Tissue morphology predicts telomere shortening in human tissues

Area of Science:

• Computational biology
• Digital pathology
• Genomics

Background:

• Telomere length is a biomarker of cellular aging and genomic instability.
• Traditional methods for telomere length measurement are often laborious and not scalable for large cohorts.
• Tissue morphology contains rich information about cellular state and biological processes.

Purpose of the Study:

• To develop and validate a deep learning framework (TLPath) for predicting bulk-tissue telomere length from routine histopathology images.
• To investigate the relationship between tissue morphology, chronological age, and telomere length.
• To explore the utility of TLPath in identifying telomere shortening in disease states.

Main Methods:

• Utilized a deep learning framework (TLPath) trained on over 5,000 whole-slide images from 919 individuals across 18 organs.
• Extracted morphological features from H&E stained histopathology images.
• Correlated predicted telomere length with chronological age and measured telomere length in 11 tissues.
• Interpreted model findings to identify senescence markers associated with telomere shortening.
• Applied TLPath to GTEx biopsies to assess telomere length in individuals with diabetes.

Main Results:

• TLPath accurately predicts telomere length (correlation r = 0.51) in 11 tissues, outperforming chronological age.
• Morphological features extracted by TLPath naturally separate individuals by age group, indicating age-related architectural tissue changes.
• TLPath identified young tissues with short telomeres and old tissues with preserved telomeres based on morphology.
• Model interpretation highlighted senescence markers (e.g., increased nuclear-to-cytoplasmic ratio) linked to telomere shortening.
• Shortened telomeres were detected in multiple tissues of individuals with type 1 and type 2 diabetes.

Conclusions:

• Histopathology image-based tissue morphology can predict bulk-tissue telomere length.
• TLPath offers a scalable approach for large-scale telomere biology studies using routine H&E images.
• Morphological analysis of tissue architecture provides insights into aging and disease-related telomere dynamics.
• This framework enables the identification of individuals with atypical telomere length for their age and the assessment of telomere status in disease contexts like diabetes.