Daily Endocrinology Research Analysis
Three impactful studies advance endocrine-metabolic science: (1) a human GR variant (rs6190) mechanistically elevates cholesterol and atherosclerosis via hepatic PCSK9/BHLHE40 with sex-specific effects; (2) periportal THRSP–MIF–CD74+ lipid-associated macrophage crosstalk drives MASH and is druggable with a small-molecule inhibitor; (3) a multimodal, explainable AI integrating fundus imaging and clinical data predicts 5‑year CKD risk in type 2 diabetes across binational cohorts.
Summary
Three impactful studies advance endocrine-metabolic science: (1) a human GR variant (rs6190) mechanistically elevates cholesterol and atherosclerosis via hepatic PCSK9/BHLHE40 with sex-specific effects; (2) periportal THRSP–MIF–CD74+ lipid-associated macrophage crosstalk drives MASH and is druggable with a small-molecule inhibitor; (3) a multimodal, explainable AI integrating fundus imaging and clinical data predicts 5‑year CKD risk in type 2 diabetes across binational cohorts.
Research Themes
- Sex-specific genetic mechanisms driving dyslipidemia and atherosclerosis
- Immunometabolic crosstalk and spatial zonation in fatty liver disease (MASLD/MASH)
- Explainable multimodal AI for chronic kidney disease risk prediction in type 2 diabetes
Selected Articles
1. The human glucocorticoid receptor variant rs6190 increases blood cholesterol and promotes atherosclerosis.
A common GR coding variant (rs6190) elevates circulating cholesterol and atherosclerosis risk via hepatic transactivation of PCSK9 and BHLHE40, suppressing LDLR/HDLR. The effect shows sex specificity: attenuated by corticosterone/testosterone in males and additive with estrogen loss in females; mechanisms validate in CRISPR-edited human hepatocyte-like cells.
Impact: It uncovers a GR-dependent, targetable pathway linking a prevalent human variant to dyslipidemia and atherosclerosis with clear sex-specific biology, integrating human population data with mechanistic in vivo and in vitro validation.
Clinical Implications: Highlights PCSK9/BHLHE40 as mediators of GR-driven hypercholesterolemia, supporting intensified lipid-lowering (e.g., PCSK9 inhibition) in genetically at‑risk women and prompting consideration of sex hormones’ modulatory roles.
Key Findings
- rs6190 associates with higher cholesterol in women in UK Biobank and All of Us.
- SNP-genocopy mice show hepatic GR-driven transactivation of Pcsk9 and Bhlhe40, elevating all lipoprotein fractions and atherosclerosis.
- Liver knockdown of Pcsk9/Bhlhe40 abrogates atherogenesis; CRISPR-edited human hepatocyte-like cells recapitulate the mutant program.
- Corticosterone/testosterone mitigate, while estrogen loss augments, the mutant GR lipid program.
Methodological Strengths
- Convergent evidence from large human cohorts (UK Biobank, All of Us), SNP-genocopy mice, and CRISPR-edited human iPSC-derived hepatocytes.
- Causal pathway interrogation via in vivo liver knockdown of Pcsk9 and Bhlhe40 with sex-hormone modulation experiments.
Limitations
- Translational extrapolation from mouse hAPOE*2/*2 background to human pathophysiology may not fully capture complexity.
- Population associations, while robust, are observational and do not quantify clinical benefit of targeted interventions.
Future Directions: Assess clinical lipid/atherosclerosis outcomes by rs6190 genotype under PCSK9 inhibition and evaluate sex hormone interactions; explore BHLHE40 as a therapeutic target.
