Project title: Rapid microscale evaluation of the impact of fermentation conditions on inclusion body formation, solubilisation and protein refolding yields. Hypothesis: That protein refolding steps can be performed and optimized at a microlitre scale and that the technique, once established, can rapidly evaluate the impact of earlier bioreactor stages on whole bioprocess performance. Significance and Background: Microscale processing techniques offer the potential to speed up the delivery of new drugs to market to reduce development costs and thereby increase patient benefits. We and others have shown that in selected cases the study of bioprocess unit operations in microwell plate formats and the use of automation can significantly enhance experimental throughput and facilitate the parallel evaluation of a large number of process conditions (eg Nealon et al. 2005, Jackson et al., 2006, Lacki, 2007). While the majority of microscale studies have focused on microbial fermentation, by comparison little work has been done on downstream processing operations. This potential bottleneck requires considerable attention if significant step-wise process enhancements are to be gained. In particular the impact of fermentation conditions on whole process performance must be understood if downstream processing is not to become rate limiting (Micheletti and Lye, 2006). A particular example that could benefit from this approach is the refolding of recombinant protein from inclusion bodies (IB). Protein refolding yields at industrially relevant concentrations are restricted by aggregation of protein upon dilution of the denatured form. A number of studies have investigated chemical (Buswell and Middleberg, 2002, Mannall et al., 2007a) as well as physical (Mannall et al., 2005) factors affecting the dilution refolding in small (20-200ml) bioreactors, however for the majority of proteins a large number of refold conditions usually need to be tested in order to optimize this processing step. The use of microwell plates as a format in which to perform protein refold experiments has recently been preliminary investigated (Mannall et al., 2007b). Not only has it been shown that refold reactions scaled well between microwell and bench scale operations but it was also demonstrated that a significant amount of information can be gained in a short period of time using a small amount of the valuable IB-derived protein. In this project we propose to radically enhance the microwell approach for the rapid optimization of the protein refolding step by: 1) adopting a whole process approach to optimization which investigates the effect of fermentation conditions on subsequent refolding yields and product quality 2) establish the automation of the refolding step, both in terms of liquid handling operations and associated analytical methods, to speed up the investigation of multiple variables under different mixing conditions. Research Program: 1) Establish and demonstrate an automated microwell-based dilution-refolding system using an industrially relevant strain provided by Avecia Biologics 2) validate the dilution-refolding system by comparing the yields obtained to standard bench scale operations 3) generate E. coli fermentation broths using different carbon sources, induction time and window and harvest time strategies and then use the microwell format in investigating the effect of the upstream parameters on IB solubilisation and refolding yields. Jackson et al. (2005) J. of Membrane Sci., 276, 31-34. Nealon et al. (2006) Chem. Eng. Sci., 61, 4860-4870. Lacki et al. (2007) ACS National Mtg., Boston BIOT-472. Micheletti and Lye (2006) Curr. Opinion in Biotechnol., 17, 611-618. Buswell and Middleberg (2002) Biotech. Progress, 18, 470-475. Mannall et al. (2007a) Biotech. Bioeng., 97, 1523-1534. Mannall et al. (2006) Biotech. Bioeng., 93, 955-963. Mannall et al. (2007b) Biotech. Bioeng., submitted.

The broad theme of the research training addresses the most rapidly developing parts of the bio-centred pharmaceutical and healthcare biotech industry. It meets specific training needs defined by the industry-led bioProcessUK and the Association of British Pharmaceutical Industry. The Centre proposal aligns with the EPSRC Delivery Plan 2008/9 to 2010/11, which notes pharmaceuticals as one of the UK's most dynamic industries. The EPSRC Next-Generation Healthcare theme is to link appropriate engineering and physical science research to the work of healthcare partners for improved translation of research output into clinical products and services. We address this directly. The bio-centred pharmaceutical sector is composed of three parts which the Centre will address:- More selective small molecule drugs produced using biocatalysis integrated with chemistry;- Biopharmaceutical therapeutic proteins and vaccines;- Human cell-based therapies.In each case new bioprocessing challenges are now being posed by the use of extensive molecular engineering to enhance the clinical outcome and the training proposed addresses the new challenges. Though one of the UK's most research intensive industries, pharmaceuticals is under intense strain due to:- Increasing global competition from lower cost countries;- The greater difficulty of bringing through increasingly complex medicines, for many of which the process of production is more difficult; - Pressure by governments to reduce the price paid by easing entry of generic copies and reducing drug reimbursement levels. These developments demand constant innovation and the Industrial Doctorate Training Centre will address the intellectual development and rigorous training of those who will lead on bioprocessing aspects. The activity will be conducted alongside the EPSRC Innovative Manufacturing Research Centre for Bioprocessing which an international review concluded leads the world in its approach to an increasingly important area .

