Weekly Endocrinology Research Analysis
This week’s endocrinology literature emphasizes shifts from mechanism to clinic: (1) a large trial analysis shows prediabetes remission — even without weight loss — reduces progression to type 2 diabetes, highlighting glycemic targets and adipose redistribution as key mediators; (2) mechanistic JCI work implicates SAM-dependent H3K27me3 as a mediator of SGLT2 inhibitor renal protection, opening biomarker and target opportunities; and (3) a randomized trial demonstrates a Bayesian decision-suppor
Summary
This week’s endocrinology literature emphasizes shifts from mechanism to clinic: (1) a large trial analysis shows prediabetes remission — even without weight loss — reduces progression to type 2 diabetes, highlighting glycemic targets and adipose redistribution as key mediators; (2) mechanistic JCI work implicates SAM-dependent H3K27me3 as a mediator of SGLT2 inhibitor renal protection, opening biomarker and target opportunities; and (3) a randomized trial demonstrates a Bayesian decision-support system safely lowers HbA1c for adults with type 1 diabetes on multiple daily injections, illustrating scalable AI-assisted insulin titration. Together these studies argue for expanding clinical endpoints beyond weight, integrating molecular biomarkers into drug action models, and adopting decision-support tools to improve glycemic control.
Selected Articles
1. Prevention of type 2 diabetes through prediabetes remission without weight loss.
Post hoc analyses of a large multicenter randomized dataset (PLIS), replicated in the US DPP, show that achieving prediabetes remission—even without weight loss or with modest weight gain—reduces progression to type 2 diabetes. Mechanistic correlates included improved insulin sensitivity, enhanced β-cell function and GLP-1 responsiveness, and redistribution of adipose tissue toward subcutaneous depots rather than visceral expansion.
Impact: Challenges the weight-centric paradigm for diabetes prevention by showing glycemic remission itself is protective and mechanistically distinct, which could reshape guideline targets and preventive strategies.
Clinical Implications: Prevention programs should incorporate explicit glycemic remission targets and strategies that improve insulin sensitivity and β-cell GLP-1 responsiveness, not solely weight-loss goals; clinicians may track remission as an independent endpoint.
Key Findings
- Prediabetes remission occurred without weight loss and conferred protection against incident T2D.
- Responders demonstrated improved insulin sensitivity, beta-cell function, and beta-cell GLP-1 sensitivity.
- Responders had adipose redistribution toward subcutaneous depots, while nonresponders increased visceral adiposity; findings were replicated in the US DPP.
2. SGLT2 inhibition protects kidney function by SAM-dependent epigenetic repression of inflammatory genes under metabolic stress.
Preclinical work demonstrates that SGLT2 loss/inhibition elevates renal S-adenosylmethionine (SAM) under diabetogenic conditions, increasing H3K27 trimethylation at NF-κB–related loci and repressing inflammatory gene expression, which is necessary for observed nephroprotection. Pharmacologic inhibition of MAT2A (reducing SAM) abrogated protection, positioning SAM/MAT2A as mechanistic mediators and candidate biomarkers for SGLT2 effects.
Impact: Provides a concrete, testable epigenetic mechanism linking SGLT2 activity to renal protection beyond hemodynamic effects and proposes measurable metabolite/epigenetic biomarkers and druggable nodes (MAT2A).
Clinical Implications: Motivates prospective clinical studies to measure renal SAM and H3K27me3 signatures in patients on SGLT2 inhibitors, to stratify responders and explore adjunct MAT2A/SAM-modulating strategies to enhance renoprotection.
Key Findings
- SGLT2-deficient mice under diabetogenic stress showed increased renal SAM correlated with improved kidney function.
- Increased H3K27me3 at NF-κB–related gene loci accompanied transcriptional repression of inflammatory pathways.
- MAT2A inhibition (reducing SAM) abolished the protective phenotype, supporting causality.
3. A Bayesian decision support system for automated insulin doses in adults with type 1 diabetes on multiple daily injections: a randomized controlled trial.
A 12‑week randomized controlled trial in adults with type 1 diabetes on multiple daily injections found that a Bayesian decision-support system providing weekly basal and prandial dose recommendations reduced mean HbA1c by 0.4% versus a non-adaptive bolus calculator without severe hypoglycemia or DKA, demonstrating feasible, safe glycemic benefit from an algorithmic titration tool integrated with CGM.
Impact: Demonstrates clinically meaningful, safe HbA1c improvement using an implementable algorithmic titration tool in MDI users — a high-impact step toward scalable diabetes care where closed-loop therapy is unavailable.
Clinical Implications: Clinicians and diabetes services should consider integrating validated decision-support systems to support weekly insulin titration for MDI patients with suboptimal control, as a scalable alternative to pump/closed-loop therapy.
Key Findings
- Bayesian DSS reduced HbA1c by 0.40% versus control over 12 weeks (treatment effect −0.40%, p=0.025).
- No severe hypoglycemia or diabetic ketoacidosis events occurred in either arm.
- Weekly algorithm-driven basal and prandial dosing using CGM data was operationally feasible.