Day 3 – Workshops and Plenaries
Designing Proteins for Competitive (Bio)Processes
Miguel Alcade from Madrid applied the technique of directed evolution to develop variants of laccases with a high redox potential which retain full functionality in human plasma and blood. He said that „this proof-of-concept opens the path for implantable 3D-nanobiodevices in humans“. Algade was also able to resurrect an ancient enzyme from plants that has been changed by natural evolution and exhibits properties no longer found in modern plant variants. He emphasized: „Ancestral enzyme resurrection helps us to study evolution on a molecular scale“.
Robert Kourist, currently still at the Ruhr University Bochum but already appointed new professor at the Institute of Molecular Biotechnology at the Graz University of Technology, presented results about arylmalonate decarboxylase (AMDase). Although the natural function is still not known, he was successful in expanding its natural sequence space to achieve (S-) selectivity instead of (R-) selectivity of the natural variant. The problem was that substrate malonates are unstable. In a collaboration with Andreas Liese (TUHH) he was able to perform a reaction engineering for an industrial interesting process for the generation of the arylpropionates (S)-flurbiprofen and (S-)Naproxen.
Additionally, he showed an example of developing a two-step one pot chemo-enzymatic synthesis of bio-based antioxidants such as resveratrol-derivatives that are even more potent than the natural variant found in e.g. red wine.
Anthony Rees, emeritus professor of Bath University, CEO of Rees Consulting AB and co-founder of Synergy BioTherapeutics, gave a comprehensive talk on antibody engineering for therapy. „The growth in therapeutic antibody approvals is not exponential yet, but will certainly move this way. By 2020, it will be a 125 billion dollar market. In 1980, it took almost one day to generate a computer model which gave you the opportunity of a lot of coffee breaks, but technology greatly approved since“. There was a constant improvement of „humanness“ in antibodies which have formerly been mouse IgG, then changed to chimeric antibodies and later humanized antibodies. He introduced a method of „stealth engineering“ just modifying the surface of a mouse antibody in order to make it look like a human antibody to avoid being recognized by the human immune system. As a current trend, he described the ‚de minimus‘ construction of antibody-like proteins from single chain and single domain antibodies to diabodies (small bivalent and bispecific antibody fragments). „A particularly interesting development was the discovery that camel, llama and shark antibodies have a different antibody construction. For example, ‚camelids‘ comprise only heavy chain antibodies (without lighter chains) which can be released to form single domains operating without light chains which are independently operable units and can have a broad range of specificities like a human antibody itself“.
The technology developed by acib key researchers Florian Rüker and Gordana Wozniak-Knopp in the company F-star to generate bi- or even tri-specific antibodies was mentioned, too, with compounds now being in clinical trials. „Smarter than nature“ engineers now apply synthetic (phage) libraries, transgenic technology (e.g. human antibodies from transgenic mice) or antibodies as partnering bridges (e.g. BiTE = Bispecific T-cell Enganger; a bispecific antibody recognizing tumor antigens and engaging T-cell simultaneously so that T-cells can release killer substances to kill the tumor cells).
Arne Skerra from the University of Munich and founder of XL-protein GmbH, presented PASylation – the biological alternative to PEGylation for Plasma Half-Life Extension and Beyond. „A big and fundamental problem in the development of proteins for biopharmaceutical applications is the plasma half-life. Especially small proteins have a very short plasma half-life of two to four hours in humans due to rapid renal clearance. If a patient returns home after an injection, he has probably lost already half of the active substance.“ PEGylation, i.e. the conjugation of the protein of interest with polyethylene glycol to increase the size and thereby reduce renal clearance, has been extensively used in the past to increase the half-life of biopharmaceuticals. However, PEGylation is costly (only three manufacturers worldwide are producing clinical grade PEG – worth much more than gold) and could lead to PEG deposits even in the brain leading to a warning of use of PEGylated drugs in children as long-term effects are still unknown. PASylation is a biological alternative. It consists of natural amino acid chains made from the amino acids proline, alanine, and serine. It is a polypeptide which is stable in plasma but fully bio-degradable in contrast to synthetic PEG. 100-1000 PAS-residues can be used to tailor the desired half-life.
