By Monica Heger
Using next-gen sequencing to follow up on previous genome-wide association studies, researchers at the National Human Genome Research Institute have identified regulatory elements in the genome of pancreatic islet cells that may be important for type 2 diabetes. According to the researchers, the approach could become increasingly useful for figuring out how variants in noncoding regions of the genome could impact disease.
"A lot of GWAS have linked regions of the genome to disease risk," said Jason Lieb, an associate professor of biology at the University of North Carolina who has developed a different sequencing method known as FAIRE-seq to study how regulatory regions impact diabetes (IS 2/9/2010). "But for a lot of the regions that are identified as being linked to disease risk, there's no functional info …, so it's difficult to know where to go from there."
For example, said Michael Stitzel, lead author of the paper, one region on chromosome 9 has been identified by GWAS as being linked to diabetes. "We know there's a variant predisposing to type 2 diabetes within 10 kilobases of sequence, but there's no known coding anything in that region. It's in a gene desert," Stitzel said. Researchers hypothesized that the region must be a regulatory region, he said.
In a study published last week in Cell Metabolism, the NHGRI researchers used DNase-seq, a sequencing technique that targets regulatory regions of the genome, and ChIP-seq, which targets transcription sites, from nine samples of human pancreatic islets to try and make sense out of the noncoding regions that previous GWAS had identified as being associated with type 2 diabetes. By first obtaining a picture of non-disease samples, the authors will be able to then apply the same approach to disease samples to pinpoint specific variants of interest.
The researchers first looked for areas in human pancreatic islets that were sensitive to DNase I, an enzyme that cuts DNA preferentially at open chromatin, which would point to potential regulatory regions. They also looked for histone modifications and genomic areas where CTCF — an insulator protein — binds in pancreatic islet cells.
They performed DNase-seq on three pancreatic islet samples on the Illumina Genome Analyzer, using a single-end sequencing strategy and read lengths of 20 base pairs. The researchers also performed ChIP-seq on six different samples using a single-end sequencing strategy and read lengths of 36 bases. For each sample, they sequenced regions sensitive to DNase, enriched for specific histone modifications, or bound by CTCF. For those regions they achieved between 4 million and 35 million aligned reads for each sample. Then, they focused their analysis on regions important for islet function and those identified by GWAS as being associated with type 2 diabetes.
[ pagebreak ]
One surprising result, said Stitzel, was that the insulin locus was "barren from the marks that we would have expected to be there," such as activating histone tail modifications. Looking further, the researchers also found the same absence of histone tail modifications at other islet hormone-encoding loci. "These are the two hormone genes that are very highly expressed, and they both lacked the modifications we expected to see," Stitzel said. "It seems that there is some type of unique gene regulatory mechanism going on for these highly expressed genes."
Additionally, in the 18 regions that have been pinpointed by GWAS as having an association with type 2 diabetes, the team found six potential regulatory elements, four of which appeared to be gene enhancers, Stitzel said.
The next step, said Stitzel, is to further study those highly expressed hormones — insulin, glucagon, and others — that have atypical regulation. They will also evaluate the six potential regulatory elements to try and figure out which genes they regulate.
"It's a really exciting time that we can start to address the role of noncoding regulatory elements and start to assign some function to this black-box genome and assess what variants may do to these different elements," Stitzel said.
The study was an important first step in identifying how epigenetic modifications could impact disease, said Lieb. "If we can first know what regions of the genome are being used in a regulatory capacity in a tissue we know is important for disease — such as the pancreatic islets for type 2 diabetes — then we can ask what SNPs occur in those regions."
Robert Hegele, director of the Blackburn Cardiovascular Genetics Lab at the Robarts Research Institute at the University of Western Ontario, agreed. "Epigenomic changes will be relevant to many other diseases," he said. "Looking at purely genomic changes, so far, those explain very little of what's going on genetically." Hegele has also used sequencing and GWAS together, to study hypertriglyceridemia, a condition characterized by high blood levels of fatty acid (IS 7/27/2010).
He added that it will be necessary to look at more samples, both from healthy individuals as well as individuals with the disease before any definitive conclusions can be drawn. "It's still early days for these studies," he said.
However, he predicted that in the future, GWAS and sequencing will be used as complementary approaches more frequently. Rather than pitting the common variant theory against the rare variant theory of disease, the two types of experiments — GWAS and sequencing — can be used in conjunction with one another.
"Each method gives you so much data. The logical thing is to look at the overlap between the two," he said.
Have topics you'd like to see covered in In Sequence? Contact the editor at mheger [at] genomeweb [dot] com