DEE-FENSE! DEE-FENSE!: Preparing for Pandemic Flu

Vaccination to prevent viral and bacterial diseases is modern medicine’s most cost-effective intervention. Were a vaccine to be available quickly after the onset of the widely predicted pandemic from an H5N1 strain of avian influenza, it might save scores of millions of lives worldwide. But that option is not feasible.

Why can’t a country that developed the atomic bomb (60 years ago) and the polio vaccine (50 years ago) and put a man on the moon (almost 40 years ago) now produce an appropriate vaccine? The answer is an unfortunate confluence of biology and public policy.

During the past several years, an especially virulent strain of avian flu, designated H5N1, has ravaged flocks of domesticated poultry in Asia and spread to migratory birds and (rarely) to humans and other mammals. It has been detected in much of Europe, Asia, Africa, and the Middle East, and it continues to spread with each seasonal migration of wild birds. Since 2003, there have been over 200 cases of H5N1 infection in humans, more than half of whom have died—a shockingly high mortality rate for an infectious disease.

Public health experts and virologists are concerned about the potential of this strain because it already possesses two of the three characteristics needed to cause a pandemic: It can jump from birds to humans, and it produces a severe and often fatal illness. If additional genetic evolution makes H5N1 easily and sustainably transmissible among humans—the third characteristic of a pandemic strain—a devastating worldwide outbreak could become a reality. The ease and frequency of worldwide travel could give rise to the first true jet-age flu pandemic.

Although it is not possible to predict the timing of that last evolutionary step, because the genetic changes that would give rise to it are wholly random molecular events, mutations occur each time the virus replicates, so the more H5N1 viruses are produced, the more likely it is that the event will occur. As avian flu spreads and more birds are infected, there are trillions more virus particles in existence every day. Flu can also evolve when both human and animal strains of flu infect a person or animal simultaneously, offering an opportunity for swapping segments of nucleic acid that code for viral proteins. That process, too, is favored by the presence of more viral particles in more locations around the world.

Some background is necessary to understand the threat and the possible public health, economic, and political consequences of a flu pandemic. The exterior of the flu virus consists of a lipid envelope from which project two surface proteins: hemagglutinin and neuraminidase (N). The virus constantly mutates, which may cause significant alterations in either or both of these proteins, enabling the virus to elude detection and neutralization by the human immune system. A minor change is called genetic drift; a major one, genetic shift. The former is the reason why flu vaccines need to be updated from year to year; an example of the latter was the change in subtype from H1N1 to H2N2 that gave rise to the 1957 pandemic. This new variant was sufficiently distinct that people had little immunity to it. The rate of infection with symptomatic flu that year exceeded 50% in urban populations, and 70,000 people died from it in the United States alone.

Ordinary seasonal flu, which is marked by high fever, muscle aches, malaise, cough, and sore throat, is itself a serious illness that on average kills 36,000 annually in the United States, but the pandemic strains are often both qualitatively and quantitatively worse. The H5N1 strains of bird flu have a predilection for infecting the tissues of humans’ lower respiratory tract; that is, deep down in the smaller airways and in the tissues where oxygen exchange takes place, where it elicits hemorrhage and “cytokine storm,” an outpouring of hormone-like chemicals that causes huge amounts of fluid to accumulate in the lungs. In this way, these pandemic strains of flu may kill within 24 to 48 hours of the onset of symptoms.

By contrast, seasonal flu most often kills not directly but via secondary bacterial infections that follow the initial viral infection of the upper respiratory tract. (Seasonal strains’ affinity for the upper respiratory tract also helps to explain why they spread so rapidly: The virus particles are readily expelled by coughing and sneezing, and when other people are thereby exposed, the viruses need only travel a short distance in the body in order to attach and infect.) A strain of bird flu that is able to infect both the upper and lower respiratory tracts, similar to one isolated from a patient in Hong Kong in 2003, would have the potential to cause a highly lethal pandemic.

Misguided policies

The problems related to the biology of the flu virus have been compounded by public policy decisions by Congress and the government’s executive branch that ensure low return on investment and high exposure to legal liability for vaccines. Why should companies make products that are only marginally profitable and whose sale, even in the absence of any negligence or wrongdoing, carries the threat of huge financial risks from lawsuits?

Several kinds of policies are responsible for our vaccine quagmire. The Vaccines for Children Program, for example, was an innovation of the Clinton administration that disrupted market forces and dealt a blow to vaccine producers. Established in 1994, it created a single-buyer system for children’s vaccines, making the government by far the largest purchaser of childhood vaccines, at a mandated discount of 50%. Try extorting that kind of discount from manufacturers of vehicles for the U.S. Postal Service or of Meals Ready to Eat for the Department of Defense, and see how long they bid on customer contracts.

