Green Plants as Biofactories for Drugs

Genetic engineering has added a new pathway to the manufacturing of fine and specialty chemicals.¹ The idea is as simple as it is overwhelming. Instead of a factory-based chemical or biochemical reactor, you can custom-design a green plant to produce a desired product in its seed, leaves, fruit, root, or sap and grow acres of these in a farm field. Each plant functions independently as a miniature biofactory.

Despite its appeal, this idea has met with considerable opposition from public advocates, environmentalists, and food industry groups over health and safety issues. A major concern is keeping medical crops from contaminating food supplies. Past incidents of food crop contamination by foreign proteins that were not approved for food use (StarLink and Prodigene), resulted in significant costs and liabilities, even when the contamination was not a defined health threat.

These costs are borne not just by the technology developer, but also by businesses in the supply chain between the producer and the consumer. Unless technology developers address these concerns proactively, there is a real likelihood that regulations — rather than what is scientifically doable — will control which plants, technology, or product combinations are to be commercialized, how quickly and at what cost. Fortunately, there is evidence that developers are taking steps to prevent such events.

Plants have long been a source of human therapeutics — prominent examples include morphine from poppies, birth control agents from the Mexican yam, digitalis from the foxglove plant and paclitaxel from the Pacific yew tree. Producing foreign proteins was demonstrated less than 20 years ago, when human growth hormone was expressed in tobacco cells in 1986.² We now have a host of such proteins and chemical products — antibodies, vaccines, and other metabolites (Table 1).

Human therapeutics comprises the single largest category, by far, of transgenic plant products in development, with somewhat more than 150.⁴ We will look more closely at the products, technology, and production economics of biopharmaceuticals in green plants and at the risks and liabilities of commercializing the technology.

The advantages of plant biofactories relative to other bioproduction technologies have been detailed elsewhere.^5-7 Some of the key ones are:

• Lower capital and operating costs are possible by avoiding a costly bioreactor.

• Production can be scaled up with little additional investment by adding acreage.

• Grain that contains protein can be stored for long periods without loss of activity, thereby eliminating cold-storage maintenance facilities.

• Protein products fold and assemble correctly, just like their counterparts in mammalian cell culture.

• Plants do not host human or animal pathogens.

PRODUCTION PLATFORMS

There are many production platforms available for manufacturing biopharmaceuticals.

⁸

Microbial fermentation using

E. coli

or yeast and the culturing of animal (mammalian) cells are the most common for producing recombinant proteins. Transgenic plants and animals (including insects) are in development for producing recombinant proteins. The choice of platform for a given protein revolves around such technical issues as oxidative processing, protein folding, multimeric assembly, and glycosylation.

Recombinant proteins produced by mammalian cell culture are delivered glycosylated. (In the process of their biosynthesis, sugar molecules are added that affect their bioactivity.) This also occurs when plant cells produce proteins, but the resulting glycosylation of the protein produced is structurly different, which can lead to a different immune response and therapeutic activity in humans. Efforts are underway to engineer transgenic plants that yield the same glycosylation pattern as from animal cells, but the desired results have not yet been achieved. Plant Research International, in collaboration with Dow Chemical, is developing therapeutic proteins with mammalian-like glycan structures in transgenic plants.⁹

The choice of platform also can depend on the commercial volumes planned for a given protein. Microbial fermentation is suitable when the need is for annual volumes in the vicinity of multiple metric tons, whereas animal cell culture is more suitable where the volumes are significantly lower than 1,000 kg/yr. Transgenic plants can span both the low and the high end of the volume spectrum, with significant savings from not building the multiple bioreactors that are required for microbial and animal cell culture technology.

Plants can be employed in different ways as a production platform. Though much of the work underway has focused on technology for growing transgenic plants for pharmaceutical production in a somewhat isolated field or greenhouse setting, considerable work exists in open field cultivation involving tens to hundreds of acres. The secretion of products from the roots of transgenic tobacco plants in special bioreactors by Phytomedics is one example. Another is UniCrop's synthesis of recombinant proteins during the germination (sprouting) of dicotyledonous barley seeds in an airlift bioreactor.¹⁰ Aquatic transgenic plants and plant cells are being cultured in reactors in the same way that fermentation products are made.

Table 1. Categories of Transgenic Plant Products

The technology for producing proteins in plants, though still emerging, is clearly technically feasible. ProdiGene has commercialized two corn-derived research and diagnostic enzymes, trypsin and β-glucuronidase. No drugs from transgenic plants are on the market, but clinical trials are being conducted by Planet Biotechnology on tobacco-derived antibodies for fighting tooth decay and the common cold. Planet's tooth decay product, CaroRX, has received European approval and could be on the market as early as 2005. A treatment for lipase deficiency in multiple sclerosis patients, gastric lipase, is currently in phase 2 clinical trials by the French company Meristem.

