Plant-produced biopharmaceuticals: A case of technical developments driving clinical deployment

See allHide authors and affiliations

Science  16 Sep 2016:
Vol. 353, Issue 6305, pp. 1237-1240
DOI: 10.1126/science.aaf6638


The ability to express heterologous proteins in plants has led to the concept of using plants as “bioreactors” or “biofactories” for the production of pharmaceutical proteins. Although initial studies were promising, the pathway to commercialization and deployment in a clinical setting has proven to be a somewhat rocky road. This Review examines the technical developments that have led to the current increase in interest in the use of plants for the production of pharmaceutical proteins, particularly in the context of clinical trials.

The development in the early 1980s of effective means of expressing heterologous (nonplant) proteins in plants offered the prospect of using plants as bioreactors for the production of pharmaceuticals at reduced cost for both human and animal health. In addition, plant cell cultures are not susceptible to mammalian viral pathogens and, conversely, plant viruses do not infect human cells (1). An early indication that mammalian proteins produced in plants would retain their desired pharmaceutical properties was provided by the expression, in transgenic tobacco plants, of a correctly assembled and functional human antibody (2). This initial success was quickly followed by the successful expression, also in transgenic plants, of hepatitis B virus-like particles (VLPs) (Fig. 1) (3), more complex antibody derivatives such as secretory immunoglobulin A (sIgA) (4), and orally immunogenic proteins (5). Simultaneously, vectors based on plant viruses were used to express antigenic peptides, usually as fusions to the viral coat protein. In several instances, assembled particles were able to protect animals from challenge with the pathogen from which the epitope was derived (6, 7). This article will review how these initially promising studies have progressed during the past 25 to 30 years toward their ultimate goal of producing practical pharmaceuticals, what technical advances have helped toward this goal, and what still remains to be achieved.

Fig. 1 The green vaccine machine.

This image shows an artistic representation of the use of plants to produce vaccines; in this case, virus-like particles (VLPs). Here, the VLPs and the leaf are not on the same scale. [Credit: © Eva Thuenemann]

Advances in expression technologies

Stable transformation technologies that lead to regenerated plants play a seminal role in the development of plants as bioreactors, have the potential for large-scale production, and are applicable to a wide range of plant species. Still, these technologies have the disadvantage of being labor-intensive and time-consuming. Thus, much of the early work in this area focused on the production of molecules that had previously been made successfully in other systems. Additionally, the long production timelines mean that these technologies are not suitable for rapidly producing pharmaceuticals to combat emerging diseases.

“The major shift in gene delivery technology has involved the use of Agrobacterium tumefaciens to deliver genes to somatic tissue via infiltration…”

The major shift in gene delivery technology has involved the use of Agrobacterium tumefaciens to deliver genes to somatic tissue via infiltration, as well as reliance on transient expression to produce the desired proteins in a matter of days. Because the introduced sequence is not heritable, this is essentially a batch process; scaling up production involves infiltrating more tissue. Thus, it is essential to maximize expression, either by using replicating viral RNAs to increase mRNA levels and/or ensuring that mRNA is maximally stabilized and translated (8, 9). Additionally, it was important to develop technologies that could be applied on a large scale. The plant species routinely used for transient expression is Nicotiana benthamiana—a wild relative of tobacco originally from Australia—which has a fast growth rate, is particularly amenable to infiltration, and appears to have a defective RNA silencing system (10). Yields using this approach can reach 1 g of product per kilogram of leaves within 5 to 7 days, though the levels are protein-dependent.

As an alternative to using whole plants, there has been considerable progress recently in the use of plant cell cultures such as tobacco BY-2 cells and carrot suspension cultures (11). Such suspension approaches offer several potential advantages in terms of the control of growth conditions and a greater similarity to more conventional methods of pharmaceutical production, such as mammalian, insect, and yeast cell cultures. Conversely, the cost of these approaches may obviate the financial advantages of plant-based production. To reduce this disadvantage, a fully disposable system for the large-scale production of plant cell suspension cultures (11) has been developed. Methods have also been developed for the production of cell packs from suspension cultures (12). In addition to using suspension cultures from higher plants, the possibility of using cultures of simpler or unicellular plants has been explored.

The above may give the impression that, in terms of the methods used, the field of plant-based expression is quite fragmented. Whereas this has certainly been the case in the past, there has been a convergence within the past few years, and though the details of the various technologies may differ, many of the tools can be used interchangeably. Thus, many Agrobacterium binary vectors can be used for both stable and transient expression as well as to introduce sequences into either whole plants or suspension cultures. Small-scale transient expression in leaves is often used to express a new molecule before embarking on the production of transgenic lines (Fig. 2).

Fig. 2 The process of vacuum infiltration on a small scale.

In this case, a single N. benthamiana plant is inverted and the aerial parts immersed in a suspension of A. tumefaciens in a vacuum desiccator. The pressure above the suspension is reduced using a vacuum pump (seen in the background), which results in the loss of air from the intercellular spaces. When the vacuum is released, this air is replaced by the A. tumefaciens suspension.

