- Email: [email protected]
Expression of a human anti-rabies virus monoclonal antibody in tobacco cell culture
BBRC Biochemical and Biophysical Research Communications 345 (2006) 602–607 www.elsevier.com/locate/ybbrc
Expression of a human anti-rabies virus monoclonal antibody in tobacco cell culture q Loı¨c Ste´phane Girard a, Marzena Jolanta Fabis a, Maryse Bastin b, Didier Courtois b, Vincent Pe´tiard b,*, Hilary Koprowski a,* a
Thomas Jefferson University, 1020 Locust Street, Biotechnology Foundation, Philadelphia, PA 19107, USA b Centre de Recherche Nestle´-Tours, 101 Avenue G. Eiffel, BP 49716, 37097 Tours Cedex 2, France Received 15 March 2006 Available online 19 April 2006
Abstract A Nicotiana tabacum cv. Xanthi cell culture was initiated from a transgenic plant expressing a human anti-rabies virus monoclonal antibody. Within 3 months, plant cell suspension cultures were established and recombinant protein expression was examined. The antibody was stably produced during culture growth. ELISA, protein G purification, Western blotting, and neutralization assay confirmed that the antibody was fully processed, with association of light and heavy-chains, and that it was able to bind and neutralize rabies virus. Quantification of antibody production in plant cell suspension culture revealed 30 lg/g of cell dry weight for the highest-producing culture (0.5 mg/L), 3 times higher than from the original transgenic plant. The same production level was observed 3 months after cell culture initiation. Plant cell suspension cultures were successfully grown in a new disposable plastic bioreactor, with a growth rate and production level similar to that of cultures in Erlenmeyer flasks. Ó 2006 Elsevier Inc. All rights reserved. Keywords: Antibody; Disposable plastic bioreactors; Plant cell culture; Rabies; Recombinant; Xanthi
Plant-based systems are increasingly used for the production of recombinant proteins, because of several advantages over other production systems, such as the ability to carry out necessary post-translational modifications not available in bacterial systems, as well as greater safety and lower production costs than those provided by animal-based systems. Plant-based technology has recently been extensively reviewed, with a full description of plants commonly used, a listing of proteins expressed thus far [1–5], and the potential economic superiority of these systems for the recombi-
q
Abbreviations: cv., cultivar; DW, dry weight; ER, endoplasmic reticulum; FW, fresh weight; HC, heavy-chain; LC, light-chain; mAb, monoclonal antibody. * Corresponding authors. Fax: +33 0 2 47 49 14 14 (V. Pe´tiard); +1 215 923 6795 (H. Koprowski). E-mail addresses: [email protected] (V. Pe´tiard), hilary.koprowski@jefferson.edu (H. Koprowski). 0006-291X/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2006.03.219
nant protein production [5,6]. Protein stability during production and storage is an important determinant of the production cost. Indeed, degradation by proteases can be avoided by the addition of a ‘‘Lys-Asp-Glu-Leu’’ (KDEL) signal to the protein sequence, which leads to storage of the heterologous protein inside the endoplasmic reticulum (ER). This ER targeting also results in a glycosylation pattern closer to the human one, by preventing addition of plant-specific glycans, fucose, and xylose to the protein [7] thus avoiding a potential source of unwanted immunogenicity in the recombinant protein while injected [8]. Many proteins relevant in human disease treatment have already been expressed in plant-based systems (reviewed in [2,9]), mainly in the form of edible vaccines [10,11] and monoclonal antibodies (mAb) [5,12]. A safe and cost-effective source of mAb is particularly desirable in the treatment of human rabies virus infection, since current protocols rely heavily on anti-rabies immunoglobulins prepared mainly
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
603
from horse serum; not only are such immunoglobulins expensive to prepare, accounting for their limited supply worldwide, but they present the risk of cross-contamination with known or unknown horse pathogens [13]. Considering that rabies virus exposure causes an economic burden in the US of about $1 billion per year for vaccination and immunoglobulin administration [14], a plantbased system of immunoglobulin production, as recently developed in tobacco plants [15], is attractive. However, mAb expression levels in field plants vary depending on plant maturation stage and climatic influences [16 and personal observations], whereas plant cell cultures are based on undifferentiated cells growing in a fully controlled environment, ensuring a constant level of expression and preventing contamination of the environment. Here, we demonstrate for the first time the expression of a human anti-rabies virus glycoprotein mAb in plant cell culture. A low-cost biomass scale-up was performed with a new disposable bioreactor.
Lyophilized cells [100 g dry weight (DW), equivalent to 1.45 kg FW] were mixed with 1.5 L of extraction buffer (20 mM sodium phosphate, pH 7.2) and disrupted using a French press. Soluble proteins were recovered by centrifugation (6000g, 20 min, 4 °C) and filtered on fiberglass (from 2.7- to 0.22-lm pore sizes). Cross-filtration on a 30-kDa membrane was used to concentrate the supernatant to 300 mL. Soluble protein extracts ¨ kta purifier equipped with a HiTrapä protein G were purified with an A high-performance affinity column (Amersham Biosciences) according to the manufacturer’s recommendations. mAb was eluted with 0.1 M glycine–HCl (pH 2.7) and pH was neutralized with 1 M Tris–HCl (pH 9.0).
Materials and methods
ELISA
Plant material A stable (T3) transgenic tobacco plant (Nicotiana tabacum cv. Xanthi) expressing an anti-rabies virus mAb [15] was used and is designated herein as R12. Briefly, the heavy-chain (HC) sequence, fused with a sequence encoding the KDEL ER retention signal, is placed under control of the cauliflower mosaic virus 35S promoter with duplicated upstream B domains (Ca2p); the light-chain (LC) sequence is under the control of the potato proteinase inhibitor II promoter (Pin2p) [15]. A well-established wild-type BY2 tobacco cell suspension (N. tabacum) was used for comparison to newly established tobacco cell suspensions.
Binding of transgenic mAb to rabies virus was assessed by ELISA as described [18], with the following modifications: 3 lg/mL of ERA rabies virus was used to coat the plates, followed by loading with 95 lL of serial 3-fold dilutions of soluble extracted proteins obtained from 7 mg of lyophilized plant cell extract disrupted in 1 mL of sodium phosphate extraction buffer. Human anti-rabies virus mAb (1 lg/mL) was used as a positive control. Recombinant bound mAb were detected by successive additions of biotin-conjugated anti-human IgG1 (1:1000; BD PharMingen, San Diego, CA), avidin alkaline phosphatase-conjugated antibody (Sigma), and p-nitrophenyl phosphate (Sigma) as substrate. Absorbance was read at 405 nm.
