Daily Endocrinology Research Analysis
A pragmatic phase 3 RCT in JAMA shows a fully automated AI-led Diabetes Prevention Program is noninferior to human coaching at 12 months, with higher initiation rates. Translational work in Hypertension identifies stress-responsive NR4A2 as a driver of aldosterone-producing cell cluster formation, refining primary aldosteronism biology. A transcriptomic study in European Journal of Endocrinology classifies prolactinomas and somatotropinomas into molecular subtypes with distinct resistance to dop
Summary
A pragmatic phase 3 RCT in JAMA shows a fully automated AI-led Diabetes Prevention Program is noninferior to human coaching at 12 months, with higher initiation rates. Translational work in Hypertension identifies stress-responsive NR4A2 as a driver of aldosterone-producing cell cluster formation, refining primary aldosteronism biology. A transcriptomic study in European Journal of Endocrinology classifies prolactinomas and somatotropinomas into molecular subtypes with distinct resistance to dopamine agonists and somatostatin analogs, informing precision therapy.
Research Themes
- AI-enabled diabetes prevention
- Adrenal pathophysiology in primary aldosteronism
- Molecular subclassification of pituitary adenomas and therapy resistance
Selected Articles
1. An AI-Powered Lifestyle Intervention vs Human Coaching in the Diabetes Prevention Program: A Randomized Clinical Trial.
In a 12-month pragmatic noninferiority RCT (n=368), referral to a fully automated AI-led DPP achieved the primary composite outcome in 31.7% vs 31.9% with human-led DPP, meeting the prespecified noninferiority margin. Program initiation was higher in the AI group (93.4% vs 82.7%), and results were consistent across composite components and sensitivity analyses.
Impact: This trial provides high-level evidence that an AI-led DPP can match human coaching effectiveness while improving uptake, addressing scale and access barriers in diabetes prevention.
Clinical Implications: Health systems can consider referring eligible adults with prediabetes to AI-led DPPs to expand access without compromising effectiveness, potentially reducing costs and workforce burden while maintaining weight, HbA1c, and activity outcomes.
Key Findings
- Primary composite outcome achieved in 31.7% (AI-led) vs 31.9% (human-led); noninferiority met with 1-sided 95% CI lower boundary of the risk difference at -8.2% (margin -15%).
- Program initiation after referral was higher with AI-led DPP (93.4%) than with human-led DPP (82.7%).
- Findings were consistent across composite components (weight loss, HbA1c reduction, physical activity) and in sensitivity analyses.
Methodological Strengths
- Phase 3 pragmatic multicenter randomized noninferiority design with prespecified margin
- Objective measurement of physical activity via actigraphy and standardized composite endpoint
Limitations
- Conducted at two US sites, which may limit generalizability
- Open-label referral comparison with 12-month follow-up; intervention delivery was external to the study team
Future Directions: Evaluate long-term diabetes incidence, cost-effectiveness, and equity impacts of AI-led DPP at scale across diverse health systems and populations.
IMPORTANCE: Prediabetes is common, yet evidence-based lifestyle interventions are underutilized. OBJECTIVE: To determine whether referral to an exclusively artificial intelligence (AI)-led lifestyle intervention based on the Diabetes Prevention Program (DPP) is noninferior to referral to a human-led DPP in achieving recommended thresholds for weight loss, hemoglobin A1c (HbA1c) reduction, and weekly physical activity among adults with prediabetes and overweight or obesity. DESIGN, SETTING, AND PARTICIPANTS: This phase 3, parallel-group, pragmatic, noninferiority randomized clinical trial was conducted from October 11, 2021, to December 16, 2024 (last follow-up) at 2 US clinical sites in Baltimore, Maryland, and Reading, Pennsylvania. Adults 18 years or older with prediabetes and overweight or obesity were enrolled. INTERVENTIONS: Participants were randomized in a 1:1 ratio to receive either a referral to an AI-powered DPP lifestyle intervention delivered via a mobile app and Bluetooth-enabled digital scale or a referral to a human coach-led DPP lifestyle intervention delivered remotely. Both interventions were delivered independently of the study team over a 12-month period. MAIN OUTCOMES AND MEASURES: The primary outcome was a composite of maintaining an HbA1c less than 6.5% throughout the study and achievement of at least 5% weight loss, at least 4% weight loss plus at least 150 minutes of weekly physical activity (assessed with actigraphy), or an absolute reduction in HbA1c of at least 0.2 percentage points at 12 months.
2. Role of Stress-Responsive NR4A2 in Aldosterone-Producing Cell Cluster Formation.
Integrated spatial transcriptomics and single-cell RNA-seq of matched human adrenal tissues revealed that APCCs are a distinct, ZG-like population. In silico perturbation and in vitro experiments indicated that stress-induced activation of NR4A2 in ZG cells promotes their progression to APCCs, with findings validated in additional patients.
Impact: This study proposes a mechanistic link between stress signaling and APCC formation via NR4A2, advancing understanding of primary aldosteronism pathogenesis and highlighting a potential therapeutic target.
Clinical Implications: NR4A2-driven APCC formation suggests stress pathways as candidates for disease modification in primary aldosteronism and may inform biomarkers or stratification for future interventional studies.
Key Findings
- APCCs form a distinct adrenal cell population whose transcriptomic profile is closer to zona glomerulosa than to aldosterone-producing adenomas.
- Stress-responsive transcription factor NR4A2 is activated in ZG cells under stimuli such as ACTH, promoting ZG-to-APCC progression (supported by in silico and in vitro data).
