Vaccine Manufacturing Reborn

This article features an additional bonus interview with Matthew Stober, global head of technical operations at Novartis.

Not so very long ago, vaccine development and manufacturing was something of a backwater in pharmaceutical R&D and manufacturing programs. Now, suddenly it's at the forefront. After all, Wyeth's (Madison, NJ) vaccine program was one of the attractions making it an acquisition target for Pfizer (New York). What's more, several analyst reports place vaccines as the highest growth sector of biopharmaceutical programs. What's happened lately to catapult vaccines into the spotlight?

(PHOTOS: PAUL TEARIE, STOCKBYTE, GETTY IMAGES SAM SARKIS, PHOTOGRAPHER'S CHOICE, GETTY IMAGES)

A few things. The outbreak in recent years of two widespread and potentially fatal strains of flu elevated influenza from medical inconvenience to public-health emergency. Second, well-publicized delays in producing vaccine against the H1N1 pandemic strain in 2009 highlighted how ineffectual is the traditional manufacturing paradigm in responding to sudden large-scale outbreaks of influenza. Both events placed new pressures on manufacturers to find novel, more responsive, more scalable methods for vaccine production.

Indeed just this past August, the President's Council of Advisors on Science and Technology (PCAST) issued a report to the president on ways to reengineer the entire influenza vaccine manufacturing enterprise to meet the challenges of a pandemic influenza outbreak. At the same time, the Department of Health and Human Services (HHS) issued a report describing how government and industry could work together to tackle pandemic influenza, and according to a recent press briefing by HHS Secretary Kathleen Sebelius, the Obama administration proposed a $2 billion strategy to create a more "nimble and flexible" system for responding to novel pathogens (see Washington Report).

But switching to new manufacturing paradigms will not come easily or cheaply. For one thing, the traditional platform already has regulatory approval. But FDA released a guidance in February 2010 that paves the way for novel formulations and manufacturing approaches. Even if regulatory roadblocks were removed, however, there remains the stark economic reality that traditional influenza vaccine is a low-margin business, which leaves little incentive for innovation. Notes Rahul Singhvi, president and CEO of Novavax (Rockville, MD), "If you looked at the vaccine landscape today, it would be tough to say that influenza is a good area to go into given that five companies dominate the market and that the vaccine is basically a commodity." The trick for companies looking to move the technology forward is to find a solution as economical as it is effective.

A long tradition

Influenza vaccine has traditionally been manufactured by a pretty dowdy process, one that has remained largely unchanged for the past 50 years. The process begins when the Centers for Disease Control and Prevention (CDC) identifies the specific strain of virus causing a particular outbreak and makes "seed" virus available to vaccine manufacturers.

In the most basic manufacturing scenario, seed virus is introduced into fertilized chicken eggs, where it reproduces in large numbers and is secreted into the allantoic fluid—essentially the white of the egg. The fluid is collected after a few days and purified, and this purified fluid forms the basis of the vaccine. (Because some egg-related matter remains in this formulation, people with egg allergies often experience adverse effects with this vaccine, an additional motivation to move away from egg-based manufacturing.)

The traditional vaccine contains either live, attenuated, or killed virus particles. Live, attenuated virus can replicate in the vaccine recipient, and can confer immunity without causing disease (at most, it may cause some limited respiratory symptoms). In contrast, killed virus cannot replicate in the recipient, but still confers immunity.

For the most part, this process with few innovations has reliably produced influenza vaccine for decades—as long as the demand for vaccine is predictable. Most experts estimate that it takes one to two eggs to manufacture each dose of vaccine, so producing some 100,000 doses of vaccine requires a lot of eggs and a lot of advanced planning. Amassing that number eggs can be difficult, to say the least. "It's a scramble every year," says James Gombold, director of testing services for Charles River Biopharmaceutical Services (Wilmington, MA).

The difficulties don't stop there. The seed virus is replicated many times in the course of manufacturing. But viruses are notoriously sloppy in their replicative fidelity, so the virus in the vaccine can potentially differ from the infective strain—which can result in a less protective vaccine than required

"It's a dirty little secret," says Alan Shaw, chief scientific officer and chairman of the board of VaxInnate (Cranbury, NJ) "that you get mutants during passage."

