OR WAIT 15 SECS
Volume 2011 Supplement, Issue 4
A plant-based expression system could provide greater speed and capacity at a comparatively low cost.
Medicago has developed a proprietary plant-based manufacturing platform that produces vaccines made of influenza antigen scaffolds (i.e., viruslike particles) that are currently in clinical trials in Canada and the United States. This platform produces vaccine doses within a month of the disclosure of hemagglutinin sequences from emerging strains. It is now being automated and scaled up in North Carolina to bring its surge capacity to more than 10 million doses/month. The platform will be housed in a facility that will be built within 12 months for less than $35 million.
The recent swine H1N1 influenza pandemic revealed the limitations of the current influenza-vaccine manufacturing technologies. Although Wyman and colleagues had predicted in 2007 that egg-based manufacturing would be able to supply at least 60 million vaccines doses within five months of the declaration of a pandemic, the actual vaccine output in the 2009–2010 porcine H1N1 (pH1N1) pandemic was much lower (1). In fact, only three million doses were available five months after the identification of the causal strain. The first doses of split vaccine became available shortly thereafter, but at levels far below expectations. Fortunately, the pH1N1 strain had a low mortality rate. Had the pH1N1 pandemic had a higher severity index, the global human cost of the delays in egg-based vaccine production could have been catastrophic. Nevertheless, this outbreak has underscored the need for a faster, simpler, and more flexible manufacturing technology to respond to an emerging threat and to protect as many people as possible rapidly.
To meet the surge capacity required for the production of pandemic vaccines, recombinant technologies are being actively developed with various expression hosts, such as bacteria, yeast, insect, or mammalian cells. Medicago developed a plant-based expression system designed to provide great surge capacity and safety, at low cost. This article will describe how this new manufacturing platform is combined with a nanoparticle antigen-presentation scaffold (i.e., viruslike particles or VLPs) that is synthesized in plants, with an extraction approach that separates the nanoparticles from the host cell components early in the process, and with a purification scheme that yields highly purified particles as drug substance.
The authors have demonstrated that this platform efficiently produces VLPs from the sole expression of recombinant influenza hemagglutinin (HA), and that administration of these purified influenza VLPs was well tolerated and triggered an immune response of great strength and breadth in animals and humans (2, 3). The speed at which influenza VLPs can be produced following the identification of a new HA gene sequence was assessed in the context of the pH1N1 pandemic. The first doses of a plant-made H1 VLP candidate vaccine against the pH1N1 strain were produced within three weeks of the disclosure of the new HA sequence by the Centers for Disease Control and Prevention (see Figure 1).
Figure 1: H1 viruslike particle (VLP) experimental vaccine production timeline. (FIGURES ARE COURTESY OF THE AUTHORS)
The manufacturing platform based on transient expression, and the VLP-based antigen-presentation platform were developed in the context of the pandemic threat. They now have been applied successfully to the development of a trivalent seasonal vaccine. Vaccine products for these two indications have been produced under good manufacturing practice (GMP) conditions and currently are in clinical trials in Canada and the United States.
The transient expression system uses the capacity of a bacterium, Agrobacterium tumefaciens, to infect plant cells and transfer essential genetic information in the form of a mobile DNA fragment to the nucleus of the infected cell. Because this mobile DNA fragment only remains viable for a few days in the nucleus and is not integrated in the plant genome, the cells are only transiently transformed. Despite the transient nature of the transformation, this technology results in extremely high expression of the gene of interest for days after cell infection, and enables synthesis of large quantities of the targeted protein.
Medicago's transient expression system starts with the assembly of an expression cassette where the target gene is positioned within transcription and translation components adapted for expression in plants. The expression cassette is inserted within mobile DNA borders of a plasmid, which is transferred into the bacterial transfection vector A. tumefaciens. Transfection of the mobile DNA copy containing the expression cassette to Nicotiana benthamiana leaf cells is performed by soaking healthy plants upside down in a vacuum tank containing a liquid suspension of A. tumefaciens, the inoculum. As vacuum is applied, the air-filled cavities inside the leaves are emptied, thus leaving space for infiltration of the bacterial inoculum upon release of the vacuum. As the inoculum is infiltrated, the A. tumefaciens establish contact with parenchyma cells, which constitute the vast majority of the leaf active cells. Because vacuum infiltration is instantaneous and highly invasive, infection of leaf cells is also instantaneous, which helps to synchronize cell infection and the resulting transfer of mobile DNA copies containing the expression cassette. Transient expression is extremely fast and results in high accumulation of the recombinant protein within three to six days.
In the life cycle of the influenza virus, formation of the viral particle involves the budding of a portion of the host plasma membrane saturated with the viral surface proteins. HA is the major surface protein of the influenza virus with a transmembrane domain that spans through the viral envelope. The authors demonstrated that transient expression of the HA surface protein in plant cells results in the formation of VLPs budding from microdomains of the plasma membrane saturated with recombinant HA.