CHOOSING A PLANT

Virtually any plant can serve the purpose of growing a therapeutic protein. Tobacco, alfalfa, and cereal crops like corn, rice, and barley are among the choices of plants to serve as biofactories for biopharmaceuticals (Table 1). Vegetables and fruits are being researched for edible vaccines. Even trees are candidates. The hevea tree, from which industrial rubber is derived, has been engineered to yield human serum albumin in the rubber latex that is harvested.

¹¹

The leaves of poplar trees have also been targeted for ligninase production at the University of Minnesota.

Biofactory plants are engineered to accumulate proteins in their grain or oilseed, in storage tissues (tubers, stems), or, in the case of tobacco and alfalfa, the leafy biomass. Every plant choice offers advantages and disadvantages. For cereal grains like corn, rice, and barley, there is a well-established commercial infrastructure for their growth, harvest, transportation, and storage. The same can be said for oilseeds like soybeans, canola, safflower, and tubers — all targeted for therapeutic protein production. Grains and oilseeds are readily stored for long periods without refrigeration and without loss of protein activity from degradation or enzyme hydrolysis. Soybeans and alfalfa, because they are legumes, have the advantage that they can fix nitrogen from the atmosphere and thus require less chemical input for growth.

Using tobacco as a production platform for pharmaceutical proteins is a well-established technology. Typically, a desired protein is expressed in the tobacco leaf, although expression via its roots (rhizosecretion) or other secretory pathways and in tobacco cells in suspension culture has been demonstrated. Because of its high biomass yield, harvesting proteins from tobacco requires handling and processing large quantities of leaf biomass in a timely fashion once the leaf is cut.

Alfalfa leaves are of particular interest as biofactories since they can yield up to 20% total protein on a dry weight basis, which is higher than corn or tobacco. Because multiple harvests per year are possible with alfalfa, the protein yield from the same amount of planted acreage well exceeds the yield of other crops. Medicago calculates that 9 kg/yr of purified protein can be produced from alfalfa using a single 1,300 m² greenhouse.¹²

Fruits and vegetables (plants that produce edible leaf, root, and fruit bodies) are choices for vaccines. Tomatoes, bananas, potatoes, peas, melons, cabbage, spinach, and broccoli have all been targeted as edible vaccine-delivery vehicles and as other therapeutic proteins. Controlled dosing of the protein is a key issue in this application, along with protecting the vaccine from thermal degradation in storage.

Protein yields that are possible from a given plant vary. Yield is usually defined as the percentage of the total soluble protein (%tsp) in a specific plant that is foreign to the plant. This yield is the heterologous protein for which the plant is genetically modified. It can vary from a low of 0.1% to 10% or more, depending on the plant, wherever the protein is expressed in the plant, the type of expression system employed, and other factors. The yield of protein per acre is the product of this percentage, times the percent of the plant biomass that is soluble protein, times the amount of biomass harvested per acre, and times the number of harvests made per year (typically one with cereal grains, two or more with tobacco, and three to nine with alfalfa). Table 2 estimates the acreages needed to produce a metric ton of protein from various crops using a range of typical parameters for %tsp and plant soluble protein levels.

Table 2. Acreage Requirements to Produce 1000 kg/yr of Protein

PRODUCTION OPERATIONS

For normal field crops, the steps of planting, growing, harvesting, transporting, storing, and processing the plant components are part of a well-defined and cost-effective commodity agriculture infrastructure. Plant biofactories for biopharmaceuticals are different. The acreage required for most biopharmaceutical products is relatively small and thus cannot benefit fully from commodity practices, which involve millions of acres. For this reason, new operating requirements and technologies must be superimposed on traditional agricultural production systems and practices.

For example, in open field operations involving transgenic plants, confinement measures are taken to isolate the plants from the surrounding environment, through the implementation of isolation distances or male sterile plants. Alternatively, if the acreage requirement is not too large, parts of the production operations can be totally contained — isolated from the environment by the use of green houses or underground mines. Whether confinement or containment is employed, added capital and operating costs are incurred.

When drugs are the product, rigorous quality assurance and monitoring systems are required in all aspects of seed and crop production, harvesting, post-harvest handling, storage, transportation, and beyond. These systems use dedicated equipment, often custom designed, and follow specific protocols — the agricultural equivalent of Good Manufacturing Practices (GMPs). As noted by Crosby, FDA has drafted Good Guidance Practices (GGPs) for plant biofactories.¹³ However, developers still face many issues, for instance, how agricultural harvesting equipment can be made cleanable in order to prevent batch-to-batch contamination.