To circumvent this problem, stringent analytical testing is required—DNA and protein analysis, primarily—to demonstrate homology between the virus in the vaccine and the infective strain, says Gombold, who directs the testing program for Charles River. Additional testing is required to ensure that avian pathogens lurking in the eggs do not end up in the vaccine, although Gombold says that this is probably more a precautionary measure than a response to any known incidence of contamination. Nevertheless, the additional testing adds time and costs to egg-based manufacturing.

Eggs have been the preferred system for vaccine manufacturing because in general, the virus grows well in them—but not always. H1N1, for example, did not, in spite of what Shaw calls manufacturers' "heroic effort...to make vaccine quickly and appropriately."

But even if the virus does grow well, says, Shaw, egg-based manufacturing can take longer than expected. "It's an egg-by-egg process," he says. "There's not much of a surge capacity for egg-based business. You can't ramp up by a factor of 20, which you essentially need to do [in a public health emergency]. The point is, the egg system has its days numbered." Many people think cell-based systems may take its place.

A tough sell for cells?

Cells are the natural targets for viruses, so it makes sense to exploit that relationship to manufacture them. Viruses cannot reproduce on their own. They contain the requisite genetic information, but not the replicative machinery. Cells, on the other hand, do. In order to reproduce, viruses penetrate and co-opt the replicative machinery of dividing cells. Fertilized eggs contain dividing cells, but in limited numbers. Dividing cells don't occupy in the entire volume of the egg, so space isn't optimized, in contrast to cells in a bioreactor.

A bioreactor containing enough cells to produce 50,000 doses of vaccines occupies far less space than do the number of eggs required to produce a comparable batch. Notes George Kemble, senior vice-president and head of research for MedImmune (Gaithersburg, MD), whose FluMist vaccine is currently manufactured in eggs, "Try to make 2000 L of cell culture fluid versus 2000 L of egg fluid. Not to mention the personnel needed."

Furthermore, the cells used—most frequently an immortalized canine kidney cell line—can be characterized in advance and stored frozen until large quantities of vaccine are needed, where they can be quickly thawed and amplified. Rapid amplification is the most compelling advantage of cells, say most vaccine makers. The first dose of vaccine can be produced within a week of infection with seed virus in both cells and eggs. "It's the time to produce the 100,000th dose that differs," says Kemble.

Among manufacturers, Novartis was the first to use this cell-based system. The company already produces two vaccines in cells in its Marburg, Germany, facility. Recently, Novartis opened a cell-based facility in Holly Springs, North Carolina, in which says Matthew Stober, Novartis's global head of technical operations, the company hopes to be able to produce up to 150 million doses of pandemic influenza vaccine as early as 2013.

"The regulatory pathway to achieve FDA licensure for vaccines is a long and complex one, particularly if clinical studies are involved. However, because Novartis accelerated the investment of this facility under the current pandemic threat, bulk prepandemic vaccine [against H5N1] could be produced from the site as early as late 2011 under an emergency use authorization," says Stober. Other drug manufacturers, including GlaxoSmithKline (GSK, London) and MedImmune are exploring cell-based programs.

So why aren't all manufacturers embracing cell-based facilities? For one thing, manufacturers with facilities geared toward egg-based manufacture can't retrofit them to produce vaccine in cells, says Gombold. But the real hesitation is economic. A dose of vaccine currently sells for under $10 says Shaw, and that's "not good for a producer that has to stay in business." Nor does it provide the kind of revenue that justifies the investment required to build a cell-based facility. "Only an external subsidy makes it feasible," he says.

Indeed, the US government provided approximately 40% of the $1 billion for Novartis's Holly Springs facility. MedImmune can consider the move because it already has bioreactors. "It would be different if we had to build the factory anew," says Kemble. Like Novartis, MedImmune has also received government funds that Kemble says, "help enormously." Once a facility is built, cell-based production should cost much less on a per unit basis than egg-based manufacturing, he adds.

GSK currently produces its vaccine in eggs and will for the foreseeable future, says Martine Wettendorff, vice-president, Influenza Vaccines Development Leader. But the company is not insensitive to the problem of scale up for a pandemic. GSK's answer is to make the existing vaccine more potent through the use of adjuvants, thereby reducing the amount of antigen required for each dose.

