One of the most important goals of studying genetics is to find genes that are involved with diseases. The problem is that most diseases aren’t caused by a single gene or mutation. They’re the result of complex interactions among dozens, if not hundreds or thousands of genes, plus environmental factors, lifestyle, and a host of other variables. That flood of genes creates a needle-in-a-haystack problem.

A growing view among geneticists holds that nearly every gene active in the relevant tissue plays some part in a disease, but the vast majority act only indirectly and from a distance, nudging a much smaller set of "central" genes that sit at the heart of the disease. Those central genes are the ones that directly drive the biology and therefore are the ones most worth targeting with drugs. Until now, though, scientists had no reliable way to pick them out of the crowd and experimentally test them.

An interdisciplinary team of researchers at UChicago and Columbia University developed a new computational tool that could make the challenge of finding genes most directly related to disease much easier. In a paper published in Cell¹, they showed how this tool was able to identify 21 genes related to asthma, most of which hadn’t been discovered by other methods. The researchers also used both CRISPR gene-editing screens and mouse models to validate that these genes lead to asthma phenotypes and demonstrated that two of the genes are in the same pathway involved in fatty acid metabolism and protein palmitoylation, which hasn’t yet been studied for asthma. 

The new tool, called DANDELION, focuses on a process known as trans-gene regulation. In complex diseases like asthma, many genetic variants may contribute to disease by changing the expression of other genes. This has a cascading effect where one variant changes the expression of a nearby gene, and then that gene changes the expression of another, and so on. This creates what’s called a gene regulatory network that ultimately drives the development of disease.

Existing approaches like genome-wide association studies (GWAS) instead focus on genes that are often in the periphery of the gene regulatory network, however, and only indirectly affect disease. 

“What's unique about our method is that we believe the disease genes are not just next to the genetic variants. They're embedded in this gene regulatory network, and the actual disease-driving gene is downstream of those associated variants, maybe on different chromosomes. So, they're on the receiving end of a genetic effect that is very far away,” Liu said.

Liu named the tool DANDELION in reference to the puffy heads of dandelion flowers once they go to seed. The puffball resembles an interconnected, branching network of genes, ultimately pointing to the center of the core disease genes (called disease-proximal genes, or DPGs).

Liu analyzed a large set of data from the human transcriptome and the UK Biobank, a repository of health and genetic data from more than 500,000 volunteers. She used DANDELION to search for DPGs for asthma and found 21 candidates, 19 of which have not been discovered before using tools like GWAS. She showed the data to Marcelo Nóbrega, MD, PhD, Chair of the Department of Human Genetics at UChicago, who has developed experimental platforms to manipulate the expression of genes in human cell types that are relevant to asthma, such as epithelial cells from the lining of the airways, inflammatory cells, and immune T cells.

At first, he was skeptical. “Xuanyao showed me a list of genes, and we didn't recognize almost any of them. I thought that either this is going to be really cool and groundbreaking or it's going to be wrong. But we had the experimental validation system running, so I thought, ‘Let's test them all,’” he said.

Nóbrega’s team, led by postdoctoral scholar Isabella Salamone, PhD, conducted a series of experiments to test the effects of the genes predicted by DANDELION. Surprisingly, most of the genes Liu identified had a direct, measurable impact on the function of asthma-related cell types, producing phenotypes that model those seen in asthma at a much higher rate than the distal genes, or any other genes in the genome that they also tested.

When Salamone looked more closely, she saw that two genes had opposite effects. Knocking out one gene called SLC27A3 protected against the effects of asthma in both epithelial and T cells, while knocking out another gene, SCD, contributed to disease. Looking at the effects of mutations of these genes in a large human cohort of almost half a million people, they found that mutations in SLC27A3 are protective of asthma, supporting their findings in their cellular phenotyping screens.

“We saw this really striking pattern. Knocking out SLC27A3 had the strongest protective effect of all the genes we tested, and knocking out SCD was very detrimental to whatever cellular function we assayed,” Salamone said. “When we dug into patient data that had been collected by other labs, we saw the same pattern repeat itself—expression of SLC27A3 is increased in lung cells of patients with severe asthma, and SCD expression is decreased.” 

Intriguingly, both SLC27A3 and SCD are involved in the same biochemical pathway for fatty acid metabolism. To understand how this might be linked to asthma, they turned to chemical biologist Hening Lin, PhD, the James and Karen Frank Family Professor of Medicine and Professor of Chemistry at UChicago, who is a world-leading expert on the process, especially its role in protein palmitoylation, the addition of a long-chain fatty acyl group to proteins that regulate protein activity. Lin helped them confirm that both genes are involved in palmitoylation, and that reducing palmitoylation by knocking out SLC27A3 causes lung epithelial cells to dampen several immune-related and inflammatory processes.

“My lab has been working on the role of protein palmitoylation in immune signaling. We know many immune signaling pathways are regulated by palmitoylation, but I am still amazed by the finding that disrupting a lipid metabolic protein, SLC27A3, could offer protection in asthma models at least in part via affecting protein palmitoylation,” Lin said.

Finally, to test whether these genetic, biochemical, and cellular findings ultimately translate into asthma susceptibility, Salamone developed mice in which SLC27A3 or SCD were inactivated. They found that SLC27A3 knock-out mice are protected against allergy-induced lung inflammation, while SCD knock-out mice are more prone to lung inflammation compared to control mice, demonstrating that the new pathway the team uncovered is indeed capable of changing susceptibility to asthma.

Liu said she is encouraged by this initial success and looks forward to testing it with other diseases like inflammatory bowel disease or Type 2 diabetes. Historically, one of the limitations of drug development has been finding the true protein targets for treatment. DANDELION has the potential to overcome this challenge by identifying new and more effective drug targets.

“We're really excited about this direction because for these diseases, GWAS has identified tons of signal, but we still don't know the actual disease-driving genes that we can target for therapies,” Liu said. “I think our collaboration has been really powerful because we closed the gap at both the computational level and the experimental level.”

Nóbrega emphasized the importance of this collaboration as well, especially the advantage of being able to confirm their findings with a leading expert like Lin. “None of this would have come to fruition if any one of us were working on this alone,” he said. “We would have three papers buried in separate journals and virtually nobody would know how to put these stories back together. So, the power of having this complementary expertise across the division is really important.” 

Reference:

1) https://www.sciencedirect.com/science/article/pii/S0092867426008664

(Newswise/HG)