Arbitrary and excessively risk-averse regulation is another obstacle. The Food and Drug Administration (FDA) has been especially tough on vaccines, continually raising the bar for approval. The agency required huge clinical trials—more than 72,000 children (and another 44,000 in post-marketing studies)—of a recently approved vaccine against rotavirus (a common, sometimes fatal gastrointestinal infection in children) in order to be able to detect even very rare side effects before approval. In fairness, one does need to be concerned about a new vaccine that is intended for large numbers of healthy people; even a rare but serious side effect in a drug administered to hundreds of millions of people could have significant impacts. Thirty years ago, the federal government attempted to administer “swine flu” vaccine to the 151 million Americans age 18 and over, but the program was halted after a small number of individuals suffered generalized paralysis after vaccine administration.

The challenge for regulators is to find an appropriate balance of pre- and postmarketing clinical trials that demonstrate a vaccine’s ability to reduce the incidence of actual community-acquired infections, or that measure efficacy by means of surrogate endpoints such as laboratory measures of antibody-mediated and cellular immunity. In recent years, U.S. regulators have frequently been overly conservative in their requirements. They have rejected evidence of safety and efficacy from European and Canadian vaccine approvals and prematurely withdrawn lifesaving products from the market because of mere perceptions of risk.

It is difficult to guess how long the required clinical trials and data analysis might take for a new vaccine against pandemic flu, but if regulators fail to use surrogate endpoints appropriately, it could be years; a catastrophic delay in the event of a pandemic.

The FDA’s recent announcement of policies intended to streamline the development and approval of annual and pandemic flu vaccines offers some cause for optimism, but even this advance must be qualified. The agency published “guidance,” not “guidelines,” for vaccine developers. The critical difference (in regulation-speak) is that guidelines bind the agency to a certain action (the licensing of a vaccine, for example) if the conditions specified in the document are met, whereas guidance is only advisory. In fact, lest that distinction go unnoticed, every page of the guidance documents admonishes in bold font,“Contains Nonbinding Recommendations.” This provides the FDA with commodious wiggle room. Another disappointment is the conspicuous absence of any mention of reciprocity with approvals by foreign regulators such as the European Medicines Evaluation Agency. Reciprocal approvals would obviate the need for companies to meet the slightly different but largely redundant requirements of many different regulatory agencies.

As the result of our flawed public policy, innovation has suffered and vaccine producers have abandoned the field in droves, leaving only four major manufacturers and a few dozen products. We are woefully short of capacity for the production of a vaccine against a pandemic strain of flu, which cannot actually begin until we have it in hand and have performed various genetic manipulations so that it doesn’t kill the chicken embryos in which flu vaccines currently are grown. An optimistic estimate is that there is sufficient flu vaccine capacity worldwide for approximately 450 million people, but that calculation assumes that two intramuscular inoculations of 15 micrograms each would confer protection. Recently developed experimental vaccines against H5N1 required two doses of 90 micrograms. That suggests that the true capacity might be closer to only 75 million people, or a little more than 1 of every 100 people on the planet.

Another worry is that when a pandemic strain of H5N1 avian flu appears, virtually all of the world’s flu vaccine development and production capacity might shift to producing a vaccine against it, which will leave us vulnerable to the nonpandemic strains that cause the usual annual or seasonal flu. As Anthony Fauci, director of the U.S. National Institute of Allergy and Infectious Diseases, has observed, “The biggest challenge unequivocally is vaccine production capacity.”

Remedying that will not be easy. Currently, it requires five to six years (and a massive investment) to build and validate a new manufacturing plant to the satisfaction of regulators. Moreover, the currently available vaccines are made using half-century-old technology: the cultivation of live virus in scores of millions of fertilized chicken eggs.

Hope from the lab

Some good news concerning vaccines is emanating from research laboratories. Several recent advances suggest ways to induce a potent immune response to the H5N1 strain, but our ability to translate these findings into commercial products is a long way off.

Using genetically engineered common-cold viruses, two separate U.S. laboratories have successfully vaccinated mice against various strains of H5N1 bird flu. Both teams used adenoviruses that were genetically modified to incorporate the gene that expresses hemagglutinin, a surface protein of H5N1, so that they are unable to replicate. In effect, these are adenovirus-influenza hybrids.

Injected into mice, various versions of the vaccines generated a potent immune response that consisted of both antibodies and activated white blood cells and that protected the animals against a challenge by high doses of H5N1. Significantly, the vaccines were able to protect against viruses that did not precisely match the strains from which the hemagglutinin was derived.