Major categories of pharmaceutical proteins successfully produced in plants

Until now, the main classes of pharmaceutical proteins produced in plants have been antibodies, subunit vaccines and VLPs, and therapeutic enzymes. However, other types of molecules, such as toxins, have also been expressed. A wide range of antibodies and antibody derivatives—including sIgA and IgM, as well as numerous IgG molecules (1315)—has been successfully expressed. Likewise, numerous VLPs derived from a variety of mammalian viruses have been produced. These include VLPs consisting of only a single type of capsid protein, more complex VLPs consisting of several types of capsid proteins, VLPs that require a maturation step for their formation, and VLPs from enveloped viruses (16). Though there are few examples in which VLPs expressed in different systems have been directly compared, work with hepatitis B VLPs indicates that plants are particularly adept at assembling high-quality homogeneous particles (17). Several therapeutic enzymes aimed at treating metabolic disorders such as Gaucher disease and Fabry disease have also been successfully expressed (18, 19). In 2012, taliglucerase alfa for the treatment of Gaucher disease became the first plant-produced biopharmaceutical to obtain approval for human administration by the Food and Drug Administration (20).

Clinical development of plant-produced biopharmaceuticals

Vaccines intended for oral administration were the first generation of plant-produced biopharmaceutical proteins to reach clinical development. The concept of edible vaccines was proposed at a time when production systems were dominated by transgenic plants, the idea being that antigens could be expressed in transgenic fruits, tubers, or seeds and administered orally without purification, alleviating the need for parenteral administration, formulation, and (possibly) refrigeration. Edible vaccines thus had the potential to be affordable, easily administered, and accessible to the developing world.

The first human study of a plant-produced pharmaceutical evaluated the immune responses to ingestion of transgenic potato tubers containing Escherichia coli heat-labile enterotoxin B subunit (21). This study demonstrated, for the first time, that oral administration of antigens in edible plants induces mucosal and systemic immune responses. However, the results highlighted the short duration of the immune response. Similar modest and short-lived responses were also observed in subsequent studies [for reviews of this early work, see (68)]. Despite the early challenges faced by edible vaccines, the idea of delivering biotherapeutics in edible plant tissues remains alive. The most recent proposal, passive oral immunotherapy, builds on the capacity of plants to produce sIgA, a primary component of mucosal immunity (22).

Parenteral administration in humans of a plant-produced biotherapeutic was first performed at the turn of the millennium by Large Scale Biology Corporation (23), who used transient expression via a replicating vector. This phase 1 clinical study, conducted over several years, evaluated purified idiotypic vaccines [in the form of a ~30-kDa single-chain variable fragment (scFv) of the idiotype] for the treatment of non-Hodgkin’s lymphoma. Purification of the antigens allowed the group to control the dosage of antigen administered. It also enabled the evaluation of adjuvanted vaccine doses, using granulocyte-macrophage colony-stimulating factor as an adjuvant.

Protalix Biotherapeutics has invested in the alternative strategy of using plant cell cultures as a means of producing biopharmaceuticals, including taliglucerase alfa (β-glucocerebrosidase). Optimal bioactivity of therapeutic glucocerebrosidase requires that the enzyme is taken up by macrophages via binding to their mannose receptors. Hence, the glycosylated enzyme needs to carry N-glycans with terminal mannose residues (24). Shaaltiel et al. achieved in planta maturation of N-glycans to a paucimannose structure (a short form with two terminal mannoses) by targeting the enzyme to storage vacuoles. This approach was efficient and eliminated the need for postpurification maturation of N-glycan (18), thus improving the cost-efficiency of the process compared with that of imiglucerase (Cerezyme), the competitor CHO-produced β-glucocerebrosidase. Pivotal clinical trials of taliglucerase alfa showed no serious adverse events and demonstrated efficacy results similar to those of other enzyme replacement therapies for Gaucher disease (25).

The swine (H1N1) influenza outbreak that highlighted the limitations of traditional influenza vaccine–manufacturing technologies led Medicago to use the speed of transient expression to produce a pandemic H5N1 vaccine based on influenza VLPs (Fig. 3). The adjuvanted vaccine was well tolerated at all doses, and immunogenicity studies revealed a definite dose response (26). Subsequent clinical studies with monovalent and quadrivalent influenza vaccines have confirmed the strong antibody responses to influenza and have shown that plant-derived influenza VLPs elicit pronounced, long-lasting, and cross-reactive polyfunctional CD4+ T cell responses, even in absence of an adjuvant (27, 28). Of the 280 subjects receiving either the H1 or H5 VLP vaccine in phase 1 and 2 studies, none have developed an allergic or hypersensitive response, despite the plant-specific N-glycosylation profile on the viral surface glycoprotein, hemagglutinin, used as an antigen (29). An ongoing phase 2 clinical trial of the quadrivalent seasonal influenza vaccine in healthy adults and elderly populations is aiming to define the optimal dose, establish potential competitive advantages over current influenza vaccines in terms of protection against heterologous strains, and support the design of future studies (30).