Plant cell culture
In vitro neutralization assay
Callus induction. Primary explants were cut from 3-week-old wild-type (WT) or R12 transgenic seedlings and placed on Murashige and Skoogbased medium, supplemented with 1 mg/L naphthalene acetic acid (NAA), 0.1 mg/L kinetin, 30 g/L sucrose, and 8 g/L agarose (MAK), under photoperiod conditions (16 h light/8 h dark), at 25 °C. The medium was autoclaved before explants for 20 min at 115 °C. Resulting calli were subcultured monthly on Petri dishes. Initiation of suspension culture. Randomly picked calli (20 g) were inoculated into 50 mL MAK liquid medium and placed under photoperiod conditions at 25 °C on an orbital shaker (100 rpm). After 2 weeks, both transformed and non-transformed tobacco cell lines were grown in 250-mL Erlenmeyer flasks containing 100 mL MAK medium and subcultured every 20 days at an initial density of 40 g of cells [fresh weight (FW)]/L.
Neutralizing activity of purified antibodies was assessed using a standard 20-h rapid fluorescent focus inhibition test (RFFIT) as described [18]. Briefly, diluted purified mAb produced by plant cell culture was incubated with rabies virus (CVS-11, CVS-N2C or ERA). Baby hamster kidney (BHK) cells were added to plates, incubated, and fixed. FITC-anti-rabies nucleoprotein monoclonal globulin (Centocor, Malvern, PA) was added and plates were examined under a fluorescence microscope to detect infected cells. The original mAb produced in human hybridoma cells was used as positive control. Both mAb were used at the same concentration to compare their neutralization ability.
1/10,000 dilution] conjugated to horseradish peroxidase (Jackson ImmunoResearch). Whole-plant-derived mAb was used as a positive control. Suspension culture in disposable bioreactors A new type of plastic bag-based bioreactor (Nestle´ Research Center, Tours) was used to scale-up the cell culture from 500 mL to 10 L. After sterilization of the disposable bioreactor, sterile medium was inserted into the bioreactor and cells were inoculated [40 g/L (FW)]. Small-scale purification of plant cell-derived antibodies
Results Morphological features of plant cell culture
Protein extraction and analysis Protein extraction, SDS–PAGE, and Western blotting were carried out as described [17], with the following modifications. Twenty-day suspension cultures were filtered (50-lm nylon mesh) to separate cells and medium. Cell disruption was performed on 1 g FW of cells, using a stainless-steel mixer mill (Retsch, Model MM301, Newtown, PA) at 30 hits/s, for 2 min. The resulting mash was homogenized with 300 lL DB2X buffer (40 mM Tris–HCl, pH 8.0, 20% glycerol, 2% SDS, and 4% b-mercaptoethanol) containing 0.1% Triton X-100 and a complete protease inhibitor mixture (Roche Diagnostics). For Western blot analysis, HC and LC were detected with goat anti-human antibodies [Fc c- and F(ab 0 )2 fragment specific, respectively,
Primary calli initiated from WT tobacco seedlings and R12 transgenic seedlings expressing human anti-rabies virus mAb were grown on MAK medium and after 3 weeks appear morphologically similar (Fig. 1A). The yellow color turned to green after 3 months. Western blot analysis of eight WT and eight R12 calli revealed HC and LC protein production only in the R12 calli (data not shown). Three weeks after inoculation, WT and R12 suspension cultures appeared healthy in MAK medium (Fig. 1B), with the same yellow color and the same morphology as a stabi-
604
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
Fig. 1. Morphological features of calli and plant cell suspension cultures grown in MAK medium. (A) Three-week-old calli, obtained by callogenesis on Nicotiana tabacum (cv. Xanthi) plant leaves grown on MAK medium (left, wild-type calli; right, R12 calli). (B) Comparison of wild-type (left) and R12 transformed (right) 10-day-old Nicotiana tabacum cv. Xanthi suspension cell culture. (C) Filtered cells from 20-day-old plant cell suspension. Note the greenish color and clumped growth.
lized N. tabacum BY2 culture. One month later, cells began to grow as clumps and turned greenish (Fig. 1C); cell suspension cultures retained these features during further subcultures. Heavy- and light-chain expression in plant suspension cultures and pellets Western blot analysis (Fig. 2B) of soluble protein cell extracts from five R12 suspension cultures and their remaining pellets (after a second extraction) revealed HC and LC bands migrating at 55 and 30 kDa, respectively. The LC appeared to be expressed as two different forms (Fig. 2B, lanes 2 and 4). Two other bands were also detected in the pellet at 60 and 75 kDa (Fig. 2B). The 60-kDa band appears to be a dimer of LC since this band was not detected by HC-specific antibody (data not shown). The 75-kDa band might be an LC–HC dimer resulting from partial denaturation of the mAb. None of these specific bands was detected in WT soluble protein
extracts (Fig. 2B, lanes 1 and 3), loaded at a similar protein concentration as shown by SDS–PAGE and Coomassie blue staining (Fig. 2A). Analysis of four successive extractions from the R12 remaining pellet revealed the presence of HC and LC bands (data not shown). Expression of heterologous proteins was observed in all R12 suspensions. Increased monoclonal antibody production in plant cell suspension culture ELISA results for mAb expression in the transgenic plant [15] indicated levels at 3 lg/g leaf FW, corresponding to 9 lg/g leaf DW (assuming 70% water content in leaves), whereas similar analysis in plant cell suspensions using lyophilized cells indicated 10 lg/g DW for four suspension cell cultures and 30 lg/g DW for the highest-yield culture. In vitro neutralizing activity of monoclonal antibodies The neutralization activity of mAb produced by plant cell cultures was as efficient as the mAb produced by hybridoma cells in a standard neutralization assay against cell culture-adapted rabies virus strains (CVS-11, CVSN2C, and ERA). The same quantity of each mAb was required to neutralize virus. Kinetics of plant cell growth and monoclonal antibody expression
Fig. 2. Analysis of protein extracts of WT and R12 cells by SDS–PAGE stained with Coomassie blue (A) and Western blotting to detect human anti-rabies virus monoclonal antibody light- and heavy-chains (B). Lanes 1 and 2 correspond to soluble protein extracts of wild-type and R12 cells from suspension cell cultures, respectively; lanes 3 and 4 correspond to proteins extracted from wild-type and R12 remaining pellets, respectively.
