COURTESY OF VITO
The one-size-fits-all approach to therapies is quickly becoming a thing of the past, as drug developers begin stepping up to the challenge of personalized medicine, and regulatory agencies scramble to keep up. As the search for new biological indicators of disease heats up, researchers are looking far and wide for new markers of disease and of response to treatment.
Technologies that probe genes, proteins, the brain, and the body have become increasingly sensitive and efficient, expanding the search for new biomarkers beyond the components within blood and urine, and into diverse new areas, from breath to saliva to brain activity.
Scientists studying traumatic brain injury need a biomarker that is more timely than the proteins detected in postmortem exams. Others are examining people’s breath, which is thought to...
The Scientist talked to researchers on the hunt for unique types of biomarkers and simple tests to detect them. Here’s what we found:
Greet Schoeters, program manager, with Karolien Bloemen, researcher, Environmental Toxicology Unit, Flemish Institute for Technological Research, Mol, Belgium
Investigating the effects of environmental pollution on asthma and other respiratory diseases in children
In the past few years, researchers have started probing the breath of people with asthma, cancer, and other diseases to see whether exhaled breath contains distinct molecules. So far, the best marker available for asthma is nitric oxide, which appears elevated when the airway is inflamed. But other respiratory problems can also cause a nitric oxide boost. Researchers needed a more specific, earlier biomarker of asthma.
Schoeters’ group compared the protein content in the exhaled breath condensate of 40 children with asthma with that of 30 healthy controls.
Unlike protein measurements in blood or urine, analysis in breath is challenging because total protein concentrations in exhaled breath condensate are about 1 µg/ml, which is at or near the detection limits for most conventional protein analysis techniques. The group spent several years optimizing sample-collection and preparation methods and making sure their analysis was sufficiently sensitive.
The researchers had children breathe into a collection tube for 15 minutes while watching a distracting movie. Researchers collected about 1 ml of condensate, got rid of the water, and ended up with only 1 to 2 µl of concentrated protein. They then digested the proteins and attached them to beads so that the peptides could be separated and detected using liquid chromatography and mass spectrometry.
That method revealed a handful of proteins that differ between children with and without asthma (Clin Exp Allergy, 41:346-56, 2011), which they now plan to confirm in a larger sample.
It’s important to minimize the number of steps leading to sample analysis, because proteins are lost in each step, Bloemen notes. It’s also crucial to check samples for saliva contamination because saliva and breath contain different types and concentrations of proteins. Bloemen checked for a-amylase, an enzyme that is plentiful in saliva but not in breath, and simply excluded those samples. Commercial kits available for the enzyme are not that sensitive, Bloemen says, so the team used mass spectrometry.
Andrew Leuchter, professor of psychiatry and biobehavioral sciences, University of California, Los Angeles
Using quantitative electroencephalography (QEEG—a noninvasive measure of the brain’s electrical activity) to probe for biomarkers that track treatment responsiveness in depression
It can take eight or more weeks to determine whether a person will benefit from a particular antidepressant therapy. Evidence from several decades ago suggested a link between electrical patterns of brain activity and how well people fare in treatment. But that relationship didn’t show much predictive power until the mid-1990s, when Leuchter and his group developed Cordance, an algorithm patented by UCLA that uses QEEG data to predict treatment response with 75–85 percent accuracy.
Still, the group was using whole-head EEGs, which took over an hour to conduct and were not practical for patients coming in for treatment. They wanted a faster and easier version of the test.
Looking through their results, Leuchter’s team realized that their best data came from the front of the head, and in a narrow band of frequencies between 4 Hertz (Hz) and 12 Hz. “We said, ‘Okay, what if we just put a few electrodes on the forehead and we focus on this one frequency band—is there enough information there to predict treatment response?’” Leuchter recalls. From this simpler set of data they were able to generate what they dubbed the antidepressant treatment-response index (ATR). The QEEG data needed to calculate the ATR can be gathered in 10 to 15 minutes.
In a clinical trial examining EEG patterns in individuals right before they started an antidepressant, and then again one week later, the scientists were able to predict which participants were likely to have responded to treatment after 8 weeks with about 74 percent accuracy (Psychiatry Res, 169:124-31, 2009).
Although EEG has been used for more than a century, its results can be difficult to interpret. “We worked for a long time before we were certain of the meaning of the EEG signals we were seeing,” Leuchter says, adding that they were eventually able to relate changes in the signals to changes in blood flow in the brain.
