July 2012: Volume 38, Number 7
A Test for Imminent Heart Attack
How Close is the Science to Clinical Reality?
By Genna Rollins
Since the 1970s, a series of discoveries and innovations have steadily narrowed the time needed to diagnose myocardial infarction (MI). Today, the gold standard is a cardiac troponin (cTn) measurement exceeding the 99th percentile of a normal reference population, with rising or falling values within 6–9 hours of symptom onset key in discerning the condition. New high-sensitivity cTn assays now being implemented in many facilities have narrowed even further the window for detecting the change value of this analyte, thereby enabling still earlier diagnosis of MI. Despite these inroads, heart disease remains the leading cause of death in the U.S., claiming nearly 600,000 lives each year. Particularly troubling to clinicians is the not-infrequent situation in which a patient presents with chest pain and is cleared after a full evaluation, only to experience MI shortly thereafter. Researchers at Scripps Translational Science Institute in San Diego recently published research that one day might lead to a diagnostic test that identifies such patients on the verge of a heart attack. If expanded upon and validated in other studies, their findings could transform management of these individuals.
“We’ve known in recent years that cracks in arterial plaque that lead to heart attack don’t occur where there’s a lot of plaque, but actually where there’s a minor amount. That means there are a lot of people out there who are susceptible to having a heart attack. But we don’t have a way to identify them because a stress test only picks up a tight narrowing and blood tests like CKMB and ultra-sensitive troponin are markers of necrosis, meaning the heart attack has already occurred. So we do not do a good job of picking up this heart attack in progress, this precursor,” said Eric Topol, MD, director of the Scripps Translational Science Institute and chief academic officer of Scripps Health. “Our idea was to pick up on a finding published in 1999 that the endothelial cells from the artery that are being shed in advance of the heart attack by days up to a couple of weeks, could be detected in the blood. In 1999, there really weren’t the tools to say for certain if those cells were coming from the artery supplying the heart muscle, so we went back to that with a large National Institutes of Health grant and basically went after it.”
Characterizing Rare Cells
Topol and his colleagues at Scripps tested levels of circulating endothelial cells (CECs) in 50 patients with EKG-confirmed ST-elevation MI (STEMI), the classic massive heart attack that occurs after acute arterial plaque rupture (Sci Transl Med 2012;4:126ra33). These patients presented for emergency care at four regional medical centers and had blood samples taken in the cardiac catheterization lab via an arterial sheath but before catheter insertion for coronary angiography or angioplasty. The researchers recorded initial, but not subsequent, creatine kinase (CK-MB) and cTn results.
Using a method for isolating CECs developed by Veridex, a Johnson & Johnson company, the investigators found that in comparison to 44 normal controls, levels of CECs in the STEMI patients were significantly elevated—more than 400% higher—with median levels 19 cells/mL versus 3.8 cells/mL. A receiver operating characteristic curve based on logistic regression showed the CEC count to have an area under the curve of 0.95. The investigators used a 10-fold cross-validation process to evaluate the sensitivity and specificity of CECs in identifying STEMI, and found that with a threshold of 9 CECs/mL, the sensitivity was 90% and specificity 93%. A second measurement of CECs 2 months later in 24 of the healthy controls found their values essentially unchanged, suggesting that CECs remain stable over time in healthy individuals, according to the researchers.
As expected, STEMI patients had significantly elevated baseline levels of both cTn and CK-MB: 5.7 ng/mL and 27.9 ng/mL, respectively. However, statistical analyses revealed no correlation between these biomarkers and CEC counts, with a Spearman rank correlation p-value of 0.02 for CK-MB and 0.03 for cTn. These findings suggest that CEC counts appear to be independent indicators of arterial injury in patients with acute MI, according to the authors.
The CECs of STEMI patients also had quite a striking morphology in comparison to healthy controls, patients with documented peripheral artery disease, and patients with non-STEMI. CECs of controls generally were small and elongated with only an occasional two nuclei associated with one cell body. In contrast, those of STEMI patients were considerably larger and grossly misshapen (See Figure, below). They had large nuclei and often had multiple nuclei associated with each cell body region, possibly representing either multicellular clusters or multinuclear cells. Overall, about 25% of CECs from STEMI patients contained two or more nuclei, compared with 5–10% of controls.
“Certainly the big new finding from our study was that these cells were so sick and that no matter how we looked at it, every heart attack individual’s cells were so remarkably different than those of healthy individuals. If the controls had these rare cells at all they were the typical elliptical cells with one nucleus, whereas the heart attack individuals had these fat cells with multiple nuclei. They were distinctly different, and that had not been reported before,” said Topol.
Crucially, the authors believe this deranged morphology is associated with a particular phenotype. They reported that gene expression analysis revealed that in comparison to controls, STEMI patients had elevated endothelin and von Willebrand factor. The team is in the process of validating a “striking” gene expression profile, according to Topol. Should these efforts prove successful, a rapid molecular expression profile test could be in the offing.
