Every good captain knows that when your ship is in distress, you send out an SOS. Now, two new papers now leave us wondering if that tactic is borrowed from diseased tissues. Many diseases are characterized by cell death that releases circulating cell-free DNA (cfDNA) into the blood stream, which can now be analyzed in a ‘liquid biopsy’ as a tissue-specific SOS message.
cfDNA and Liquid Biopsy
While not yet perfect, liquid biopsy analysis of genetic variation in cfDNA offers a lot of potential for preventative medicine, due to the non-invasive nature of obtaining blood plasma. One particular advantage of screening cfDNA is that it allows therapeutic action to be taken before the disease progresses to the easier to detect, but much harder to treat, later stages. A large number of publications have shown the clinical potential of using cfDNA sequencing for tumor detection.
However, we’ve come to know that there are a number of diseases and disorders where sequence variation isn’t definitive. A few talented teams have now begun to notice that epigenomic (and thus cell-type specific) information can be detected by examining cfDNA. These studies are based on the rationale that cfDNA contains an epigenomic footprint that can be used to home in on the dying tissue of origin before the situation goes overboard.
Tissue-Specific Nucleosome Positioning Influences cfDNA Fragmentation Patterns
In January of 2016, Snyder et al. examined the nucleosome positioning of cfDNA. They reasoned that during cell death, genomic DNA is fragmented by enzymes that prefer to cut in the unexposed sections that aren’t protected by nucleosomes. Therefore, the cfDNA fragments released are those protected by nucleosomes and transcription factors, whereas the naked DNA is preferentially degraded.
They found that:
- The average fragment size was 167 bp. This suggests the DNA was associated with chromatosomes (nucleosome core particle + linker histone).
- There is a 10.4 bp periodicity in other fragments as well, which is indicative of the helical pitch of DNA bound to the nucleosome core particle.
- Both nucleosome and transcription factor (CTCF) occupancy can be inferred from the patterns of 120-180 bp and 35-80 bp fragments, respectively.
- This is done by analyzing the distribution of aligned fragment endpoints, which cluster around the boundaries of nucleosomes and are also depleted in the nucleosome.
- The patterns can be used to identify cell type of origin by comparing to known nucleosome positioning profiles, where the spacing of nucleosomes differs according to chromatin state and gene expression.
- The cfDNA footprint of healthy people is of lymphoid and myeloid cells. These represent hematopoietic lineages and their presence makes sense for healthy blood.
- In cancer patients, the cfDNA footprint includes non-hematopoietic lineages that often happen to align with the cancerous cell type.
Senior author Jay Shendure shares, “Our findings suggest it is possible to identify tissues contributing to cell-free DNA by looking at these fragmentation patterns, instead of looking for specific mutations in the DNA. The test examines the ends of each fragment of DNA, and tries to identify hotspots, or parts of the DNA that get cut more frequently than others.”
cfDNA Methylation as a Marker of Tissue Type
Meanwhile, in March of 2016, Lehmann-Werman et al. took the cfDNA methylation route to investigate tissue-specific patterns. The team profiled a number of tissues using the 450K array, which measures methylation at its tutular number of genomic loci, to find tissue-specific CpGs. Then, they examined cfDNA with a targeted NGS assay to look for the identified tissue-specific CpG along with several surrounding CpGs.
Using the above strategy, they examined the blood of patients and identified tissue-specific patterns of methylation in CpGs related to:
- The INS promoter of pancreatic β-cells in type 1 diabetes patients.
- MBP3 and WM1 in oligodendrocytes of relapsing multiple sclerosis patients.
- The CG09787504 locus (Brain1) in brain cells of patients after traumatic or ischemic brain damage.
- CUX2 and REG1A in exocrine pancreas cells of patients with pancreatic cancer or pancreatitis.
In order for their strategy to work, the single tissue-specific CpG from the 450K array must be analyzed alongside additional flanking CpGs, since none of the CpGs are definitive individually. This promising work drives home the point that examining the methylation of select genes opens up a number of diseases, in addition to cancer, which have the potential to benefit from the early and non-invasive detection capabilities of liquid biopsies.
In terms of future outlook, co-senior author Benjamin Glaser shares that, “In the long run, we envision a new type of blood test aimed at the sensitive detection of tissue damage, even without a-priori suspicion of disease in a specific organ. We believe that such a tool will have broad utility in diagnostic medicine and in the study of human biology.”
The Future of cfDNA Epigenomics
Overall, these two papers highlight the emerging tissue-specific potential of analyzing cfDNA epigenetics and suggest that cfDNA harbours much more information than sequence variation. Furthermore, the two manuscripts open up the possibility of integrating nucleosome positioning and DNA methylation patterns from cfDNA to develop a suite of epigenomic assays for the next generation of liquid biopsies.