The company markets pandemic influenza vaccines in Europe and Canada that contain its proprietary ASO3 adjuvant and plans to seek FDA approval to supply an adjuvanated H5N1 vaccine in the US. Nevertheless, the company is researching cell-based technologies, which Wettendorff describes as "not yet a fully mature technology. The cost of goods remains a challenge," she says, but GSK is "always looking around at what's happening to lower that cost as well as at recombinant technologies." She reflects the attitudes of many who feel the cell-based technology is not yet an economically viable proposition.

Neither fish nor fowl

"This is where we come in," says VaxInnate's Shaw. "We can ramp up very quickly and cost effectively." The possibility of bringing new ideas to an old product has produced what Björn Lundgren, vaccine marketing manager for GE Healthcare Bio-Sciences (Uppsala, Sweden), a producer of equipment and reagents for vaccine manufacturing, calls an "R&D playground." Several biotech firms are banking on recombinant technologies as the future for influenza vaccine production.

Rather than reproduce entire virus particles, recombinant technologies like VaxInnate's produce selected viral proteins from the infectious strain, the sequence for which can be obtained from CDC. VaxInnate produces the viral surface protein hemagglutinin (HA) in the bacterium Escherichia coli, which Shaw calls "a prodigious producer of protein." Prodigious enough, in fact, that at the current selling price, the company can realize a profit, he says. However, until recently, E. coli seemed an odd choice in which to express the highly glycosylated HA protein, because the bacterium does not glycosylate its proteins.

As it turns out, says, Shaw "[HA] doesn't have to be glycosylated." VaxInnate harvests its protein and refolds it in urea, he says. When combined with an adjuvant, the protein seems effective in the company's early-stage clinical trials.

Medicago (Quebec, Canada) produces virus-like particles—including recombinant HA—in plant cells, which do glycosylate their proteins but in a manner different from mammalian cells. That hasn't seemed a problem, though, says Nathalie Charland, the company's product portfolio director, who says that initial clinical studies uncovered no deleterious immune responses to the plant-glycosylated protein. Both companies believe they can produce pandemic-scale quantities of their vaccine within three weeks of receiving the seed-virus sequence from CDC.

Recombinant proteins do have one drawback, however. They tend to be less immunogenic than are whole viruses, and therefore require some kind of adjuvant to boost potency. Says Mark Tomai, head of vaccine business development for 3M (St. Paul, MN), "The type of vaccine dictates whether and what type of adjuvant is required."

Both the Medicago and VaxInnate formulations include adjuvants. The Medicago adjuvant is part of the plant-cell membrane that envelopes the virus as it buds off from the plant cell. VaxInnate uses flagellin, a bacterial protein that binds to toll-like receptors on immune cells and activates them. There has been some concern, notes Tomai, that toll-like-receptor ligands distribute throughout the body causing nonspecific and unwanted immunological effects. VaxInnate circumvents that by coupling adjuvant and antigen, which, says Shaw, prevents adjuvant molecules from "swimming around on their own," and instead directs the immune response specifically against the antigen.

Novavax manufactures a trio of viral proteins inside insect cells, producing a virus-like particle, which Singhvi says closely resembles the actual virus in electron micrographs. He adds that without adjuvant, the preparation has elicited a strong immune response in Phase II clinical trials. Like the others, Novavax's production costs are competitive with traditional vaccine manufacturing, says Singhvi. "That's why we believe we can enter the market without being pushed out based on cost," he says.

Beyond recombinant proteins there are DNA vaccines, such as the one under development by San Diego-based biotech company Vical. Plasmid DNA containing a viral gene sequence and formulated with Vaxfectin, the company's proprietary lipid-based adjuvant, is injected, and the recipient's own cells manufacture the immunogenic proteins. The vaccine, says, Vijay Samant, Vical's president and CEO, provided a lasting immune response to H5N1 in Phase I clinical trials.

"It's hard to predict which technology will be the winner," says Lundgren. Choosing one and standardizing it, he thinks will be the real key to manufacturing an effective, scalable, and low-cost influenza vaccine.

See an online bonus interview with Novartis's Matthew Stober, global head of technical operations.