Elevated cholesterol poses cardiovascular risks. The glucocorticoid receptor (GR) harbors a still undefined role in cholesterol regulation. Here, we report that a coding SNP in the gene encoding the GR, rs6190, is associated with increased cholesterol in women according to UK Biobank and All of Us (NIH) datasets. In SNP-genocopying mice, we found that the SNP enhanced hepatic GR activity to transactivate Pcsk9 and Bhlhe40, negative regulators of LDL and HDL receptors, respectively. In mice, the SNP was sufficient to elevate circulating cholesterol across all lipoprotein fractions and the risk and severity of atherosclerotic lesions on the proatherogenic hAPOE*2/*2 background. The SNP effect on atherosclerosis was blocked by in vivo liver knockdown of Pcsk9 and Bhlhe40. Also, corticosterone and testosterone were protective against the mutant GR program in cholesterol and atherosclerosis in male mice, while the SNP effect was additive to estrogen loss in females. Remarkably, we found that the mutant GR program was conserved in human hepatocyte-like cells using CRISPR-engineered, SNP-genocopying human induced pluripotent stem cells. Taken together, our study leverages a nonrare human variant to uncover a GR-dependent mechanism contributing to atherogenic risk, particularly in women.
2. MIF-mediated crosstalk between THRSP + hepatocytes and CD74 + lipid-associated macrophages in hepatic periportal zone drives MASH.
Spatial transcriptomics pinpoints periportal THRSP-high hepatocytes as hubs recruiting CD74+ LAMs through MIF, driving MASH. THRSP augments palmitate via de novo lipogenesis and prevents FASN ubiquitination (FASN–TRIM21 disruption). A small-molecule THRSP inhibitor (C6) significantly ameliorates MASH in mice.
Impact: It elucidates spatially resolved immunometabolic crosstalk underpinning MASH and validates THRSP as a druggable node with a proof‑of‑concept small-molecule inhibitor that reverses disease in vivo.
Clinical Implications: Supports targeting THRSP–MIF–CD74+ LAM axis for MASH therapy, prioritizing periportal processes and offering a new class of agents beyond metabolic/anti-inflammatory standards.
Key Findings
- Periportal zones in MASH show increased myeloid cells and THRSP-high hepatocytes.
- THRSP drives MASH by MIF-mediated recruitment of CD74+ lipid-associated macrophages.
- THRSP increases palmitate via de novo lipogenesis and blocks FASN ubiquitination by disrupting FASN–TRIM21.
- A THRSP inhibitor (C6) significantly ameliorates MASH in mice.
Methodological Strengths
- Spatial transcriptomics with ligand–receptor inference (CellPhoneDB) and in situ co-localization.
- Mechanistic gain/loss-of-function with pathway dissection and in vivo therapeutic testing of a novel inhibitor.
Limitations
- Preclinical models; human causal validation and safety/PK of C6 are pending.
- Quantitative contribution of periportal vs pericentral zones in human MASH progression remains to be defined.
Future Directions: Advance THRSP inhibitors to IND-enabling studies; biomarker development for periportal THRSP/MIF/CD74 activity; patient stratification by zonation signatures.
BACKGROUND AND AIMS: Spatial location of steatosis is closely related to the progression of metabolic dysfunction-associated steatohepatitis (MASH), and reports suggest lipid-associated macrophages (LAMs) facilitate this progression. However, the underlying mechanisms remain elusive. APPROACH AND RESULTS: Spatial transcriptomics (ST) data revealed a significant increase in myeloid cells and MASH-related genes in the hepatic periportal (PP) zone of MASH mice, suggesting a vital role of the PP zone in MASH progression. THRSP (SPOT14), involved in fatty acid synthesis, was markedly elevated in the livers of MASH patients and mice. Notably, CellPhoneDB analysis identified strong interactions between CD74 and macrophage migration inhibitory factor (MIF) within the Thrsp -high zone. Furthermore, Thrsp , Cd74 , Mif , Col3a1 , and LAMs markers were prominently colocalized in the hepatic PP zone of MASH mice, suggesting that Thrsp -mediated crosstalk in this region played a crucial role in MASH progression. Thrsp overexpression/knockout experiments confirmed that THRSP drove MASH progression by recruiting CD74 + LAMs mediated by MIF. Mechanistically, THRSP promoted hepatic palmitic acid (PA) synthesis, mainly by promoting hepatic de novo lipogenesis, and disturbing the binding of FASN-TRIM21, thereby inhibiting FASN ubiquitination. CD74 + LAMs were activated and recruited by chemokine-like MIF secreted from hepatocytes and macrophages stimulated with PA. Additionally, the compound C6, identified as a THRSP inhibitor, significantly ameliorated MASH in mice. CONCLUSIONS: Our study demonstrates that THRSP drives MASH progression by recruiting CD74 + LAMs mediated by MIF in the hepatic PP zone, providing novel insights into the spatial zonation and crosstalk between lipogenic hepatocytes and LAMs that may result in novel therapies for MASH.