Plant biotech companies have devised a variety of strategies to address the many issues that affect economic production, confinement, containment and downstream processing operations.¹⁴ Six examples are presented in Table 3. Each of the Table 3 strategies is an agriculturally based version of the fermentation and cell-culture operations typical of current biopharmaceutical practice. Biolex, the last entry in the table, is an exception. Its production strategy does not involve growing plants in an open field nor in a greenhouse or underground plot of soil.

In each system of Table 3, recovery of a crude extract from plant tissue is accomplished in a contained environment. The next task is to isolate a specific protein from the mixture and to purify it sufficiently for clinical trials or packaging for sale as a drug formulation. This assumes that the protein must be recovered in pure form for formulation. This is standard technology in drug firms, who routinely recover and purify proteins from mixtures of other proteins in microbial and mammalian cells.

Table 3. Typical Upstream Operational Strategies for Protein Biofactories

A special problem of tobacco as the host crop for a drug is that a large amount of biomass must be handled and processed to recover a relatively small amount of active product. This processing introduces the need for new equipment and operations for the transportation, handling, and size reduction of the biomass before extraction of the active ingredient. Companies like LSBC and ProdiGene have developed considerable experience with these processing steps that are not typical of drug industry practice.

COMMERCIALIZATION

The last hurdle is going through those drug industry practices necessary for a successful commercialization effort. These include demonstrating consistent lot-to-lot expression of the active ingredient, both plant-to-plant and over the host plant's life cycle, positive clinical trial results, drug formulation, regulatory approval, packaging, and marketing. These practices are not necessarily familiar to plant biotechnology startups and must either be learned from scratch or obtained via relationships with others knowledgeable in pharmaceutical commercialization efforts.

OPPORTUNITY TARGETS

Steiner found, by sampling 69 (of the 80 or so total) human therapeutic proteins on the market in 2003, that most sell at prices over $10,000/g and have a market demand of below 10 kg/yr.

¹⁵

Only a half dozen have a market demand exceeding 1,000 kg/yr. One-half of the 69 products are produced by fermentation, the rest being mammalian cell culture products. The few products with market volumes near 1,000 kg/yr sell at an average of $300/kg, yielding a market of $300 million/yr.

Table 1 alluded to a few therapeutic products in the pipeline for production via green plants. Table 4 expands the list for several key categories of products in feasibility studies. Many of these products are targeted for commercialization. Others are, or have been, the research focus of various universities and research institutes, often with corporate sponsorship.

Of the Table 4 categories, many developers view MAbs as the best opportunity for production via plant biofactories because of their structural complexity, large volume requirement, and market size. There are currently more than 200 MAbs under development by drug companies. Nearly all are being produced via mammalian cell culture. Among the factors driving this opportunity is the fact that most are going to be needed in large quantities at a time when there is a projected shortage of biopharmaceutical production capacity for such products.¹⁶ The stakes are high in that the five or six non-plant derived MAbs that have been commercialized so far represent markets of several hundred million dollars each. With a $20 billion projected market for MAbs by 2011, six to eight mammalian cell culture-derived product launches per year can be expected over the next seven years.¹²

Table 4. Biopharmaceuticals from Plant Biofactories

A key issue faced by developers of plant-based biopharmaceuticals is how quickly they can progress through the various stages of scale up. These stages extend from initial seed production to the establishment of stable transgenic plants in quantities necessary for full-scale field production, followed by validation of a purification scheme and the securing of regulatory approval for undertaking commercial production. Experts estimate that these steps could take up to four or more years unless developers find ways to accelerate their efforts. This is a somewhat longer time than for mammalian cell culture products, but much less time when compared to a transgenic animal system.⁴ Meanwhile, proponents of mammalian cell culture will be working to improve their time advantage, regulatory agencies' acceptance, and the cost competitiveness of current MAb production technology.

ECONOMICS

Generally, production costs are only a small contributor to the cost of commercializing a therapeutic protein. Even so, plant biofactories have an advantage over traditional fermentation or mammalian cell culture. A creditable case has been made that the capital cost of a facility to produce and purify 1,000 to 2,000 kg/yr of a therapeutic protein via mammalian cell culture is about $600 million. About half of this investment is for biosynthesizing the protein and the remainder for its recovery, separation from other proteins, and final purification. If one replaces the synthesis step with a plant biofactory, the capital of the biosynthesizing step drops to about $3 million, while the downstream recovery and purification facility costs remain unchanged.

