NEW YORK (GenomeWeb) – A prospective analysis of the genetic, epigenetic, and gene expression features found in pre-invasive lung cancer lesions has helped to spell out the molecular features that appear in those lesions that go on to become invasive lung squamous cell carcinomas.
"If we can use this new understanding of cancer development to create new diagnostic tests, it may one day be invaluable in picking up cancer early, enabling people to access treatment much earlier in the disease process," co-first author Adam Pennycuick, a respiratory researcher at University College London, said in a statement.
Pennycuick and colleagues from UCL, the Wellcome Trust Sanger Institute, and elsewhere used whole-genome sequencing, gene expression arrays, and methylation arrays to profile genomic, transcriptomic, and epigenomic features in lung lesion samples collected over time from dozens of individuals with carcinoma in situ (CIS). The findings, published online this week in Nature Medicine, highlighted molecular features linked to CIS progression — from methylation shifts to a chromosomal instability signature.
The team also noted that the broader suite of mutation, copy number, and other molecular features plotted out in CIS samples, including those that progressed to invasive squamous cell carcinoma, is expected to reveal clues to the biological pathways and processes that go awry in pre-invasive and invasive forms of lung cancer.
"We are now continuing our research to further understand how these genes are driving cancer progression, and to see which ones could be targeted by new drug treatments," co-first author Vitor Teixeira, another researcher with UCL's Lungs for Living Research Centre, said in a statement.
Prior research suggests invasive forms of lung squamous cell carcinomas are the culmination of "step-wise evolution," the authors explained, moving from low-grade lung lesions, such as mild or moderate dysplasia, to high-grade lesions, including pre-invasive CIS. Even so, they noted, strategies for treating pre-invasive lung lesions often remain murky since features of biopsied samples "cannot distinguish lesions that will ultimately progress to invasive tumors from those that will spontaneously regress."
For the new analysis, the team followed 85 CIS patients over time, collecting 129 samples with autofluorescence bronchoscopy, biopsy, and computed tomography scanning analyses at four months and 12 months. Cases that either regressed or progressed to invasive cancer were then prioritized for multi-omic analyses.
In a subset of 51 formalin-fixed, paraffin-embedded samples from 42 of these individuals, the investigators used Illumina or Affymetrix arrays to profile gene expression. They also performed array-based methylation analyses of 87 samples from 47 individuals with CIS and used whole-genome sequencing on 39 fresh frozen samples from 29 of the patients.
The team compared these progressing and regressing CIS samples to one another and to lung squamous cell carcinomas previously profiled for the Cancer Genome Atlas project. Together, these analyses provided a look at the gene expression, methylation, somatic mutation, and copy number features in the CIS samples that ultimately regressed or progressed to become invasive lung cancers.
Both CIS and lung squamous cell carcinomas were frequently affected by mutations in the TP53, CDKN2A, SOX2, and AKT2 genes, the researchers reported, and frequently had a mutational signature associated with tobacco use. On the other hand, the CIS tumors were less prone to other gene mutations, including in EGFR and NOTCH1, when compared to the invasive squamous cell carcinomas. A handful of CIS that regressed to normal or to low-grade dysplasia had far fewer mutations or copy number changes.
"Our study helps to understand the earliest stages of lung cancer development, by figuring out what’s going on inside these cells even before they become cancerous," senior and corresponding author Sam Janes, a respiratory and thoracic medicine researcher affiliated with UCL and University College London Hospital, said in a statement, adding that the results may eventually lead to new screening tests and treatment strategies.
When the researchers integrated mutation, methylation, and expression features in samples from a couple dozen progressive or regressive CIS cases, for example, they came up with gene and methylation signatures for distinguishing between progressive and regressive CIS cases — signatures that also showed promise for separating invasive squamous cell carcinomas in TCGA from control samples.
Still, the authors cautioned that "although our predictive signatures demonstrate the power of molecular data in guiding management decisions, a prospective clinical trial using predictors derived from our data will be required before clinical use."