May 12, 2008

Volume 86, Number 18

pp. 38-40

Heparin Undone

A consortium of scientists raced against the clock to identify the cause of adverse reactions

Jyllian Kemsley

ON NOV. 19, 2007, St. Louis Children's Hospital noticed the first of several acute allergic reactions in children undergoing dialysis. The hospital reported the reactions to the Missouri Department of Health & Senior Services, which in turn notified the Centers for Disease Control & Prevention on Jan. 7. CDC soon amassed reports of roughly 50 reactions in adult dialysis patients across at least six states. It alerted the Food & Drug Administration to the situation on Jan. 9.

Meanwhile, at the end of December 2007, Baxter International observed an increase in reports of reactions to one of its products—heparin, a common anticoagulant often given to dialysis and heart surgery patients. The reactions included a sudden drop in blood pressure and other symptoms of anaphylaxis—flushing, difficulty breathing, nausea, abdominal pain, and fainting. Baxter investigated the manufacturing and quality-control records for the drug lots implicated in the reactions; all raw materials and final products met testing specifications. On Jan. 11, Baxter contacted FDA. On Jan. 17, the company began to recall products.

Once Baxter announced the recall, the pressure was on both Baxter and FDA to figure out what was going on—and quickly. People weren't just reacting, they were dying: From October 2007 through February, FDA received reports of 66 deaths in which patients had received heparin and showed an allergic reaction, compared with a total of three for all of 2006. Although Baxter had recalled lots tied to adverse events, no one knew exactly what was wrong. Until scientists could pinpoint the problem, no one was certain that the rest of the heparin supply was safe.

During the intense weeks that followed, researchers raced against the clock to figure out what the problem was and how to ensure the safety of a drug critically important to health care worldwide. That effort culminated on April 23 with the publication of two peer-reviewed papers: one laying out the analytical methods used to identify oversulfated chondroitin sulfate (OSCS), a contaminant in heparin derived from animal cartilage (Nat. Biotechnol., DOI: 10.1038/nbt1407), and the other demonstrating a causal link between OSCS and the symptoms seen in patients (N. Engl. J. Med., DOI: 10.1056/NEJMoa0803200).

Notably, Baxter scientists were not authors on the papers even though Baxter and FDA say they worked closely together on the investigation. Baxter also declined to have its scientists interviewed for this story, citing their current workload and looming litigation.

The drug at the center of all the scrutiny is heparin, a variably sulfated glycosaminoglycan polysaccharide composed of alternating D-glucosamine and uronic acid residues. Heparin has multiple functions physiologically; its clinical role is to inhibit blood coagulation during dialysis and other procedures by binding to antithrombin III (AT-III). Upon binding to heparin, AT-III undergoes a conformational change that allows it to bind and inhibit either thrombin, which plays a critical role in blood clot formation, or factor Xa, which produces thrombin.

At first, FDA did not look at the drug itself. Rather, it began its investigation into the cause of the patients' reactions by looking at the clinical settings of the adverse events, suspecting that perhaps the reactions were due to how the drug was administered. But reactions were not confined to one procedure or clinical setting, says Moheb Nasr, director of the Office of New Drug Quality Assessment in FDA's Center for Drug Evaluation & Research. FDA concluded that the problem was not how clinicians used the drug and so turned to examining the drug material itself.

FDA next did a risk assessment to evaluate what was most likely to cause a quality problem that would lead to reactions. At the top of the list was the possibility that there could be something in the heparin solution that shouldn't be there: Perhaps an impurity was copurified with heparin or was generated as a by-product of processing or from degradation in storage. Alternatively, a contaminant could be present—one that was put there accidentally or deliberately.

But again, the raw materials and finished lots had passed standard regulatory assays mandated by U.S. Pharmacopeia (USP), the pharmaceutical standards-setting organization in the U.S. FDA scientists realized they needed to look in different directions, Nasr says, and explore more specific bioassay methods, advanced separation methodologies, and detailed structural analyses.

THEY BEGAN with a bioassay, working with researchers at Washington University, St. Louis, to expose heparin to heparinases. Heparinases cleave the 1,4-linkages between uronic acid and D-glucosamine and should hydrolyze pure heparin mostly into disaccharides, but the contaminated lots didn't entirely succumb to the enzymes. "Obviously, that was completely unexpected," Nasr says. The test results made it clear that an unknown substance was present.

FDA also evaluated chromatography as a means to separate components in a heparin solution. Chromatographic separation of heparin is difficult, Nasr notes, because heparin itself is a complex mixture with a molecular weight distribution from 5,000 g/mol up to 15,000 g/mol. "You don't get well-defined chromatographic peaks using traditional chromatography," Nasr says. At the same time, gel electrophoresis does not allow sufficient separation of small amounts of something like a protein contaminant, he adds.