¹⁵

Operating costs for the production and purification of MAbs via mammalian cell culture, depending on the scale of operations, can range from $100/g to more than $300/g.^{17, 18} About 40% of these costs are incurred for production of the protein and the rest for its recovery and purification. The cost of producing a drug in a transgenic crop, in contrast, is estimated at $12-15/g,¹³ which is higher than typical commodity corn production costs but much lower than a factory-based fermentation or cell culture system.

It is clear that there can be significant savings in both capital and operating costs by using a plant biofactory. However, these savings are likely to be reduced as developers begin to factor in the costs of addressing the many environmental, health, and safety concerns of the public. These costs are associated with such issues as the need for containment, confinement, testing, monitoring, special dedicated equipment, training, security, regulatory documentation and liability insurance. Manufacturing costs are only a small portion of the overall costs of bringing a therapeutic to market. The jury is still out as to whether such savings, of themselves, will motivate drug companies to adopt plants as drug biofactories.

RISKY BUSINESS

Reference was made earlier to the StarLink and ProdiGene instances of contamination involving food crops. Both involved proteins that have not been approved for food use and, as such, illustrate some of the risks and liabilities associated with plants as biofactories. Regulatory perceptions based on these historical incidents may be the greatest impediment to commercialization.

The StarLink case involved the inadvertent mixing of GMO corn containing a foreign pesticide protein (Cry9C) with non-GMO corn destined for food use. The foreign protein was approved for use as animal feed but not for people. In 2001 it was first discovered in a processed food product and in the general supply of corn grain. This created high liabilities and threw the commodity corn market into turmoil. This affected not only US domestic sales, but also corn export markets. The annual stake was 50 million metric tons of export. This involved hundreds of businesses around the world in trade, grain processing, transportation, food processing, restaurants, and more. The problem became an international issue with trade stoppages. The US entered into high-level negotiation with Japan, the largest export market. It took over a year for the turmoil to subside. Export levels from the US to Japan and Korea have not rebounded to earlier levels. Damages from the StarLink incident ran to hundreds of millions of dollars.

Before and since the StarLink incident, though no scientific evidence of harm to humans has been uncovered, expert panels have not ruled out the possibility of a human allergenicity risk. FDA does not consider the protein a threat to human health but has not undertaken to establish a safe threshold level for it.¹⁹

Routine testing for the presence of the protein is now standard procedure for corn arriving at dry grind corn mills and corn destined for certain overseas markets. Jim Bair, vice president of the North American Millers Association, said that small biotech companies should not be developing plant-based pharmaceuticals without first demonstrating a financial liability capability.²⁰ He questioned whether the FDA should establish a reasonable minimum acceptable level of the protein in corn targeted for food use.

The ProdiGene case was much smaller in scope. It arose when the company failed to adequately clear a field of transgenic corn engineered to produce a swine vaccine during the next growing season. Soybeans were planted in that field the next spring, but transgenic corn left in the field after the previous harvest also germinated, leading to commingling of transgenic corn and soybeans. The problem was contained when the soybeans were destroyed before being marketed, but repercussions from the incident continue.

The liabilities incurred by ProdiGene were high for a small company, involving costly USDA sanctions, the need to purchase and destroy the soybean crop, the dismissal of its CEO, and a furor that continues to impact public debate on GMO crops and the costs of doing business. Because the problem was found early and successfully managed, it is argued that the new regulations put in place as a result of StarLink worked.¹⁴ Whether true or not, both incidences sensitized developers of the technology, the grain industry, and food processors to the liabilities that can result from failure to monitor and manage the implementation of plant-derived drug products.

BIO has developed but not released its plan called a Containment Analysis and Critical Control Point (CACCP) Plan for PMP production. The central idea is to ensure that no plant material from plant-made pharmaceutical operations enters the food chain.²¹ And, the USDA is working on a revised review process for pharma applications.

REGULATION, ISOLATION, AND TECHNOLOGY

Equally important are the regulatory developments that will define the conditions under which the plant-made pharmaceuticals business can grow. Regulatory developments will not only define what is possible and what is not possible, but they will define a whole range of costs involved in proceeding. There are many unknowns in the regulatory environment, and these pose risks for companies whose participation downstream of the technology is necessary to develop a supply system. Issues related to liability will always be a risk factor.

Technology developers using a food or feed crop in open-field production may make their case based on production systems designed to keep pharmaceutical crops scrupulously separate from conventional crops, or propose defining permissible levels for the adventitious presences of selected pharma crops in conventional crops. We can be sure they will still face stiff opposition not only from consumer advocacy groups but, more importantly, from the US food industry and its affiliates. At risk is the market for US crops and food and the liabilities that go along with both human and management failures once the technology is implemented.

