RamanOmics Maps the Molecular Fingerprint of Senescent Cells in Aging Tissue
A new multimodal platform identifies senescent cells without labels, linking biochemical lipid signatures to transcriptional aging programs across tissues.
Summary
Scientists at Harvard and MIT developed RamanOmics, a label-free imaging framework that combines Raman spectroscopy with single-nucleus RNA sequencing and spatial transcriptomics to profile senescent cells in aging tissue. Applied to young and old mouse lung and skin, the platform revealed that senescence looks different depending on the tissue — lung senescent cells are dominated by extracellular matrix remodeling and TGF-β signaling, while skin senescence involves epidermal differentiation genes. Crucially, a conserved lipid-linked Raman signal at 1,131–1,135 cm⁻¹ was found across all tissues, marking p21-positive senescent cells. A machine learning barcode built from this multimodal data can identify senescent cells nondestructively in intact tissue. In wound-healing models, the same epidermal genes and lipid signatures reappear in senescent cells, hinting at a repair role. This tool could transform how researchers study and target cellular senescence.
Detailed Summary
Cellular senescence — the state where cells stop dividing but resist dying — is a central driver of aging and age-related disease. Identifying senescent cells in living tissue has historically required molecular labels that destroy the tissue, limiting the ability to study senescence dynamically or spatially. A new platform called RamanOmics, developed by researchers at Massachusetts General Hospital, Harvard Medical School, and MIT, aims to solve that problem.
RamanOmics integrates label-free hyperspectral Raman imaging with single-nucleus RNA sequencing and spatial transcriptomics. Raman spectroscopy detects the vibrational frequencies of molecules in cells without staining or destroying the sample. By layering this biochemical readout onto transcriptomic data at single-cell spatial resolution, the platform links what a cell looks like chemically to what genes it is expressing — and where in the tissue it sits.
Applied to young and old mouse lung and skin, the framework uncovered tissue-specific senescence programs. Lung senescent cells were enriched for extracellular matrix remodeling pathways and TGF-β signaling, while skin senescent cells were dominated by epidermal differentiation genes including Krt10, Lor, and Sbsn. Despite these tissue-specific differences, a conserved lipid-associated Raman signature at 1,131–1,135 cm⁻¹ emerged as a universal marker of p21-positive senescent cells across both tissues.
Using machine learning, the team built a multimodal molecular barcode capable of identifying senescent cells nondestructively in situ — a potential step toward real-time senescence mapping in intact tissue. In a mouse wound-healing model, the same epidermal differentiation genes and lipid Raman signatures reappeared in senescent cells, suggesting senescence plays an active and context-dependent role in tissue repair rather than being purely degenerative.
The clinical implications are significant. If this platform can be adapted for human tissue, it could accelerate senolytic drug development, improve biomarker discovery for aging biology, and enable direct testing of anti-aging interventions at the tissue level. Limitations include the animal model focus and abstract-only availability of the full data.
Key Findings
- A conserved lipid Raman signal at 1,131–1,135 cm⁻¹ marks p21+ senescent cells across lung and skin tissue.
- Lung and skin senescent cells have distinct transcriptional programs — TGF-β versus epidermal differentiation genes.
- Machine learning built a multimodal barcode enabling nondestructive senescence identification in intact tissue.
- In wound healing, senescent cells reactivate epidermal differentiation genes, suggesting a repair function.
- RamanOmics integrates Raman imaging, single-nucleus RNA-seq, and spatial transcriptomics at single-cell resolution.
Methodology
RamanOmics was applied to young and aged mouse lung and skin tissue, combining label-free hyperspectral Raman imaging with single-nucleus RNA sequencing and spatial transcriptomics. A mouse wound-healing model was also used to examine senescence dynamics during tissue repair. Machine learning was applied to multimodal data to generate a senescence identification barcode.
Study Limitations
This summary is based on the abstract only, as the full paper is not open access. All experiments were conducted in mice, and translation to human tissue remains to be demonstrated. The clinical utility of the Raman senescence barcode in human disease settings has not yet been established.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: