The profusion of genome-wide association studies (GWAS) in recent years have uncovered hundreds of genetic variants linked to neurodegenerative diseases. Yet the wealth of sequences hasn’t yielded an abundance of treatments. Until researchers understand the biochemical pathway between inheriting a risk variant and developing a disease, new therapies remain largely out of reach.

Now, researchers at the University of Pennsylvania have shown how one risk locus alters chromatin structure, kick-starting the path to dementia. The team, led by Alice Chen-Plotkin, associate professor of neurology, set out to characterize genetic variants associated with frontotemporal lobe dementia (FTLD).  Inevitably fatal, FTLD strikes between the ages of 45–64, causing changes in personality and behavior. The disease is closely related to amyotrophic lateral sclerosis (ALS), and in some cases, causes similar motor deterioration.

Dozens of genetic variants had been identified that correlate with expression of TMEM106B, a gene associated with FTLD. “This was a completely unknown gene in 2010, and over the years it’s become a hot topic,” says neuroscientist Rosa Rademakers, Ph.D., who studies FTLD in her role as a professor at the Mayo Clinic in Florida. People with FTLD have increased TMEM106B mRNA levels, and excess TMEM106B causes the lysosomes to malfunction, allowing toxic waste to build up in the neuron.

Because blocks of risk variants are generally inherited together, it’s impossible to know which individual SNP might be the culprit, influencing TMEM106B expression. To narrow down the field, the researchers compared results from different ethnic groups to see which variants persisted across all populations. This gave them a pool of 84 variants—which is still a dauntingly large number to try to test for biological function.

While they pondered their options, new data from ENCODE project came along with a timely answer: by overlaying the SNP sequences against the ENCODE data, they winnowed their options even further. “It allowed us to zero in on the most plausible candidate regulatory elements,” Chen-Plotkin tells GEN. “That was a huge help to us, because when you’re faced with 84 [variants], you can test all of them in just a superficial way, or you can pick and choose your favorites, but that’s a little non-objective.”  They soon eliminated all but one promising candidate, and that sequence turned out to bind a chromatin-organizing protein called CTCF.

While the study on TMEM106B was first published on BioRxiv in June 2017, it was subsequently published online in AJHG in October 2017.

“As a personal scientific journey, it was so fun,” recalls Chen-Plotkin. “We had to learn how to do lots of things when we found the genetic risk variant was differentially recruiting CTCF.” Chen-Plotkin reached out to experts in chromatin architecture, and her team began learning and adapting techniques to study CTCF and chromatin looping.

CTCF regulates expression by binding to the DNA and also to itself, forming homodimers that loop the chromatin and bring far-flung regulatory elements into contact with each other. “It’s a very important factor and it does many different things,” says Albert La Spada, M.D., Ph.D., director of the Duke Center for Neurodegeneration and Neurotherapeutics. “That’s what’s always fascinating about it to me.” Sometimes, CTCF acts as an insulator, blocking interaction between the enhancer and the promoter. In this case, CTCF enables interaction between the promoter and enhancer regions, boosting TMEM106B transcription. Though CTCF chromatin manipulations aren’t commonly associated with neurological disorders just yet—preliminary analysis suggests that risk variants for other diseases, such as Parkinson’s and Alzheimer’s, may affect CTCF binding sites.

“There are important cellular processes that probably are being regulated very carefully and very finely through changes in chromatin remodeling,” says La Spada, “so it stands to reason this would be particularly important in the nervous system, where cells have many demands placed on them.”

As GWAS data keep adding to the mountain of genetic disease-associated loci, it will become increasingly important to have strategies to understanding just how variants in noncoding regions influence expression.

“It’s definitely a really nice study,” Rademakers concludes. “I think it will be an example for other people trying to figure out what variants in noncoding regions may be doing.”


Caroline Seydel is a contributor to GEN.


Previous articleStress Accompanied by Epigenetic Add-Ons, Many at “Neuropsychiatric” Loci
Next articleUntangling Spider Biology