Both WT and R12 cell suspensions were grown for 30 days and examined every 4 days for fresh weight (data not shown) and dry weight (Fig. 3A). A well-established N. tabacum BY2 suspension culture was used for comparison. WT and R12 cell cultures showed similar growth curves, with no exponential phase, starting from 4.5 g/L
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
605
Fig. 4. Western blot analysis of human anti-rabies virus monoclonal antibody light- and heavy-chains in tobacco cell extracts after protein G purification. FT, flowthrough; W, washing step; ELU, elution step; T+, positive control (mAbP purified from plant on protein A column).
Fig. 3. Monoclonal antibody production in plant cell cultures grown in Erlenmeyer flasks. (A) Growth curve of wild-type (-h-) and R12 (-j-) Nicotiana tabacum cv. Xanthi, and wild-type Nicotiana tabacum BY2 (-d-). Data are the average of four samples and error bars correspond to the resulting standard deviation. (B) Western blot detection of human antirabies virus LC expressed in R12 cells with respect to growth curve.
and reaching 18 g/L final DW in 20 days. The BY2 cell culture reached 14 g/L in 10 days, starting from 0.6 g/L. Western blot analysis of HC (data not shown) and LC (Fig. 3B) expression in R12 cells revealed the same level of expression at all time points examined. ELISA of extracellular medium precipitated with 70% ammonium sulphate showed no significant difference between WT and R12 suspension culture with respect to the presence of mAb chains, and Western blot revealed only the LC in the R12 medium, as a monomer (30 kDa) and a dimer (60 kDa). Bioreactor scale-up of monoclonal antibody production Cells were grown for 15 days in Erlenmeyer flasks or a disposable plastic bioreactor and harvested for analysis of biomass and mAb expression. In both systems, biomass production was on average of 13 g/L DW. mAb expression as determined using ELISA was also comparable, with 10 lg/g DW in flasks and 13 lg/g DW in the bioreactor. Small-scale purification of plant-produced monoclonal antibodies R12-produced mAb was purified on a protein G column, and flowthrough, washing step, and elution fractions were examined for the presence of the HC and LC by Western blotting (Fig. 4). WT cells were used as negative control. No HC or LC was detected in the WT sample or in the flowthrough or the washing step phases of the R12 extract, confirming the high specificity of the column. Both LC and HC were detectable in the elution phase, in the same range of intensity, and at the same molecular weight position as
mAb purified from whole plants (Fig. 4). From 100 g DW of cells, 250 lg of mAb was collected, which represents 25% of the measured production. Discussion In an effort to increase the safety and reproducibility of human anti-rabies virus mAb production in plant-based systems, a transgenic tobacco plant expressing this recombinant protein was grown as an undifferentiated cell culture. WT and R12 calli were rapidly obtained on MAK medium, and LC and HC expression was readily detected in R12 calli by Western blotting. All calli were morphologically healthy and identical, indicating that mAb expression does not adversely affect cell metabolism. Moreover, LC and HC expression proved that both the Ca2p and Pin2p promoters are also functional in undifferentiated plant cells. Six weeks later, plant cell cultures were initiated in MAK medium and examined for mAb production. Expression of the mAb appeared to remain constant throughout the growth periods, with no effect on cell growth rate. However, the presence of clumps, the lack of an exponential growth phase, and the slow growth rate indicated the need for improved growth conditions. Indeed, N. tabacum BY2 cells are about 10 times more efficient than Xanthi cells for biomass production, reflecting the decades of optimization of BY2 cell growth conditions while Xanthi cells were rarely used. Despite the relatively low growth rate of Xanthi cell cultures, a biomass scale-up was performed with a new disposable plastic bioreactor and showed similar growth rate and mAb expression levels to those in Erlenmeyer flask cultures. Disposable bioreactors up to a size of 150 L for biomass production of BY2 cell culture also showed similar results (personal communication). These data demonstrate that inexpensive and easy-to-use disposable bioreactors are suitable for biomass production and for the production of recombinant proteins. Analysis of the mAb concentration from plant cell suspension cultures by ELISA revealed a production level as
606
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
high as and even higher than that in whole transgenic plants (up to 30 lg/g DW compared to 9 lg/g DW, respectively). The production level in plant cells is probably higher than 30 lg/g DW since several successive extractions revealed some mAb remaining in the pellet, probably trapped in the ER. Since the ELISA used is based on rabies virus recognition, the plant cell-produced mAb was presumably fully processed, i.e., with LC and HC association, and functional, i.e., able to bind the virus. This resulting mAb was also as effective as human hybridoma cell-produced mAb in neutralizing three different rabies virus strains, preventing infection of BHK cells. This result showed that undifferentiated plant cells were able to produce mAb with the same neutralization activity as whole transgenic plants, also compared to human hybridoma cell-produced mAb [15]. The glycosylation pattern of the mAb and several other characteristics, such as immunogenicity, remain to be examined. However, a high mannose-glycosylation like that for whole-plant-produced mAb is expected due to the KDEL ER-retention signal [7,8]. This hypothesis is supported by the presence in Western blots of both LC and HC at the same molecular weight (30 and 55 kDa, respectively) as the glycosylated form of mAb produced by the whole plant [8]. The LC might present two glycosylation patterns since a second band was observed. The KDEL signal expressed on HC also prevented mAb secretion into the medium. The small amount of mAb detected in the R12 medium might be released by dying cells. Only LC was detected outside of the cells, probably secreted via the ‘‘default pathway’’ when not bound to HC. The amount of human anti-rabies virus mAb recovered from 100 g DW of R12 cells was 4-fold less than in pilot experiments, probably due to mAb precipitation during the long extraction process [17]. In the present study, the main compounds responsible for mAb precipitation are likely the polyphenols, as described by Haslam [19]. These compounds form aggregates, possibly corresponding to the dark precipitates observed in the supernatant even within a few minutes after filtration through 0.22-lm filters. These compounds imposed multiple filtration steps, several times during the purification, increasing the processing time to 4 h. Another phenomenon has been described [20,21], whereby 80% of a mAb dissolved in an Erlenmeyer flask with pure Murashige and Skoog-medium can ‘‘disappear’’ in less than 2 h by anchoring onto the glassware wall. Such adsorption of the mAb might also explain the low recovery. Further studies are needed to overcome these obstacles to optimal mAb recovery. Plant cell culture is a developing technology for recombinant protein production and presents great potential in competing with other production systems. Many features remain to be optimized such as the expression level of the product and the purification processes. Currently, the plant cell culture system is more suitable for the production of recombinant proteins needed in small quantity and presenting high activity.
