Skip to main content
Premium Trial:

Request an Annual Quote

AI-Assisted Proteomic Analysis Uncovers Biomarkers for Eye Aging, Disease

Human Eye Iris Girl

NEW YORK – New research suggests it is possible to tap into the protein and gene expression patterns in eye fluid cells to better understand eye aging and disease, including the effects of neurodegenerative, metabolic, or autoimmune conditions on the eyes.

In Cell on Thursday, researchers from Stanford University, Aarhus University, and elsewhere presented a proteomic clock approach to tracking aging, eye diseases, and conditions such as Parkinson's disease using proteomic profiling and artificial intelligence (AI) modeling of eye fluid samples.

Using almost 7,600 protein-specific aptamers, the researchers conducted aptamer-based proteomic assay profiling of 5,953 proteins in aqueous humor and vitreous samples collected from 120 individuals during eye surgeries. They also performed single-cell RNA sequencing on more than 82,000 individual cells from the same samples in an effort to narrow in on cell type-linked molecular markers of aging, health, and disease.

"Compared to the past, we can measure several thousand more proteins in the eyes of living humans," senior and corresponding author Vinit Mahajan, a surgeon and ophthalmology researcher at Stanford and the Veterans Affairs Palo Alto Health Care System, said in an email.

With the help of its "tracing expression of multiple protein origins" (TEMPO) approach, the research team determined which cells the proteins they found came from and then went on to model the liquid biopsy-based proteomics and single-cell transcriptomics data with AI, flagging markers related to everything from eye aging to maladies such as retinitis pigmentosa as well as retinal symptoms related to Parkinson's disease or diabetes.

"[F]or the first time, our TEMPO data analysis method tells us which cells inside the eye made the proteins," Mahajan said. "With this information, we can monitor the molecular health of individual retinal cells essential for sight."

Along with protein markers reflecting specific immune, vascular, or neuronal cell types, for example, the team described a proteomic clock based on cell type-specific protein patterns — a suite of molecular markers that made it possible to see faster-than-usual molecular aging alongside eye conditions such as proliferative diabetic retinopathy that are not typically linked to age.

"This protein clock is very different from other biological clocks where DNA in the blood is used but cannot predict the age of specific organs," Mahajan said. "We found strong evidence that different eye diseases accelerate aging of the eye [by] as much as 30 years, pointing to a potential need for 'anti-aging' therapies."

The team's data also highlighted Parkinson's disease-related retinal degradation features, while uncovering distinct sets of informative cell types during different stages of diabetic retinopathy.

More broadly, a similar approach may be applicable to fluid samples associated with other organ types, ranging from the brain or breast to kidney, lungs, joints, or cysts, the authors suggested, noting that the broad strategy used "has the potential to transform molecular diagnostics and prognostics while uncovering new cellular disease and aging mechanisms."

"By connecting high resolution protein data to specific cells, we can develop drugs and design better clinical trials with greater precision," Mahajan explained. "We expect researchers will apply this example of TEMPO in the eye to other human organs and diseases."