Skip to main content

Mount Sinai Team Developing Proteomic, Transcriptomic Tests for Aneurysms, Stroke


NEW YORK – Christopher Kellner, a neurosurgeon at New York City's Mount Sinai Hospital, is developing RNA and protein biomarker panels aimed at improving detection of aneurysms and the treatment of strokes.

In an initial pilot study, Kellner identified a panel of eight plasma proteins that could prove useful for detecting aneurysms in asymptomatic patients. He has applied for a patent covering these markers and is now applying for funding from the National Institutes of Health to follow up these findings with studies in larger cohorts.

Additionally, Kellner plans to launch a multicenter study backed by medical device firm Penumbra to explore whether transcriptomic and proteomic characterization of blood clots taken from stroke patients could help identify where in the body they originated, which he noted could inform treatment.

The aneurysm work originated with a hypothesis that because aneurysms are an inflammatory process, circulating RNA or protein markers linked to inflammation might be able to detect them in asymptomatic patients, Kellner said.

Aneurysms are relatively common, present in between 1 and 2 percent of the population, he noted. Of people with aneurysms, around 1 percent will suffer a rupture, and among patients who have a rupture, roughly half will die and roughly half of those who do survive will be severely impaired.

When detected, aneurysms can almost always be treated effectively, Kellner said, but most are found incidentally during imaging, following, for instance, an accident or during an investigation of unrelated symptoms.

He suggested that a blood test like the one he is working on could find use as a screening tool, initially, perhaps, in people with a family history of aneurysms and ultimately in the general population. Because aneurysms tend to develop slowly, such a test could potentially be given every five or so years, he said.

In the pilot study, Kellner used Olink's 92-protein inflammation panel to measure proteins in blood samples from 28 patients with unruptured aneurysms and 28 controls, identifying a set of eight proteins that distinguished between the two groups with an area under the curve of .97.

He now aims to follow up that work with a larger study that will help determine how well the signature performs across more diverse groups of patients and how specific it is for aneurysm-related inflammatory processes as opposed to other inflammatory processes occurring throughout the body.

"This is not by any means a final biomarker test," he said. "This is just a pilot project with a very small number of patients and a lot of bias going into it, so I think the next phase of research related to this will be to do a larger pilot with a much more controlled population."

He noted, for instance, that in the pilot some of the blood used was venous while some was arterial, which might have impacted the findings.

Kellner added that he plans to use mass spectrometry as opposed to the Olink platform for protein analysis in the next stage of the effort, which he said will enable a more unbiased look at the proteome that could pull in markers not measured by the Olink panel.

He also has data correlating the expression of specific proteins with aneurysm size, which suggests plasma protein panels could also be useful for following patients with known aneursyms, though he noted that more research is required to establish the feasibility of this approach.

Kellner is also moving ahead with work using RNA-seq and mass spec-based proteomics to analyze blood clots taken from stroke patients with the goal of developing biomarker panels that could tell clinicians where in the body a clot developed.

Around 30 percent of strokes are cryptogenic, meaning doctors are unable to determine where they originated, Kellner said. Knowledge of the site of origin is important, however, to treating stroke patients effectively.

"If you find out that it came from the heart, for instance, then you know that the patient might be having some intermittent irregular heartbeats that are allowing the clot to form there and they need to be on some anticoagulates to prevent that from happening," he said.

In recent years, technology for removing blood clots from stroke patients has improved significantly, and removal of these clots, Kellner said, improves patient outcomes. It also provides researchers with an opportunity to study these clots to determine whether they can find characteristics that indicate where in the body they came from.

"People have started looking at histology, taking the clot, staining it, looking at it under the microscope, and using that characterization to try to predict what the etiology of the stroke was," he said. "And we are learning a little bit and making some correlations, but it's not a diagnostic test by any means."

To refine these clot characterization efforts, Kellner and his colleagues have turned to what he called "clotomics," using RNA-seq and proteomics to characterize them. Over the last two years his lab has collected clots from 60 patients. Last week it sent them out for sequencing and mass spec analysis.

While data from these samples has yet to come back, Kellner said he has secured funding from Penumbra, which makes the catheters used in removing clots from stroke patients, to launch a multi-center, 1,000-patient study using mass spec-based proteomics and RNA-seq to profile blood clots with the goal of developing signatures that can help doctors identify the origins of cryptogenic clots.

"We're in the final stages of planning the [study] protocol and enlisting other sites," he said.