„This technology works like plug-and-play,“ Arne Skerra was happy to state, „most of the compounds of our in-house pipeline have already shown efficacy and we have also an increasing number of collaborations with pharma companies including big-pharma, which I cannot name right now.“ Through the n-terminal amino group, which is the only reactive group, a targeted elongation is relatively easy achievable and leading to monodisperse materials. An example of compounds using PASylations are Fab-fragments, which normally circulate only for a short time. By attaching PAS-chain(s), plasma half-life could be drastically improved from 1.2 hours in mice to more than 24 hours. „As plasma half-life is faster in smaller organisms this would mean we would have a plasma half-life of one to two weeks in humans,“ Arne Skerra explained. An interesting case study was also leptin, inhibiting hunger and in absence is responsible for extreme cases of obesity. By PASylation, the plasma half-life of therapeutic leptins could be greatly increased 50-fold and could reduce the weight of overweight mice by just one dosage per week (while non-PASylated leptin had no effect due to rapid renal clearance). After four shots of PASylated leptins mice with twice the size of normal mice, experiencing diabetes and other obesity-related symptoms, have turned back to normal weight and got rid of obesity-symptoms.
In-silico Analysis and Design
Some time ago microorganisms have been described as “programmable manufacturing facilities”. Based on that assumption, Lars Nielsen, Professor of biological engineering at the Australian Institute for Bioengineering and Nanobiotechnology, and his group are designing new production cells using algorithms to virtually direct the flux in the desired direction: “Our approach starts with the design in-silico. Based on the results we build cells, test them, learn about reality and transfer the knowledge into a new design.” The challenge is the improvement of the learning process and the integration of the new data into an improved in-silico model that finally leads to reliable results.
Nielsen demonstrated the power of in-silico methods by explaining an industrial approach to producing limonene. The substance is already used as an additive in jet fuels and has the potential to substitute classical petrol based jet fuels. By applying approximate Bayesian computation, considering factors like enzyme kinetics and thermodynamics, and using proteomics, metabolomics, and fluxomics, Nielsens group figured out how to improve limonene production. Therefore, they tested detailed models of the pathways of interest with different assumptions (without regulation, with inhibition or positive acceleration or both influences).
Mats Akesson, a researcher at Biogen, talked about a more traditional approach. He concluded that modeling plays nearly no role in the industry because there are other tools for improvement (like process- or medium development), because of a lack of resources or time for hanging around at a computer or because of a lack of company culture. “Modeling is something for academia,” said Akesson. However, he admitted that industry definitely uses models but doesn’t talk about it. In-silico approaches are used for platform medium development, for the characterization of high or low producers, and for cell line development.
Industrial Biotechnology 2.0
Defining the future of biotechnology was a challenging task for Helene Faustrup Kildegaard, senior researcher at the Novo Nordisk Foundation, Rainer Schneider, associate professor of biochemistry at the University of Innsbruck, and Nigel Titchener-Hooker, professor and dean of the faculty of Engineering Sciences at the University College London. According to Faustrup Kildegaard, next generation genomes belong to the 12 disrupting technologies of our days (among mobile internet or renewables). The latest disruptive technology in biotech is CRIPSR/CAS, a fast easy and accurate way to alter production cells.
Traditional host engineering with Chinese hamster ovary cells (CHO) uses overexpression (via random integration), down-regulation (using RNAs), or knock our methods (mutagenesis or HR mediated). Since the introduction of CRISPR/CAS, either genes could be knocked out or desired genes could be introduced highly specific. “The new bottleneck is the characterization of clones”, says Faustrup Kildegaard, “CRIPSR/CAS facilitates targeted cell engineering a lot.” Cell line development this way was reduced from one year to just three months. Nevertheless, more than this technique is required. The researcher mentioned the optimization of genome editing, characterized expression vector systems, or enhanced genome stability.
Rainer Schneider’s progressive approach is an in-vivo evolution and selection system that has been established in bacteria. The method excels because of an “extremely large mutation spectrum”, an over night selection process and the fact that only new variants with a good stability and solubility survive. The method can be used as a tailored biotech tool for tag-removing or to produce alternatives to antibody therapeutics to treat neurodegeneration or inflammation. Schneider pointed out the chance to be able to produce antibodies with E. coli with strain development times of just a few days.
Nigel Titchener-Hooker talked about dramatic titer increases while downstream processes remain limited. The industry’s scope is to speed up processes using approaches like ultra stage-down plus bioprocess modeling in an early stage of drug evolution. As an alternative to transitional methods like centrifugation and depth filtration, Titchener-Hooker demonstrated the power of tangential flow microfiltration and tangential flow ion exchange chromatography. This way, 3 L bench reactors can replace 50 L-systems and process time shrinks from five hours to five minutes. Perfusion systems take over the role of (more stable) fed-batch reactors which leads to saving 20 % of costs of goods.
Finally, using new production technologies it would be possible to manufacture targeted personalized drugs in a sustainable way by 2025 – if there were no regulations and approvals. However, based on current technology and knowledge, offering targeted drugs every user would be a participant in a study, a test case without exactly knowing the effects of a treatment.