Star role for bacteria in controlling flu pandemic?

from Nature Reviews Drug Discovery

David Bradley

Engineering microbes could be our best hope for creating enough antiviral drugs.

Alternatives to star anise are needed to allow enough Tamiflu to combat a potential flu pandemic to be produced.

As fears spread about a potential flu pandemic, Roche's antiviral drug Tamiflu (oseltamivir phosphate) has found itself thrust under the spotlight. With the current supply thought to cover just 2% of the world population, health officials and researchers are asking how Roche can create enough supplies to be stockpiled and help control an outbreak until a vaccine can be made. Of all the proposals suggested, it is ironic that the best solution to controlling a flu pandemic with an antiviral drug could lie within bacteria.

To make Tamiflu, you start with shikimic acid, which is naturally abundant in the oriental spice Illicium anisatum, or star anise. A complex multi-step chemical synthesis procedure then converts shikimic acid into the drug. Producing large amounts of Tamiflu not only takes months to complete, but is also hazardous. Some of the steps in the synthesis require careful handling and relatively mild reaction conditions, as they involve the use of potentially explosive azide chemistry.

The problem doesn't seem to be replicating this demanding synthetic process. For instance, Cipla, the Indian generics company, says it could produce a version of Tamiflu within months (authorized or unauthorized), despite Roche's claims that it would take a newcomer 2–3 years to start from scratch and produce the drug.

Cipla says it could create generic Tamiflu as it has experience of making the HIV drug AZT, which relies on similar chemistry. John Frost, a bioorganic chemist at Michigan State University, says any competent organic chemist can follow the literature and make Tamiflu in gram quantities. “But making a ton of drug and manipulating tons of azide [intermediate] is quite a different matter,” says Frost.

Making large-scale quantities of the drug, of course, relies on having enough starting material. But star anise is harvested only from March to May, and there aren't enough supplies to produce the number of doses needed worldwide. With no other agricultural sources, other than perhaps the leaves of the gingko biloba tree, the pharmaceutical industry needs to find an alternative sustainable supply.

Shikimic acid, however, is also a natural intermediate in the formation of microbial amino acids. Researchers could boost production of shikimic acid in strains of the bacterium Escherichia coli by ramping up enzyme activity and feeding the microbe a diet of carbohydrates.

John Frost is hoping to create large-scale quantities of shikimic acid for Tamiflu production. Michigan State University

Roche already uses engineered bacteria to boost shikimic acid production. “Today, around two thirds of the shikimic acid used for Tamiflu is gained from star anise and Roche also has access to the remaining shikimic acid through fermentation,” says company spokesperson Martina Rupp, “This gives us more flexibility, and in particular allows us to become more independent from events such as bad harvests.” Roche hopes to scale up its fermentation capacities over the next few years, but there is no firm timeline as to when this will happen.

In theory, there is no limitation on how much shikimic acid can be produced by fermentation. The bioproducts industry currently produces 750,000 tons of lysine from glucose by microbial fermentation for use in products such as animal feed. If needed, the same tanks could be used to produce shikimic acid from glucose. “Globally, there is ample fermentor-tank capacity that can be re-deployed to make shikimic acid,” says Frost, whose lab has developed an E. coli strain that produces shikimic acid.

There's another advantage to using fermentation. Microbes could be engineered to make variations on the Tamiflu starting material and so allow production of Tamiflu analogues with slightly different pharmacological properties that could potentially target emerging strains other than H5N1. Should a pandemic hit, it might be valuable to focus on accelerating such analogues through testing procedures to cope.

Frost, who consulted with Roche on how to grow his E. coli strain and purify the shikimic acid on a commercial scale, has started up a company called Draths Industries to help boost supplies of shikimic acid. Draths already has an agreement with one bioproducts firm to contract-manufacture 200 tons of shikimic acid by fermentation, but it could produce more, according to Frost.

“Draths Industries feels that there is enough capacity to access 1,000 tons of shikimic acid in 14 months,” says Frost. Roche is on record as claiming 0.13 grams of shikimic acid is needed per Tamiflu capsule. So, 1,000 tons should provide enough feedstock for more than 20 million capsules a month, or enough doses for 2 million people. This would fulfill Roche's current annual production but would be inadequate for its planned production of 300 million courses of Tamiflu in 2007.

Bradford Frank, of the University of Buffalo School of Medicine and Biomedical Sciences, is not confident that we could cope with a pandemic, even with microbial assistance. He suggests that the infection rate could be as high as 50% of the global population — 3.2 billion people. Half of those people would have inapparent or asymptomatic infections — as is often the case with viral diseases — but the other 1.6 billion people could become seriously ill. Jeff Levi at the Trust for America's Health adds, “It is correct that many are now saying that a severe pandemic, along the lines of 1918, would result in a 2% fatality rate among those who become ill.” Frank explains that even if only half of those needed treatment with Tamiflu, this would still require around 30 billion capsules.

Tamiflu, however, can be synthesized from other products in a microbial fermentation tank. Quinic acid, found in cinchona bark, is also in limited supply, but Frost and his team have already engineered a microbe to make quinic acid under fermentor-controlled conditions. Another starting material that is ripe for conversion into Tamiflu is aminoshikimic acid, says Frost. This starting material is easier to isolate from fermentor broth compared with shikimic acid, which could allow Tamiflu to be made in higher yield, he says. “Introducing a nitrogen atom into the starting material would eliminate the need to deal with azide and organoazide intermediates,” Frost adds.

The sad truth is that all of these efforts will prove to be academic if a pandemic strikes within months. Even if we wait a year before a breakout of the virus, which should give manufacturers enough time to make reasonable supplies of Tamiflu, our ability to cope with a putative bird flu pandemic shouldn't hinge on small molecules, says Frank. “Within the next few years, no matter what is done, there won't be enough Tamiflu to use as we like for everyone who needs it,” says Frank.

Fear of a global avian flu pandemic is putting a strain on antiviral drug supplies.

Frank suggests that the inhaled drug Relenza (zanamivir; GlaxoSmithKline), which is manufactured from a readily available starting material, N-acetyl-neuraminic-acid, should be included in pandemic contingency plans. It is, he says, “just as good as Tamiflu, but not too many people even have this on their radar screen.”