Capturing Circulating Endothelial Cells
On the top are normal, healthy circulating endothelial cells (CEC). On the bottom are CECs from heart attack patients which appear abnormally large, misshapen, and with multiple nuclei, the latter marked with asterisks.
Image courtesy of Scripps Translational Science Institute.
Old Method, New Use
Prior to launching this study, the Scripps team worked with Peter Kuhn, PhD, an associate professor of cell biology at Scripps, to isolate CECs. However, his method “wasn’t something we could rev up and wasn’t commercially available as a platform,” Topol recalled. This led the team to Veridex, which already had a kit to measure CECs, an outgrowth of its primary interest in circulating tumor cells. “We’ve been looking at circulating endothelial cells as part of our oncology effort,” said Mark Connelly, PhD, scientific director of cellular research at Veridex. “Angiogenesis is an important part of tumor growth, so we’re very interested in and have been looking at circulating endothelial cells for a number of years because of their potential role in angiogenesis in cancer. The opportunity came up to collaborate with Scripps, and it seemed like a natural extension of our rare cell technology, which can be adapted for the exploration of many different kinds of rare cells.”
The Veridex method involves isolating enriched CECs from blood using immunomagnetic beads coated with CD146, a cell surface endothelial cell glycoprotein. After this step, the cells are immunostained for CD105, an endothelial cell marker, and CD45, a leukocyte marker, followed by imaging with fluorescence microscopy. “It’s pretty straightforward. We use small magnetic particles, about 200 nanometers in size, bind them to the target cells of interest—the antibodies on the surface of the magnetic particles—making the cells magnetic. Using magnets we enrich those cells, reduce the volume, stain the cells with specific antibodies, and stain for the cells we don’t want,” explained Connelly. “All of that is then presented to a computer-controlled fluorescent scanner that scans the cartridge and identifies and takes pictures of all the cells that are present. The whole process is automated, all the way up to the end where the user gets a selection of computer-identified images of the cells of interest. From this, the lab counts how many are there.”
Even though the process is an automated, scalable improvement over Kuhn’s method, it still takes 4–5 hours, far too long to be useful in the setting of patients presenting to the emergency department with suspected imminent heart attack. However research still in process involving a gene expression profile of impending STEMI could result in a whole blood test with a less-than-1-hour turnaround time, according to Topol.
The State of the Science
The authors’ methods and findings, published in Science Translational Medicine, have been hailed as a promising means of predicting imminent heart attack. But are they? Researchers not involved in the project expressed interest in this line of investigation, but cautioned that the science needs to be fleshed out considerably before it has any significance clinically.
“This is exciting to me as a researcher, but not as much as a clinical cardiologist because it’s far from ready for application in clinical practice. However, if the authors take their findings through all the appropriate steps of development, then they could end up with a new test that’s either an early marker of plaque rupture or plaque instability, or a diagnostic marker for MI in the absence of conventional laboratory or EKG findings,” said L. Kristin Newby, MD, MHS, professor of medicine at Duke University School of Medicine in Durham, NC and director of the Duke Clinical Research Institute’s biosignatures advanced biomarkers group. “That’s the Holy Grail we’re all trying for—a subclinical marker of vulnerable plaque or plaque that’s about to rupture and cause a heart attack.” Newby also served on the National Academy of Clinical Biochemistry’s (NACB) Laboratory Medicine Practice Guidelines (LMPG) Committee on Biomarkers of Acute Coronary Syndrome and Heart Failure.
The researchers are the first to agree that their findings are preliminary. However, they argue that the totality of their discoveries amount to a smoking gun, crucially implicating CECs in impending heart attack. “This is a seminal work proving that circulating endothelial cells are a biomarker of acute plaque rupture. How we get from here to clinical use still has to be determined. Obviously, more work needs to get done as this was a very small study looking at very pure phenotypes,” said lead author, Samir Damani, MD, PharmD, a cardiologist and founder of MDRevolution, a medical practice in San Diego that uses genomic medicine and mobile health technologies to individualize healthcare. “We validated that people with heart attack have high numbers of endothelial cells, that these cells are morphologically abnormal, and that they’re not correlated with muscle death markers. Given previous autopsy data showing that endothelial cell erosion occurs in the weeks to months preceding sudden death, the implications are big and include the fact that circulating endothelial cells could very well serve as a predictive biomarker for heart attack in the near future.”
