Diabetes Can Age the Eye by 30 Years, Finds Protein Study

A team of scientists at Stanford University has taken a look inside the human eye with a new level of detail by using cutting-edge molecular analysis.

The study, published in the journal Cell, examined drops of liquid taken from the eye, revealing over 6,000 proteins. The data was analyzed with artificial intelligence (AI) to create a “proteomic clock” of the eye that allowed researchers to determine how the eye was aged by different diseases, like diabetic retinopathy and uveitis. The work even showed that the eye was degenerated by Parkinson’s disease – a link not previously identified that could improve diagnosis.

Proteomics – the analysis of proteins, molecules that action nearly all of the biological processes that happen within our body – has been hugely accelerated by molecular techniques that enable scientists to study these tissues at the level of the single cell. But these studies require samples of our cells to work. Certain cell populations of our body, like the tissue of the brain and eye, can’t be biopsied in this way without catastrophic damage. Enter the “liquid biopsy”.

A look through liquid

During cataract surgery, a few drops of topical anesthetic render the eye numb. To Stanford University Professor Vinit Mahajan, this was the perfect opportunity to conduct a liquid biopsy of the eye. His research team extracted fluid from 120 patients that would “otherwise have been thrown away” during the surgery, he tells Technology Networks. The process is painless and fast, he says, but using a new technique can reveal of world of information about the eye.

The sampled liquid came from two sources – the aqueous and vitreous humors. These are the liquids found in the front part of the eye, and between the lens and retina, respectively.

Liquid biopsies like these have been possible in previous studies but to a limited resolution. Mahajan’s team developed a technique called TEMPO (Tracing Expression of Multiple Protein Origins), which powered up the biopsy analysis to an unprecedented level. By matching the proteins identified using TEMPO to transcripts – chunks of RNA that are converted to proteins – recorded in a previous analysis, the team was able to record 5,953 proteins expressed by all 57 types of cells in the eye. This depth was a surprise to Mahajan – the proteins, floating in the eye’s humors, are like “needles in haystacks”, he says.

This is a significant improvement over previous analyses of the eye – a 2016 study that reviewed 90 years’ worth of eye research found only 4,000 proteins in total. “Today's technology is so much better,” says Mahajan, “especially with being able to measure so many more proteins in such a small volume.”

Predicting eye age

A subset of these proteins was then used to train an AI machine-learning model. By teaching the model how the level of these proteins varied with age, the model gained the ability to predict the age of healthy eyes. The study then analyzed how this biological age was accelerated in the eyes of people with diseases like diabetic retinopathy. “We found that diseases like diabetes accelerated the biological age of the eye - by as much as 30 years. Our eye clock is different from other types of clocks for other parts of the body which appear to age at different rates,” says Mahajan.

TEMPO was also able to detect abnormal proteins linked to Parkinson’s disease. Normally, these proteins are found after death. “These molecular markers could open up new ways to diagnose and follow patients with Parkinson’s,” Mahajan adds.

Perhaps most excitingly, the research may offer improved routes to achieving personalized medical interventions that target unique molecular changes happening in patients’ bodies. “We can now measure protein molecules for specific cells in living patients without ever removing the tissues themselves. Since most drugs target proteins, we can now be very precise in selecting drugs that match the proteins active in a person’s eye,” says Mahajan. “We can identify new drug targets and design the right drugs. We think our methods can be applied to other organs like the brain, lungs, liver and kidney and diseases like cancer.”

Reference: Wolf J, Rasmussen DK, Sun YJ, et al. Liquid-biopsy proteomics combined with AI identifies cellular drivers of eye aging and disease in vivo. Cell. 2023;186:1–17. doi: 10.1016/j.cell.2023.09.012