Abstract: SA-PO0005
Differential Glomerular Pathomics and Spatial Transcriptomics in Short and Long-Looped Nephrons: Insights from Human and Mouse Models
Session Information
- Intelligent Imaging and Omics: Phenotyping and Risk Stratification
November 08, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Artificial Intelligence, Digital Health, and Data Science
- 300 Artificial Intelligence, Digital Health, and Data Science
Authors
- Yin, Mengmeng, Shanghai Jiao Tong University School of Medicine Affiliated Ninth People's Hospital, Shanghai, China
- Yang, Yuechen, Vanderbilt University, Nashville, Tennessee, United States
- Wang, Yu, Vanderbilt University Medical Center, Nashville, Tennessee, United States
- Deng, Ruining, Weill Cornell Medicine, New York, New York, United States
- Liu, Jing, Vanderbilt University Medical Center, Nashville, Tennessee, United States
- Zhao, Shilin, Vanderbilt University Medical Center, Nashville, Tennessee, United States
- Fogo, Agnes B., Vanderbilt University Medical Center, Nashville, Tennessee, United States
- Huo, Yuankai, Vanderbilt University, Nashville, Tennessee, United States
- Yang, Haichun, Vanderbilt University Medical Center, Nashville, Tennessee, United States
Background
Nephrons in the kidney include short-looped (SLN) and long-looped nephrons (LLN). SLN and LLN glomeruli differ in cortical location, size, mesangial morphology, and disease susceptibility. This study analyzes these differences using pathomics and spatial transcriptomics from human samples and diseased mouse kidneys.
Methods
We examined 105 non-tumor nephrectomy samples. These included patients without comorbid conductions as well as individuals with hypertension, diabetes, and/or aging conditions. Mouse models included normal mice, angiotensin II + high salt-treated mice (model of hypertension), db/db eNOS-/- mice (model of T2 diabetes), and aging mice. An AI-driven pipeline was used for the segmentation and pathomics quantification and comparison of SLN and LLN glomeruli. Spatial transcriptomics was performed on diabetic human kidney samples and db/db eNOS-/- mice.
Results
In all human samples, SLN glomeruli showed greater glomerulosclerosis and mesangial intensity (a pathomics feature related to pixel density) than LLN. Both hypertension and aging showed more probability of glomerulosclerosis in SLN and LLN glomeruli vs those without conditions, whereas diabetes was linked to glomerular hypertrophy in both SLN and LLN. In mice, aging increased mesangial expansion and intensity in SLN, but not in LLN glomeruli. Hypertension and diabetes induced similar pathomic changes in SLN and LLN glomeruli compared to control mice. Spatial transcriptomics identified 742 differentially expressed genes (DEGs) in diabetic SLN glomeruli vs normal samples, whereas only 24 DEGs were found in LLN glomeruli. DEGs in SLN glomeruli were associated with endothelial injury and matrix regulation pathways. In db/db eNOS-/- mice, SLN showed more DEGs than LLN at week 10, an early stage of diabetic nephropathy, but similar DEGs in both types by week 18, a later stage.
Conclusion
Our AI-driven pathomics analysis revealed distinct morphological responses in SLN and LLN glomeruli across various diseases. Spatial transcriptomics highlighted molecular changes related to endothelial injury and matrix regulation preceding morphological alterations and further differentiated SLN and LLN responses.
Funding
- NIDDK Support