD48, COX7A1, IKZF1, and TNFRSF1B in the M2 group, and EGF, EGFR, and CDC6 in the M4 group, suggesting novel regulatory mechanisms.

Limitations and Future Work

  • Benign Nodules vs. Normal Tissue: The study used benign nodules instead of normal lung tissue, which can be challenging to differentiate from malignant ones in CT scans.
  • Small Preinvasive Cohort: The AAH+AIS group was relatively small (n=32) and lacked epigenetic data, limiting comprehensive analysis of this stage.
  • Lack of In Vitro/Vivo Experiments: The study did not include experimental validation of molecular functions, such as the interaction between EGFR and estrogen/androgen pathways.
  • Patient-Specific Tracking: It's challenging to track individual patients from nodule appearance to cancer development and obtain samples at all stages.
  • Future Directions: The authors suggest future studies to validate findings in larger cohorts, conduct in vitro/vivo experiments to explore molecular mechanisms, and investigate the identified genes as potential diagnostic biomarkers and therapeutic targets for early lung cancer interception.

Summary updated

Abstract

To tackle misdiagnosis in lung cancer screening with low-dose computed tomography (LDCT), we aimed to compile a genome atlas for differentiating benign, preinvasive, and invasive lung nodules and characterize their molecular pathogenesis. We collected 432 lung nodule tissue samples from Chinese patients, spanning benign, atypical adenomatous hyperplasia (AAH), adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (IA). We performed comprehensive sequencing, examining somatic variants, gene expressions, and methylation levels. Our findings uncovered EGFR and TP53 mutations as key drivers in - early lung cancer development, with EGFR mutation frequency increasing with disease progression. Both EGFR mutations and EGF/EGFR hypo-methylation activated the EGFR pathway, fueling cancer growth. Transcriptome analysis identified four lung nodule subtypes (G1-4) with distinct molecular features and immune cell infiltrations: EGFR-driven G1, EGFR/TP53 co-mutation G2, inflamed G3, stem-like G4. Estrogen/androgen response was associated with the EGFR pathway, proposing a new therapy combining tyrosine kinase inhibitors with antiestrogens. Preinvasive nodules exhibited stem cell pathway enrichment, potentially hindering invasion. Epigenetic regulation of various genes was essential for lung cancer initiation and development. This study provides insights into the molecular mechanism of neoplastic progression and identifies potential diagnostic biomarkers and therapeutic targets for lung cancer.
BibTeX
@inproceedings{peng2024development,
  author = {Liang Peng and Minhua Peng and Jinsheng Tao and Bo Wang and Jinwang Wei and Lixuan Lin and Bo Cheng and Shan Xiong and Jianfu Li and Caichen Li and Ziwen Yu and Chunyan Li and Jun Wang and Hui Li and Zhiwei Chen and Jian‐Bing Fan and Wenhua Liang and Jianxing He},
  title = {{Development of a genome atlas for discriminating benign, preinvasive, and invasive lung nodules}},
  booktitle = {MedComm},
  year = {2024},
  doi = {10.1002/mco2.644},
}
Figure