Acknowledgments We thank Dr. K. Ko for providing seedlings and mAb controls, and Dr. D.C. Hooper for fruitful discussion. We also thank C. Brimer, R. Kean, P. Hilaire, and B. Terrier for technical assistance. This work was supported by a grant from USDA to Biotechnology Foundation (to H.K.). References [1] S. Hellwig, J. Drossard, R.M. Twyman, R. Fischer, Plant cell cultures for the production of recombinant proteins, Nat. Biotechnol. 22 (2004) 1415–1422. [2] N.P. Teli, M.P. Timko, Recent developments in the use of transgenic plants for the production of human therapeutics and biopharmaceuticals, Plant Cell Tiss. Org. 79 (2004) 125–145. [3] T.J. Menkhaus, Y. Bai, C. Zhang, Z.L. Nikolov, C.E. Glatz, Considerations for the recovery of recombinant proteins from plants, Biotechnol. Prog. 20 (2004) 1001–1014. [4] S. Schillberg, R.M. Twyman, R. Fischer, Opportunities for recombinant antigen and antibody expression in transgenic plants—technology assessment, Vaccine 23 (2005) 1764–1769. [5] E. Stoger, M. Sack, L. Nicholson, R. Fischer, P. Christou, Recent progress in plantibody technology, Curr. Pharm. Des. 11 (2005) 2439– 2457. [6] S. Nandi, D. Yalda, S. Lu, Z. Nikolov, R. Misaki, K. Fujiyama, N. Huang, Process development and economic evaluation of recombinant human lactoferrin expressed in rice grain, Transgenic Res. 14 (2005) 237–249. [7] A. Triguero, G. Cabrera, J.A. Cremata, C.T. Yuen, J. Wheeler, N.I. Ramirez, Plant-derived mouse IgG monoclonal antibody fused to KDEL endoplasmic reticulum–retention signal is N-glycosylated homogeneously throughout the plant with mostly high-mannose-type N-glycans, Plant Biotechnol. J. 3 (2005) 449. [8] Y. Tekoah, K. Ko, H. Koprowski, D.J. Harvey, M.R. Wormald, R.A. Dwek, P. Rudd, Controlled glycosylation of therapeutic antibodies in plants, Arch. Biochem. Biophys. 426 (2004) 266–278. [9] J.K. Ma, E. Barros, R. Bock, P. Christou, P.J. Dale, P.J. Dix, R. Fischer, J. Irwin, R. Mahoney, M. Pezzotti, S. Schillberg, P. Sparrow, E. Stoger, R.M. Twyman, Molecular farming for new drugs and vaccines—current perspectives on the production of pharmaceuticals in transgenic plants, EMBO Rep. 6 (2005) 593–599. [10] H. Koprowski, Vaccines and sera through plant biotechnology, Vaccine 23 (2005) 1757–1763. [11] S.J. Streatfield, Plant-based vaccines for animal health, Rev. Sci. Tech. 24 (2005) 189–199. [12] R. Valdes, L. Gomez, S. Padilla, J. Brito, B. Reyes, T. Alvarez, O. Mendoza, O. Herrera, W. Ferro, M. Pujol, V. Leal, M. Linares, Y. Hevia, C. Garcia, L. Mila, O. Garcia, R. Sanchez, A. Acosta, D. Geada, R. Paez, J. Luis Vega, C. Borroto, Large-scale purification of an antibody directed against hepatitis B surface antigen from transgenic tobacco plants, Biochem. Biophys. Res. Commun. 308 (2003) 94–100. [13] WHO, Consultation on a monoclonal antibody cocktail for rabies post-exposure treatment. www.who.int/rabies/vaccines/en/mabs_final_report.pdf Geneva, 23–24 May 2002 (2002) 1–8. [14] ACIP, Recommendations of the Advisory Committee on Immunization Practices. Centers for Disease Control and Prevention. Human rabies prevention-United States, Morbidity & Mortality Weekly Rep. 48 (1999) 1–21. [15] K. Ko, Y. Tekoah, P.M. Rudd, D.J. Harvey, R.A. Dwek, S. Spitsin, C.A. Hanlon, C. Rupprecht, B. Dietzschold, M. Golovkin, H. Koprowski, Function and glycosylation of plant-derived antiviral monoclonal antibody, Proc. Natl. Acad. Sci. USA 100 (2003) 8013– 8018.
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607 [16] L.H. Stevens, G.M. Stoopen, I.J. Elbers, J.W. Molthoff, H.A. Bakker, A. Lommen, D. Bosch, W. Jordi, Effect of climate conditions and plant developmental stage on the stability of antibodies expressed in transgenic tobacco, Plant Physiol. 124 (2000) 173–182. [17] L.S. Girard, M. Bastin, D. Courtois, Expression of the human milk protein sCD14 in tobacco cell culture, Plant Cell Tiss. Org. 78 (2004) 253–260. [18] J.M. Champion, R.B. Kean, C.E. Rupprecht, A.L. Notkins, H. Koprowski, B. Dietzschold, D.C. Hooper, The development of monoclonal human rabies virus-neutralizing antibodies as a
607
substitute for pooled human immune globulin in the prophylactic treatment of rabies virus exposure, J. Immunol. Methods 235 (2000) 81–90. [19] E. Haslam, Plant Polyphenols—Vegetable Tannins Revisited, Cambridge University Press, Cambridge, 1989, pp. 167–192. [20] B.M. Tsoi, P.M. Doran, Effect of medium properties and additives on antibody stability and accumulation in suspended plant cell cultures, Biotechnol. Appl. Biochem. 35 (2002) 171–180. [21] P.M. Doran, Loss of secreted antibody from transgenic plant tissue cultures due to surface adsorption, J. Biotechnol. 122 (2006) 39–54.