Conventional flu vaccines induce only antibodies, a limitation that requires new vaccines to be developed constantly to keep up with the mutating, evolving virus. The dual response and the cross-protection seen in these experimental vaccines is important because it increases the likelihood that they will be at least partially effective against newly arising variants of H5N1.

Another recent development offers a possible generic method to enhance the immunogenicity of many different vaccines. Researchers at the University of British Columbia used genetic engineering techniques to incorporate into various viral vaccines two proteins that help cells of the immune system to process foreign antigens. They found that these proteins act as a potent booster, inducing the immunized recipient to produce greater numbers of immunologically active cells against foreign antigens also contained in the vaccines. In their animal model, in which a challenge of a potentially lethal dose of virus was administered after vaccination, one of their engineered vaccines “provided protection against a lethal challenge . . . at doses 100-fold lower” than controls that did not have the modification. Although these experiments involved viruses other than influenza, the technique should be applicable as well to flu and to adenovirus-flu hybrids.

Scientists are also working on ways to boost the immunogenicity of vaccines by adding chemical ingredients known as adjuvants, which make it possible to use lower doses of the vaccine antigens themselves. Adjuvants are not specific to particular antigens but act in various ways to activate one or more components of the immune system. They may help to display vaccine antigens to appropriate antigen-presenting-cell (APC) types; to target particular intracellular APC compartments for optimal antigen presentation; or to induce appropriate APC maturation steps that increase the stimulation of T lymphocytes, activate antibody production, and induce immune memory.

France’s Sanofi-Pasteur and Australia’s CSI have begun trials of candidate pandemic vaccines that use adjuvants made of alum, an aluminum salt, the only adjuvant currently approved for use in humans in the United States. California-based Chiron Corporation might have a more promising candidate. In clinical trials of an adjuvant called MF59, which has been incorporated into a candidate vaccine being tested for protection against avian flu strain H5N1, vaccine containing adjuvant was significantly better than vaccine alone at eliciting antibodies to H5N1. An important potential advantage of this adjuvant-containing vaccine is the discovery that it may offer protection against H5N1 even if the virus’ cell-surface proteins change, or “drift,” in a way that makes them slightly different immunologically.

That suggests a viable, if not optimal, strategy to prepare for the pandemic: Stockpile vaccine against the current avian flu H5N1 strain, with adjuvant added to boost the immune response. Although it would not be a perfect match to the pandemic strain, it might be useful as a first “priming” dose that could afford some protection until vaccine against the actual pandemic flu strain is available.

An obstacle to this approach is that MF59, used in European vaccines since 1997, has never been approved for use in a vaccine sold in the United States, at least partly because R&D on vaccines has become so unprofitable and unattractive that there has been little incentive for vaccine developers to perfect a technology to boost their efficiency or to perform the expensive clinical testing necessary to license what regulators would regard as a new vaccine technology. Also, the addition to existing vaccines of an adjuvant, even one with a long history, would make a previously approved vaccine a “new drug” that would require exhaustive testing (especially given that the products would be administered to very large numbers of healthy people).

Various research groups are studying alternatives routes of administration of vaccines (intradermally, for example, instead of via the usual intramuscular route) in order to be able to use smaller dosages and/or to elicit heightened immune responses. One study found that the dose of flu vaccine administered intradermally could be reduced to 40% of the usual intramuscular dose without compromising the immune response.

Opting for a conservative strategy, British health authorities have ordered sufficient conventionally produced vaccine against the actual pandemic strain to treat every person with the needed two doses. The limitation of this approach is that because production cannot begin until the pandemic begins and the virus is in hand, there will be a substantial lag: perhaps nine months or more until the vaccine is available. Why so long? The producer must demonstrate that the vaccine can be manufactured in batch after batch to high levels of purity and potency as well as conduct and analyze clinical trials. Thus, although this approach is the most definitive in the long run, it would leave the population vulnerable to the first wave of the pandemic.

In sum, the good news is that once we know the genetic sequence of the pandemic strain, we can reverse-engineer flu virus and get a candidate vaccine into animal trials rapidly; various chemicals can be used to enhance the immune response; and we have good early prototype “subunit” vaccines that use only a single gene from the flu virus, can be grown in cultured cells instead of chicken eggs, induce both antibody-based and cellular immunity, and show a high degree of effectiveness at protecting mice against challenge with a variety of H5N1 strains.

The bad news is that there are prodigious obstacles to translating these developments into a clinical setting, let alone into a commercial human vaccine.

First, adenovirus infections are extremely common in children and adults, and if recipients have previously been infected with the particular adenovirus strain used in the vaccine (there are dozens of different ones that infect humans), they may be “immune” to the vaccine. In other words, they’ll ward it off before it can carry out the “controlled infection” necessary to elicit immunity to the engineered adenovirus-flu hybrid.

