Influenza Vaccine Manufacture: Keeping Up with Change

Influenza virus causes annual epidemics of respiratory disease that affect all age groups. Manifestations of influenza vary by age and underlying risk condition and, on a population basis, result in substantial morbidity, mortality, and lost productivity. Inactivated vaccines against influenza virus have been licensed for >60 years. These vaccines have demonstrated efficacy and effectiveness across broad age groups and among different populations and have a strong record of safety [1–3]. Live attenuated influenza vaccines have been licensed in the United States since 2003 and likewise have demonstrated efficacy and safety [1, 4, 5]

Despite many years of use, influenza vaccines remain challenged by many factors, first and foremost being the characteristics of the virus itself. Currently licensed influenza vaccines are designed to elicit immunity predominantly against the virus hemagglutinin (HA), a surface glycoprotein critical for the attachment of virus to host cells. Antibodies specific to HA are believed to be the best correlate of protection against influenza virus infection and are the primary end point used to evaluate vaccine immunogenicity. However, the influenza virus—and particularly the HA glycoprotein—undergoes constant genetic and antigenic change. As generally defined, the term antigenic drift refers to changes to the HA within a given subtype that contribute to a reduced ability of circulating antibodies to a previous HA to bind effectively to a newly circulating strain of influenza virus (eg, H1 or H3 annual antigenic drift). From 1977 (when influenza A/H1N1 viruses reemerged in humans) until 2009, the H1N1 component of the seasonal influenza vaccine has been changed 8 times to accommodate HA drift, including 3 changes in the last decade. For influenza A/H3N2 viruses, the antigenic drift has been even greater, with 5 vaccine strain changes in the last decade alone [6]. The term antigenic shift generally refers to the emergence in humans of an entirely new HA type, typically resulting from a reassortment event (eg, the emergence of H2N2 in 1957 and H3N2 in 1968). However, the current novel influenza A/H1N1 outbreak among humans challenges these usual categories. Although this novel virus does not represent a new influenza A subtype in humans, it likely arose from a series of reassortment events in pigs and is both genetically and antigenically distinct from other recent A/H1N1 viruses circulating among humans [7, 8]. In a recent study, there was littledetectable cross-reactivity to the novel A/H1N1 virus among serum antibodies from individuals who were vaccinated with recent (2005–2009) seasonal influenza virus vaccines [8]

The outbreak of the novel A/H1N1 strain highlights the challenges experienced with influenza vaccine production. Vaccine production for the 2009–2010 season in the Northern Hemisphere was well under way when this novel virus was first identified, and the 2009 Southern Hemisphere vaccination season was already in progress. The most appropriate course of action was difficult to determine. Could and should a quadrivalent vaccine that included preselected seasonal strains and the novel strain be developed?. Should a monovalent vaccine directed against the novel A/H1N1 strain be produced?. How quickly could such actions be taken?. Would manufacturers have sufficient egg supply at the end of the seasonal production cycle to initiate production of an additional vaccine in desired quantities?

An understanding of the current vaccine production process illustrates why the emergence of a new strain late in an influenza season cannot be easily accommodated. Generally, strain selection for the Northern Hemisphere influenza vaccine occurs in February. Strains are selected early enough to allow for the production of 3 strains that comprise the trivalent formulation, for release and distribution of vaccine, and for vaccine delivery to recommended populations before the onset of the influenza season. Although in recent years the peak influenza season has most commonly occurred in February in the United States, the start of the influenza season cannot be predicted in advance and has occurred as early as October [1]. Ideally, vaccine would be given 2 weeks before the start of the season for optimal immunogenicity; for children receiving influenza vaccine for the first time, 2 doses must be delivered. Thus, each year delivering influenza vaccine into the arms of the people who need it is literally a race against the clock