Expression of a human anti-rabies virus monoclonal antibody in tobacco cell culture q Loı¨c Ste´phane Girard a, Marzena Jolanta Fabis a, Maryse Bastin b, Didier Courtois b, Vincent Pe´tiard b,*, Hilary Koprowski a,* a
Thomas Jefferson University, 1020 Locust Street, Biotechnology Foundation, Philadelphia, PA 19107, USA b Centre de Recherche Nestle´-Tours, 101 Avenue G. Eiffel, BP 49716, 37097 Tours Cedex 2, France Received 15 March 2006 Available online 19 April 2006
Abstract A Nicotiana tabacum cv. Xanthi cell culture was initiated from a transgenic plant expressing a human anti-rabies virus monoclonal antibody. Within 3 months, plant cell suspension cultures were established and recombinant protein expression was examined. The antibody was stably produced during culture growth. ELISA, protein G purification, Western blotting, and neutralization assay confirmed that the antibody was fully processed, with association of light and heavy-chains, and that it was able to bind and neutralize rabies virus. Quantification of antibody production in plant cell suspension culture revealed 30 lg/g of cell dry weight for the highest-producing culture (0.5 mg/L), 3 times higher than from the original transgenic plant. The same production level was observed 3 months after cell culture initiation. Plant cell suspension cultures were successfully grown in a new disposable plastic bioreactor, with a growth rate and production level similar to that of cultures in Erlenmeyer flasks. Ó 2006 Elsevier Inc. All rights reserved. Keywords: Antibody; Disposable plastic bioreactors; Plant cell culture; Rabies; Recombinant; Xanthi
Plant-based systems are increasingly used for the production of recombinant proteins, because of several advantages over other production systems, such as the ability to carry out necessary post-translational modifications not available in bacterial systems, as well as greater safety and lower production costs than those provided by animal-based systems. Plant-based technology has recently been extensively reviewed, with a full description of plants commonly used, a listing of proteins expressed thus far [1–5], and the potential economic superiority of these systems for the recombi-
q
Abbreviations: cv., cultivar; DW, dry weight; ER, endoplasmic reticulum; FW, fresh weight; HC, heavy-chain; LC, light-chain; mAb, monoclonal antibody. * Corresponding authors. Fax: +33 0 2 47 49 14 14 (V. Pe´tiard); +1 215 923 6795 (H. Koprowski). E-mail addresses: [email protected] (V. Pe´tiard), hilary.koprowski@jefferson.edu (H. Koprowski). 0006-291X/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2006.03.219
nant protein production [5,6]. Protein stability during production and storage is an important determinant of the production cost. Indeed, degradation by proteases can be avoided by the addition of a ‘‘Lys-Asp-Glu-Leu’’ (KDEL) signal to the protein sequence, which leads to storage of the heterologous protein inside the endoplasmic reticulum (ER). This ER targeting also results in a glycosylation pattern closer to the human one, by preventing addition of plant-specific glycans, fucose, and xylose to the protein [7] thus avoiding a potential source of unwanted immunogenicity in the recombinant protein while injected [8]. Many proteins relevant in human disease treatment have already been expressed in plant-based systems (reviewed in [2,9]), mainly in the form of edible vaccines [10,11] and monoclonal antibodies (mAb) [5,12]. A safe and cost-effective source of mAb is particularly desirable in the treatment of human rabies virus infection, since current protocols rely heavily on anti-rabies immunoglobulins prepared mainly
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
603
from horse serum; not only are such immunoglobulins expensive to prepare, accounting for their limited supply worldwide, but they present the risk of cross-contamination with known or unknown horse pathogens [13]. Considering that rabies virus exposure causes an economic burden in the US of about $1 billion per year for vaccination and immunoglobulin administration [14], a plantbased system of immunoglobulin production, as recently developed in tobacco plants [15], is attractive. However, mAb expression levels in field plants vary depending on plant maturation stage and climatic influences [16 and personal observations], whereas plant cell cultures are based on undifferentiated cells growing in a fully controlled environment, ensuring a constant level of expression and preventing contamination of the environment. Here, we demonstrate for the first time the expression of a human anti-rabies virus glycoprotein mAb in plant cell culture. A low-cost biomass scale-up was performed with a new disposable bioreactor.
Lyophilized cells [100 g dry weight (DW), equivalent to 1.45 kg FW] were mixed with 1.5 L of extraction buffer (20 mM sodium phosphate, pH 7.2) and disrupted using a French press. Soluble proteins were recovered by centrifugation (6000g, 20 min, 4 °C) and filtered on fiberglass (from 2.7- to 0.22-lm pore sizes). Cross-filtration on a 30-kDa membrane was used to concentrate the supernatant to 300 mL. Soluble protein extracts ¨ kta purifier equipped with a HiTrapä protein G were purified with an A high-performance affinity column (Amersham Biosciences) according to the manufacturer’s recommendations. mAb was eluted with 0.1 M glycine–HCl (pH 2.7) and pH was neutralized with 1 M Tris–HCl (pH 9.0).
Materials and methods
ELISA
Plant material A stable (T3) transgenic tobacco plant (Nicotiana tabacum cv. Xanthi) expressing an anti-rabies virus mAb [15] was used and is designated herein as R12. Briefly, the heavy-chain (HC) sequence, fused with a sequence encoding the KDEL ER retention signal, is placed under control of the cauliflower mosaic virus 35S promoter with duplicated upstream B domains (Ca2p); the light-chain (LC) sequence is under the control of the potato proteinase inhibitor II promoter (Pin2p) [15]. A well-established wild-type BY2 tobacco cell suspension (N. tabacum) was used for comparison to newly established tobacco cell suspensions.
Binding of transgenic mAb to rabies virus was assessed by ELISA as described [18], with the following modifications: 3 lg/mL of ERA rabies virus was used to coat the plates, followed by loading with 95 lL of serial 3-fold dilutions of soluble extracted proteins obtained from 7 mg of lyophilized plant cell extract disrupted in 1 mL of sodium phosphate extraction buffer. Human anti-rabies virus mAb (1 lg/mL) was used as a positive control. Recombinant bound mAb were detected by successive additions of biotin-conjugated anti-human IgG1 (1:1000; BD PharMingen, San Diego, CA), avidin alkaline phosphatase-conjugated antibody (Sigma), and p-nitrophenyl phosphate (Sigma) as substrate. Absorbance was read at 405 nm.
