Summary: Study reveals a link between cholesterol metabolism and a subtype of autism. The association appears to have a genetic component. Mothers with lipid abnormalities are 16% more likely to have a child diagnosed with ASD, and fathers with abnormal lipid levels were 14% more likely. Individuals on the autism spectrum were twice as likely to have lipid abnormalities than those without ASD. Among those with ASD and abnormal lipid levels, conditions such as ASD, epilepsy, and sleep disorders were more common than in those with normal levels. Findings suggest dyslipidemia may alter neurodevelopment and result in other medical conditions, such as anemia and vitamin D deficiency.
Researchers at Harvard Medical School, Massachusetts Institute of Technology and Northwestern University have identified a subtype of autism arising from a cluster of genes that regulate cholesterol metabolism and brain development.
The researchers say their findings, published Aug. 10 in Nature Medicine, can inform both the design of precision-targeted therapies for this specific form of autism and enhance screening efforts to diagnose autism earlier.
The team identified the shared molecular roots between lipid dysfunction and autism through DNA analysis of brain samples–findings that they then confirmed by examining medical records of individuals with autism. Indeed, both children with autism and their parents had pronounced alterations in lipid blood, the analysis showed.
The results of the study, the researchers said, raise many questions; key among them are: Just how do lipid alterations drive neurodevelopmental dysfunction and could normalizing lipid metabolism affect disease outcomes?
The new findings set the stage for future studies to answer these questions and others.
“Our results are a striking illustration of the complexity of autism and the fact that autism encompasses many different conditions that each arise from different causes–genetic, environmental or both,” said study senior investigator Isaac Kohane, chair of the Department of Biomedical Informatics in the Blavatnik Institute at Harvard Medical School. “Identifying the roots of dysfunction in each subtype is critical to designing both treatments and screening tools for correct and timely diagnosis–that is the essence of precision medicine.” A Google map of autism Autism and autism-spectrum disorders, estimated to affect one in 54 children in the United States, are among the most complex heritable conditions. Thousands of gene variants, both rare and common, have been implicated in autism, likely through an intricate and not-well understood interplay between genetic and environmental factors–both before and after birth.
The new study findings not only underscore this complexity but also demonstrate the critical importance of defining the various subtypes of the condition and developing treatments that target subtype-specific anomalies.
Achieving a meaningful level of specificity in the study of a vastly complex disorder such as autism, however, is not easy. To do so, the researchers used a novel approach based on the interlacing of multiple layers of data, including whole exome sequencing, patterns of protein expression, medical records and health insurance claims.
“Think of a Google map and how it overlays various types of information on top of one another–cities, streets, parcels, land use, electrical grids, elevations–for a more detailed representation,” said Yuan Luo, who co-led the study with Alal Eran, a Harvard Medical School lecturer on pediatrics at Boston Children’s Hospital.
“This is what we did with our data to get a complete view of genes that have multiple regulatory functions and are implicated in autism,” said Luo, who started working on the research while at MIT’s Computer Science & Artificial Intelligence Lab and continued the work at Northwestern University, where he is now associate professor of preventive medicine at the Feinberg School of Medicine.
The team started out by analyzing patterns of gene expression from brain samples contained in two large national brain banks, focusing on genes that work in tandem during prenatal and postnatal brain development. Because autism is four times more common in males than females, they further focused on genes that had the largest male-to-female differences during development. Within those, they homed in on exons–the protein-coding parts of genes–to seek out mutations that occurred more often in patients with autism. Through this progressive zooming in, the researchers identified a previously unrecognized node of shared function–a cluster of exons regulating both neurodevelopment and fat metabolism.
Protein to person
To confirm whether the molecular link between autism and lipid metabolism was borne out in actual patients, the team turned to two vast clinical record repositories. In one that contained more than 2.7 million records of patients seen at Boston Children’s, including more than 25, o00 children with autism, the researchers identified notable lipid alterations in children with autism, including changes in levels of their bad cholesterol (LDL), good cholesterol (HDL) and triglycerides.
The other dataset contained medical records of more than 34 million individuals seen at multiple U.S. medical institutions. Of those, more than 80,700 individuals had diagnoses of autism. Overall, 6.5 percent of those who had an autism diagnosis also had abnormal lipid levels. Individuals with autism were nearly twice as likely to have abnormal lipid tests results as those without autism. There was also a pronounced familial link. Mothers with lipid abnormalities were 16 percent more likely to have a child with autism than mothers without lipid abnormalities. The risk for having a child with autism among fathers with lipid abnormalities was 13 percent greater than in males with normal lipid levels. And within families with more than one child, children diagnosed with autism were 76 percent more likely to have abnormal lipid profiles than their siblings.
Among individuals with autism and abnormal lipid levels on their blood work, conditions such as epilepsy, sleep disorders and attention deficit hyperactivity disorder were markedly more common than among those without elevated lipid levels–a finding that suggests dyslipidemia may alter neurodevelopment in general, the researchers said. Individuals with autism and dyslipidemia were also more likely to have certain hormonal and metabolic conditions including anemia, hypothyroidism and vitamin D deficiency.
The autism-dyslipidemia link persisted even when the researchers accounted for the possible influence of drugs commonly used in people with autism, some of which are known to affect lipid levels. In fact, lipid abnormalities were more common among people with autism who were not taking such medications.
The newly found link offers a molecular explanation to the well-established observation that a mutation in a gene involved in cholesterol metabolism is also found in people with Rett syndrome, a neurodevelopmental disorder closely related to autism. Another striking observation that may be explained by the newly found link is that between 50 and 88 percent of children born with Smith-Lemli-Opitz syndrome, caused by a defect in cholesterol synthesis, also have autism.
The researchers say their approach–based on integrating multiple data modalities– could be adapted to other similarly genetically complex conditions as a way to precision-profile subtypes of disease.
For example, the ability to identify disease subtypes in cancer in the past two decades has propelled the field of oncology forward and led to the development of many targeted cancer treatments, researchers said.
“Our findings can help design precision-targeted treatments that home in on the specific defect underlying the development of dyslipidemia-related autism,” Kohane said. “Conceptually, this is the same framework that we can apply in complex inherited neurodevelopmental disorders like autism and beyond. Our multimodal approach combining multiple types of data demonstrates that this is not only possible but imminent.”
Co-investigators on the study included Nathan Palmer, Paul Avillach, Ami Levy-Moonshine and Peter Szolovits.
Funding: The work was supported by the National Institute of Health (grants 1R21LM012618, 5UL1TR001422, P50MH106933, U01HG007530, OT3OD025466, OT3HL142480, U54HG007963, 1U01TR002623-01 and 1U54HD090255-01), Israeli Ministry of Science and Technology (grant 17708), and Precision Link Biobank for Health Discovery at Boston Children’s. Palmer received funding support from Aetna Life Insurance.
About this autism research article
Source: Harvard Contacts: Ekaterina Pesheva – Harvard Image Source: The image is in the public domain.
A multidimensional precision medicine approach identifies an autism subtype characterized by dyslipidemia
The promise of precision medicine lies in data diversity. More than the sheer size of biomedical data, it is the layering of multiple data modalities, offering complementary perspectives, that is thought to enable the identification of patient subgroups with shared pathophysiology. In the present study, we use autism to test this notion. By combining healthcare claims, electronic health records, familial whole-exome sequences and neurodevelopmental gene expression patterns, we identified a subgroup of patients with dyslipidemia-associated autism.