Chloroplasts: Transforming Biopharmaceutical Manufacturing

Since the 1980s, the use of human proteins to treat and cure diseases has comprised an increasingly significant sector of the pharmaceutical industry. In 2003, an estimated $39 billion of protein therapeutics were sold in the US. In the next ten years, domestic sales of protein pharmaceuticals are estimated to grow to $359 billion.¹

Karen K. Oishi

While traditional drugs are chemically synthesized, living organisms produce therapeutic proteins. The current generation of microbial and mammalian cell bioreactor-based systems will not be able to handle future demand. Expensive small-capacity units are too big a financial burden to produce and operate in the quantities needed. Manufacturing systems that use plants as bioreactors can meet the growing demand of the healthcare industry for therapeutic proteins at far lower cost.

Plant chloroplast transformation technology (CTT) can deliver quality proteins at low cost. Several of the advantages of this production platform are listed in Table 1. CTT in tobacco is now being validated as we explain below.

Chloroplasts

Green plant cells contain three distinct genetic entities; a single nuclear genome and two plastomic organelle genomes.

²

These organelles are the chloroplast and the mitochondria, each with multiple copies of a circular double-stranded DNA genome. Each organelle contains replication, transcription, translation and processing machinery that is functionally and physically separate from the rest of the cell. Since the late 1980s the light-harvesting chloroplast has been the focus of intense genetic engineering research, which has established that plants with transformed chloroplasts can potentially provide low-cost, stable, quality proteins.

^3-8

Photo courtesy of the Kentucky Research and Development Center, University of Kentucky, Lexington, KY.

The chloroplast is a metabolic organelle designed to synthesize abundant proteins for energy generation and plant growth and development. The chloroplast's genome is a self-replicating circular double-stranded DNA molecule ranging in size from 110 to 220 kilobase pairs depending on the species of plant. In the majority of plants, the chloroplast genome has two inverted regions.² With up to 100 genomes per chloroplast and up to 100 chloroplasts per cell, the integration of a transgene through homologous recombination into the inverted repeat region can generate up to 20,000 copies of the transgene per cell.

The chloroplast generates proteins in a manner similar to a prokaryotic cell, creating polycistronic messages and accepting large DNA inserts with multiple genes and repetitive elements without inducing silencing pathways. The chloroplast has been demonstrated to properly fold and assemble disulfide-bonded proteins, which can be sequestered away from cytosolic proteases during growth and development.⁹ However similar to bacterial expression systems, chloroplasts do not have the post-translational processing machinery for glycosylation of proteins.

Source of a Million Seeds

As shown in Figure 1, single or multiple genes can be integrated and expressed in transformed chloroplast as polycistronic mRNA. The translational regulatory sequence, or ribosome-binding site (RBS), then directs high levels of target protein synthesis. Dr. Henry Daniell, the technical founder of Chlorogen, developed universal transformation vectors, which direct integration of the transgene to the inverted repeated region of plant chloroplasts by homologous recombination.

¹⁰

Table 1. Chloroplast Transformation Technology Advantages

The chloroplast transformation-expression vector is introduced into the chloroplast by particle bombardment of plant tissue, or polyethylene glycol treatment of protoplast.^3,8 Transformation is initiated by transgene integration into a few copies of chloroplast genome within a single cell, followed by antibiotic selection and transgene replication during the next 15 to 20 cell divisions, which results in a homogeneous plastid genome population. Integration of the transgene expression cassette into one of the inverted repeat regions will initiate a copy number correction resulting in the duplication and insertion of another transgene cassette into the second inverted repeat region.

After two rounds of antibiotic selection, the resulting plant will contain 100% transformed chloroplasts (Figure 2). A working seed bank for production lines will be generated ten to eleven months post-bombardment and will be ready for field propagation. Because the chloroplast genome is maternally inherited in higher plants, 100% of the seed and seedlings will carry the transformed chloroplast genome in each subsequent generation.

More than 30 recombinant proteins have been expressed using CTT (Table 2), including both therapeutic proteins, and as proteins that improve agronomic traits in a variety of plants.

