After many starts and stops, hype and disappointment, foreign protein expression in plants is now routine and biopharmaceuticals produced in green plants will soon be with us.
After many starts and stops, hype and disappointment, foreign protein expression in plants is now routine and biopharmaceuticals produced in green plants will soon be with us. Plants have been part of medicine for thousands of years and pharmaceuticals still rely heavily on plant-based materials: more than half of our pharmacopoeia is extracted from plants or related to plant compounds.
Plants have been part of medicine for thousands of years, but bioengineering plants to produce therapeutic proteins is changing the shape of biotech. Plant-made biopharmaceuticals are cheaper to develop and cheaper to produce. Will they take over the field?
Plant-made biopharmaceuticals (PMBs), however, break the traditional mold of plants and medicines in two ways: plants are bioengineered, and the resulting protein drugs are foreign to it - just as they are to recombinant yeast or bacteria or to transfected mammalian cells.
PMBs are preparing the way for the "next wave" in biotech which will parallel the evolution of biotechnology itself. According to a University of Florida study, more than 300 PMBs are under investigation worldwide.1 Most of this work goes on at academic or government institutions. Eventually, this undercurrent of technology will provide the foundation for PMB companies, just as it did for biotech 20 years ago.
PMBs could have a big impact on markets for therapeutic recombinant proteins. The argument for PMBs is compelling. Datamonitor, a business information company, estimates the world market for therapeutic proteins at about $30 billion, and predicts demand will rise to $59 billion by 2010.2 Leading the way are monoclonal antibodies, with 2003 sales exceeding $3 billion and growth approaching 60% yearly. The market for therapeutic vaccines, although of much smaller volume, is also growing quickly, with 61% growth on 2002 sales of $31 million. Table 1, from Datamonitor's report, lists leading marketed therapeutic proteins.
Moreover, what PMB companies lack in capital funding they more than make up for with innovation and vision. Larry Grill, Ph.D., chief scientific officer at Vacaville, CA-based Large Scale Biology Corp. (LSBC), predicts "at least a couple of dozen" approved plant-based biotech products by 2010. Most managers at PMB firms share his optimism.
Optimistically, PMBs' lower capital costs and greater production flexibility could simply steal the show. Realistically, PMBs eventually may dominate specific therapeutic protein markets or monopolize biogenerics. Or they could fail.
Figure 1. These rice plants, developed by Ventria Bioscience, will grow into mature plants and express therapeutic proteins.
On paper at least, PMBs offer numerous advantages over traditional fermentation and cell culture. Among these are
Compared to traditional biomanufacturing, PMBs offer lower capital investment and lower ongoing manufacturing costs. Brick-and-mortar plants take from two to seven years to build and cost anywhere from $100 million to $500 million, and this only includes the cost of the plant, not ongoing manufacturing and other functions. The Dow Chemical Co. and Sigma Aldrich estimate capital reduction at 75% to 80%, with manufacturing costs slashed 50% to 60%. Neil Cowen, Ph.D., vice president for product development at Epicyte Pharmaceutical in San Diego believes ongoing cost savings for manufacturing might even reach 75%.
Figure 2. Ventria's rice-based PMBs illustrate the basic method of protein expression in plants.
Cost advantages for downstream PMB operations depend on the final dosage form. As with all therapies, injectibles tend to be high-cost products, while edible or topical PMBs will be cheaper to develop.
Still, potentially significant cost savings are possible downstream. Because of the large biomass expected from PMB processes, batches will be large enough for developers to consider continuous or semi-continuous processing, which biotech companies traditionally avoid. A side benefit of large batches is lower quality assurance and quality control requirements since these activities are performed per batch, not per volume or weight of product.
PMB manufacturing resembles food processing more than pharmaceutical production. So where do good agricultural practices (GAP) end and good manufacturing practices (GMP) begin? According to Andrew Baum, chief executive officer of SemBioSys Genetics and head of the Biotechnology Industry Organization's working group on PMBs, GAPs apply while the crop is in the field. GMPs, however, extend across the entire enterprise, overlapping with and supplementing GAPs. For example, the United States Department of Agriculture (USDA) governs isolation of engineered crops, whereas FDA regulates materials and purity, equipment, and manufacturing - encompassing everything from seed stock to packaging.
Figure 3. SemBioSys fuses proteins with safflower seed oil bodies.
USDA and FDA, which jointly regulate PMB projects, state that the field is part of the manufacturing "facility" and require that PMB operations be treated like any other pharmaceutical process. Therefore, the appropriate safety and quality controls, dedicated equipment if necessary, contamination safeguards, and trained staff to assure safe, consistent product must be in place. "Companies involved in this work must apply the same types of regulatory and quality standards, from planting to harvesting, as they would for a fermentation," says Epicyte's Cowen.
LSBC claims to be the only "plant" company with large-scale GMP extraction capability. As a result, LSBC extracts proteins for other PMB firms on a contract basis. Says LSBC's Grill, "We can extract 6,000 pounds per hour of plant tissue, for 24 hours per day, under GMP."