- Validation in tissue from two additional patients supports the generalizability of NR4A2-associated APCC formation; APCCs retain ZG-like stress responsiveness distinct from adenomas.
Methodological Strengths
- Integrated spatial transcriptomics and single-cell RNA sequencing on matched human adrenal tissues
- Convergent in silico perturbation and in vitro validation with additional patient tissue replication
Limitations
- Small number of patient samples limits generalizability
- Lack of in vivo functional manipulation to establish causality and no clinical intervention tested
Future Directions: Test NR4A2 modulation in preclinical in vivo models and evaluate NR4A2-associated signatures as biomarkers for stratifying primary aldosteronism.
BACKGROUND: Aldosterone-producing cell clusters (APCCs), which share some transcriptomic features with aldosterone-producing adenomas, are frequently observed in the adrenal cortex of aged individuals and those with primary aldosteronism. However, the mechanisms driving APCC formation remain poorly understood. METHODS: We performed an integrated analysis using spatial transcriptomics and single-cell RNA sequencing of APCC cells, aldosterone-producing adenoma cells, and zona glomerulosa (ZG) cells present in the same adrenal gland of a patient with unilateral primary aldosteronism, with validation analyses conducted on tissue samples from 2 additional patients. RESULTS: APCC cells exhibited a distinct cellular population with a gene expression profile more similar to that of the ZG cells than that of aldosterone-producing adenoma cells. In silico perturbation and in vitro studies suggest that the stress-responsive TF (transcription factor) NR4A2 is activated in ZG cells in response to a variety of stresses, such as ACTH (adrenocorticotropic hormone) stimulation, thereby promoting the progression from ZG to APCC cells. CONCLUSIONS: Our findings suggest that NR4A2 activation in ZG cells functions as a key molecular mediator in stress-induced APCC formation at least in some cases.
3. Transcriptomic classification of prolactinomas and somatotropinomas identifies subtypes with variable resistance to treatment.
Unsupervised transcriptomics of 46 prolactinomas and 58 somatotropinomas identified discrete tumor subtypes with distinct drug sensitivities. DRD2-high prolactinomas clustered among dopamine agonist–sensitive cases, while other subtypes showed resistance linked to cAMP/mitochondrial/ribosomal or immune gene programs. Somatotropinomas segregated into subtypes with variable somatostatin analog response, with SSTR2 expression predictive only in sparsely granulated tumors.
Impact: Defines molecular subtypes that explain heterogeneous medical therapy responses in pituitary adenomas, enabling movement toward transcriptome-guided treatment selection.
Clinical Implications: Molecular profiling may help select dopamine agonists or somatostatin analogs more effectively and identify patients needing alternative agents (e.g., pegvisomant or pasireotide) based on subtype-associated resistance programs.
Key Findings
- Four prolactinoma subtypes showed variable dopamine agonist sensitivity; DRD2-high tumors were enriched among sensitive cases, while resistant clusters showed cAMP, mitochondrial/ribosomal, or immune gene enrichment.
- Sparsely granulated somatotropinomas formed a separate molecular entity; other somatotropinomas split into five subtypes with differing somatostatin analog response, including subgroups defined by GNAS mutation, PIT1/SF1 coexpression, or SOX2.
- SSTR2 expression predicted somatostatin analog response in sparsely granulated somatotropinomas (P=0.022), but not in other somatotropinomas (P=0.923).
Methodological Strengths
- Unsupervised transcriptomic classification linked to histology and clinical treatment response
- Relatively large combined cohort (n=104) across two tumor types enhancing comparative insights
Limitations
- Observational design with potential treatment heterogeneity and lack of prospective validation
- Generalizability may be limited without external cohorts and standardized therapeutic protocols
Future Directions: Prospective validation of subtype-guided therapy and development of clinically deployable assays to classify tumors pre-treatment.
OBJECTIVE: Dopamine agonists and somatostatin analogs are commonly used in prolactinomas and somatotropinomas. Response to these therapies is heterogeneous. Whether resistance is linked to specific tumor subtypes or to shared mechanisms of resistance is not known. The aim was to explore distinct molecular subtypes of prolactinomas and somatotropinomas and their association with response to treatment. METHODS: The transcriptome of 46 prolactinomas and 58 somatotropinomas was analyzed. Unsupervised classifications were generated and tested for association with histological and clinical data, including response to treatment. RESULTS: Four subtypes of prolactinomas were identified, with variable sensitivity to dopamine agonists (P < 10-4). Sensitive tumors accumulated in the subgroup with the highest DRD2 expression, while resistant tumors accumulated in the 3 remaining ones, enriched in genes related to cAMP metabolism, to mitochondrial and ribosomal activity, and to immunity, respectively. Sparsely granulated somatotropinomas (N = 16) presented a separated molecular entity, mixing tumors resistant and sensitive to somatostatin analogs. The remaining somatotropinomas were classified into 5 subtypes, with variable sensitivity to somatostatin analogs (P < 10-4). Sensitive tumors accumulated in 3 subgroups, characterized by GNAS somatic mutation, PIT1 and SF1 coexpression, and the stem-cell marker SOX2 expression, respectively. Resistant tumors accumulated in the 2 remaining ones, enriched in genes related to cell cycle and to mesenchymal differentiation, respectively. Somatostatin receptor 2 (SSTR2) expression was associated with response to somatostatin analogs in sparsely granulated somatotropinomas (P = .022), but not in other somatotropinomas (P = .923). CONCLUSIONS: Prolactinomas and somatotropinomas were classified into distinct transcriptomic groups. Resistance to medical therapies was linked to distinct tumor subtypes, suggesting distinct mechanisms of resistance.