At maturity, these VLPs have morphological features similar to those of the influenza virus, but only contain its immunogenic determinants, with none of the viral RNA, which is packaged within the enveloped particles of true viruses. Compared with isolated soluble antigens, antigens at the surface of VLPs have the double advantage of adopting a conformation that is identical to those found on the cognate virus, and of being presented in a multivalent structure similar to that of the native virus. These features enable VLPs to achieve enhanced stimulation of the immune response with balanced humoral and cellular components. This benefit of VLPs over soluble recombinant antigens is believed to be particularly strong in vaccine products against enveloped viruses because VLPs present the surface antigens in their natural, membrane-bound state (4). As the influenza viral particle, influenza VLPs are enveloped, pleiomorphic, and 100 to 150 nm in size (see Figure 2).
Figure 2: Structural characteristics of influenza virus and viruslike particles, including (a) a comparison of surface attributes, (b) a cross-section showing internal distinctions, and (c) transmission electron microscopy images of influenza viral particles and influenza viruslike particles.
HA expression plasmid, master, and working cell bank of transformed Agrobacteria
The recombinant influenza HA gene is built in such a way that native HA is expressed, thus yielding a mature protein product identical to the wild type influenza HA protein. Expression is maximized by the integration of elements that result in high transcription of mRNA and hypertranslation of the product. In addition to these elements, a suppressor of silencing is coexpressed with the HA protein to counteract specific HA mRNA degradation (i.e., mRNA silencing) that is triggered by the significant accumulation of HA mRNA in plant cells.
Before their transfer into A. tumefaciens, the plasmids are fully sequenced to ensure sequence integrity. Upon confirmation of integrity, plasmids are introduced into A. tumefaciens, selected clones are grown in a master cell bank (MCB) that generates the working cell bank (WCB). The MCB and WCB are tested for purity and plasmid integrity.
A liquid culture of the bacterial vector is produced from the WCB and used to prepare the A. tumefaciens inoculum for infiltration.
Plant biomass production
N. benthamiana is a wild Australian relative of cultivated tobacco (i.e., Nicotiana tabacum), but has no agronomic or food use. N. benthamiana plants are grown in controlled and contained conditions. A high-quality, soilless, sterilized root-substrate medium provides physical support and optimal water retention. Apart from fertilization and substrate moisture, which are adjusted during plant development, all growth conditions remain unchanged during the production cycle (i.e., 40 days). All components used for biomass production are released under standard operating procedures. Special care is taken to control the presence of heavy metals in all liquids and solids that contact the biomass. Fungal, microbial, and algae growth are strictly controlled during all phases of growth.
Plant development is monitored from seeding to infiltration to ensure proper and reproducible biomass development. Plants are characterized during their growth and must meet specific criteria to be accepted for inoculum infiltration.
Inoculum infiltration and incubation
The authors' pilot facility in Québec City, Canada, has an inoculum infiltration capacity of 1200 plants per day. Infiltration is performed in a stainless-steel tank, 15 plants at a time. The vacuum infiltration lasts only 1 min, and the procedure is computer-controlled to ensure reproducibility. At a plant under construction in North Carolina, the infiltration unit will accommodate 15,000 plants per day. Transportation of the biomass to and from the infiltration unit will be completely automated. Following infiltration, plants are again maintained in controlled and contained conditions during the three to six days required for transient expression to take place.
Harvest, extraction, and conditioning
Plant leaves are collected and mixed with a depolymerization solution. Medicago has developed a proprietary technology that treats the biomass with cell-wall depolymerization agents. This treatment acts on structural or cross-linking components of the cell wall, which loosens the rigid extracellular matrix of the leaf cells, and subsequently releases the VLPs into the solution. The procedure releases more than 90% of intact VLPs with less than ~5% of the host-cell components (i.e., protein and DNA), and is thus an extremely efficient primary recovery step.
Because VLPs are large particles (130 nm in diameter), they can be concentrated by tangential-flow filtration (TFF). This concentration step also eliminates most of the remaining host-cell components, including low molecular-weight compounds. Diafiltration at this stage conditions the VLPs in preparation for chromatography.
Chromatography and formulation
VLPs are purified by ion-exchange chromatography, concentrated, and diafiltered against the formulation buffer through a final TFF. Chromatography also allows the removal of residual DNA and endotoxins.
The purified influenza VLP vaccines used in clinical trials were released under the criteria used for licensed influenza vaccines and have met all standards for purity and identity. DNA and endotoxins were always below acceptable limits, and these preparations were tested for sterility and subjected to General Safety Testing in guinea pigs before being released.
Table I: Characteristics of purified H5 viruslike particles (VLPs).
Mass spectrometry helps to identify trace protein contaminants. Most of the contaminants are naturally associated with plant plasma membrane and are neither immunogenic nor induce an allergic response in animal and human models. Because this product arises from a new platform, an extensive complementary physicochemical characterization was performed on purified preparations. Table I presents the measured attributes of the pandemic H5 VLP vaccine product analyzed on multiple lots. The narrow range of measurements for mean particle size, HA purity, and lipid–protein ratio demonstrates the high lot-to-lot reproducibility of the manufacturing process. The Coomasie-stained SDS-PAGE and Western blot analysis of the pandemic H5 VLP vaccine product in Figure 3 further illustrate the purity level of the product (i.e., 98–99% purity).