Because they come from plants, PMBs are believed to be inherently safer than recombinant proteins from micro-organisms or cells. All mammalian cells carry viral genomes that spawn potentially pathogenic viruses. By contrast, food plant viruses are not pathogenic to humans.
Incredible, Edible Vaccines
Much has been written about post-translational modifications (PTMs) - glycosylations, phosphorylations, and others - the finishing steps in eukaryotic protein-making. No two organisms glycosylate in quite the same way, and some, like bacteria, don't glycosylate at all. PTM differences have not proven to be a source of undesirable immunogenecity in PMBs. Epicyte, for example, has failed to condition an immune response in mice with plant-derived mouse antibodies. Similarly, circulating half-life, clearance rate, and other pharmacokinetic properties do not differ between native and plant-derived proteins. "Worries over glycosylation differences were overrated," says Grill.
Since plants cannot transmit prions, the current scourge of the beef industry, replacing animal-derived proteins with plant-derived copies seems promising. For example, LSBC makes the protease inhibitor aprotinin in a tobacco-like plant instead of extracting it from cow lungs, the traditional source.
Environmental concerns with PMBs include altered plants contaminating wild strains and human exposure to plant material containing potent drugs. Government agencies are well aware of these issues and appear to have them under control. In August, USDA's Animal and Plant Health Inspection Service (APHIS) amended its regulations for genetically engineered plants that make drugs, requiring a standard permit for field testing. Previously, APHIS allowed transgenic plant work under an expedited permit.
Figure 4. Minichromosome technology from Chomatin allows multiple gene insertion via a single operation.
PMB companies generally operate in a manner that protects both workers and the environment. For example, Ventria Bioscience exclusively uses self-pollinating crops because their reproductive mechanism keeps genes within the plant, a plus for what Chief Executive Officer Scott Deeter calls "environmental stewardship." The USDA-mandated setback distances for barley and rice, which self-pollinate, are 50 and 100 feet, respectively, compared to one mile for wind-pollinated maize.
PMBs are generally easier to express upstream - and about as difficult to purify downstream - as conventional biopharmaceuticals. For some dosage forms, such as oral and topical, PMB purity is primarily driven not by safety concerns, since food crops are already recognized as safe, but for purposes of standardizing formulations. Lower sterility requirements and the generally large quantities of PMBs available from food crops could improve downstream economics, allowing development of high-dose, non-sterile formulations that would be too expensive for proteins manufactured from cell culture or fermentation.
Downstream innovations will almost certainly arise from the huge biomass volumes expected from manufacturing-scale PMB operations and the potential for continuous processing. Another source will be expression methods that engineer ease of purification into raw products. A third avenue, process naivete, stems from the fact that PMB developers must experiment with purification techniques in ways that traditional bioprocessors do not.
As Chenming "Mike" Zhang, Ph.D., assistant professor of biological systems engineering at Virginia Tech, points out, "[The PMB] industry needs to develop unique purification processes for each product. Generic processes are not realistic for plant-derived proteins."
Zhang's specialty, aqueous two-phase extraction (ATPE), is an example of a separation technique considered feasible for PMBs but which is generally untouchable for conventional biopharmaceuticals. A single ATPE operation produces 87% pure egg white lysozyme from ground tobacco leaves using two aqueous phases - one pure water and the other spiked with additives like poly(ethylene glycol), dextrin, and salts.3 The only drawback is that Zhang never knows which aqueous layer will dissolve the product (and therefore, where the protein will end up), a rather small disadvantage considering the complexity of typical downstream protein purifications and ATPE's ease of scale-up.
Table 1. Classes and values of marketed therapeutic proteins in 2000 and 2001.
SemBioSys' expression of proteins attached to oil bodies of safflower seeds is a good example of expression-level purification. After harvesting and grinding, company engineers use continuous centrifuges to isolate proteins attached to oil bodies. "By addressing the entire manufacturing issue, we offer the same expression benefits of other systems but change the economics of purification," says Chief Executive Officer Andrew Baum. SemBioSys calls its process the StratoSome Biologics Production System.
Since StratoSome proteins are inexpensive, they could potentially be used orally or topically with minimal processing and without regard to unfavorable pharmacokinetics. "In cases where the oil-body assembly is the product," says Baum, "you're done." For injectibles, proteins are cleaved from the oil bodies and subjected to another centrifugation, followed by column chromatography.
A related oil-body technology, StrataCapture Purification, uses affinity-modified oil bodies to purify proteins generated through more conventional expression systems.
Technologies that help PMB developers quickly demonstrate feasibility also can aid in developing orphan drugs or advancing personalized medicine.
LSBC recently completed a sixteen-patient phase 1 non-Hodgkin's lymphoma vaccine study in which personalized vaccines were generated from individual patients' B-cell surface epitopes. After cloning into a viral vector, the product was expressed in Nicotiana benthamiana, a greenhouse plant suitable for rapid transfection and protein turnover. A similar vaccine from hybridoma cells works, but its cost is prohibitive and its production timeline - six months to a year - is impractical. "With our Geneware technology, you can use a small number of plants and within a few weeks obtain enough protein - less than two milligrams - to immunize each patient against their own cancer," says Grill.