Second, some adenoviruses are thought to have the potential to induce malignancies, which will likely elevate the threshold for regulatory approval.

Third, the adenovirus-flu hybrid vaccines rely on the flu gene that expresses the viral surface protein hemagglutinin, which is notorious for the antigenic drift or shift that enables the flu virus to elude vaccines. Thus, even with the advantage of being able to elicit cellular (T lymphocyte–mediated) immunity as well as antibodies, it’s unclear how effective a vaccine against the current largely bird-specific H5N1 would be against an emergent pandemic strain.

Fourth, mice are not little humans, and it is difficult to extrapolate with confidence the results of mouse experiments to humans. (Our ability to predict efficacy in humans would have been greater had the investigators used transgenic mice engineered to have a human immune system.)

Fifth, a published analysis of thousands of bird flu samples taken from across southern China illustrates the difficulty of mounting an effective vaccine strategy against a possible pandemic strain (or even the bird-specific strains) of H5N1 avian flu. The authors concluded that the region, a reservoir of the virus for nearly a decade, has spawned divergent strains that have been spread as far as Russia and Eastern Europe by migratory birds. That diversity makes choosing a vaccine strain(s) problematical.“The antigenic diversity of viruses currently circulating in Southeast Asia and southern China challenges the wisdom of reliance on a single human-vaccine candidate virus for pandemic preparedness,” the authors wrote.

Sixth, getting a vaccine production facility (which is very different in design from one that produces conventional small-molecule drugs) up and running to the satisfaction of regulators currently requires five to six years and a huge financial investment, and we will need vast amounts of flu vaccine.

Seventh, in the absence of an actual pandemic or of government-guaranteed vaccine purchases or payments for meeting R&D milestones, that investment might be for naught.

Finally, regulatory obstacles, especially where new technologies are involved, are daunting.

In summary, government and private-sector funding of high-quality research projects is bearing fruit, but the recent research advances leave us far from real-world solutions.

Government both giveth and taketh away, and in recent years, the latter has dominated public policy. We need incentives for industry to develop the products that we need, and the FDA’s gatekeeper function for new medicines should not be permitted to delay clinical progress unduly. The agency needs to develop and implement a plan for active collaboration on and rapid review of candidate vaccines. As has not been the case since World War II’s Manhattan Project to develop the atomic bomb, we need a robust government–university–private-sector partnership (with cooperation on issues much broader than just vaccine development) to counter a universal and dire threat.

Which strategies should we adopt? My answer is a wide variety, simultaneously and as expeditiously as possible. Just as the Manhattan Project pursued at least three methods to enrich uranium for the needed isotope U-235 on independent parallel tracks, we need to set in motion many research approaches. The Manhattan Project was arguably the most ambitious and successful R&D undertaking in history, and the threat of an avian flu pandemic argues for a similar approach: numerous parallel strategies pursued on many fronts. There is some acknowledgement of this philosophy in the Bush administration’s National Strategy for Pandemic Influenza Implementation Plan, published by the Homeland Security Council in May 2006, but its incentives for R&D are both tentative and vague. The details that are relevant to vaccines and anti-influenza drugs are unimpressive, consisting primarily of contracts for decidedly inadequate stockpiles of drugs and un-adjuvanted, probably ineffective vaccines against pre-pandemic strains of H5N1.

Vaccines are widely acknowledged to have high social value, but compared to therapeutic drugs, their economic value to pharmaceutical companies is low. Because governmental policies have caused market failures in vaccine R&D, government actions now must be an integral part of the solution.

A variety of incentives is needed to revitalize the portion of the private sector that has so long been so beleaguered by policymakers and regulators. Public policy must reward both inputs on vaccine R&D (via grants, tax credits, and the waiver of regulatory registration fees) and outputs of products (with guaranteed purchases, milestone payments when regulatory approval of new vaccines is granted, indemnification from liability claims, waiver of FDA user fees for vaccine reviews, and reciprocity between U.S. regulatory approvals and those in certain foreign countries). This effort should include aggressive funding of “proof-of-concept” R&D on various new technologies and approaches to making flu vaccine, to boosting the immune response to vaccines, and to creating greater reserve capacity for the commercial production of vaccines (including alternatives to vaccine production in eggs). And instead of being a major cause of the problem, regulators must become part of the solution. For example, they should work with companies to approve development approaches and manufacturing facilities in advance of the actual production of pandemic flu vaccine. In effect, the infrastructure would be ready to plug in the actual pandemic strain when it appears and to facilitate testing and regulatory approval.

Federal officials are largely responsible for the current lack of societal resilience needed to combat a flu pandemic. Now they must do more than fiddle while flu fulminates.