Fig. 3 Transient expression for the large-scale manufacturing of vaccines.

Medicago’s facility in North Carolina is scaled to produce 10 million doses of pandemic influenza vaccine per month. (Left) Plants are grown in a greenhouse before infiltration. (Right) An automated device handles tables of 128 plants for infiltration.

Another important clinical development came with the study of transiently expressed idiotypic vaccines (31). Building on the previous demonstrations of the capacity of plants to produce personalized vaccines against B cell follicular lymphoma, and learning from the challenges faced with the first-generation products (23), Icon Genetics studied a second generation of idiotypic vaccines, in the form of full-size human IgG1 coupled to keyhole limpet hemocyanin, produced with an improved transient expression platform (8). For this specific product, the platform provided the robustness, yields, speed, cost-effectiveness, and quality needed for production and human administration of individualized tumor-targeted idiotypic vaccines. The study demonstrated that the idiotypic vaccine was well tolerated, as no severe adverse events related to vaccination were observed. Of the patients who could be evaluated at the end of the study, 82% displayed a vaccine-induced, idiotype-specific cellular and/or humoral immune response (31).

Although several plant-produced antibodies have been assessed in animal studies, only a few have made it to clinical development. The field of plant- and algae-produced antibodies was recently reviewed by Yusibov et al. (32). Early clinical developments of plant-made antibodies include a hybrid secretory IgA-IgG antibody targeting Streptococcus mutans for the treatment of dental caries (33) and Avicidin, a full-length IgG specific for epithelial cell adhesion molecule (EpCAM) (a marker of colorectal cancer), produced in maize (34). Both reached phase 2 trials. In addition, the plant-expressed IgG, 2G12, which is intended for topical application to prevent the transmission of HIV infections, has also undergone phase 1 clinical testing (35).

Whereas the first clinical studies were performed with antibodies produced in transgenic plants—the only available technology at the time—transient expression is now generally used, not only because of the advantages described above but also because glycoengineered N. benthamiana plants have been developed that enable the production of antibodies with improved effector functions. The ZMapp cocktail of antibodies developed for the treatment of Ebola virus infection and administered to 15 patients under compassionate-use protocols during the recent Ebola virus outbreak in West Africa represents a good example of plant-produced antibodies with improved potency via glycoengineering. The cocktail, resulting from a collaboration involving Mapp Biopharmaceutical and LeafBio (San Diego, California), Defyrus (Toronto, Canada), the U.S. government, and the Public Health Agency of Canada, is produced in N. benthamiana plants with knocked-down plant-specific glycosyltransferases (β1,2-xylosyltransferase and α1,3-fucosyltransferase) (36, 37), as studies with previous generations of anti-Ebola antibodies and antibody cocktails have indicated that nonfucosylated antibodies show increased potency, most likely via improved effector functions (38, 39). Earlier this year, LeafBio, the commercial arm of Mapp Biopharmaceutical, released the results of the first human study with the ZMapp treatment. The results obtained showed that the product was well tolerated and that the mortality rate in the treated group appeared to be 40% lower than in the nontreated participants, although the difference did not reach statistical significance in such a small sample size.

Future prospects

To paraphrase the 17th-century English poet John Donne, no branch of science is an island entire of itself. Thus, the direction that the field of plant-based pharmaceutical production takes in the future is likely to be governed not just by developments within the field but also by advances in other areas of bioscience. This has been illustrated recently by genome editing technologies, particularly the CRISPR-Cas9 system, which was originally developed in nonplant organisms but is now being increasingly deployed to modify the host plants used for expression (40). Developments in entirely different branches of science are also likely to influence the attractiveness of plants as an expression system (e.g., the recent developments in light-emitting diode technology that have reduced the cost of providing the lighting necessary for efficient plant growth). The clinical developments mentioned in this Review (see Table 1 for a summary) clearly demonstrate the capacity of plant-based production systems to deliver effective biopharmaceuticals of high quality for a wide diversity of products, ranging from prophylactic vaccines to therapeutic human enzyme replacement therapies. Today, the speed of deployment enabled by transient expression technologies, combined with the application of the above transformational technologies, is further accelerating platform and product developments to meet the growing need for novel and innovative therapeutics.

Table 1 Clinical development of plant-produced biopharmaceuticals mentioned in this article.

View this table:

References and Notes

  1. Acknowledgments: G.P.L. is supported by the U.K. Biotechnological and Biological Sciences Research Council Institute Strategic Programme grant “Understanding and Exploiting Plant and Microbial Secondary Metabolism” (BB/J004596/1) and the John Innes Foundation. G.P.L. also declares that he is a named inventor on granted patent WO 29087391 A1, which describes a transient expression system used for some of the work described in this Review. M.-A.D. is an employee of Medicago and is a co-inventor on several patent applications and granted patents pertaining to the platform and products of the company.
View Abstract

Navigate This Article