6 June 2024 in Nature Communications
The I148M variant of PNPLA3 is closely associated with hepatic steatosis. Recent evidence indicates that the I148M mutant functions as an inhibitor of PNPLA2/ATGL-mediated lipolysis, leaving the role of wild-type PNPLA3 undefined. Despite showing a triglyceride hydrolase activity in vitro, PNPLA3 has yet to be established as a lipase in vivo. Here, we show that PNPLA3 preferentially hydrolyzes polyunsaturated triglycerides, mobilizing polyunsaturated fatty acids for phospholipid desaturation and enhancing hepatic secretion of triglyceride-rich lipoproteins. Under lipogenic conditions, mice with liver-specific knockout or acute knockdown of PNPLA3 exhibit aggravated liver steatosis and reduced plasma VLDL-triglyceride levels. Similarly, I148M-knockin mice show decreased hepatic triglyceride secretion during lipogenic stimulation. Our results highlight a specific context whereby the wild-type PNPLA3 facilitates the balance between hepatic triglyceride storage and secretion, and suggest the potential contribution of a loss-of-function by the I148M variant to the development of fatty liver disease in humans.
13 April 2023 in Journal of Hepatology
28 August 2022 in Cardiovascular Digital Health Journal
Sarcoidosis with cardiac involvement, although rare, has a worse prognosis than sarcoidosis involving other organ systems. We used a large dataset to train machine learning models to predict in-hospital mortality among sarcoidosis patients admitted with heart failure (HF). Utilizing the National Inpatient Sample, we identified 4659 patients hospitalized with a primary diagnosis of HF. In this cohort, we identified patients with a secondary diagnosis of sarcoidosis using International Statistical Classification of Disease, Tenth Revision (ICD-10) codes. Patients were separated into a training group and a testing group in a 7:3 ratio. Least absolute shrinkage and selection operator regression was used to select variables to prevent model overfitting or underfitting. For machine learning models, logistic regression, random forest, and XGBoosting were applied in the training group. Parameters in each of the models were tuned using the GridSearchCV function. After training, all models were further validated in the testing group. Models were then evaluated using the area under curve (AUC) score, sensitivity, and specificity. A total of 2.3% of sarcoidosis patients died in HF admission. Our machine learning model analysis found the RF model to have the highest AUC score and sensitivity. Feature analysis found that comorbid arrhythmias and fluid electrolyte disorders were the strongest factors in predicting in-hospital mortality. Machine learning methods can be useful in identifying predictors of in-hospital mortality in a given dataset.
8 January 2026 in Nature Metabolism
Several adipose depots, including constitutive bone marrow adipose tissue, resist conventional lipolytic cues. However, under starvation, wasting or cachexia, the body eventually catabolizes stable adipocytes through unknown mechanisms. Here we developed a mouse model of brain-evoked depletion of all fat, including stable constitutive bone marrow adipose tissue, independent of food intake, to study this phenomenon. Genetic, surgical and chemical approaches demonstrated that catabolism of stable adipocytes required adipose triglyceride lipase-dependent lipolysis but was independent of local nerves, the sympathetic nervous system and catecholamines. Instead, concurrent hypoglycaemia and hypoinsulinaemia activated a potent catabolic state by suppressing lipid storage and increasing catecholamine-independent lipolysis via downregulation of cell-autonomous lipolytic inhibitors including G0s2. This was also sufficient to delipidate classical adipose depots and was recapitulated in tumour-associated cachexic mice. Overall, this defines unique adaptations of stable adipocytes to resist lipolysis in healthy states while isolating a potent catecholamine-independent neurosystemic pathway by which the body can rapidly catabolize all adipose tissues.
24 May 2025 in npj Digital Medicine
SRSivaraman Rajaganapathy
+7
The complexities inherent in EHR data create discrepancies between real-world evidence and RCTs, posing substantial challenges in determining whether a treatment is likely to have a beneficial impact compared to the standard of care in RCTs. The objective of this study is to enhance the prediction of efficacy direction for repurposed drugs tested in RCTs for heart failure (HF). To achieve this, we propose an efficacy direction prediction framework that integrates drug-target predictions with EHR-based Emulation Trials (ET) to derive surrogate endpoints for prediction using HF prognostic markers. Our validation of the proposed novel drug-target prediction model against the BETA benchmark demonstrates superior performance, surpassing existing baseline algorithms. Furthermore, an evaluation of our framework in identifying 17 repurposed drugs—derived from 266 phase 3 HF RCTs—using data from 59,000 patients at the Mayo Clinic highlights its remarkable predictive accuracy.