Single‐cell third‐generation sequencing‐based multi‐omics uncovers gene expression changes governed by ecDNA and structural variants in cancer cells

Lei Chang, En-Ze Deng, Jun Wang, Wei Zhou, Jian Ao, Rong Liu, Dan Su, Xiaoying FanClinical 2023 · DOI

TL;DR

This paper introduces scGTP-seq, a novel single-cell multi-omics method that leverages third-generation sequencing (TGS) to simultaneously profile both the genome and transcriptome. This breakthrough allows researchers to precisely link extrachromosomal DNA (ecDNA) and structural variants (SVs) to gene expression changes at the single-cell level, offering unprecedented insights into cancer heterogeneity and progression.

What Problem Does This Paper Address?

Cancer cells are notoriously complex, often characterized by large-scale genomic alterations like extrachromosomal DNA (ecDNA) and structural variants (SVs). These genomic changes are strongly implicated in cancer initiation and progression, but understanding how they influence gene expression in the diverse cell populations within a tumor has been a major challenge. Existing single-cell sequencing methods, primarily based on next-generation sequencing (NGS), struggle to accurately detect these large genomic variations due to their short read lengths. This limitation prevents a comprehensive understanding of how ecDNA and SVs contribute to cancer cell phenotypes at a single-cell resolution.

Methods

The core innovation is scGTP-seq, a single-cell multi-omics sequencing method. It combines:

  • Third-Generation Sequencing (TGS): Specifically, Nanopore and PacBio sequencing platforms are used. TGS offers long reads, which are crucial for accurately identifying large genomic structures like ecDNA and SVs, a significant advantage over traditional NGS.
  • SMOOTH-seq adaptation: The authors adapted a TGS-based single-cell whole-genome sequencing (scWGS) method called SMOOTH-seq.
  • Physical Separation of DNA and RNA: Carboxylic acid magnetic beads are used to physically separate DNA and RNA within each single cell. This allows for independent amplification and sequencing of both modalities.
  • Smart-seq2 for RNA: The separated RNA is processed using the established Smart-seq2 protocol for transcriptome analysis.
  • Custom Bioinformatics Pipeline (ecDNAFinder): A novel pipeline, ecDNAFinder, was developed to specifically identify ecDNAs from TGS reads, addressing a gap in existing analytical tools.
  • STRT-based method for RNA (in clinical samples): To improve throughput for analyzing clinical samples, an STRT-based method was used for single-cell RNA sequencing.

Datasets

The study utilized the following datasets:

  • Cultured Cell Lines: HEK293T, U2OS, COLO320DM, and PC3 cell lines were used for method development and validation.
  • Clinical Hepatocellular Carcinoma (HCC) Sample: A clinical sample of HCC was analyzed to demonstrate the method's applicability to real-world patient data.
  • Human Embryonic Stem Cells (hESCs): Used for validation of ecDNA candidates.

No new benchmarks were explicitly introduced, but the study aimed to overcome limitations of existing NGS-based single-cell methods for detecting ecDNA and SVs.

Key Insights and Ideas

  • TGS is Key for Large Genomic Variants: The long reads from TGS platforms are essential for accurately detecting ecDNA and SVs at the single-cell level, a capability largely missing in NGS-based single-cell methods.
  • ecDNA and SVs Directly Impact Gene Expression: The scGTP-seq method allows for the direct correlation of genomic variations (ecDNA and SVs) with transcriptomic changes within the same cell, revealing how these alterations drive gene expression.
  • ecDNA Hotspots Exist: The study identified specific genomic regions that are frequently amplified as ecDNAs, suggesting "ecDNA hotspots" in cancer cells.
  • Copy Number Isn't Everything: The relationship between ecDNA copy number and gene expression is complex, with some low-copy number ecDNA genes showing high expression, indicating the involvement of other regulatory mechanisms.
  • ecDNAs are Present in Non-Cancer Cells: Surprisingly, ecDNAs were also detected in non-cancerous cell types (TECs and TAMs) within the HCC sample, suggesting they might represent accumulated genomic mutations or be influenced by the tumor microenvironment.
  • ecDNA Genes as Prognostic Biomarkers: Certain ecDNA genes showed a correlation with overall survival in HCC patients, highlighting their potential as biomarkers for cancer diagnosis and prognosis.

Key Results

  • Method Validation: scGTP-seq demonstrated high sensitivity and accuracy for both RNA and DNA detection, comparable to existing methods for RNA and offering novel capabilities for genomic variants.
  • ecDNA Detection: The ecDNAFinder pipeline successfully identified ecDNAs in cultured cell lines and a clinical HCC sample.
  • Gene Expression Correlation: Genes located on ecDNAs consistently showed higher expression levels than non-ecDNA genes.
  • SV Impact: Deletion and insertion SVs were found to lead to altered transcripts, and this was experimentally validated.
  • Clinical Application: In HCC samples, scGTP-seq identified ecDNAs in cancer cells (HPCs) as well as non-cancer cells (TECs and TAMs), providing cell-type-specific ecDNA profiles.
  • Survival Correlation: Genes like TRMT112 and BANF1, found on ecDNAs, showed a significant correlation with overall survival in HCC patients.

Limitations and Future Work

  • Sequencing Cost and Throughput: The high cost and limited read output of TGS platforms currently restrict the number of cells that can be analyzed. The authors suggest that future improvements in TGS technology and automation will be crucial for increasing sample size.
  • Distinguishing ecDNA from Tandem Duplications: While the pipeline has filters, distinguishing ecDNA from large duplication events, especially for long ecDNA candidates, can still be challenging.
  • Low Copy Number Detection: Low copy number ecDNAs might be missed due to sequencing depth limitations.
  • ecDNA Integration: The study did not capture evidence of ecDNA reintegration into the genome, possibly due to limitations in detection sensitivity.
  • Non-Cancer Cell Role: Further research is needed to fully elucidate the role of ecDNA in different non-cancer cell types.
  • Future Directions: The authors envision scGTP-seq as a powerful tool for studying ecDNA biogenesis and its role in cancer progression, with future work likely focusing on refining the method for higher throughput and broader applications.

Summary updated

Abstract

We demonstrated that our method can efficiently detect ecDNA and SVs and illustrated how these variations affect transcriptomic changes in various cell lines. Finally, we applied and validated this method in a clinical sample of hepatocellular carcinoma (HCC), demonstrating a feasible way to monitor the evolution of ecDNA and SVs during cancer progression.
BibTeX
@inproceedings{chang2023singlecell,
  author = {Lei Chang and En-Ze Deng and Jun Wang and Wei Zhou and Jian Ao and Rong Liu and Dan Su and Xiaoying Fan},
  title = {{Single‐cell third‐generation sequencing‐based multi‐omics uncovers gene expression changes governed by ecDNA and structural variants in cancer cells}},
  booktitle = {Clinical},
  year = {2023},
  doi = {10.1002/ctm2.1351},
}
Figure