Daily Endocrinology Research Analysis
Analyzed 127 papers and selected 3 impactful papers.
Summary
Analyzed 127 papers and selected 3 impactful articles.
Selected Articles
1. Hypothalamic POMC neurons regulate intestinal glucose absorption via a gut-brain circuit.
PKA signaling in hypothalamic POMC neurons is activated after feeding and GLP-1–based therapies and drives a vagal efferent program that suppresses SGLT1-dependent intestinal glucose absorption. Despite insulin resistance and obesity caused by pituitary effects, POMC-specific PKA activation improved glucose tolerance via reduced absorption and increased fecal glucose loss, defining a POMC–vagal–gut SGLT1 regulatory axis.
Impact: This work links central melanocortin signaling to intestinal glucose transport via a defined vagal circuit and explains part of GLP-1 therapy’s systemic effects, opening avenues to target SGLT1 via brain–gut pathways in insulin-resistant states.
Clinical Implications: Suggests potential to augment glycemic control by neuromodulating the POMC–vagal pathway or combining GLP-1 therapies with intestinal SGLT1 modulation; motivates studies assessing whether vagal or central interventions reduce postprandial glycemic excursions.
Key Findings
- Feeding and GLP-1–based agents activate PKA signaling in hypothalamic POMC neurons.
- POMC-specific PKA activation improved glucose tolerance by reducing intestinal glucose absorption and increasing fecal glucose excretion, despite obesity and insulin resistance.
- Mechanistically, a POMC–vagal efferent pathway suppresses SGLT1-dependent glucose absorption in the upper intestine.
Methodological Strengths
- Genetic, cell-type-specific manipulation of PKA signaling in POMC neurons with in vivo physiological readouts.
- Circuit-level mechanistic linkage from hypothalamus to vagal efferents and intestinal transporter function.
Limitations
- Pituitary corticotroph PKA activation confounds systemic phenotype (obesity, hypercortisolism).
- Preclinical mouse study; human translatability and safety of neuromodulation remain untested.
Future Directions: Map synaptic nodes and neurotransmitters in the POMC–vagal circuit; test pharmacologic and device-based neuromodulation to modulate intestinal SGLT1; examine synergy with GLP-1 receptor agonists in large animals and early human studies.
Hypothalamic proopiomelanocortin (POMC)-producing neurons are essential for maintaining energy balance and glucose homeostasis. We show that cAMP-dependent protein kinase A (PKA) signaling in these neurons is activated postprandially and upon the administration of glucagon-like peptide-1-based antiobesity/antidiabetic agents. To investigate the metabolic regulatory role of PKA signaling in hypothalamic POMC neurons, we generated mice with POMC-specific constitutive PKA activation by depleting the PKA regula
2. Cell-type-specific proximity labeling of organ secretomes reveals energy balance-dependent proteomic remodeling.
A Cre-dependent, ER-targeted TurboID system labels secreted and membrane proteins in specific cell types in vivo, enabling organ secretome profiling. Application to hepatocytes, adipocytes, and B cells at baseline and under fasting, inflammation, and diet-induced obesity revealed tissue- and perturbation-specific ER proteome remodeling tied to systemic energy balance.
Impact: This is a methodological advance that enables cell-type-resolved in vivo mapping of secretory/membrane proteomes across metabolic states, accelerating discovery of endocrine mediators, biomarkers, and therapeutic targets.
Clinical Implications: Provides a pipeline to identify tissue-of-origin biomarkers and drug targets relevant to obesity, diabetes, and inflammatory metabolic diseases; may support precision endocrinology by linking secreted proteomes to systemic phenotypes.
Key Findings
- Developed a Cre-dependent, ER-targeted TurboID proximity labeling system for temporally controlled, cell-type-specific labeling of secreted and membrane proteins in vivo.
- ER proteomes of hepatocytes, adipocytes, and B cells undergo tissue- and perturbation-specific remodeling under fasting, inflammation, and diet-induced obesity.
- The method is broadly applicable and supports biomarker and therapeutic target discovery across metabolic tissues.
Methodological Strengths
- Genetic cell-type specificity with temporal control in vivo across multiple metabolic tissues.
- Application under diverse physiological and pathological perturbations (fasting, inflammation, dietary obesity).
Limitations
- Mouse ES cell-derived model; translation to human tissues remains to be shown.
- ER-targeted labeling may underrepresent non-ER secretory routes; requires Cre-driver availability.
Future Directions: Extend to human organoids/primary tissues; integrate with longitudinal proteomics in metabolic disease cohorts; validate candidate endocrine factors as biomarkers/targets.
Intercellular communication is critical for maintaining organismal metabolic homeostasis. Here, we develop a method enabling temporally controlled, cell-type-specific labeling of secreted and membrane proteins in key metabolic tissues. The method employs a genetically encoded proximity-labeling strategy by targeting a Cre-dependent TurboID ligase to the endoplasmic reticulum (ER) in ES cell-derived mice. The expression of TurboID in hepatocytes, adipocytes, and B lymphocytes enabled the characterizat
3. Development and prospective evaluation of a real-time deep learning model for inpatient hypoglycemia prediction.
Using time-series EHR data from 143,124 admissions, a real-time LSTM predicted 24-hour inpatient hypoglycemia with F1=0.30, precision=0.23, recall=0.44, and AUPRC=0.23, outperforming logistic regression, dense NN, and XGBoost. Prospective daily validation in live EHR streams showed stable performance and SHAP-based interpretability identified clinically meaningful predictors.
Impact: First large-scale, prospectively validated, deployable real-time hypoglycemia predictor leveraging multi-hospital EHR time series, positioning AI for proactive inpatient glycemic safety.
Clinical Implications: Supports proactive glycemic stewardship: earlier insulin dose adjustments, meal/insulin coordination, and risk-triggered monitoring; next steps include randomized implementation to test reductions in hypoglycemia and alarm burden.
Key Findings
- An LSTM using 5-day, 4-hourly EHR windows achieved F1=0.30, precision=0.23, recall=0.44, AUPRC=0.23 at threshold 0.7, outperforming LR, dense NN, and XGBoost.
- Prospective daily validation on live EHR extracts demonstrated stable model performance over time.
- SHAP explanations identified recent insulin administration and prior hypoglycemia as top temporal predictors; performance was similar across demographic subgroups.
Methodological Strengths
- Large multi-hospital dataset with time-series predictors and prospective deployment evaluation.
- Transparent interpretation via SHAP and comparison against strong machine-learning baselines.
Limitations
- Moderate precision may create alert burden; no randomized trial of clinical impact.
- External generalizability beyond three hospitals not yet established; threshold selection affects metrics.
Future Directions: Conduct randomized implementation trials measuring hypoglycemia reduction and workflow outcomes; site-specific calibration, fairness audits, and integration with closed-loop protocols.
Inpatient hypoglycemia is associated with increased morbidity, mortality, length of stay, and healthcare costs, yet current management remains reactive due to the lack of real-time prediction tools. We developed, validated, and prospectively evaluated a real-time long short-term memory (LSTM) model to predict hypoglycemia within 24 h using electronic health record (EHR) data from 143,124 adult inpatient admissions across three hospitals between 2014 and 2025. Eligible patients were ≥ 18 years old, hospi