Genetic analysis has been successful in identifying causative mutations for individual cardiovascular risk factors. Success has been more limited in mapping susceptibility genes for clusters of cardiovascular risk traits, such as those in the metabolic syndrome. We identified three large families with coinheritance of early-onset coronary artery disease, central obesity, hypertension, and diabetes. We used linkage analysis and whole-exome sequencing to identify the disease-causing gene. A founder mutation was identified in DYRK1B, substituting cysteine for arginine at position 102 in the highly conserved kinase-like domain. The mutation precisely cosegregated with the clinical syndrome in all the affected family members and was absent in unaffected family members and unrelated controls. Functional characterization of the disease gene revealed that nonmutant protein encoded by DYRK1B inhibits the SHH (sonic hedgehog) and Wnt signaling pathways and consequently enhances adipogenesis. Furthermore, DYRK1B promoted the expression of the key gluconeogenic enzyme glucose-6-phosphatase. The R102C allele showed gain-of-function activities by potentiating these effects. A second mutation, substituting proline for histidine 90, was found to cosegregate with a similar clinical syndrome in an ethnically distinct family. These findings indicate a role for DYRK1B in adipogenesis and glucose homeostasis and associate its altered function with an inherited form of the metabolic syndrome. (Funded by the National Institutes of Health.)

A rare mutation in LRP6 has been shown to underlie autosomal dominant coronary artery disease (CAD) and metabolic syndrome in an Iranian kindred. The prevalence and spectrum of LRP6 mutations in the disease population of the United States is not known. Two hundred white Americans with early onset familial CAD and metabolic syndrome and 2,000 healthy Northern European controls were screened for nonconservative mutations in LRP6. Three novel mutations were identified, which cosegregated with the metabolic traits in the kindreds of the affected subjects and none in the controls. All three mutations reside in the second propeller domain, which plays a critical role in ligand binding. Two of the mutations substituted highly conserved arginines in the second YWTD domain and the third substituted a conserved glycosylation site. The functional characterization of one of the variants showed that it impairs Wnt signaling and acts as a loss of function mutation.

Genetic variation in the FAM13A (Family with Sequence Similarity 13 Member A) locus has been associated with several glycemic and metabolic traits in genome-wide association studies (GWAS). Here, we demonstrate that in humans, FAM13A alleles are associated with increased FAM13A expression in subcutaneous adipose tissue (SAT) and an insulin resistance-related phenotype (e.g. higher waist-to-hip ratio and fasting insulin levels, but lower body fat). In human adipocyte models, knockdown of FAM13A in preadipocytes accelerates adipocyte differentiation. In mice, Fam13a knockout (KO) have a lower visceral to subcutaneous fat (VAT/SAT) ratio after high-fat diet challenge, in comparison to their wild-type counterparts. Subcutaneous adipocytes in KO mice show a size distribution shift toward an increased number of smaller adipocytes, along with an improved adipogenic potential. Our results indicate that GWAS-associated variants within the FAM13A locus alter adipose FAM13A expression, which in turn, regulates adipocyte differentiation and contribute to changes in body fat distribution.

Red teaming, the practice of adversarially exposing unexpected or undesired model behaviors, is critical towards improving equity and accuracy of large language models, but non-model creator-affiliated red teaming is scant in healthcare. We convened teams of clinicians, medical and engineering students, and technical professionals (80 participants total) to stress-test models with real-world clinical cases and categorize inappropriate responses along axes of safety, privacy, hallucinations/accuracy, and bias. Six medically-trained reviewers re-analyzed prompt-response pairs and added qualitative annotations. Of 376 unique prompts (1504 responses), 20.1% were inappropriate (GPT-3.5: 25.8%; GPT-4.0: 16%; GPT-4.0 with Internet: 17.8%). Subsequently, we show the utility of our benchmark by testing GPT-4o, a model released after our event (20.4% inappropriate). 21.5% of responses appropriate with GPT-3.5 were inappropriate in updated models. We share insights for constructing red teaming prompts, and present our benchmark for iterative model assessments.

Mutations in Dyrk1b are associated with metabolic syndrome and nonalcoholic fatty liver disease in humans. Our investigations showed that DYRK1B levels are increased in the liver of patients with nonalcoholic steatohepatitis (NASH) and in mice fed with a high-fat, high-sucrose diet. Increasing Dyrk1b levels in the mouse liver enhanced de novo lipogenesis (DNL), fatty acid uptake, and triacylglycerol secretion and caused NASH and hyperlipidemia. Conversely, knockdown of Dyrk1b was protective against high-calorie-induced hepatic steatosis and fibrosis and hyperlipidemia. Mechanistically, Dyrk1b increased DNL by activating mTORC2 in a kinase-independent fashion. Accordingly, the Dyrk1b-induced NASH was fully rescued when mTORC2 was genetically disrupted. The elevated DNL was associated with increased plasma membrane sn-1,2-diacylglyerol levels and increased PKCε-mediated IRK^T1150 phosphorylation, which resulted in impaired activation of hepatic insulin signaling and reduced hepatic glycogen storage. These findings provide insights into the mechanisms that underlie Dyrk1b-induced hepatic lipogenesis and hepatic insulin resistance and identify Dyrk1b as a therapeutic target for NASH and insulin resistance in the liver.