Figure 3: Purity and identity of an H5 viruslike-particle (VLP) vaccine: Coomassie-stained SDS-PAGE of (1) molecular-weight marker and (2) 2.5-Î¼g of influenza H5 VLP vaccine, and Western blot analysis using polyclonal antibodies against influenza H5N1 virus (strain A/Indonesia/5/05) of (3) 50 ng and (4) 500 ng of the influenza H5 VLP vaccine.
The pandemic H5 VLP vaccine product has been tested for safety and immunogenicity in a completed Phase I clinical trial and an ongoing Phase II trial in Canada. The Phase I results demonstrated that the vaccine product was safe and well tolerated and did not induce allergenic responses. No increase in the level of naturally occurring serum antibodies to plant N-glycans was observed. The evaluation of immune response to VLP immunization by hemagglutination-inhibition, single-radial hemolysis, and microneutralization assays were well correlated and showed a clear dose-response (3). In the Phase II trial, the preliminary results showed 65% seroprotection and 65% seroconversion after the administration of two 20-μg doses to 18 to 49 year-old subjects.
Today, the globalization of trade and the amount of international travel provide many opportunities for infectious diseases to spread rapidly and globally. As seen with the outbreak of the 2009 A–H1N1 influenza, within two days of the first confirmed cases from Mexico, the virus was reported in five additional countries. Because of difficulties in producing the vaccine, initial doses became available only after 26 weeks in the US, and supplies for the protection of all US citizens would have taken almost a year to produce. The President's Council of Advisors on Science and Technology stated that more than 2000 lives—with an average age of less than 40—could have been saved if vaccination had begun even one month earlier (5).
This increasing risk is posing new challenges to public-health agencies around the globe and stressing the need for effective vaccines sourced from highly flexible manufacturing facilities that can be mobilized rapidly and cost-effectively. The greatest potential to meet this challenge lies in the use of recombinant DNA technologies.
Medicago is pursuing a promising alternative approach that combines the speed of transient-expression technology with the efficacy of VLPs as antigen-presentation scaffolds. The company's manufacturing platform offers surge capacity and cost advantages over recombinant technologies, such as mammalian and insect cells. The technology can deliver a vaccine for testing in less than a month after the disclosure of genetic sequences from a pandemic strain, as demonstrated by the first doses of plant-made H1 VLP candidate vaccine that were produced at the onset of the 2009 pandemic. This production time frame thus has the potential to allow the vaccination of a population before the first wave of a pandemic strikes, and therefore to supply large volumes of vaccine to the world market.
In the past five years, Medicago has successfully developed this technology from the bench scale to the pilot scale according to cGMP. The compnay is now transitioning into large-scale manufacturing. Since 2008, Medicago has operated a 24,000-ft2 pilot clinical cGMP facility located in Québec City. This facility includes 10,000 ft2 of high-tech contained plant biomass production space, as well as 14,000 ft2 of cGMP manufacturing units for plant manipulation, product recovery, and purification.
In August 2010, Medicago began building a 97,000-ft2 cGMP facility in Research Triangle Park, North Carolina. Designed to produce 10 million doses of pandemic influenza vaccine per month, this facility will include a fully automated and contained plant biomass production facility and a state-of-the-art extraction-and-purification unit. On an annual basis, this facility would have a production capacity of 40 million doses of seasonal-trivalent influenza vaccine or 120 million doses of pandemic-influenza vaccine. The facility will be built in a 12-month timeframe at a cost of less than $35 million.
The capacity, speed of construction, and low cost of Medicago's large-scale facility clearly illustrates the promises of the company's manufacturing platform. These factors, in addition to the interim Phase II clinical results of Medicago's first enveloped VLP vaccine, could form the basis for a fast, high-quality, and cost-effective manufacturing solution to address the needs for improved 21st-century vaccines.
LOUIS-P. VÉZINA* is vice-president and chief scientific officer, MARC-ANDRÉ D'AOUST is director of innovative technology, NATHALIE LANDRY is vice-president of product development, MANON M.J. COUTURE is director of new products, NATHALIE CHARLAND is director of product portfolio, FRÉDÉRIC ORS is vice-president of business development, BRIGITTE BARBEAU is vice-president of operations, and ANDY J. SHELDON is president and chief executive officer, all at Medicago, 1020 Route de l'Eglise, Suite 600, Québec, QC, Canada, G1V 3V9, firstname.lastname@example.org.
1. O. Wyman and the Program for Appropriate Technology in Health, "Influenza Vaccine Strategies for Broad Global Access," (PATH, Washington, DC, 2007).
2. M.A. D'Aoust et al., Plant Biotechnol. J. 6 (9), 930–940 (2008).
3. N. Landry, PLoS ONE 5 (12), e15559 (2010).
4. E.V.L. Grgacic and D.A. Anderson, Methods 40 (1), 60–65 (2006).
5. President's Council of Advisors on Science and Technology, "Report to the President on Reengineering the Influenza Vaccine Production Enterprise to Meet the Challenges of Pandemic Influenza" (Washington, DC, August 2010).