Plant cell culture
In vitro neutralization assay
Callus induction. Primary explants were cut from 3-week-old wild-type (WT) or R12 transgenic seedlings and placed on Murashige and Skoogbased medium, supplemented with 1 mg/L naphthalene acetic acid (NAA), 0.1 mg/L kinetin, 30 g/L sucrose, and 8 g/L agarose (MAK), under photoperiod conditions (16 h light/8 h dark), at 25 °C. The medium was autoclaved before explants for 20 min at 115 °C. Resulting calli were subcultured monthly on Petri dishes. Initiation of suspension culture. Randomly picked calli (20 g) were inoculated into 50 mL MAK liquid medium and placed under photoperiod conditions at 25 °C on an orbital shaker (100 rpm). After 2 weeks, both transformed and non-transformed tobacco cell lines were grown in 250-mL Erlenmeyer flasks containing 100 mL MAK medium and subcultured every 20 days at an initial density of 40 g of cells [fresh weight (FW)]/L.
Neutralizing activity of purified antibodies was assessed using a standard 20-h rapid fluorescent focus inhibition test (RFFIT) as described [18]. Briefly, diluted purified mAb produced by plant cell culture was incubated with rabies virus (CVS-11, CVS-N2C or ERA). Baby hamster kidney (BHK) cells were added to plates, incubated, and fixed. FITC-anti-rabies nucleoprotein monoclonal globulin (Centocor, Malvern, PA) was added and plates were examined under a fluorescence microscope to detect infected cells. The original mAb produced in human hybridoma cells was used as positive control. Both mAb were used at the same concentration to compare their neutralization ability.
1/10,000 dilution] conjugated to horseradish peroxidase (Jackson ImmunoResearch). Whole-plant-derived mAb was used as a positive control. Suspension culture in disposable bioreactors A new type of plastic bag-based bioreactor (Nestle´ Research Center, Tours) was used to scale-up the cell culture from 500 mL to 10 L. After sterilization of the disposable bioreactor, sterile medium was inserted into the bioreactor and cells were inoculated [40 g/L (FW)]. Small-scale purification of plant cell-derived antibodies
Results Morphological features of plant cell culture
Protein extraction and analysis Protein extraction, SDS–PAGE, and Western blotting were carried out as described [17], with the following modifications. Twenty-day suspension cultures were filtered (50-lm nylon mesh) to separate cells and medium. Cell disruption was performed on 1 g FW of cells, using a stainless-steel mixer mill (Retsch, Model MM301, Newtown, PA) at 30 hits/s, for 2 min. The resulting mash was homogenized with 300 lL DB2X buffer (40 mM Tris–HCl, pH 8.0, 20% glycerol, 2% SDS, and 4% b-mercaptoethanol) containing 0.1% Triton X-100 and a complete protease inhibitor mixture (Roche Diagnostics). For Western blot analysis, HC and LC were detected with goat anti-human antibodies [Fc c- and F(ab 0 )2 fragment specific, respectively,
Primary calli initiated from WT tobacco seedlings and R12 transgenic seedlings expressing human anti-rabies virus mAb were grown on MAK medium and after 3 weeks appear morphologically similar (Fig. 1A). The yellow color turned to green after 3 months. Western blot analysis of eight WT and eight R12 calli revealed HC and LC protein production only in the R12 calli (data not shown). Three weeks after inoculation, WT and R12 suspension cultures appeared healthy in MAK medium (Fig. 1B), with the same yellow color and the same morphology as a stabi-
604
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
Fig. 1. Morphological features of calli and plant cell suspension cultures grown in MAK medium. (A) Three-week-old calli, obtained by callogenesis on Nicotiana tabacum (cv. Xanthi) plant leaves grown on MAK medium (left, wild-type calli; right, R12 calli). (B) Comparison of wild-type (left) and R12 transformed (right) 10-day-old Nicotiana tabacum cv. Xanthi suspension cell culture. (C) Filtered cells from 20-day-old plant cell suspension. Note the greenish color and clumped growth.
lized N. tabacum BY2 culture. One month later, cells began to grow as clumps and turned greenish (Fig. 1C); cell suspension cultures retained these features during further subcultures. Heavy- and light-chain expression in plant suspension cultures and pellets Western blot analysis (Fig. 2B) of soluble protein cell extracts from five R12 suspension cultures and their remaining pellets (after a second extraction) revealed HC and LC bands migrating at 55 and 30 kDa, respectively. The LC appeared to be expressed as two different forms (Fig. 2B, lanes 2 and 4). Two other bands were also detected in the pellet at 60 and 75 kDa (Fig. 2B). The 60-kDa band appears to be a dimer of LC since this band was not detected by HC-specific antibody (data not shown). The 75-kDa band might be an LC–HC dimer resulting from partial denaturation of the mAb. None of these specific bands was detected in WT soluble protein
extracts (Fig. 2B, lanes 1 and 3), loaded at a similar protein concentration as shown by SDS–PAGE and Coomassie blue staining (Fig. 2A). Analysis of four successive extractions from the R12 remaining pellet revealed the presence of HC and LC bands (data not shown). Expression of heterologous proteins was observed in all R12 suspensions. Increased monoclonal antibody production in plant cell suspension culture ELISA results for mAb expression in the transgenic plant [15] indicated levels at 3 lg/g leaf FW, corresponding to 9 lg/g leaf DW (assuming 70% water content in leaves), whereas similar analysis in plant cell suspensions using lyophilized cells indicated 10 lg/g DW for four suspension cell cultures and 30 lg/g DW for the highest-yield culture. In vitro neutralizing activity of monoclonal antibodies The neutralization activity of mAb produced by plant cell cultures was as efficient as the mAb produced by hybridoma cells in a standard neutralization assay against cell culture-adapted rabies virus strains (CVS-11, CVSN2C, and ERA). The same quantity of each mAb was required to neutralize virus. Kinetics of plant cell growth and monoclonal antibody expression
Fig. 2. Analysis of protein extracts of WT and R12 cells by SDS–PAGE stained with Coomassie blue (A) and Western blotting to detect human anti-rabies virus monoclonal antibody light- and heavy-chains (B). Lanes 1 and 2 correspond to soluble protein extracts of wild-type and R12 cells from suspension cell cultures, respectively; lanes 3 and 4 correspond to proteins extracted from wild-type and R12 remaining pellets, respectively.