FDA THEREFORE turned to capillary electrophoresis (CE) and found not one but two peaks outside of heparin. One was familiar???a peak indicative of dermatan sulfate (DS), another glycosaminoglycan that can copurify with heparin. DS is composed of alternating N-acetylgalactosamine and iduronic acid residues. DS has been known to turn up at low levels in heparin, but the amounts in the tested samples were higher than expected. Nevertheless, "we have enough pharmacologic knowledge" about dermatan sulfate "to know that it most likely was not the root cause of the observed reactions" in patients, Nasr says. Attention thus turned to the second chromatographic peak, which elutes just before heparin.

On Feb. 29, Nasr called Ram Sasisekharan, a professor of biological engineering at Massachusetts Institute of Technology, to ask for Sasisekharan's help in doing a detailed structural analysis of the unknown species. Sasisekharan in turn solicited the expertise of other researchers at MIT along with scientists at the G. Ronzoni Institute for Chemical & Biochemical Research, in Italy; Momenta Pharmaceuticals; and Rensselaer Polytechnic Institute (RPI).

One critical element of the work, Sasisekharan notes, was that all the researchers worked with blinded samples—they didn't know which were control or suspect lots—that were provided by FDA. This approach kept the focus on the data, Sasisekharan says, rather than allowing scientists' biases to creep in. He also developed a strategy to tackle the problem systematically: To avoid unnecessary duplication but to ensure reproducibility, not all of the labs did the same tests, but the protocol ensured that two groups independently confirmed each result. The labs worked separately to test the samples and analyze the data, then came together to determine whether they were all reaching the same conclusions.

The groups used a combination of methods, including one-dimensional nuclear magnetic resonance (NMR), two-dimensional NMR, mass spectrometry, and ion-pair reverse-phase high-performance liquid chromatography. The first thing they noticed was an unusual N-acetyl signal in 1-D proton NMR that is not observed for heparin or DS. Carbon-13 NMR also showed unusual signals indicating a highly O-substituted N-acetylgalactosamine residue—signals again distinct from those for heparin and DS.

In 2-D NMR, plotting ¹H versus ¹³C signals yielded 10 signals observed in contaminated lots that were not present in pure heparin. The scientists obtained even more detail through correlation spectroscopy and rotating-frame nuclear Overhauser effect spectroscopy. The growing consensus was that the contaminant was a polymer of N-acetylgalactosamine β-linked to glucuronic acid.

The researchers tried and discarded several hypotheses as to the identity of the contaminant. In the end, the data converged on chondroitin sulfate (CS), which like DS is an alternating chain of N-acetylgalactosamine linked to glucuronic acid. CS is isolated from cartilage and is commonly used in the U.S. as a dietary supplement to treat arthritis. But the material in heparin was not typical CS—it had four sulfate groups per disaccharide, while natural CS generally has three or fewer.

Heparin is currently isolated from the mucosal lining of pig intestines. Because the raw material for contaminated lots was traced to China, Sasisekharan wondered whether the contaminant was actually an impurity resulting from the "blue ear" virus, or porcine reproductive and respiratory syndrome, which is ravaging Chinese pigs. Such viruses are known to trigger the secretion of highly sulfated chondroitin E, perhaps leading to the material being copurified with heparin from pig intestines. But that couldn't account for the quantities seen in the contaminated drug: The lots tested by Sasisekharan and colleagues contained 20 to 30% contaminant, and FDA says some lots contained as much as 50%.

The eventual conclusion was that the CS had to have been a natural, commercially available product that was chemically altered to incorporate additional sulfate groups. Such material was familiar to one member of the team investigating the contaminated heparin: More than a decade ago, Robert J. Linhardt, a chemistry professor at RPI, had worked with CS as part of an effort to develop new anticoagulants. In particular, Linhardt and colleagues had sulfated the polysaccharide and found that the extra sulfate groups enhanced its anticoagulant activity. It was surreal and depressing, Linhardt says, to realize that his former work may have been used to contaminate the drug supply.

On the basis of the team's experiments, FDA announced on March 14 that all heparin imported into the U.S. must be tested by NMR and CE. On March 17, Sasisekharan presented the final analytical results to FDA, and on March 19, the agency announced to the public that the contaminant was OSCS. On March 21, the researchers submitted their work to Nature Biotechnology for peer review. Following the recalls and the implementation of NMR and CE, as of press time, FDA had reported no deaths of patients receiving heparin in the months of March or April.

But there was still an outstanding question: Was OSCS, in fact, capable of causing the hypotension and anaphylaxis reactions seen in patients? Again in collaboration with FDA, Sasisekharan recruited a second team of researchers to address this question, turning again to other researchers at MIT along with scientists at Momenta, and adding researchers from Harvard Medical School and Virginia-Maryland Regional College of Veterinary Medicine, Virginia Polytechnic Institute & State University.