Road Blocks to PMBs
Another of LSBC's "orphan" projects is a treatment for Fabry disease, whose sufferers lack the enzyme alpha galactosidase. Genzyme currently makes alpha galactosidase in chinese hamster ovary cells; the drug costs hundreds of thousands of dollars per year. A PMB version could cost substantially less and still be profitable, says Grill. "FDA is supportive of having these types of drugs made in plants because that's a way to get the cost down," he adds.
Many scientific, regulatory, economic, and production barriers to PMB development remain.
Expression levels depend on many factors, including expression system. The most common expression elements are Agrobacterium tumefaciens, genetically engineered bacteria that cause tumor-like growths on plants while transferring genes. Another technique is to shoot DNA-coated gold microparticles or silicon carbide fibers into plant cells. Genes that hit their target - the nucleus - are taken up during mitosis.
Expression techniques aim for protein expression in abundant, convenient tissues like seeds or leaves. Some companies target even more specific tissues, for example SemBioSys and oil bodies.
According to Scott Deeter of Ventria, volume can never compensate for poor expression levels. "Expression levels drive production costs for plant-based production, just as they do for cell culture and fermentation."
In other words, plants' native protein expression levels do not guarantee transgenic results. Soybeans, otherwise phenomenal protein expressors, don't process foreign genes well. That is why Ventria works feverishly to increase protein expression levels in rice from twelve grams per kilogram of biomass to about fifteen grams.
"A tenfold improvement in expression means planting and processing facilities need only be one-tenth the size," Deeter notes.
Timing of the plant growth cycle is an underappreciated risk. While not as costly and unpredictable as building a half million square foot facility, PMB projects are limited by the time it takes plants to reach maturity, as well as exogenous factors such as climate. If all goes well, the time was well invested. If not, developers are back to square one.
"Nobody wants to wait a year or two only to learn that an expression system is not producing the material they want," says Mich Hein, Ph.D., chief executive officer of Chromatin, based in Chicago, IL. "Because of inherent slowness in plants, it takes about two years to isolate the genes, design a transformation vector, get the genes into the plants, and multiply the plants," he says. "The equivalent processes using mammalian cell culture take about nine months, so plants are disadvantaged by about one year."
Despite the overall rosy scientific picture, the business concept of PMBs remains unproven. "The suitability of green plants for protein manufacture is still not 100% resolved," says Andrew Baum of SemBioSys. "The economics seem compelling, and every trend points towards feasibility," he says. "But skeptics and outsiders still have questions."
Financing may be PMB's biggest woe and the source of their ultimate undoing. PMBs began attracting scientific attention, Baum notes, during a brutal investment climate.
Although regulators in the United States and Europe encourage PMB development, companies tremble at the vast regulatory unknown. This uncertainty is probably what keeps away many risk-averse large pharmas, a potential source of PMB financing. "Large pharmaceutical companies don't want to be the first to take plant-produced recombinant proteins through the regulatory process," says Larry Grill of LSBC. "We've pretty much had to carry out all our science and development on our own nickel."
However, companies like Dow Plant Biopharmaceuticals, based in Midland, MI, a business unit of The Dow Chemical Co., are getting involved. Dow, which doesn't develop its own PMBs, entered the plant biopharmaceutical business about three years ago as a contract manufacturer. In this capacity, Dow helps companies bridge the material gap between development and preclinical and early-phase clinical testing.
Because plants are slow-growing, getting proteins from them takes about a year, which is normally longer than developers can wait. Dow shortens the development timeline by using plant tissue culture, rather than intact plants, to prepare evaluation protein rapidly. While producing small quantities from plant tissue culture, the company uses the same gene transfer methods to produce plants suitable for large-scale production. "This allows us to fill the gap in the beginning of a contract manufacturing relationship," says Kerr.
Since entering the business, Dow has signed manufacturing agreements with Epicyte and Centocor and entered a research collaboration with Plant Research Institute of the Netherlands.
"Plant transgenics makes sense with proteins that don't express well in traditional systems, are given in large doses, or for which cost of production would make them too expensive to bring to market," notes Kerr Anderson, Ph.D., technical director for Dow Plant Biopharmaceuticals.
"Despite the hype, bioengineered plants will not solve all protein manufacturing problems and address all the inherent issues," says Kan Wang, director of the Plant Transformation Center at the University of Iowa's Plant Science Institute. "Plants are just one possibility among many for manufacturing therapeutic proteins."
(1) University of Florida Institute of Food and Agricultural Science. Critical thinking in plant biotechnology. Available from: URL:
(2) Datamonitor. Therapeutic proteins: strategic market analysis and forecasts to 2010. New York; 2002 Aug. Report reference code DMHC 1803.
(3) Balasubramaniam D, Wilkinson C, Van Cott K, Zhang C. Tobacco protein separation by aqueous two-phase extraction. Journal of Chromatography A 2003, 989:119-129. BPI