Both WT and R12 cell suspensions were grown for 30 days and examined every 4 days for fresh weight (data not shown) and dry weight (Fig. 3A). A well-established N. tabacum BY2 suspension culture was used for comparison. WT and R12 cell cultures showed similar growth curves, with no exponential phase, starting from 4.5 g/L
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
605
Fig. 4. Western blot analysis of human anti-rabies virus monoclonal antibody light- and heavy-chains in tobacco cell extracts after protein G purification. FT, flowthrough; W, washing step; ELU, elution step; T+, positive control (mAbP purified from plant on protein A column).
Fig. 3. Monoclonal antibody production in plant cell cultures grown in Erlenmeyer flasks. (A) Growth curve of wild-type (-h-) and R12 (-j-) Nicotiana tabacum cv. Xanthi, and wild-type Nicotiana tabacum BY2 (-d-). Data are the average of four samples and error bars correspond to the resulting standard deviation. (B) Western blot detection of human antirabies virus LC expressed in R12 cells with respect to growth curve.
and reaching 18 g/L final DW in 20 days. The BY2 cell culture reached 14 g/L in 10 days, starting from 0.6 g/L. Western blot analysis of HC (data not shown) and LC (Fig. 3B) expression in R12 cells revealed the same level of expression at all time points examined. ELISA of extracellular medium precipitated with 70% ammonium sulphate showed no significant difference between WT and R12 suspension culture with respect to the presence of mAb chains, and Western blot revealed only the LC in the R12 medium, as a monomer (30 kDa) and a dimer (60 kDa). Bioreactor scale-up of monoclonal antibody production Cells were grown for 15 days in Erlenmeyer flasks or a disposable plastic bioreactor and harvested for analysis of biomass and mAb expression. In both systems, biomass production was on average of 13 g/L DW. mAb expression as determined using ELISA was also comparable, with 10 lg/g DW in flasks and 13 lg/g DW in the bioreactor. Small-scale purification of plant-produced monoclonal antibodies R12-produced mAb was purified on a protein G column, and flowthrough, washing step, and elution fractions were examined for the presence of the HC and LC by Western blotting (Fig. 4). WT cells were used as negative control. No HC or LC was detected in the WT sample or in the flowthrough or the washing step phases of the R12 extract, confirming the high specificity of the column. Both LC and HC were detectable in the elution phase, in the same range of intensity, and at the same molecular weight position as
mAb purified from whole plants (Fig. 4). From 100 g DW of cells, 250 lg of mAb was collected, which represents 25% of the measured production. Discussion In an effort to increase the safety and reproducibility of human anti-rabies virus mAb production in plant-based systems, a transgenic tobacco plant expressing this recombinant protein was grown as an undifferentiated cell culture. WT and R12 calli were rapidly obtained on MAK medium, and LC and HC expression was readily detected in R12 calli by Western blotting. All calli were morphologically healthy and identical, indicating that mAb expression does not adversely affect cell metabolism. Moreover, LC and HC expression proved that both the Ca2p and Pin2p promoters are also functional in undifferentiated plant cells. Six weeks later, plant cell cultures were initiated in MAK medium and examined for mAb production. Expression of the mAb appeared to remain constant throughout the growth periods, with no effect on cell growth rate. However, the presence of clumps, the lack of an exponential growth phase, and the slow growth rate indicated the need for improved growth conditions. Indeed, N. tabacum BY2 cells are about 10 times more efficient than Xanthi cells for biomass production, reflecting the decades of optimization of BY2 cell growth conditions while Xanthi cells were rarely used. Despite the relatively low growth rate of Xanthi cell cultures, a biomass scale-up was performed with a new disposable plastic bioreactor and showed similar growth rate and mAb expression levels to those in Erlenmeyer flask cultures. Disposable bioreactors up to a size of 150 L for biomass production of BY2 cell culture also showed similar results (personal communication). These data demonstrate that inexpensive and easy-to-use disposable bioreactors are suitable for biomass production and for the production of recombinant proteins. Analysis of the mAb concentration from plant cell suspension cultures by ELISA revealed a production level as
606
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607
high as and even higher than that in whole transgenic plants (up to 30 lg/g DW compared to 9 lg/g DW, respectively). The production level in plant cells is probably higher than 30 lg/g DW since several successive extractions revealed some mAb remaining in the pellet, probably trapped in the ER. Since the ELISA used is based on rabies virus recognition, the plant cell-produced mAb was presumably fully processed, i.e., with LC and HC association, and functional, i.e., able to bind the virus. This resulting mAb was also as effective as human hybridoma cell-produced mAb in neutralizing three different rabies virus strains, preventing infection of BHK cells. This result showed that undifferentiated plant cells were able to produce mAb with the same neutralization activity as whole transgenic plants, also compared to human hybridoma cell-produced mAb [15]. The glycosylation pattern of the mAb and several other characteristics, such as immunogenicity, remain to be examined. However, a high mannose-glycosylation like that for whole-plant-produced mAb is expected due to the KDEL ER-retention signal [7,8]. This hypothesis is supported by the presence in Western blots of both LC and HC at the same molecular weight (30 and 55 kDa, respectively) as the glycosylated form of mAb produced by the whole plant [8]. The LC might present two glycosylation patterns since a second band was observed. The KDEL signal expressed on HC also prevented mAb secretion into the medium. The small amount of mAb detected in the R12 medium might be released by dying cells. Only LC was detected outside of the cells, probably secreted via the ‘‘default pathway’’ when not bound to HC. The amount of human anti-rabies virus mAb recovered from 100 g DW of R12 cells was 4-fold less than in pilot experiments, probably due to mAb precipitation during the long extraction process [17]. In the present study, the main compounds responsible for mAb precipitation are likely the polyphenols, as described by Haslam [19]. These compounds form aggregates, possibly corresponding to the dark precipitates observed in the supernatant even within a few minutes after filtration through 0.22-lm filters. These compounds imposed multiple filtration steps, several times during the purification, increasing the processing time to 4 h. Another phenomenon has been described [20,21], whereby 80% of a mAb dissolved in an Erlenmeyer flask with pure Murashige and Skoog-medium can ‘‘disappear’’ in less than 2 h by anchoring onto the glassware wall. Such adsorption of the mAb might also explain the low recovery. Further studies are needed to overcome these obstacles to optimal mAb recovery. Plant cell culture is a developing technology for recombinant protein production and presents great potential in competing with other production systems. Many features remain to be optimized such as the expression level of the product and the purification processes. Currently, the plant cell culture system is more suitable for the production of recombinant proteins needed in small quantity and presenting high activity.
