Novel Blood Thinners, Antibiotics Clued by Bacteria

It is rare that basic scientists can illuminate two vastly different medical issues at once. Yet two new studies show unexpected similarities between the human protein target of the widely used blood thinner warfarin and a bacterial molecule crucial for the growth of the tuberculosis bacterium.

The findings are fueling a larger collaboration to search for new compounds in two categories: antibiotics against tuberculosis and other bacteria, as well as anticoagulants to prevent blood clots in people with atrial fibrillation, artificial heart valves, deep venous thrombosis and pulmonary embolism.

The unlikely connection involves the bacterial version of a human enzyme called VKOR for short. In people and other mammals, VKOR works in the liver by recycling vitamin K to make clotting factors that are secreted into the blood, where they amass at sites of vascular injury. Warfarin thwarts VKOR’s enzymatic manipulations, leaving clotting factors adrift.

“We’d have to eat a lot of [vitamin K–rich] spinach, indeed, if we didn’t have VKOR,” said Bruce Furie, HMS professor of medicine and chief of the Division of Hemostasis and Thrombosis at Beth Israel Deaconess Medical Center. The two new studies, which focus on bacterial VKOR, are intriguing because the human protein is a major therapeutic target, said Furie, who was not a coauthor but is collaborating on the follow-up drugscreening studies.

Interest is in the eye of the beholder. Tuberculosis microbiologist Eric Rubin, HSPH professor of immunology and infectious diseases, values the work for revealing a potential new antibiotic target, bacterial VKOR. New antibiotics will be necessary to combat the global rise in drugresistant TB, said Rubin, a coauthor on one of the papers and continuing collaborator. “Warfarin isn’t a drug you can use to treat TB, but there’s a huge advantage in starting a project where so much is already known.”

The origins of warfarin date back to a bleeding problem affecting dairy cattle in Wisconsin caused by spoiled sweet red clover hay. In the midst of the Great Depression of the 1930s, researchers isolated a compound that inhibited the action of vitamin K, determined its atomic structure, and synthesized more than 100 analogues.

One particularly potent compound, later named warfarin, was first used as a rat poison (and still is). It was famously used to treat President Eisenhower after he had a heart attack. It has remained a clinically useful anticoagulant for the past 50 years. Last year, warfarin ranked as the 16th most prescribed drug, with more than 35 million prescriptions, according to IMS Health.

The recent studies began with an esoteric question posed in the lab of Jon Beckwith, the American Cancer Society professor of microbiology and molecular genetics at HMS. Bacterial toxins wreak havoc in the body in part because they are locked in peak fighting form by disulfide bonds. Twenty years ago, Beckwith’s group discovered two major components of one disulfide bond–making pathway.

More recently in the Beckwith lab, Rachel Dutton, a graduate student, and Dana Boyd, a lecturer in microbiology and molecular genetics, asked if there were other ways for bacteria to make disulfide bonds. Using bioinformatics and benchwork, they found that TB and other bacteria relied on VKOR. And, they reported, bacterial VKOR was strangely similar to the human protein required for blood clotting in people.

Then Dutton asked, Does warfarin inhibit VKOR activity in bacteria too? Again, the answer was yes (at very high doses), according to the lab’s paper in the Jan. 5 Proceedings of the National Academy of Sciences.

In search of a detailed structure of VKOR to gain insight into its molecular interactions, they turned to Howard Hughes investigator Tom Rapoport, HMS professor of cell biology.

“They were not optimistic,” Dutton said, but Rapoport and postdoctoral fellow Weikai Li were nonetheless interested. VKOR is a membrane protein, a wet and greasy mishmash of oil- and waterrepellant parts that are notoriously difficult to analyze at the atomic level. Or as Li puts it, “50,000 water soluble proteins have a known structure; fewer than 200 membrane protein structures have been solved.”

In a well-reasoned technical feat, co–first authors Li and Sol Schulman, an MD–PhD student, grew an array of crystallized molecules for X-ray measurement of the atomic angles and features. The structure is published in the Jan. 28 Nature.

Meanwhile, Dutton and Markus Eser, a postdoc in the Beckwith lab, have spent the past year developing a screening assay for new anticoagulants and antibiotics. It turns blue when disulfide- bond formation is inhibited by candidate molecules, an indication that VKOR is no longer active. The follow-up translational research is being funded by grants from Harvard Catalyst and stimulus dollars from the federal Recovery Act.

For more information, students may contact Tom Rapoport at tom_rapoport@hms.harvard.edu or Jon Beckwith at jbeckwith@hms.harvard.edu.

Conflict Disclosure: The authors declare no conflicts of interest.

Funding Sources: For the Beckwith study—the National Institute of General Medical Sciences, Burroughs Wellcome Fund, National Institutes of Health, and American Cancer Society. For the Rapoport study—the National Institutes of Health; National Institute of General Medical Sciences; Charles King Trust; National Heart, Lung, and Blood Institute; and Howard Hughes Medical Institute.