Research Roundup: March 28, 2022

Boosting barrier tissues, lifestyle vs. genetics in heart disease, countering cancer drug resistance, and more

Susanna M. Hamilton
Credit: Susanna M. Hamilton

Welcome to the March 28, 2022 installment of Research Roundup, a recurring snapshot of recent studies published by scientists at the Broad Institute and their collaborators.

Screening organoids for barrier boosters

Molecules that change the cellular makeup of the body’s barrier tissues — the skin, the airway, and the gut — could help treat diseases in which certain cell types are depleted or dysfunctional. A team led by Ben Mead, Kazuki Hattori (HMS/MIT), associate member Jeffrey Karp in the Chemical Biology and Therapeutics Science Program and Brigham and Women’s Hospital, and institute member Alex Shalek in the Klarman Cell Observatory used 3D intestinal organoids to find a molecule that boosts the numbers of protective Paneth cells in the gut. Their approach could be applied to other barrier tissues in the body to study their biology and explore ways to alter their cellular composition therapeutically. Read more in Nature Biomedical Engineering and a Broad news story.

Keeping CAD at bay

A balanced diet, exercise, and other elements of a healthy lifestyle reduce risk for coronary artery disease (CAD). In JAMA Network Open, Akl Fahed, Amit Khera of the Program in Medical and Population Genetics (MPG) and colleagues explored whether this holds true for those with rare variants in familial hypercholesterolemia-related genes. Using a large gene sequencing and lifestyle dataset from the UK Biobank, the team found among carriers of those variants, an unhealthy lifestyle led to a 66.2 percent estimated risk of CAD by age 75 years, whereas a favorable lifestyle led to a 34.5 percent estimated risk. This work suggests that a healthy lifestyle can significantly lower their risk even for those who are genetically predisposed to CAD.

AI-powered rejection detection

Screening heart biopsies is the standard way to detect rejection in heart transplant patients. Jana Lipkova, Tiffany Chen, associate member Faisal Mahmood, and colleagues set out to develop tools to automate the manual examination of endomyocardial biopsy (EMB) histological slides. In Nature Medicine, they describe a deep learning model, CRANE, that detects, subtypes, and grades allograft rejection from whole-slide EMB images. The team tested their model on curated datasets from three large international cohorts, finding that CRANE was just as accurate at detection as a small number of pathologists. The team suggests that when used as an assistive tool, CRANE could boost accuracy and reduce assessment time and variability between pathologists. Learn more in Lipkova and Chen's tweetorial.

Defining drivers of heart valve disease risk

Mitral valve prolapse (MVP) is a common disease of unknown molecular origin that affects the valve between the left upper and lower chambers of the heart. Carolina Roselli, institute member and Cardiovascular Disease Initiative director Patrick Ellinor, David Milan (MGH), and colleagues merged data from six genome-wide association studies with epigenetic, transcriptional, and proteomic data to conduct a multi-omic meta-analysis of MVP in more than 4,800 cases and 434,600 controls. They identified 16 loci and six candidate genes associated with MVP, and developed a polygenic score that — combined with age, sex, and clinical factors — improved MVP risk prediction. Learn more in the European Heart Journal.

Predicting resistance-resisting drug combinations

Combining drugs could help improve the treatment of patients with advanced cancers, but identifying promising combinations and the patients who might benefit from them remains challenging. In Nature Chemical Biology, research scientist Matthew Rees of the Cancer Program, Cory Johannessen, and colleagues describe using public datasets to predict effective combinations by scanning cancer cell lines for their small molecule sensitivities and corresponding transcriptional states. To illuminate one resistance mechanism, they validated that expression of the genes MGLL or CES1 conferred resistance to the compound GSK-J4 by direct enzymatic modification. Overall, their approach could help generate preclinical evidence to inform drug-combination trials.

Exploring the links between obesity, liver inflammation, and metabolism

Doctors have long recognized that obesity triggers chronic liver inflammation, which in turn is associated with metabolic dysfunction. But the molecular mechanisms behind that connection have been unclear. Writing in Science Translational Medicine, Suraj Patel, institute member Evan Rosen, and colleagues make the case that a transcription factor called interferon regulatory factor 3 (IRF3) — known for its immune functions — is a fundamental link between the three. Among their key findings: Knocking out or suppressing liver IRF3 expression rescues glucose metabolism in a mouse model, and expression of an IRF3 target gene, PPP2R1B, correlates with blood glucose levels in people with obesity.

Mining for Huntington disease modifiers

Genome-wide association studies of Huntington disease (HD) have highlighted loci that modify the “age at onset” and suggested a two-step pathogenesis: somatic instability of the causative HTT CAG repeat, followed by neuronal damage. To further study genetic HD modifiers, Jong-Min Lee, MPG associate member James Gusella, and others used motor and cognitive measures from several large HD natural history studies to predict when participants exhibit motor symptoms, finding that genetic differences at HD modifier loci have different impacts on motor- or cognitive-related phenotypes. The team's approach could help identify therapeutic targets and elucidate mechanisms of HD pathogenesis. Read more in the American Journal of Human Genetics.

Drivers of autoimmune risk

Human leukocyte antigen (HLA) risk alleles are well known to play a role in autoimmune disease risk. Kazuyoshi Ishigaki, institute member Soumya Raychaudhuri of the MPG, and colleagues investigated the influence of HLA alleles on thymic T cell selection, which can modulate the frequency of autoimmunity-triggering T cell receptors (TCRs). The team found strong associations between HLA and CDR3, the key region of the TCR that controls antigen recognition. A particularly strong association was identified at a position known to mediate risk for multiple autoimmune diseases, including rheumatoid arthritis and type I diabetes. This finding sheds light on potential mediators of inter-individual differences in autoimmune disease risk. Learn more in Nature.


To learn more about research conducted at the Broad, visit, and keep an eye on