Acknowledgments We thank Dr. K. Ko for providing seedlings and mAb controls, and Dr. D.C. Hooper for fruitful discussion. We also thank C. Brimer, R. Kean, P. Hilaire, and B. Terrier for technical assistance. This work was supported by a grant from USDA to Biotechnology Foundation (to H.K.). References [1] S. Hellwig, J. Drossard, R.M. Twyman, R. Fischer, Plant cell cultures for the production of recombinant proteins, Nat. Biotechnol. 22 (2004) 1415–1422. [2] N.P. Teli, M.P. Timko, Recent developments in the use of transgenic plants for the production of human therapeutics and biopharmaceuticals, Plant Cell Tiss. Org. 79 (2004) 125–145. [3] T.J. Menkhaus, Y. Bai, C. Zhang, Z.L. Nikolov, C.E. Glatz, Considerations for the recovery of recombinant proteins from plants, Biotechnol. Prog. 20 (2004) 1001–1014. [4] S. Schillberg, R.M. Twyman, R. Fischer, Opportunities for recombinant antigen and antibody expression in transgenic plants—technology assessment, Vaccine 23 (2005) 1764–1769. [5] E. Stoger, M. Sack, L. Nicholson, R. Fischer, P. Christou, Recent progress in plantibody technology, Curr. Pharm. Des. 11 (2005) 2439– 2457. [6] S. Nandi, D. Yalda, S. Lu, Z. Nikolov, R. Misaki, K. Fujiyama, N. Huang, Process development and economic evaluation of recombinant human lactoferrin expressed in rice grain, Transgenic Res. 14 (2005) 237–249. [7] A. Triguero, G. Cabrera, J.A. Cremata, C.T. Yuen, J. Wheeler, N.I. Ramirez, Plant-derived mouse IgG monoclonal antibody fused to KDEL endoplasmic reticulum–retention signal is N-glycosylated homogeneously throughout the plant with mostly high-mannose-type N-glycans, Plant Biotechnol. J. 3 (2005) 449. [8] Y. Tekoah, K. Ko, H. Koprowski, D.J. Harvey, M.R. Wormald, R.A. Dwek, P. Rudd, Controlled glycosylation of therapeutic antibodies in plants, Arch. Biochem. Biophys. 426 (2004) 266–278. [9] J.K. Ma, E. Barros, R. Bock, P. Christou, P.J. Dale, P.J. Dix, R. Fischer, J. Irwin, R. Mahoney, M. Pezzotti, S. Schillberg, P. Sparrow, E. Stoger, R.M. Twyman, Molecular farming for new drugs and vaccines—current perspectives on the production of pharmaceuticals in transgenic plants, EMBO Rep. 6 (2005) 593–599. [10] H. Koprowski, Vaccines and sera through plant biotechnology, Vaccine 23 (2005) 1757–1763. [11] S.J. Streatfield, Plant-based vaccines for animal health, Rev. Sci. Tech. 24 (2005) 189–199. [12] R. Valdes, L. Gomez, S. Padilla, J. Brito, B. Reyes, T. Alvarez, O. Mendoza, O. Herrera, W. Ferro, M. Pujol, V. Leal, M. Linares, Y. Hevia, C. Garcia, L. Mila, O. Garcia, R. Sanchez, A. Acosta, D. Geada, R. Paez, J. Luis Vega, C. Borroto, Large-scale purification of an antibody directed against hepatitis B surface antigen from transgenic tobacco plants, Biochem. Biophys. Res. Commun. 308 (2003) 94–100. [13] WHO, Consultation on a monoclonal antibody cocktail for rabies post-exposure treatment. www.who.int/rabies/vaccines/en/mabs_final_report.pdf Geneva, 23–24 May 2002 (2002) 1–8. [14] ACIP, Recommendations of the Advisory Committee on Immunization Practices. Centers for Disease Control and Prevention. Human rabies prevention-United States, Morbidity & Mortality Weekly Rep. 48 (1999) 1–21. [15] K. Ko, Y. Tekoah, P.M. Rudd, D.J. Harvey, R.A. Dwek, S. Spitsin, C.A. Hanlon, C. Rupprecht, B. Dietzschold, M. Golovkin, H. Koprowski, Function and glycosylation of plant-derived antiviral monoclonal antibody, Proc. Natl. Acad. Sci. USA 100 (2003) 8013– 8018.
L.S. Girard et al. / Biochemical and Biophysical Research Communications 345 (2006) 602–607 [16] L.H. Stevens, G.M. Stoopen, I.J. Elbers, J.W. Molthoff, H.A. Bakker, A. Lommen, D. Bosch, W. Jordi, Effect of climate conditions and plant developmental stage on the stability of antibodies expressed in transgenic tobacco, Plant Physiol. 124 (2000) 173–182. [17] L.S. Girard, M. Bastin, D. Courtois, Expression of the human milk protein sCD14 in tobacco cell culture, Plant Cell Tiss. Org. 78 (2004) 253–260. [18] J.M. Champion, R.B. Kean, C.E. Rupprecht, A.L. Notkins, H. Koprowski, B. Dietzschold, D.C. Hooper, The development of monoclonal human rabies virus-neutralizing antibodies as a
607
substitute for pooled human immune globulin in the prophylactic treatment of rabies virus exposure, J. Immunol. Methods 235 (2000) 81–90. [19] E. Haslam, Plant Polyphenols—Vegetable Tannins Revisited, Cambridge University Press, Cambridge, 1989, pp. 167–192. [20] B.M. Tsoi, P.M. Doran, Effect of medium properties and additives on antibody stability and accumulation in suspended plant cell cultures, Biotechnol. Appl. Biochem. 35 (2002) 171–180. [21] P.M. Doran, Loss of secreted antibody from transgenic plant tissue cultures due to surface adsorption, J. Biotechnol. 122 (2006) 39–54.