It is important to consider ways in which the vaccine production process could be improved or the timelines shortened to permit a more rapid response to an emerging outbreak. Ultimately, an influenza vaccine that elicits broad immunity to a breadth of drifted viruses could preclude the need for the continuous chasing of evolutionary changes seen with influenza. Such improvements could address some of the limitations associated with current vaccines. If significant cross-protective immunity could be induced, influenza vaccines would be more effective against novel strains and could be manufactured and delivered throughout the year or even stockpiled for rapid use in the event of an uncontrolled outbreak of a drifted virus. These factors could ease the stress placed on vaccine manufacturers and health care providers in such situations. Unfortunately, the development of such vaccines has proven to be difficult and remains in early—albeit active—stages of development

Until the time when influenza vaccines with broad-spectrum and long-lasting immunity are available, improvements that allow for enhanced immunogenicity, speed of production, and cross-reactivity are needed. Cell culture–based production is one variable that could potentially improve influenza vaccine manufacture. Currently in the United States, all licensed influenza vaccines are manufactured in eggs. In general, 1 embryonated chicken egg yields 1 dose of a typical trivalent vaccine, composed of 3 strains and totaling 45 μg of HA. This process has been used for many decades and is dependent on egg supply, in contrast to cell-based vaccine production, for which the substrate is continuously available. Egg supply may be even more constrained during a widespread human outbreak with an avian influenza virus. In addition, virus seed strains adapted for growth in eggs must be made via time-consuming reassortment techniques followed by serial passage in eggs. The process of passaging influenza viruses through eggs to create the high-growth variant could also introduce mutations into the HA that could alter the immunogenicity of the vaccine. Furthermore, identifying viruses that grow well in eggs may delay the selection of vaccine strains. The selection process for several H3N2 candidates has been postponed in recent years, pending the identification of suitable high-growth reassortants in chicken eggs. In some years, alternate vaccine strains have been selected on the basis of their growth properties [6]. Thus, cell culture–based vaccines may yield a closer match of vaccine strains to circulating strains, eliminating the entire egg-adaptation procedure. Overall, use of cell-based rather than egg-based vaccine production could save several weeks from the time of strain identification to the quality control release and delivery of vaccine [9]. Clearly, strain selection could occur later in the year if production times could be shortened, allowing for better assessment of strains that may emerge late in the influenza season

Both Solvay Biologicals and Novartis Vaccines have licensed inactivated trivalent influenza vaccine (TIV) manufactured in a mammalian cell line—Madin-Darby canine kidney (MDCK) cells—outside the United States. In this issue of the Journal, studies by Szymczakiewicz-Multanowska et al [10] and Reisinger et al [11] compared a Novartis cell culture–derived TIV made in the MDCK cell line with a conventional egg-based TIV. It is reassuring but not surprising that the cell culture vaccines were comparable to the egg-derived vaccines with respect to the immunogenicity and reactogenicity profiles in the adult populations tested. These studies add to a body of evidence supporting the comparability of cell culture–derived and egg-derived TIVs. In addition to the traditional inactivated methodology, recombinant vaccine antigens made in mammalian or insect cell lines can be efficiently generated using innovative expression systems that rely only on the genetic sequence information from the vaccine strain [12]. A challenge remaining for all of these newer cell culture–derived vaccines will be to successfully clear important regulatory hurdles for broader licensure and to ensure that such cell lines and derived products will share the strong safety record that traditional egg-based products have. In addition, whether final yields and costs of the cell-based TIV can compare favorably with the egg-based TIV will be an important consideration

In the coming months, questions relating to the vaccine response to the novel influenza A/H1N1 virus will be addressed. US efforts in pandemic preparedness have resulted in an improved overall production capacity and manufacturer readiness, yet the ability to respond remains encumbered by the current realities of the influenza vaccine manufacturing process. It is encouraging to see improvements in vaccine substrates as well as new technologies that have the potential to be licensed to ensure a more rapid, flexible, and scalable response to the next emerging strain of influenza virus

• Received June 9, 2009.
• Accepted June 9, 2009.

• © 2009 by the Infectious Diseases Society of America