Summary: Researchers have identified multiple new risk genes for Alzheimer’s disease and progressive supranuclear palsy (PSP).
A new UCLA-led study has identified multiple new risk genes for Alzheimer’s disease and a rare, related brain disorder called progressive supranuclear palsy (PSP) by using a combination of new testing methods allowing for mass screening of genetic variants in a single experiment.
The study, published today in the journal Science, also presents a revised, new model showing how common genetic variants, while individually having a very small impact on disease, collectively may raise the risk of disease by disrupting specific transcriptional programs across the genome.
Typically, researchers have relied on genome-wide association studies (GWAS) in which they survey the genomes of a large group of people to identify genetic variants that increase risk for the disease. This is done by testing for markers along the chromosome, or loci, associated with a disease.
Each locus on average has dozens—and sometimes hundreds or thousands—of genetic markers in common that are co-inherited and therefore associated with the disease, making it difficult to identify which are actually the functional variants that cause disease.
Identifying the causal variants and the genes they impact is a major challenge in modern genetics and biomedicine. This study provides an efficient roadmap for tackling this problem.
For this this study, the authors conducted one of the first known uses of high-throughput testing to study neurodegenerative disease. The authors ran massively parallel reporter assays (MPRAs) to simultaneously test 5,706 genetic variants in 25 loci associated with Alzheimer’s and nine loci associated with PSP, a neurological disease that is much rarer than Alzheimer’s but has a similar pathology.
From that test, the authors with high confidence were able to identify 320 genetic variants that were functional. To validate the results, they ran a pooled CRISPR screen on 42 of those high-confidence variants in multiple cell types.
“We combined multiple advances that allow one to conduct high-throughput biology, in which instead of doing one experiment at a time, one does thousands of experiments in parallel in a kind of pooled format.
“This allows us approach this challenge of how to move from thousands of genetic variants associated with a disease to identifying which are functional and which genes they impact,” said Dr. Dan Geschwind, the study’s corresponding author and the Gordon and Virginia MacDonald Distinguished Professor of Human Genetics, Neurology and Psychiatry at UCLA.
Their data provided evidence implicating several new risk genes for Alzheimer’s, including C4A, PVRL2 and APOC1, and other new risk genes for PSP (PLEKHM1 and KANSL1). The authors were also able to validate several previously identified risk loci. The next steps would be studying how newly identified risk genes interact in cells and model systems, Geschwind said.
The study provides proof of principle that high-throughput testing can provide a “very efficient” roadmap for further research, Geschwind said, but he stressed that those approaches must be thoughtfully paired with more targeted experiments, as they were in this study.
“This success does not mean that we can jettison the kind of detailed, careful experimentation studying individual genes in model systems,” he said. “This just provides a key step between the GWAS and understanding disease mechanisms.”
Yonatan Cooper, the study’s lead author, said the combination of approaches the researchers took gave them greater confidence in their findings, while it also highlighted the challenge inherent in interpreting human genetic variation.
“We believe that integration of multiple methodologies will be critical for future work annotating disease-relevant variation in both the research and clinical domains,” said Cooper, who is a MD/Ph.D. candidate in the Medical Scientist Training Program at UCLA’s David Geffen School of Medicine.
The authors were also able to show in PSP at least one mechanism in which multiple loci associated with the disease acted additively to disrupt a core set of transcription factors, which essentially turn genes on and off, that are known to work together in specific cell types.
Geschwind said this indicated that common genetic variation located across the genome was affecting specific regulatory networks in specific cell types. That finding, he said, identifies new potential drug targets and suggests that rather than targeting one gene, targeting a network of genes could be an effective approach.
“We’re entering in a new stage of therapies—it’s beginning to be plausible to think about targeting networks,” Geschwind said.
About this genetics research news
Author: Press Office Source: UCLA Contact: Press Office – UCLA Image: The image is in the public domain
Functional regulatory variants implicate distinct transcriptional networks in dementia
The widespread adoption of genome-wide association studies (GWASs) has revolutionized the detection of genetic loci associated with complex traits. However, identification of the causal variants and mechanisms underlying genotype–phenotype associations poses an enormous challenge because most common susceptibility loci reside in noncoding genomic regions and are composed of many correlated polymorphisms owing to linkage disequilibrium (LD). Massively parallel reporter assays (MPRAs) permit the high-throughput functional characterization of noncoding genetic variation, yet they have not been systematically applied to neurodegenerative disease. In this study, we used MPRA coupled with CRISPR-based validation to identify likely causal genetic variants underlying two neurodegenerative conditions that are neuropathologically linked by intracellular tau protein aggregation—Alzheimer’s disease (AD) and progressive supranuclear palsy (PSP).
Neurodegenerative dementias such as AD and PSP are a major cause of morbidity and mortality worldwide. There are no disease-modifying therapeutics for either disorder, motivating major efforts to characterize disease pathogenesis. Identifying causal genetic factors and downstream associated risk genes are fundamental initial steps in developing a mechanistic understanding that would enable therapeutic development.
We tested the transcriptional regulatory activity of 5706 noncoding single-nucleotide variants, representing 25 genome-wide significant loci associated with AD and nine loci associated with PSP, by MPRA, using the neuroepithelial-like human embryonic kidney 293T (HEK293T) cell line. We identified 320 different functional regulatory variants (frVars) that affect gene expression within 27 of these loci. AD frVars were enriched within microglial enhancers, whereas PSP frVars were enriched within neuronal enhancers and, to a lesser extent, oligodendrocytes, consistent with differential cellular impact in each disease.
The majority of frVars (94%) overlapped two or more known functional annotations in human brain tissue or blood, nearly two-thirds of which were predicted to disrupt transcription factor binding, indicating their potential relevance in human disease. Forty-two high-confidence regulatory variants distributed across 15 AD loci and three PSP loci were selected for validation, using either CRISPR droplet sequencing (CROP-seq) or direct CRISPR excision in induced pluripotent stem cell–derived neurons, microglia, and astrocytes, enabling validation of 19 functional variants, implicating 20 risk genes across 11 loci.
Our data provide evidence implicating C4A, PVRL2, and APOC1 as risk genes in AD and PLEKHM1 and KANSL1 as risk genes in PSP, as well as additional validation for more than a dozen other genes. MPRA-defined functional variants preferentially disrupt predicted transcription factor binding sites (TFBSs) that converge on enhancers with differential cell type–specific activity in PSP and AD, implicating a neuronal SP1-driven regulatory network in PSP pathogenesis.
We provide systematic characterization of common variants underlying disease risk for two distinct neurodegenerative disorders, AD and PSP. Our work illustrates the utility of integrating multiplexed reporter and CRISPR assays to efficiently characterize noncoding disease-associated variation, thereby permitting the identification of risk genes, even in complex loci with high LD. These analyses support a mechanism underlying noncoding genetic risk, whereby common genetic variants drive disease risk in aggregate through polygenic cell type–specific regulatory effects on specific gene networks.