Figure 1. The Transformation-Expression Cassette for Single (X) or Multiple-Gene (X, Y, Z) Expression in Chloroplast

TOBACCO CTT

There are four major factors that impact the selection of a transgenic plant as a production system for biopharmaceuticals: cost of goods, scale-up capacity, regulatory hurdles, and time to market. The success of plant-made biopharmaceuticals or industrial protein production in tobacco is particularly sensitive to these four factors and can be evaluated through the analysis of discrete steps in the manufacturing process. The primary development model for producing human recombinant proteins in tobacco is based on green tissue (leaf) biomass processing. The major steps in green tissue processing are illustrated in Figure 3.

Master seed bank production is less intensive with tobacco than with other plants. Systems that rely on sub-cloning (for example, in potatoes and sugarcane) require time to amplify cloning stocks and are susceptible to the consequences of pest attacks or disease devastation as cloning stocks are restored. Seed-based systems rely on agronomic seed-banking practices that have been practiced for thousands of years. These provide a ready source of production plants, rapid scale-up ability, and a stable recombinant gene platform. However, a plant such as corn, which is open-pollinated, requires significant field acreage to generate production seed and large storage facilities due to the relatively large size of corn kernels. Tobacco can produce as many as one million seeds per plant, which allows for master seed bank production in a contained and controlled greenhouse environment at a very low cost.¹⁵

Field production of biomass (the target protein) provides the primary cost advantage over traditional production technologies such as mammalian cell culture and bacterial fermentation.¹⁶ The lack of facilities, bioreactors, and process technicians significantly reduces upstream costs. However, issues with gene containment — specifically pollen flow to conventional plants — have caused concern about field production of biopharmaceuticals. Using chloroplast transformation technology eliminates the concern about pollen flow because chloroplast genomes are inherited maternally and are not found in pollen. Therefore, tobacco allows for outstanding field containment, resulting in lower regulatory-related costs. The high expression levels observed in chloroplast-transformed tobacco also impacts the overall cost.

Figure 2. Generation of Tobacco Plants with Transformed Chloroplast

Chlorogen's current transgenic production cultivar is adapted to a wide range of soil types and climates. Field tobacco practices differ from traditional practices due to the extraordinarily high plant densities and growing and harvesting procedures. A field study investigated staggered biomass yields for seedlings transplanted at weekly intervals. Yields were not significantly different for the May 15 through July 15 release dates.¹⁷ The current production cultivar possesses a rapid growth rate, excellent ratooning (regeneration) efficiency, and low total alkaloid concentrations. These characteristics provide tremendous flexibility with respect to biomass generation and processing.

Harvest

Harvesting of tobacco is relatively straightforward and is generally accomplished with the use of sickle-bar or rotating-disk type harvesters. The cut biomass (approximately 10 inches above ground level) is then typically loaded into transportation bins via moving-bed conveyers. This equipment is relatively low-tech and is readily available from farm equipment vendors. Plants are harvested at the growth stage that results in maximum protein yields, generally at the pre-button stage just before flowering. Although chloroplast genes are inherited maternally and not through pollen, the harvesting of pre-flowering plants reduces even the smallest potential of cross-pollination with conventional tobacco cultivars.

A fertilizer regime is used to encourage the regeneration of multiple secondary shoots every four weeks, providing a maximum of four harvests during the growing season. Three harvests during a single growing season generated biomass up to 297,000 lbs/acre.

Table 2. Proteins Expressed Using Tobacco Transformation Chloroplast Technology4, 9-14

Biomass Production Yield

Average yields for a growing season of high-density tobacco fields are approximately 55 tons/acre or 110,000 lbs/acre. Averaged total soluble protein (TSP) per unit biomass can be used to predict the total crude target-protein yield on an acreage basis. A predictive protein-yield graph based on a range of expression levels and acreage yields is shown in Figure 4.

Published reports describing protein expression levels for a number of protein classes have identified ranges between 2 and 40% total soluble protein (TSP), depending on the specific protein.^4,9,10,12-14 If one assumes a nominal expression level of 8% TSP and a seasonal crop of 55 tons/acre, the crude protein yield specified in Figure 4 is approximately 97 lbs/acre. This is an extremely large target-protein-to-biomass ratio, which translates into several advantages during processing. A 20% TSP-producing cultivar can generate up to 265 pounds of crude recombinant protein per acre.

