No abstract

The prevalence of non-obese nonalcoholic fatty liver disease (NAFLD) is increasing worldwide with unclear etiology and pathogenesis. Here, we show GP73, a Golgi protein upregulated in livers from patients with a variety of liver diseases, exhibits Rab GTPase-activating protein (GAP) activity regulating ApoB export. Upon regular-diet feeding, liver-GP73-high mice display non-obese NAFLD phenotype, characterized by reduced body weight, intrahepatic lipid accumulation, and gradual insulin resistance development, none of which can be recapitulated in liver-GAP inactive GP73-high mice. Common and specific gene expression signatures associated with GP73-induced non-obese NAFLD and high-fat diet (HFD)-induced obese NAFLD are revealed. Notably, metformin inactivates the GAP activity of GP73 and alleviates GP73-induced non-obese NAFLD. GP73 is pathologically elevated in NAFLD individuals without obesity, and GP73 blockade improves whole-body metabolism in non-obese NAFLD mouse model. These findings reveal a pathophysiological role of GP73 in triggering non-obese NAFLD and may offer an opportunity for clinical intervention.

Weakly supervised machine reading comprehension (MRC) task is practical and promising for its easily available and massive training data, but inevitablely introduces noise. Existing related methods usually incorporate extra submodels to help filter noise before the noisy data is input to main models. However, these multistage methods often make training difficult, and the qualities of submodels are hard to be controlled. In this paper, we first explore and analyze the essential characteristics of noise from the perspective of loss distribution, and find that in the early stage of training, noisy samples usually lead to significantly larger loss values than clean ones. Based on the observation, we propose a hierarchical loss correction strategy to avoid fitting noise and enhance clean supervision signals, including using an unsupervisedly fitted Gaussian mixture model to calculate the weight factors for all losses to correct the loss distribution, and employ a hard bootstrapping loss to modify loss function. Experimental results on different weakly supervised MRC datasets show that the proposed methods can help improve models significantly.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection induces new-onset diabetes and severe metabolic complications of pre-existing diabetes. The pathogenic mechanism underlying this is incompletely understood. Here, we provided evidence linking circulating GP73 with the exaggerated gluconeogenesis triggered by SARS-CoV-2 infection. We found that SARS-CoV-2 infection or glucotoxic condition increased the cellular secretion of GP73. Secreted GP73 trafficked to the liver and kidney to stimulate gluconeogenesis through cAMP/PKA pathway. By using global phosphoproteomics, we found a drastic remodeling of PKA kinase hub exerted by GP73. Notably, COVID-19 patients showed pathologically elevated plasma GP73, and neutralization of the secreted GP73 inhibited enhanced PKA signaling and glucose production associated with SARS-CoV-2 infection. GP73 blockade also reduced gluconeogenesis and lowered hyperglycemia in type 2 (T2D) diabetic mice. Therefore, our findings provide novel insight into the roles of GP73 as a key glucogenic hormone and mechanistic clues underlying the development of SARS-CoV-induced glucose abnormalities.