NEW YORK (GenomeWeb) – Cell-free DNA (cfDNA) circulating in blood plasma can provide clues for predicting gene expression, according to a study published online today in Nature Genetics.
By tapping into nucleosome differences previously described upstream of active or inactive transcription start sites, Austrian researchers developed a method for estimating gene expression from read depth patterns in cfDNA — a method they established in blood samples from more than 100 healthy individuals and ultimately applied to samples from hundreds more individuals with metastatic cancer.
"[W]e leveraged whole-genome sequencing data to include entire promoter regions for the establishment of gene expression status," senior author Michael Speicher, a human genetics researcher at the Medical University of Graz, and his colleagues wrote. "Our study provides a new view on the genomes of cells that release their DNA into the circulation and hence expands upon currently existing options for cfDNA analyses."
The team noted that much of the cfDNA that's floating in the blood represents genetic material from apoptotic cells, including DNA that remains paired with protein complexes in nucleosomes. As such, the group reasoned that it should be possible to use DNA read depth to uncover nucleosome footprint related to active gene transcription.
In particular, the study's authors pointed to past micrococcal nuclease (MNase) assays that revealed stretches of nucleosome-free DNA upstream of transcription start sites of genes being transcribed.
With that in mind, they first set out to investigate the possibility that expression-related nucleosome occupancy at promoters leaves patterns that can be detected in cfDNA and, if so, whether such marks reflect gene expression. From there, the team expanded its analysis to search for expression profiles that might inform searches for cancer driver genes in blood samples from individuals with cancer.
In the first 179 blood samples, collected from 50 healthy men and 54 healthy women, the researchers did paired-end sequencing on Illumina MiSeq and NextSeq instruments to identify nucleosome-associated nuclear DNA in the blood plasma. They then generated single-end sequencing data that could be used to assess read depth patterns around transcription start sites, uncovered with help from existing MNase map and gene expression data from the ENCODE and FANTOM5 projects. That information, in turn, was used to develop an algorithm for predicting expression from cfDNA read depths.
After demonstrating that it was possible to pick up expression and copy number profiles by whole-genome sequencing on circulating tumor DNA in blood samples from two individuals with breast cancer — generating profiles that corresponded with RNA sequence data from primary tumor samples — the team took a look at gene expression and copy number gains in 426 blood samples from individuals with metastatic colon, breast, lung, or prostate cancer.
The approach appeared particularly promising for picking up amplifications affecting cancer driver genes, the researchers noted, though they cautioned that it may not be suitable for assessing circulating tumor DNA in a minimal residual disease context.