A high target-protein-to-biomass ratio allows for a reduction in the size of upstream processing equipment. For example, compare processing tobacco with a 1:100 target protein-to-biomass ratio versus a 1:1000 ratio. Within equal time periods, we can expect a tenfold reduction in the size of the equipment used in upstream processing, resulting in lower capital and overall costs. The initial concentration of target protein also impacts downstream processing in terms of volume reduction steps and ratios of transgene protein to host and environmental contaminating molecules — another saving.

Figure 3. Modular Components of the Manufacturing Process Using Green Tobacco Biomass

Pre-Extraction

Green biomass may need to be manipulated to prepare the material for extraction. For green biomass, pre-extraction processing generally refers to comminution or pulverizing the material into specific-sized particles that will enhance the subsequent extraction process. Comminution is generally required when increased surface area or reduced diffusion distances are required. Typical comminution methods include cutting, grinding, and milling.

¹⁴

Extraction

Extraction is based on two complementary processes running parallel to each other.

• Rinsing target protein out of disintegrated chloroplasts and plant cells

• Dissolution of target protein from plant material by diffusion (the resulting extraction solution is called the miscella)

Extraction processes can be further divided into two major groups.

• Processes that result in a concentration equilibrium between solute and target protein

• Processes in which the target protein is extracted exhaustively.¹⁷ There are many methods of extraction including ultrasonic, electromagnetic, and supercritical fluid extraction,¹⁵ but the two primary methods of extraction from tobacco are countercurrent perfusion-extraction and compression sieving.

Greenhouse production of transgenic tobacco transplants using a traditional “float tray” system

Countercurrent perfusion-extraction requires the addition of an externally applied buffer or solvent, which is added to comminuted solid matter in a continuous process. Proteins are extracted by the liquid solvent phase flowing countercurrent to the solids. In this type of extraction, there is a continuous gradient of target protein in both the solvent and solids. The solids removed from the extraction machinery at one end have low target-protein concentrations, while the solvent removed at the opposite end has a high target-protein concentration. However, the addition of solvent increases the process volume, which requires larger equipment and adds to raw material costs.

The most cost-effective extraction method is compression sieving or screw press. Plant material is fed into the screw press in a continuous manner, and compression under high pressure ruptures the plant cell walls. The compressed material is moved along a sieve plate, which separates the plant-derived extract from solids. This type of extraction equipment is low-tech and readily available from equipment vendors. Screw press extraction uses natural plant liquids, eliminating the need to pay for additional solvents.

Clarification

Once the target protein has been extracted from tobacco, residual plant solids, bioburden, oils, and other extraneous matter must be removed. Typically, the two methods employed are centrifugation and filtration. Centrifugation generally requires a larger capital investment than filtration and needs to be carefully evaluated during process modeling. Using fine and ultra-fine membrane filtration, and of course filtration with diatomaceous earth (DE) or polymeric filters can be cost effective with higher efficiencies and lower capital costs. The target protein can be separated by differential solubility by selectively tagging the protein, which induces aggregation or binding to specific media.

Table 3. Selected Component Composition of Tobacco (Green Tissue)

Purification

Generally, the first step in purification is capture. This is normally accomplished through the use of specific chromatography resins such as ion exchange or affinity type resins. Resins are used in a variety of configurations such as the standard downflow system, radial chromatography, or streamline expanded-bed technology from Amersham Biosciences. The choice of capture system typically depends on the protein of interest and the cost of the resin.

Purification may employ several further chromatography steps including ion exchange, affinity, hydrophobic interaction, low-pressure reverse phase, or size exclusion. Each subsequent chromatography step is targeted for the removal of specific classes of contaminants such as host proteins, host compounds, host DNA, endotoxin, and environmental contaminants such as residual herbicides, pesticides, and bacterial and fungal toxins. Attention to interactions between target protein and plant phenolics is especially important during the purification process. A list of some of the contaminating molecules is shown in Table 3. Water, which is 80 to 90% of the wet weight, is not counted in this analysis of the dry solids.

