PBVAB 5 Part 4
The fifth session on Day 2 was “Antibodies 1” – and who better to kick off, than Rainer Fischer (RWTH / Fraunhofer Institute, Aachen), talking about ‘Pharma-Planta – The European project to introduce plant-derived monoclonal antibodies to the clinic’.
One of the most impressive features of the FP6 PharmaPlanta project was its sheer size: 28 academic institutions were involved over 7 years, at a cost of €12 million plus €3 million from the Fraunhofer Institute in Aachen. Their mission was to move molecular farming beyond proofs of concept, and to develop candidate products. They selected the anti-HIV-1 subtype B MAb 2G12 as their final candidate, but also developed MAbs to rabies and some vaccine candidates. Importantly, their IP had a Humanitarian Use Commitment: knowledge created was made freely available for humanitarian purposes.
They had a total of 39 postdocs and 8 students trained; they produced 200 peer-reviewed publications consisting of 150 research papers and 50 reviews, and a spin-out company. The project also helped to develop a South African plant-made MAb production platform. Their plant-produced 2G12 was the first plant-made MAb in human clinical trials – and went from gene to clinic in just 7 years. They had also very materially helped the development of the regulatory regime in Europe, from the viewpoint of pharmaceutical guidelines and environmental safety for PMPs.
Rainer Fischer in full flow
The final yield figures for 2G12 were 5 g of 97% pure MAb from 240 kg of transgenic tobacco, with a recovery of 55%. The product had a better glycosylation homogeneity than CHO cell-produced 2G12. In clinical trials of the MAb used as a vaginal microbicide, the product was safe and well tolerated with no serious adverse reactions. There were no anti-Abs found in serum or in the vagina, with no systemic absorption. The MAb survived for 8 hrs in the vagina, meaning it had serious potential as a microcode.
The project resulted in great human capital, a manufacturing facility at the Fraunhofer IME, and a number of important follow-on projects. It also opened bottlenecks in regulatory practice, and in clinical trials of PMPs. There was a pipeline of additional product candidates, eg anti-rabies MAbs.
Important lessons from the project were the following: one should focus early on on the plants used, the expression technology, the threshold level of production, realistic timelines, the plant line and purification process, production issues, QC stability, regulatory contract – FIND A CLINICAL SPONSOR!, set up contractual framework, draft specifications for drugs, contact authorities in countries for manufacture and testing.
Issues such as smart product selection, synthetic biology/host cell line engineering, glycan/protease profile, hi-throughput cloning, selection of elite lines, scale-up automation / vertical farming, downstream processing, regulatory approval had also surfaced, and were important.
For the future, a fully automated vertical farm unit for seed development was going to come on stream. They would move from niche production to mainstream production, taking advantage of economies of scale. Other developments could be designing an optimal host cell line, with fully human glycosylation, and site-directed transgene integration.
Some day someone should write a book about this endeavour – and I think it should be Rainer.
Larry Zeitlin (MAPP Biopharmaceutical) spoke next, on producing monoclonals against respiratory syncytial virus (RSV): the reason for doing this is that RSV is a major pathogen among small children worldwide, and while there are MAb-based therapeutics (eg: Synagis, from MedImmune), with sales in the order of USD 1 billion annually, these cost around USD 5 000 for one treatment for one child – and premature infants or cardiac / respiratorily challenged children required 4-5 monthly doses per RSV season. Additionally, infection with RSV in the 1st year of life is associated with development of asthma later, so paediatricians were wanting to treat a much wider spectrum of children.
Accordingly, MAPP was making a Synagis equivalent via Icon vectors in N benthamiana for half the cost of goods, which had the same neutralisation ability and same affinity but a different glycosylation profile and shorter half-life. When tested in cotton rats it was identical in pharmacokinetics and worked as well as Synagis. An attempt to reduce the interaction of the IgG1-based MAb with the immune system by changing the subtype to IgG2 failed in rates even though it was neutralising, possibly due to there being less ADCC. Larry mentioned that they could engineer the Fc region with point mutations to significantly extend the half life – and then use this as a scaffold, possibly for some of their other products.