To explain the hypotension symptoms, the second group focused on the contact (or kininin) pathway of coagulation. In previous research, K. Frank Austen, an immunology professor at Harvard Medical School, had demonstrated that chondroitin sulfate isolated from squid seemed to activate that pathway.

This pathway begins with the enzyme factor XII, which, when it comes in contact with a negatively charged surface such as damaged tissue, triggers production of the enzyme kallikrein. Kallikrein in turn liberates bradykinin from a protein precursor; bradykinin is a potent vasodilator—perhaps the source of the hypotension experienced by patients. Again working with blinded samples, the researchers indeed found that contaminated heparin or pure OSCS activated kallikrein activity in human plasma, whereas control samples of pure heparin or naturally occurring (not oversulfated) CS did not.

TO SHED LIGHT on the anaphylactic reactions, the researchers turned to the anaphylaxis-inducing complement pathway of the immune system. When triggered by an event such as antigen-antibody binding, the complement pathway sets off a cascade that eventually involves several proteins working together to break open the cell membranes of pathogens.

Produced along the way are the proteins C₃a and C₅a, which are known as anaphylatoxins. They can trigger the release of substances such as histamine from mast cells or basophils; a strong release can lead to the symptoms of an allergic reaction. Once again, contaminated heparin or pure OSCS generated C₃a and C₅a in human plasma, while control samples of heparin or naturally occurring CS did not.

The plasma tests were promising, but the researchers still needed to prove that OSCS generated anaphylactic reactions in vivo, necessitating animal tests. Lab tests for OSCS-induced kallikrein activity in mouse, rat, guinea pig, rabbit, cat, dog, and horse plasma were negative. Ironically, it was pig plasma that turned up positive.

The scientists got the serum results on April 10. Sasisekharan then called Subbiah Elankumaran, a professor of biomedical sciences at Virginia-Maryland Regional College of Veterinary Medicine, to ask if the school could do the necessary studies in pigs to prove causality. Elankumaran and colleagues Kevin D. Pelzer and Nammalwar Sriranganathan worked into the night to write the protocol for the study, which was approved the next day by the school's Institutional Animal Care & Use Committee. David Moore, associate vice president for research compliance, signed off at 2 AM while traveling in India. Starting on April 12, the veterinary team dosed pigs with contaminated heparin, pure OSCS, control heparin, or naturally occurring CS. Two of the six pigs dosed with contaminated heparin and all of the pigs treated with pure OSCS experienced drops in blood pressure along with an increased heart rate; one had difficulty breathing. None of the pigs that received control heparin or naturally occurring CS had any adverse reactions.

Blood tests of the pigs showed those that had received contaminated heparin had increased kallikrein activity, even when they had no change in blood pressure or heart rate. This mimics the variation of reactions seen in humans, and the researchers propose that the variability may be due to individual variations in how bradykinin is regulated.

Sasisekharan presented these results to FDA on April 16. On April 17 and 18, FDA held closed-door meetings with international drug regulatory agencies to discuss heparin test results. By that time, 10 countries other than the U.S. had reported finding OSCS in heparin through the use of NMR and CE. Even so, Germany is so far the only other country to report an increase in reactions to the drug. Meanwhile, the second team wrote up the results and submitted the paper to the New England Journal of Medicine (NEJM) on April 21, as Chinese officials and FDA held competing press conferences to discuss the issue of causality (C&EN, April 28, page 14). NEJM published the article a mere two days later, on April 23, the same day that Nature Biotechnology published the chemical analysis. Despite the quick publication timelines, both papers underwent the same critical review process as typical articles, say representatives of their respective publications.

On April 29, in testimony before the House Energy & Commerce Subcommittee on Oversight & Investigations, Baxter Chief Executive Officer Robert L. Parkinson Jr. said that the results reported in NEJM confirmed the company's animal tests.

Research into contaminated heparin is ongoing. Elankumaran and colleagues are planning additional studies in pigs to evaluate the effect of different doses and route of administration to see if they can elucidate the variability of reactions. FDA is investigating whether OSCS was added to heparin accidentally or deliberately.

And USP and the European Pharmacopeia are working on new testing protocols to ensure future safety of the drug. Developing such tests is a difficult task, says Ganesh Venkataraman, chief scientific officer of Momenta Pharmaceuticals, which participated in OSCS analysis and is also developing two heparin-related products. "You not only have to address today's contaminant but future contaminants," he says, knowing that millions of patients worldwide are depending on the success of the endeavor.

Chemical & Engineering News

ISSN 0009-2347

Copyright © 2011 American Chemical Society