Figure 4. Crude Protein Yield Based on Tobacco Biomass per Acre, and Expressions Levels Measured as % Total Soluble Protein (TSP)

Process ECONOMICS

The manufacture of tobacco-produced proteins can be cost effective due to the low cost of field production and upstream processing, relative to the capital burden and operations-intensive processes used in traditional technologies such as mammalian cell culture. Various companies suggest that by eliminating expensive upstream equipment, the cost of therapeutic proteins made in transgenic systems may be an order of magnitude lower than the same proteins produced with traditional technologies. Others have reviewed process models, predicting that there may only be a savings of 10 to 15% over traditional cell culture systems. This is because upstream costs of traditional technologies are said to contribute less than one-half of the overall production costs, so an order of magnitude reduction in cost-of-goods using transgenic technology is not reasonable.

¹⁸

However, the traditional model uses many standard unit operation criteria for both transgenic and cell culture systems that are not always applicable.

There are many ways to reduce the downstream costs associated with protein manufacturing. Often, when cost allocations are applied to upstream comparative construction and operations models, support or ancillary cost factors are not subtracted. Applying modular construction can reduce capital costs. Modular cleanroom components are pre-fabricated and are generally assembled within a low-cost shell structure. Although the modular components may contribute a higher cost per square foot to the cleanroom structure, the overall facility cost is generally less due to the simplified construction and readily available materials.

Unlike many traditional specific-use biopharmaceutical facilities that must be constructed as fully functional and uniform monolithic structures, modular components can be installed rapidly (weeks). This allows for just-in-time capital expenditures to track the actual need. Modular cleanroom components usually come with validation documentation, are uniform in construction and finish, and can be serviced (lights, filters) from outside the clean environment. This reduces validation, assembly time, and operational cost. A lower-cost facility results in smaller depreciation allocated to an operating budget, and this can greatly decrease the cost-of-goods. Costs associated with upstream quality control, such as in-process testing, environmental monitoring and testing, component validation and re-validation, are not always subtracted from the upstream cost models when reviewing transgenic technologies, but they should be.

Other innovative manufacturing approaches such as disposable process materials; outsourced raw material (buffers and media) production; on-line, time-of-use mixing of buffers from concentrates; and fewer holding steps (tanks and piping) during the overall manufacturing process can result in significantly lower downstream costs. Other innovations such as the development of moderate-cost, custom affinity resins can also reduce costs. Many of these cost-saving technologies are difficult to incorporate into traditional processes due to standardized processes required for licensed products, technology risk avoidance, and simple inertia. On the other hand, the embryonic stage of transgenic plant protein development is well suited for the incorporation of innovative technologies, and protein production from tobacco using CTT will benefit from such innovation.

Reduced Regulatory Risk

There is uncertainty about the level of regulatory risk because no plant-made therapeutic has yet been commercialized. The review process, however, will follow the existing regulatory framework developed for approval of biopharmaceuticals and, separately, genetically engineered plants. FDA and USDA issued a joint draft guidance document in September 2002, "Drugs, Biologics, and Medical Devices Derived from Bioengineered Plants for Use in Humans and Animals,"

¹⁸

providing an interpretation of the regulations and outlining the points to address when requesting product approval.

Confinement is a key concern in the regulation and public acceptance of genetically modified crops, particularly those expressing pharmaceuticals and industrial proteins in food and feed crops that are not intended to be food or feed. The concern is that the gene or tissue expressing the protein will escape during field production through pollen movement, seed dispersal, or germination and growth of plants from released seeds in subsequent growing seasons, and will then co-mingle with food and feed crops.

In a chloroplast system, the confinement risks are minimized by several inherent properties of the technology and process. Tobacco chloroplasts are maternally inherited, thereby mitigating the risk of gene transfer via pollen to other tobacco or compatible wild species. Because leaves are used for processing, only seed used for proliferation is required, and the seed set is so prolific (up to one million seeds per plant) that sufficient amounts can be produced in enclosed greenhouses, eliminating pollen and seed disbursal. In addition, leaf material is harvested prior to flowering and seed set — which eliminates the potential for pollen movement, seed dispersal, or growth from the released seeds.

