Trawling the protein sea for disease biomarkers

By Haley Bridger, Broad Communications, June 29th, 2011
  • Researchers have developed a new approach to "catch" and identify
    protein biomarkers.
    Image courtesy of National Oceanic and Atmospheric Administration/
    Department of Commerce, photo by David Csepp, Auke Bay Lab

Diagnosing a heart attack quickly is critical: heart attacks kill more than 600,000 people in the United States every year, but if a doctor can accurately and rapidly diagnose a patient, the patient can immediately receive medication or catheter-based treatment. For years, researchers have looked for proteins or biochemicals in the blood that are released when heart cells are injured (biomarkers), but the ones that scientists have identified to date are only detectable several hours after the injury.

A team of researchers from the Broad Institute and Massachusetts General Hospital developed a new, systematic approach to detect and verify biomarkers and put this “pipeline” into practice, using the search for heart attack biomarkers as a test case. The researchers published their findings and methods in a Nature Biotechnology paper last week. Steve Carr, co-senior author of the paper, explains that what is most exciting about the research is the potential of applying this detection strategy to a variety of diseases and following up on the most promising protein candidates.

“A systematic pipeline that leverages the right samples and employs the right instrumentation and moves toward targeted, annotated measurements can deliver,” said Carr, director of the Broad Institute’s Proteomics Platform. “In many ways, I view this as pushing the reset button on the biomarker field.”

Proteins are several steps closer to influencing a person’s health than DNA or even RNA, making them tantalizing diagnostic tools. But pinpointing proteins is an incredibly complex task. Unlike DNA, which has fixed chemical properties and structure, proteins come in all shapes, sizes, and concentrations. Many key protein biomarkers are only detectable at the nanogram per milliliter level (a nanogram is one thousand millionth of a gram – a small paper clip weighs about a gram). A smattering of potential biomarkers of disease have been detected over the years, but very few have been approved for clinical use.

Robert Gerszten, co-senior author of the paper, compares the search for protein biomarkers to a fishing expedition. “To date, what most people have tried to do is create big fishing nets, but they’ve only sunk them ten feet into the water,” said Gerszten, a senior associate member at the Broad Institute and director of Clinical and Translational Research at the MGH Heart Center. “This improved platform is not yet a huge fishing net, but it’s starting to get some interesting organisms from down at the bottom of the sea.”

In order to find clear protein indicators of disease, researchers not only need precise instruments that can detect minute protein concentrations, but ideally need a distinct “before” and “after” state at which to measure these levels. A surgical technique called septal ablation offered the researchers a rare opportunity. In patients with the disease hypertrophic cardiomyopathy, a portion of the tissue in the heart’s main pumping chamber is abnormally thickened, which can lead to sudden cardiac death, especially in young athletes. In patients with severe forms of the disease, interventional cardiologists use septal ablation to relieve symptoms and improve the heart’s function. During the course of this surgical technique, the surgeon induces a planned heart attack.

By taking blood samples directly from the heart before, during, and after this surgery, the researchers could monitor changes in protein levels over time, using each patient as his or her own control. During the “discovery” phase of the project, the team analyzed blood from three patients and identified over 1,000 unique proteins, 100 of which changed dramatically following the induced heart attack. They then used a strategy known as AIMS (accurate inclusion mass screening) to create a shorter list of the most promising protein candidates found in blood samples taken from an artery in the patients’ thighs (peripheral blood).

“AIMS is a technology that is part of the bridge between discovery and clinical validation,” said Carr. “It attempts to address the following question: regardless of where discovery occurs, can we now detect those proteins in peripheral blood?”

However, AIMS is just the first step in the process of credentialing candidate biomarkers. The critical next step is to measure, with high precision and sensitivity, the candidate biomarker proteins in much larger numbers of patients. In the past this has been done using antibody reagents in immunoassays, which bind to specific molecules. The major drawback of this approach is the lack of suitable antibody reagents with which to construct immunoassays, especially for these new protein candidates. Creating new antibodies capable of being used in immunoassays is a long and expensive process. To overcome this limitation, the team at the Broad has developed mass spectrometry-based technologies for assaying the levels of proteins in patient blood (or tissues) without the need for immunoassay reagents.

“The multiple reaction monitoring-mass spectrometry (MS) technologies that we have developed enables us to configure assays for essentially any protein and measure it in blood at levels at or below 1 nanogram/mL with very high specificity,” Carr added. “Importantly, these assays can be highly multiplexed so that instead of measuring a single protein at a time we can measure 50-100 proteins at a time using these MS-based approaches.”

The team looked at the peak levels of a subset of proteins in the peripheral blood of patients who had suffered a spontaneous heart attack and found a similar pattern. Many of these proteins possessed promising characteristics, but given the small number of patients surveyed, it is too soon to draw meaningful conclusions about their usefulness in the clinic.

“The next step now is to take these proteins and assay them in large, heterogeneous populations and see if they add value on top of existing blood markers,” said Gerszten. Currently, the most commonly used heart attack biomarkers are proteins known as cardiac troponins, which appear several hours after a heart attack. Other markers, such as creatine kinase, are only elevated for a short period of time while others, such as lactate dehydrogenase, are not heart attack-specific and may result from a different underlying disease.

“We also want to explore more of the proteins that are on the candidate biomarker list,” said Carr. “We are exploring additional cardiovascular diseases, particularly ischemia [inadequate flow of blood to the heart]. We’re going to be looking in lots of additional patient cohorts over the next several years.”

Other Broad researchers who contributed to this work include Terri Addona, Hasmik Keshishian, D R Mani, Michael Burgess, Michael Gillette, and Karl Clauser.

Paper(s) cited: 
Addona TA et al. "A pipeline that integrates the discovery and verification of plasma protein biomarkers reveals candidate markers for cardiovascular disease," Nature Biotechnology. DOI: 10.1038/nbt.1899