While developers of the technology have limited control over links in a functioning supply chain, they do have control over the choice of plant species and the production systems they develop. Even at this early stage in the evolution of this technology, it is clear that this choice is likely to be a key component in its commercial viability. And, the industry can and should more actively educate and promote the value of its products to the public and to patient advocacy groups.

Researchers and developers of plants as biofactories have used both food and feed crops like corn and canola, as well as plants that are not grown commercially for these uses. Some production systems are based on open-field production, others on production in closed systems like greenhouses, underground mines, or locations remote from food crops, and still others on hybrid systems that combine elements of both. In our opinion, the green plants and production systems that are most removed from conventional commercial food and feed crops are the most likely to succeed. Although using commercial crops seemed like a good idea because it could readily make use of well-established agricultural know-how and the various functioning farm production systems, regulatory developments and opposition from other industries makes this strategy seem less likely to succeed than when the idea was first proposed.

It is clear that many different activities, players, and businesses are impacted by any effort to successfully develop, test, and implement plant biofactories in a commercial context. All have different interests, assets, cost structures, competitive strategies, priorities, and appetites for risk. Technology developers will only be successful in commercialization if what they offer fits the strategies and needs of these stakeholders so that a supply chain can actually develop.

ACKNOWLEDGEMENTS

Our thanks go to Dr. Julio Baez, senior research fellow, FibroGen, South San Francisco, CA and Dr. François Arcand. Special thank you to Dr. Paul Arnison, FAAR Biotechnology Group, Orleans, Ontario, Canada, for permission to use his Plant Factories cartoon.

REFERENCES

1. DePalma A. Is a green plant in your manufacturing future?

BioPharm International

, 2003 Nov; 16(11):24-33.

2. Barta A., et al. The expression of a nopaline synthase human growth hormone chimaeric gene in transformed tobacco and sunflower callus tissue. Plant Mol. Biol. 1986; 6:347-357.

3. Flinn JE. Food, feed, pharma and specialty chemicals via transgenic plants. Proprietary report. 2000. Naperville IL.

4. Ma JK-C, et al. The production of recombinant pharmaceutical proteins in plants, Nature Reviews 2003 Oct; 4:794-805.

5. DePalma, op.cit.

6. Zavon JA., Flinn JE. Future of pharming involves a look at the big picture. Feedstuffs, 2003; 75(25).

7. Báez J. Transgenic protein production in plants. Presentation at Risk-assessment Symposium on Corn-Produced Pharmaceuticals and Industrial, 2004. April 22; Ames, IA.

8. Ibid.

9. Plant Research International. Press release. 2002 Nov 6. Available at URL: www.plant.wageningen-ur.nl.

10. Flinn, JE. Plants as biofactories: groups active globally in transgenic plants as bioreactors. Proprietary report. 2004. Naperville IL.

11. Yeang HY, et al. Rubber latex as an expression system for high value proteins. Engineering crop plants for industrial end uses, compiled by Rubber Research Institute of Malaysia. London: Portland Press; 1998, p.55-64

12. Martel M. Large-scale production of alfalfa biomass and its processing by wet fractionation. Presentation at Plant-Made Pharmaceuticals Conference, 2003 Mar. Quebec City, Quebec, Canada.

13. Crosby L. Commercial production of transgenic crops genetically engineered to produce pharmaceuticals, BioPharm International, 2002 April 15(4):60-67.

14. Arcand F and Arnison P. Development of novel protein production systems and economic opportunities & regulatory challenges for Canada, a discussion paper. 2004 April. Available at URL: www.cpmp2003.org15. Steiner UP. Business case for plant factories. Presentation at Plant-made Pharmaceuticals Conference, 2003 Mar. Quebec City, Quebec, Canada. Available at URL: www.cpmp2003.org

16. Price B. The capacity and capability "crunch": myth or reality, Presentation at Plant-Made Pharmaceuticals Conference, 2003 Mar. Quebec City, Canada.

17. Hood E, Woodard SL, Horn ME. Monoclonal antibody manufacturing in transgenic plants — myths and realities. Current Opinions in Biotechnology. 2002. 13:630-635.

18. Morrow JK Jr. Genetic Engineering News, 2002, 22 (7):1,35,39

19. Eric Flamm. Presentation at Risk Assessment Symposium. 2004 April 22. Ames IA.

20. Bair J. Presentation at Risk Assessment Symposium. 2004 April 22. Ames IA.

21. Phillips M. Presentation at Risk Assessment Symposium. 2004 April 22. Ames, IA.