It is easy for beginners to underestimate the amount of work it takes to get a potential diagnostic tool into the hands of clinicians for testing. “We naively thought at first that if we simply did the basic work and showed it was useful, that people would automatically say, ‘That’s great, let’s get this out to practice,’” he says. Instead, Leuchter learned that creating a practical tool involves not only scientific research, but technical developments; that process, he says, was made much easier through collaboration with a medical-device company.
Masanori Aikawa, assistant professor of medicine, in collaboration with cardiologist Peter Libby and systems biologist radiologist Ralph Weissleder, Harvard Medical School
Monitoring heart-attack predictors and disease precursors within atherosclerotic plaques
To follow the progression of atherosclerosis in rabbits fed a high-cholesterol diet, Aikawa’s group was using magnetic resonance imaging, which allowed them to see atherosclerotic plaques and monitor changes in their size during drug intervention. But they knew from other studies that the number of macrophages—immune cells important in the etiology of atherosclerosis—that are found on the plaques are better predictors of heart attacks than plaque size. His group wanted a way to examine these macrophages.
Borrowing an iron oxide nanoparticle used in brain and cancer imaging, the group targeted the nanoparticle to macrophages and used it to improve MRI contrast. In a set of in vivo and in vitro experiments, the researchers verified that the iron oxide levels reflect the quantities of macrophages present in plaques. Treatment with the cholesterol-lowering drug rosuvastatin (Crestor) lowered levels of macrophages, the group found (Circulation, 122:1707-15, 2010).
Even though the group was using well-studied nanoparticles, it took several years to conduct the necessary validation experiments to show that what they were seeing was real. Keep in mind that different nanoparticles have different kinetics, and you’ll have to figure out how long to wait to image after injections, Aikawa says. In addition, because MRI techniques are sensitive to movement, they produce more artifacts in images of arteries near the beating heart. Some peripheral arteries might be a better choice for imaging (at least for now), and they might be used to speculate about the risk of heart attacks, he adds.
Alexander Lin, principle investigator in radiology, Brigham and Women’s Hospital, Boston
Studying brain trauma in former professional athletes
Chronic traumatic encephalopathy, a degenerative brain disease caused by repeated blows to the head, is associated with memory problems, depression, and impulsive and erratic behavior, but often goes undiagnosed. The best available biomarker for CTE is a buildup of abnormal tau protein—which is also known as a marker for Alzheimer’s disease—in the brain. Tau is typically found by doing a postmortem exam. “You have to wait until you’re dead before you know whether or not you have a problem,” Lin says. The researchers needed an earlier biomarker of CTE.
Lin and his colleagues decided to try magnetic resonance spectroscopy (MRS), an imaging technique that uses a magnetic field and radio waves to detect chemicals in the body other than tau that might predict CTE. (Lin’s group is in the process of determining whether they can accurately measure tau using this method.)
The team employed a specialized MRS method called two-dimensional correlated spectroscopy (2D COSY) that measured up to 35 metabolites that go undetected using conventional MRS. These chemicals can potentially add sensitivity and may spur more accurate diagnosis, Lin says.
2D COSY has long been used to analyze proteins in tissue samples in vitro, but its application to brain disorders in vivo is a new twist, Lin says. “To bring it in vivo in human beings and try to diagnose disease with it—it’s definitely breaking the mold of traditional biomarker analysis.”
The team found changes in a handful of brain metabolites and signaling chemicals in five former athletes with suspected CTE compared to five age-matched controls. One of the metabolites, choline, signals the presence of damaged tissue.
Adjusting MRS to human in vivo studies isn’t easy, Lin says. “In typical chemistry work, you can analyze your sample for hours on end,” he notes. “That sample’s not going to complain.”
The key is to maximize comfort during the exam. For instance, Lin and colleagues had to use wide bore magnets (70 cm), which have larger openings for individuals to slide into for imaging, making the exam less claustrophobic for the large former athletes. “There are benefits to that—they stay still and the results are even better,” Lin says.
Lin’s group also worked to get the exam time down to 15 minutes by tweaking the parameters they measured and focusing on select areas of the brain. They practiced extensively on themselves and volunteers. With these improvements to the exam, Lin’s group is now validating the results in a larger cohort.
Correction (11th April): When originally posted, this article misidentified Alexander Lin as an assistant professor and a radiologist. He is at present a principle investigator and describes himself as a ‘radiology scientist.’ The Scientist regrets the errors.