The economic potential of plant-made pharmaceuticals in the manufacture of biologic pharmaceuticals

Abstract

Biologic pharmaceuticals have emerged to treat a number of diseases such as cancer and immune disorders. Although these drugs provide significant benefits, they have tended to be costly to produce. Production technologies that grow biologic drugs in plants offer the potential to lower those costs. This article uses a system of simulation models to compare the economics of plant made pharmaceuticals (PMP) to incumbent productions systems. Although it is found that significant cost savings are possible, RD investment in PMPs continue to be limited, which suggests that there are a number of obstacles that limit adoption.

INTRODUCTION

After years of anticipation the biotechnology industry has begun to deliver an array of promising pharmaceuticals. These biological pharmaceuticals (biologics) offer new mechanisms of treatment with application to a wide array of diseases including cancer and immune disorders, which have otherwise been difficult to treat.

A number of these biologics, especially monoclonal antibodies (mAbs), have demonstrated ongoing success. Table 1 lists the top selling biologics in 2008. In 2009, biotechnology medicines accounted for 15.5 per cent, or US$46.5 billion,¹ of the total $300 billion drug market.² New biologics continue to emerge with nine new mAb entities ( Table 2) and 14 recombinant proteins approved by the Food and Drug Administration (FDA) between 2006 and 2009.³

Table 1 Top biologic drugs of 2008⁴

Table 2 New mAb approvals 2006–2009³

Steadily increasing R&D activity signals that biologics will offer ongoing importance as a medical tool. In 2009, the biopharmaceutical industry invested $65.3 billion in R&D.⁵ Such investment has lead to a pipeline of over 600 new drugs in development⁶ and hundreds of biologics, either in clinical trials or awaiting action by the FDA.

Although these biologics have considerable potential for treating disease, their high price tag is often cited as a major drawback. Annual treatment costs with mAbs can exceed $30 000⁷ and in some cases double of that. As many diseases require prolonged treatments, total treatment costs can reach hundreds of thousands of dollars. These comparatively high prices have been, in part, tied to the manufacturing methods used to produce the biologics. mAbs, for example, are made with transgenic cells that are carefully grown in large bioreactors and ultimately harvested and purified in an injectable product. The necessary equipment in this process can lead to manufacturing facilities costing hundreds of millions of dollars and taking as many as 5 years to erect. Operating costs are similarly high, with the multiplication of mAbs requiring expensive inputs (for example, growth media) and careful monitoring.

Advocates of plant-made pharmaceutical (PMP) technology have claimed that plants offer an opportunity to lower the manufacturing costs and expand the production capacity of biologics. Instead of amplifying biologics in expensive bioreactors, they can be grown in plants circumventing much of the fixed production costs. Less fixed costs translate into the potential for large and scalable production capacity – requiring only that more acres be brought into production to meet additional demand.

State of the PMP industry

The potential cost savings associated with PMPs have encouraged more than 20 biotech companies to begin commercial R&D programs. These companies have used a breadth of plants to produce a wide range of therapeutics. However, industry participation clearly peaked around the turn of the century and has dropped off since. As an evidence of this, in 2001 there were 19 field trial permits issued for pharmaceutical and industrial proteins, whereas in 2010 only six were planted (see Figure 1) comprising a total of 28.3 acres.⁸ Part of the divestiture after 2001 has likely come from public concerns over using open pollinated food crops to produce pharmaceuticals. Following an incident where Prodigene's PMP corn germplasm was at risk of entering the food system, the Biotechnology Industry Association (BIO) and others called for a moratorium on using open pollinated crops such as corn, which had, until that point, received the majority of R&D attention.

Figure 1

 Planted field trials of pharmaceuticals, industrials and value-added proteins.⁸

Many firms using non-food crops (for example, tobacco) have also exited the industry. In fact, in recent years only a handful of firms have conducted field trials with pharmaceutical proteins, the most active being the Ventria Bioscience and the SemBioSys, which utilizes rice and safflower, respectively.⁹ Furthermore, since 2001, companies researching PMPs have shifted their interest away from new molecular entities such as mAbs, instead developing biologics that have long been available on the market (for example, lactoferrin and insulin).

The current levels of PMP production and R&D may suggest that PMPs do not yet demonstrate an optimal manufacturing solution for biologic pharmaceuticals. At the very least, the absence of PMPs in the marketplace creates an unclear picture of their economic benefit, and ultimately their potential for commercial success.