3. A Multimodal Predictive Model for Chronic Kidney Disease and Its Association With Vascular Complications in Patients With Type 2 Diabetes: Model Development and Validation Study in South Korea and the U.K.
An ensemble multimodal model combining fundus photographs (VGG16) and clinical data (DNN) achieved AUC 0.88 internally and 0.72 externally for 5‑year incident CKD in T2D. Explainability (SHAP/Grad‑CAM) highlighted eGFR and optic disc features; higher model probabilities aligned with elevated macro- and microvascular event risks.
Impact: Demonstrates clinically actionable, explainable AI integrating imaging and clinical modalities for CKD risk stratification across diverse cohorts, linking predictions to vascular outcomes.
Clinical Implications: Supports fundus-and-clinical multimodal screening to identify high-risk T2D patients for CKD prevention, earlier nephroprotective therapy, and vascular risk management.
Key Findings
- Internal AUC 0.880 and external AUC 0.722 for 5-year incident CKD prediction in T2D using fundus images plus clinical data.
- Explainability identified eGFR and optic disc features as key predictors (SHAP, Grad‑CAM).
- Higher model probability associated with increased macrovascular (HR up to 1.64) and microvascular (HR 1.30) risks.
Methodological Strengths
- Large discovery cohort with external validation in UK Biobank, enhancing generalizability.
- Explainable AI (SHAP/Grad-CAM) clarifies imaging and clinical feature contributions.
Limitations
- External performance attenuation (AUC 0.722) indicates domain shift; prospective impact studies are needed.
- CKD definition includes codes and eGFR thresholds; potential misclassification bias.
Future Directions: Prospective deployment with clinical decision support to test outcome improvement; domain adaptation to enhance cross-population performance; integration with nephroprotective treatment pathways.
OBJECTIVE: To develop a multimodal model to predict chronic kidney disease (CKD) in patients with type 2 diabetes mellitus (T2DM), given the limited research on this integrative approach. RESEARCH DESIGN AND METHODS: We obtained multimodal data sets from Kyung Hee University Medical Center (n = 7,028; discovery cohort) for training and internal validation and UK Biobank (n = 1,544; validation cohort) for external validation. CKD was defined based on ICD-9 and ICD-10 codes and/or estimated glomerular filtration rate (eGFR) ≤60 mL/min/1.73 m2. We ensembled various deep learning models and interpreted their predictions using explainable artificial intelligence (AI) methods, including Shapley additive explanation values (SHAP) and gradient-weighted class activation mapping (Grad-CAM). Subsequently, we investigated the potential association between the model probability and vascular complications. RESULTS: The multimodal model, which ensembles visual geometry group 16 and deep neural network, presented high performance in predicting CKD, with area under the receiver operating characteristic curve of 0.880 (95% CI 0.806-0.954) in the discovery cohort and 0.722 in the validation cohort. SHAP and Grad-CAM highlighted key predictors, including eGFR and optic disc, respectively. The model probability was associated with an increased risk of macrovascular complications (tertile 1 [T1]: adjusted hazard ratio, 1.42 [95% CI 1.06-1.90]; T2: 1.59 [1.17-2.16]; T3: 1.64 [1.20-2.26]) and microvascular complications (T3: 1.30 [1.02-1.67]). CONCLUSIONS: Our multimodal AI model integrates fundus images and clinical data from binational cohorts to predict the risk of new-onset CKD within 5 years and associated vascular complications in patients with T2DM.