That the authors describe an automated and timely—if not rapid—method for isolating CECs and detecting any deranged morphology, is not in question. However, the meaning of these findings is less obvious, according to Frank Kolodgie, PhD, associate director of the CVPath Institute in Gaithersburg, MD. “They are able to recognize dysmorphic endothelial cells; there’s no question about that. However, when we look at biomarkers in particular, troponin is one of the best ones. We know exactly where it’s coming from, exactly what it means. But when it comes to these dysmorphic endothelial cells, I think it’s more of a black box until we know more about their functionality, or why they’re being released. Is it something specific or is it non-specific because there’s an ongoing MI, and the bone marrow is now activated?”
The investigators argue that several of their findings support the CECs they isolated as having emanated from mature arterial endothelial cells as opposed to activated T-cells or mesenchymal hematopoietic precursors. In particular, they cited the fact that the CECs were CD45-negative, but also expressed CD31, a marker found on vascular endothelium but not on bone marrow-based mesenchymal cells.
The Next Phase of Discovery
Other researchers saw limitations with the study design and populations, and suggested that in order to advance their findings, the authors would need to deal with these issues in subsequent rounds of investigation. “My bottom line here is that the next study is going to be critically important to convince us that this needs to be looked at closely. We need to see how circulating endothelial cells perform in general populations or subsets that are not presenting with an MI, and we need to see how they compare to high-sensitivity troponin after accounting for confounding variables like renal function,” said Fred Apple, PhD, medical director of clinical laboratories at Hennepin County Medical Center and professor of laboratory medicine and pathology at the University of Minnesota School of Medicine, in Minneapolis. He suggested that at this point, hs-cTn is the marker to beat, not only in detecting MI, but also in predicting increased risk of cardiovascular disease-related death in a community population, an issue he recently examined (Clin Chem 2012;58:930–5). Apple served with Newby on the NACB LMPG Committee on Biomarkers of Acute Coronary Syndrome and Heart Failure and has published extensively on the science behind and analytics involving cardiac biomarkers.
Topol acknowledged the imperfect study protocol, which the investigators hope to address as best they can in follow-up research. “We would love to be able to get blood samples from people who have not had a heart attack but are about to. If you can find out how to do that, let me know!” he quipped. “We did the best we could under the circumstance, but it’s like a catch-22. There’s no way to find these people and if you knew they were going to have a heart attack and you didn’t do anything about it that would be considered ethically questionable. That’s why we’re stuck with this design, which is not this sanctimonious randomized clinical trial with placebo.”
A validation study of the CEC molecular expression profile is in process. The investigators plan to recruit 40–50 STEMI patients and analyze their CECs with a whole blood genetic test. Following that step, the goal is to recruit 200 individuals who present with chest pain in various emergency departments. “We plan for the test to be used in patients who are ruled out for heart attack but in whom the physicians have enough suspicion that they run this test. If it is positive then we plan to treat them empirically with potent antiplatelet therapy and assess their platelet function to make sure it’s suppressed,” explained Topol. “We’ll rely on historical data in comparing the event rate because doing otherwise would be very difficult ethically.”
Topol believes this step can be completed before the end of 2013. If all goes well and the subsequent rounds of research build on what’s already been published, he sees a commercially available, clinically actionable gene expression test becoming available shortly thereafter.
What About Imaging?
Even as the Scripps team pursues this line of investigation, the scientific literature is abuzz with reports of various imaging technologies that might detect vulnerable plaque. Numerous imaging methods have been proposed, and the Prospective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) trial is comparing functional and anatomic imaging tests to determine which does a better job of identifying patients who have coronary artery blockages.
However, Topol and other researchers see advantages to a robust blood-based method. “The noninvasive imaging techniques have never been correlated as an early marker of plaque rupture. They in general predict risk for future heart attack but they’re not actionable at a given point in time, although work from the PROMISE trial is still going on,” said Newby. “The obvious thing with imaging is it’s expensive, requires imaging equipment, and technical expertise. If one can develop a blood-based test that’s point-of-care, anybody can perform it. So it has the advantage of wide-spread accessibility, at a lower cost than specialized imaging.”
Topol echoed those points. “There’s a critical difference between a vulnerable plaque and a vulnerable patient. In order to study a vulnerable plaque you have to do invasive imaging. You can’t do that everyday, and you can’t do it with every person who presents with chest pain. So we need a non-invasive approach, and there isn’t one to tell us there’s a vulnerable patient who’s had a crack in his artery and is shedding cells into his blood stream,” he explained. “We think our approach has a lot of merit because with the molecular signature it could be like a fluid-phase biopsy of the artery that could be done in the emergency department.”
As the Scripps team continues to press the envelope scientifically, the research community awaits further support for this intriguing, if early, line of investigation. No one could be more interested in doing so than the authors themselves. “Clearly we’re still in the discovery phase with this technology and test, but I’m very excited about all the interesting clinical questions we can now ask and investigate using these findings,” said Connelly. “There’s a lot of good medicine and science we’ll be able to do going forward, and we’re very interested and hopeful that we’ll be able to answer some of these questions.”