METHODS

This article compares the economics of manufacturing biologics in a PMP system to a more conventional fermentation system in order to evaluate the viability of PMPs. Although the production of biologics via fermentation is relatively well characterized and information about the facility/equipment can be obtained from existing facilities, this is not the case with PMP facilities. No analogous PMP facilities exist in the United States and there are few reliable estimates of the costs of producing PMPs. Commonly cited estimates suggest that PMPs do decrease the manufacturing costs of protein therapeutics ( Table 3), but the magnitude of that benefit varies significantly and key assumptions are seldom denoted.

Table 3 Cost comparison of mAb manufacturing costs ($ per gram)

Accordingly, we use a simulation model to compare two analogous systems producing identical biologics, in this case mAbs. mAbs were chosen (as opposed to fragments, enzymes, peptides, interferons and so on), as they represent a large portion of the R&D pursued by the pharmaceutical industry.

The controlled conditions of the production facility allow for a detailed static analysis, whereas the alternative environmental conditions (for example, location, weather and pests) of the field require variability to be considered. To account for these differences two separate simulation modeling methodologies were employed. SuperPro Designer by Intelligen was used as the modeling software for much of the analysis of the manufacturing facilities. Shanklin et al¹⁴ examined the competencies of SuperPro Designer in performing material balances, equipment and facility options and performing economic analysis, and found it well suited for these tasks. For those operations that occur outside of a controlled environment (that is, field production, transport), the study utilized Process and Economic Simulation of Identity Preservation, PRESIP – a model designed at the University of Missouri to test changes in operations in the corn supply chain including modules of seed production, farm production, grain elevator, feedmill and wetmill.¹⁵

The data to populate each model were obtained from companies with expertise in the respective manufacturing processes. Such information gave each piece of equipment a known cost, operating volume, operating time, energy consumption and so on. Similarly, as the raw materials pass through the equipment their dimensions are changed based on available research, reflecting any chemical, mechanical or thermal interactions. The final models were reviewed and validated by industry practitioners.

For simplicity, the R&D underlying the cell/seed lines is not included in this study. The evaluation is limited to only those steps necessary to amplify and purify the specific mAbs during a typical production year. Thus, the mammalian cell system begins as the cells leave the companies’ established cell bank to be multiplied in the operation's bioreactor train. Likewise, this analysis omits the cost of the developmental PMP seed breeding program. However, the seed breeding program operating concurrent to the PMP grain production is considered.

The scale of both the PMP and fermentation systems is simulated to produce 1000 kg of purified mAb per year – a scale approximating the demand of a successful drug.

RESULTS

PMP manufacturing process

The PMP system considered here utilizes corn as the production vehicle¹⁶ despite some skepticism of its relevance for PMP use in the future. The corn PMP production system differs from conventional corn farming and production occurs outside of the Corn Belt, where compliance with containment protocols is expected to be somewhat less difficult.

The corn PMP system begins with the specific genes being introduced into the plant germplasm. These transformed plants are developed in a breeding program involving selective inbreeding and backcrossing, which lasts for a number of years. Expression levels of different antibodies in stably transformed plants vary. Agracetus created a corn line they reported to have produced human antibodies with yields of 1.5 kg of pharmaceutical-quality protein per acre of corn. Monsanto presentations commonly cited the potential of 2 kg per acre.¹²

For the purpose of this article, we use a conservative expression level of 0.000323 g of mAb per gram of raw protein produced in the mature seed. Putting this into context, at 120 bushels of corn per acre, the protein expression is 1 kg per acre requiring a total of 1000 acres. However, inherent production risks require some degree of overproduction to mitigate potential yield and harvest losses – 20 per cent here. Over production is also needed to compensate for the 53 per cent extraction rate of the mAbs from the corn seed. In total, the facility requires the production of 2338 acres to yield 1000 kg of mAbs per year.

Compliance with genetic containment regulations increases this acreage footprint. Following BIO¹⁷ and United States Department of Agriculture Animal and Plant Health Inspection Service¹⁸ guidelines, all PMPs will require dense crop border rows and fallow rows. Assuming that 10 fields are used to produce the PMP corn, this will increase the acreage required for production to 2549 acres. In addition to this, 13 908 surrounding acres must be monitored to ensure that corn is not being produced and 27 755 surrounding acres must be monitored for seed production.

Other containment measures must also be taken including: (1) the dedication or extensive cleaning of certain equipment; (2) additional land costs necessary to buy isolation (in this case producing in California instead of in the Corn Belt); (3) the active monitoring of fields, farm operations, and grain under a traceable system; (4) the containerization of seed and grain and shipment to the facility for warehousing and cleaning; (5) the training of producers; and (6) incentive payments made to producers.