Michael McLean (Univ Guelph, Canada) described his group’s work on a HIV Ab cocktail theoretically capable of neutralising 99% of HIV strains – this was for PlantForm Corp, who had a mandate to produce biosimilars and novel biologics using plants. The HIV project was focused presently on demonstrating anti-HIV functionality, and at improving glycosylation profiles of a cocktail of b12, 2F5 and 4E10 broad-spectrum anti-HIV MAbs.
They worked with BeYDV-derived, 2-replicon vectors expressing whole MAbs, as well as their own vectors, using the Steinkellner group glycosylation pathway engineered plants. With 9 days maximum expression period they could get 1 g/kg maximum yields. All the MAbs worked fine, with similar activity in in vitro HIV pseudovirion neutralisation assays. Using the deltaFX N benth line, they get uniform glycosylation – and add Gal using their own vectors.
Shawn Chen (BioDesign Inst, Arizona State Univ) described their work on a humanized West Nile virus (WNV) therapeutic MAb which protected mice from WNV infection. They wanted blood-brain barrier (BBB)-permeable bifunctional Abs to extend efficacy, presently limited because of the barrier. They got 0.3 – 0.5 g/kg yield of a bifunctional MAb which bound the BBB endothelial receptor and virus Ag, using Icon and BeYDV vectors, and showed endocytosis into brain cells. He also mentioned that they could “tune” glycoforms to change ADCC.
Victor Klimyuk (Icon Genetics GmbH, Germany) presented on ‘Biogeneric antibodies made in plants’: these used a generic IgG1 constant region gene codon-optimised for plants, with add-on variable (V) regions derived from other Abs of different types and specificities. The first product had been the non-Hodgkin lymphoma personalized MAbs: they had done glycotyping of each NHL MAb, all with the same H but diff L chains, to show these were differently glycosylated – and that all the idiotypes were expressed at very different levels. Interestingly, expression levels had little to do with occupancy of glycosylation sites – and this occupancy could be tuned by directed point mutations.
They had made analogues of trastuzumab and herceptin, etc – and noted that herceptin analogues differed in potency, and wt plants produced lower levels than their engineered plants. Rituximab analogues were all the same as the original MAb at day 0 of treatment, but MAbs with no fucose were best at persistence – equal to the original.
Vikram Virdi (VIB, University of Gent, Belgium) described passive immunisation of piglets against enterotoxigenic E coli (ETEC) using llama-derived antibodies produced in Arabidopsis. This was useful in that it extended the maternally-derived passive immunity. Their product was a “porcinised camellid Ab” against the major adhesion molecule of ETEC, which should survive the digestive tract. They made MAbs based on a camellid Vh gene fused to IgG and IgA Fc regions, and expressed them in seeds for a piglet feed challenge. They got a maximum of 15% TSP expressed in seed, 3% of seed weight. By triple transformation with the 3 genes required for an IgA analogue (Vh:Fc, J chain and secretory component) and then selfing and breeding plant lines, they got in planta assembly of a sIgA analogue (0.2% seed weight). This worked in inhibiting attachment of bacteria, so they upscaled production and tested a cocktail of IgG vs IgA types. The latter was best, with a swift decline of bacterial shedding with a 4 x lower dose than for IgG. There was also a better weight gain for IgA treated piglets.
Thomas de Meyer (VIB-PSB/University of Gent) compared production of bivalent camellid VHH-derived MAbs in Arabidopsis, N benthamiana and Pichia pastoris, given that the VHH Fc enhanced functional affinity, and led to longer serum 1/2 life, and was a convenient protein tag. They compared VHH and VHH-Fc MAbs with 4 fusions, including anti-globulin, anti-albumin, and anti-GFP. The products were stable in seed production (with KDEL) in Arabidopsis and also N benthamiana, and Pichia secreted the products. They got yields of 1.5 – 27% TSP, 0.1 to 0.82 g/kg in plants, and with Pichia, 15 – 30 mg/l culture.
The MAbs had different size profiles from the different hosts, though all were bivalent VHH, and N benthamiana and Pichia products were fully glycosylated. Several of the Fc-type MAbs outperformed the VHHs in ELISA.
Overall, it was obvious that expression of a wide variety of antibodies in plants is a maturing technology: yields are high, of antibodies whose glycosylation and retention profiles can be handily engineered, and which perorm equivalently or better than their conventional homologues in in vitro and in vivo assays.
Go Green, he said, not quietly…B-)
ViroBlogy 