With regard to contamination of food or feed, tobacco is not grown for human food or animal feed, so it is unlikely that co-mingling would occur. Physical isolation from tobacco grown for smoking is easily achieved because tobacco is grown only in limited geographic areas, and with such a high-yielding system, commercial quantities of protein can be produced from a small acreage, resulting in lower overall environmental exposure. These inherent properties of tobacco cultivation contribute to a low regulatory risk and lead to enhanced public acceptance.^3,18,19

References

1. Global Industry Analysts, Inc. Global Marketing Data Compendium; Rockville MD. 2003.

2. Palmer JD. Comparative organization of chloroplast genomes. Annu. Rev. Genet. 1985; 19:325-354.

3. Daniell H, Ruiz ON, Dhingra A. Chloroplast genetic engineering to improve agronomic traits. Methods in Molecular Biology 2004; 286:111-137.

4. Kumar S, Dhingra A, Daniell H. Manipulation of gene expression facilitates cotton plastid transformation of cotton by somatic embryogenesis & maternal inheritance of transgenes. Plant Mol. Biol. 2004; in press.

5. Ruf S, Hermann M, Berger IJ, Carrer H, Bock R. Stable genetic transformation of tomato plastids and expression of a foreign protein in fruit. Nat Biotechnol. 2001; 19:870-5.

6. Sidorov VA, Kasten D, Pang SZ, Hajdukiewicz PT, Staub JM, Nehra NS. Technical advance: stable chloroplast transformation in potato: use of green fluorescent protein as a plastid marker. Plant J. 1999; 19:209-216.

7. Boynton JE, Gilham NW, Harris EH, Hosler JP, Johnson AM, Jones AR, et al. Chloroplast transformation in Chlamydomonas with high velocity microprojectiles. Science 1988; 240:1534-1538.

8. Daniell H, Vivekananda J, Nielsen BL, Ye GN, Tewari, KK, Sanford JC. Transient foreign gene expression in chloroplast of cultured tobacco cells after biolistic delivery of chloroplast vector, Proc. Nat'l Acad Sci. 1990; 87:88-92.

9. Daniell H. Medical Molecular Pharming: Therapeutic recombinant antibodies, biopharmaceuticals, and edible vaccines in transgenic plants engineered via the chloroplast genome. In R. M. Goodman (Ed.), Encyclopedia of Plant and Crop Science. New York: Marcel Decker; 2004.p. 704-710. .

10. Daniell H, Khan MS, Allision L. Milestones in chloroplast genetic engineering: an environmentally friendly era in biotechnology. Trends Plant Sci. 2002; 7: 84-91.

11. Daniell H. Foreign gene expression in chloroplasts of higher plants mediated by tungsten particle bombardment. Methods Enzymol. 1993; 217:536-56.

12. Daniell H. Molecular strategies for gene containment in transgenic crops, Nat Biotechnol. 2002; 20:581-586.

13. Kumar S, Dhingra A, Daniell H. Plastid expressed betaine aldehyde dehydrogenase gene in carrot-cultured cells, roots and leaves confers enhanced salt tolerance. Plant Physiol. 2004 September issue.

14. Ruiz ON, Hussein, HS, Terry N, Daniell H. Phytoremediation of organomercurial compounds via chloroplast genetic engineering. Plant Physiol. 2003; 132:1344-1352.

15. Davis D and Nelson M, eds. Tobacco - production, chemistry and technology, Malden (MA). Blackwell Science; 1999.

16. Glacken M. Plant transgenics vs. CHO bioreactor culture: an objective comparison for monoclonal antibody production, 8th International Conference on Antibody Production and Downstream Processing; 2002; San Diego, CA.

17. Mulesky M. Chlorogen Inc. internal data. 2003 Sept.

18. FDA, Guidance for Industry: Drugs, biologics, and medical devices derived from bioengineered plants for use in humans and animals. Federal Register 2002; 67: 57828, Docket No. 02D-0324.

19. Palme, H, and Winberg G., Arch Pharm, 1916; 254:537.

20. Goldschmidt-Clermont M. Transgenic expression of aminoglycoside adenine transferase in the chloroplast: a selectable marker of site-directed transformation of chlamydomonas. Nuc Acid Res 1991; 19:4083-4089.