When all the field costs and containment costs are compiled, the per acre PMP production costs are assumed to be $1492 ( Table 4) and are roughly in line with other PMP production estimates.¹³ Such per acre costs are between two and three times the cost of commodity corn production.

Table 4 PMP field costs (120 bushel/acre)

In addition to the field production costs, the PMP production system also requires a dedicated and ongoing breeding program operating in parallel to the production of grain. The low-scale economies result in an annual seed production cost of $72 707, including the cost of field production and the management of the breeding program.

When all the acreage needed to supply the PMP facility is accounted for, the cost of corn is $3.64 million dollars, which equates to $15.6 per bushel of grain used at the processing facility ( Table 5).

Table 5 PMP production costs for the modeled facility

Once this grain is harvested, it is transported to the facility, unloaded, tested and warehoused. When needed for production, the grain is cleaned and reduced to meal, where it is soaked and agitated to begin the separation of the protein portion of the corn. The mAb recovery section begins with centrifugation to fractionate proteins from the other biomass. This is followed by microfiltration, diafiltration and then chromatography steps. After the final polishing filtration, the product moves into the formulation and packaging process.

PMP manufacturing costs

The facility is a significant investment requiring $12 million in equipment purchases. The total physical cost of the facility including all buildings is $64.4 million. After indirect costs are considered, the facility could reach $118.5 million. Despite the large capital outlay the facility is only the second largest operating expense of the PMP system, comprising 26 per cent of the annual bill ( Table 6). The most significant costs are process consumables, namely filter membranes and chromatography resins, totaling 39 per cent of the total operating cost of the facility. Chromatography resins comprise the majority of this.

Table 6 PMP operating costs

Raw materials are important, comprising 19 per cent of the PMP facility operating costs. Corn comprises 23 per cent of the total raw material bill with the agents used in purification (for example, Phosphate buffered saline) making the bulk of the cost.

The total annual cost of producing 1000 kg of mAbs via PMP production is $82 million, or $82 per gram. This figure is squarely in the middle of the industry PMP cost estimates found in Table 3.

Fermentation process and costs

The primary difference between the PMP system and the fermentation system is the growth of mAbs in bioreactors instead of in the field. As the fermentation system involves similar steps of recovery and purification, it is this bioreaction section that is expected to create the majority of the differences in cost and operation.

Bioreaction starts in a train of successively larger bioreactors employed to scale up the mAb producing cells to quantities appropriate for the largest bioreactors. These main bioreactors operate in a staggered fashion owing to the length of time required to grow the cells to maturity. The total final capacity of the facility includes eight bioreactors each having a capacity of 15 000 litre. These bioreactors are assumed to have conservative titers of 1.5 g/litre. Industry reports suggest titers are frequently in the 1–2 g/litre range and occasionally much higher.¹⁹

Once the cells have reached an appropriate concentration, the antibodies are extracted from the cells and debris. The recovery section begins with a centrifugation step followed by microfiltration, diafiltration and virus inactivation. The remaining recovery and purification procedures are largely similar to those of the PMP system, from a process standpoint. Although PMPs have unique complexities to be worked through, they also enjoy some advantages such as not requiring virus inactivation steps.²⁰ The costs reported in both cases assume an optimally performing process and do not account for unforeseen difficulties, which could be significant, especially with the unproven PMP process. Here, as expected, the final reclamation yield achieved by the fermentation facility is slightly higher than the PMP facility, at 56 per cent.

The capital cost of the mammalian cell culture facility is considerably higher than the PMP facility – by a factor of almost two. This added cost is largely associated with the fermentors needed to amplify the mAbs and leads to $23 million in total equipment cost, and $120 million in total physical facility cost. When indirect costs are considered the facility reaches $222 million.

The total annual operating costs at the fermentation facility are $122 million per year ( Table 7), with the facility costs at the fermentation facility comprising 33 per cent of the total. As, with the PMP facility, consumables were the largest cost source and similarly chromatography resins were the main components. Raw materials were also a significant cost, of which media for growing the mAb producing cells comprised 64 per cent.

Table 7 Fermentation facility operating costs

The total annual cost of producing 1000 kg of mAbs via fermentation is $122 million, or $122 a gram. This figure is on the low side of the cost estimates reported in Table 3, but well within the appropriate range.

Cost comparison

This study suggests that PMPs can lower the manufacturing costs of mABs. In fact, the per unit cost for the PMP system may be two-thirds as much as those of the mammalian facility ( Table 8), saving a 1000 kg operation $44 million per year. The capital investment required for facility construction is also much less for the PMP facility – by almost half.

Table 8 Comparative costs of PMP versus fermentation systems for producing 1000 kg of mAbs

These cost relationships are not expected to remain static over time, as the technologies mature at various rates. High levels of investment in fermentation technology have had the effect of increasing the efficiency of mAb production. This is especially evident in the rising mAb titers, which decrease the cost of fermentation relative to PMPs. This analysis finds that as fermentation titers approach 9 g/litre, the cost advantage of PMPs could be largely eroded at the considered scale. The lower the cost of fermentation, the larger the production scale must be for PMPs to be viable.

DISCUSSION

Even modest potential cost savings of PMPs would, intuitively, warrant higher levels of R&D investment and adoption than are currently being experienced. It stands to reason then, that despite the potential manufacturing cost savings, PMPs may not yet offer an optimal approach for mAb production. One reason for this may be uncertainties associated with PMPs that elevate the true cost of adoption.

Problems associated with PMPs

While the pharmaceutical industry is still coming to terms with the effective regulation of biologics, PMPs add a number of new questions and uncertainties including:

• Facility regulation: the pharmaceutical manufacturing facility is regulated based on standards of control and repeatability. This is relatively straightforward when the production process is constant, but more difficult when the production process is prone to changing environmental conditions. This difficulty could result in regulations being more costly to meet.

• Drug efficacy regulation: biologic pharmaceuticals can be complicated drugs and the production process can play a large role in efficacy. It is not certain that plants can be made to express the pharmaceutical proteins accurately and maintain that efficacy consistently over time. For example, any difficulty in maintaining correct glycosylation patterns and protein folding could create significant regulatory obstacles.

• Environmental containment: much has been written about the possibility of pharmaceutical corn entering the food chain, or pharmaceutical genes introgressing into the environment. The FDA regulates these events with a policy of zero tolerance. As meeting this is not functionally possible, significant risk of liability exists in the event of a breach.

• Popular opinion: PMP crops have generated both interest and opposition. Unpopular PMPs might be more difficult to produce and be an image liability for the drug manufacturer.

The future adoption of PMPs

PMP adoption is expected to occur only in an environment, where manufacturing costs are of key concern and the perception of PMP risks is decreased. Such a scenario may not be likely in all segments of the pharmaceutical industry. For example, with an estimated $1.2 billion price tag to bring one biologic to market²¹ and a minute chance of successfully completing clinical trials, manufacturing costs are only one of many considerations for pharmaceutical firms developing innovative drugs.

In cases where the drug company wishes to manufacture a new biologic in-house, investment in the manufacturing facility may need to be made as much as 5 years in advance of commercialization, to accommodate the long lead times in construction and regulatory approval. This creates a significant investment risk, as a firm could be caught with not only a drug that failed clinical trials but also costly manufacturing capacity that is not utilized.

In part to alleviate this problem, many biotech drug companies have opted to rent manufacturing capacity. While manufacturing costs may be higher than if done in-house, risk adjusted costs may be significantly less. Drugs under patent protection shift the focus of the pharmaceutical firm away from manufacturing cost and towards quick commercial introduction and minimizing risk associated with product delays, shortages, recalls, regulatory compliance and so on. Contract manufacturing is a dominant strategy for managing these issues except where capacity is already owned. As lead times associated with PMP breeding programs can be long and associated risks are high, some PMP systems may have significant disadvantages compared to conventional contract manufacturers.

The market for off-patent drugs is governed by different dynamics. Without patent protection, firms have more incentive to compete on manufacturing cost. Risk too may be somewhat less constraining, as issues of time and brand name are less important. Such a market seems better suited to PMP firms, and recently the PMP industry appears to be targeting established biologics such as insulin and lactoferrin.

The opportunity of off-patent mAbs is, however, complicated by regulations governing generic biologics (or biosimilars). The United States currently views biosimilars as non-equitable to the originals, and requires them to undergo many of the regulatory hurdles of the original. The high cost of compliance and long lead times limit the incentive for generic manufacturers to enter the market. However, the issue is being actively debated by industry and legislators, with signs pointing to development of an equitable pathway for generic biologic drugs. A clear regulatory path for biosimilars may open up demand for low-cost, biologic manufacturing capacity even as new biologics continue to put pressure on existing fermentation capacity.

The role that PMPs might play in the pharmaceutical industry will largely depend on the ability of PMPs to effectively produce and replicate complex biologics. It will also depend on the PMP industries’ ability to evolve and continue to maintain a cost advantage over fermentation systems. Difficulties with such issues could relegate PMPs to the production of comparably simple, large volume biologic products such as enzymes. In either capacity, PMPs have the potential to have a significant impact on the pharmaceutical industry provided that they encourage adoption by demonstrating ongoing safety and efficacy.