Biopharmaceutical manufacturing continues to evolve with an increased emphasis on underpinning science and engineering. Effective deployment of contemporary knowledge in science and engineering throughout the product life cycle will facilitate manufacturing efficiencies and regulatory adherence for biopharmaceuticals. Fundamental to this paradigm shift has been the drive to adopt an integrated systems approach based on science and engineering principles for assessing and mitigating risks related to poor product and process quality. Changes have been enabled as a consequence of the regulatory authorities introducing a new risk-based pharmaceutical quality assurance system. The traditional approach to manufacture has been to accommodate product variability into the specifications and fix operational strategies to ensure repeatability. Developments in measurement technology have invited changes in operational strategy. This revised approach is based on the application of Quality by Design (QbD), underpinned by process analytical technology (PAT) to yield products of tighter quality and more assured safety. QbD is defined as the means by which product and process performance characteristics are scientifically designed to meet specific objectives. Practical improvements therefore demand a knowledge base of science and engineering understanding to identify the interrelationship between variables and integrate the learning into different manufacturing scenarios. The focus of the Centre is to address the challenges emerging from this paradigm shift and to train a new generation of students with competencies in all stages of commercial biopharmaceutical process development. Critical to this is to ensure they have the skills to work at the discipline interfaces in the areas of biosystem development, upscaled upstream process engineering, and the engineering and development of downstream processing. The training will be formulated around three elements that form the backbone of achieving an enhanced understanding of the process. The three elements are (i) Measurement, Data and Knowledge Management, (ii) Enhance Available Knowledge and (iii) Use Knowledge More Effectively. The power of the approach being adopted is that it is equally applicable to established bioprocesses based on microbial and animal cell culture, as well as emerging areas including stem cells, marine biotechnology and bio-nanotechnology. The rationale for proposing a Centre in this area is to address a well recognised problem, a lack of appropriately trained personnel, who will deliver the next generation of biopharmaceutical development. These issues have been clearly articulated in a series of reports. SEMTA reported that over a quarter of bioscience companies do not have sufficient science skills. 39% of bioscience/pharmaceutical companies have long-term vacancies; with 22% having skill shortages in the science arena (five times that for other sectors). Lord Sainsbury, concerned at the rapidly changing nature of the bioscience business, set up the BIGT and commissioned Bioscience 2015. One of the strong messages raised was the serious shortfall in trained staff. Furthermore a quantitative assessment of the increase needed of trained people entering the sector was made by bioProcessUK. They estimated an increase of 100 trained personnel was required on top of the current 150 doctoral level candidates graduating per year. It is not simply a matter of increasing the number of trained persons. The Centre will also address the limitations of the current UG training of engineers, chemists and biologists which does not prepare them for the challenge of working in process development distinguished by disciplinary interfaces. The proposed programme will address a strategic shortfall and produce a new generation of graduates with the appropriate inter-disciplinary skills to drive both the research agenda and knowledge transfer of underlying concepts into industry.

As the number and volume of therapeutic proteins produced commercially has risen rapidly in recent years the problem of protein misfolding and aggregation have come to be more widely recognised. A common biopharmaceutical expression system is E coli which in high productivity systems can frequently express the recombinant protein as an inclusion body. Recovery of native structure requires solubilisation and then refolding of the protein. Due to competition between the processes of refolding and aggregation these can be low yielding processes carried out at high dilutions. Additionally during protein processing and in formulation protein aggregation can be manifest arising from concentration and solution conditions. While there has been much work in this area, much of the current work reported has a largely empirical emphasis due to both the experimental complexities and restricted to individual cases making it difficult to draw general learning and scope for new approaches Despite the enormous number of configurations of a polypeptide chain, often the protein folding into the native state can be rapid and robust. However, subtle and poorly understood variations in the forces between the solvent (as a function of ionic strength and pH) with the different part of the polypeptide backbone chain and it's side chains leads to misfolding and aggregated proteins. Surprising small variations can lead to substantially different outcomes! The pathways by which polypeptides misfold and aggregate, and the stability of these aggregates relative to the natively folded protein are of considerable industrial relevance and academic importance. The research work proposed here will be based on an experimentally focussed plan exploiting improvements in current research instrumentation for following protein aggregation using Dynamic Light Scattering (DLS), Analytical Ultracentrifugation (AUC), Dynamic Surface Tension/Surface Rheology (DSTSR), Fourier Transform Infrared Spectroscopy (FTIR) and Size Exclusion Chromatography (SEC). DLS especially has shown significant improvements in performance in recent years and is well suited for studying both steady state protein size as well as the kinetics of protein aggregation in solution. As a CMO Avecia potentially has access to a number of both model and industrially relevant protein systems and part of the pre-initiation activities of the project will be to determine the most useful set of proteins to include in the study. A suggested initial model is rPA (anthrax protective antigen) a 80kD single subunit protein with a known 3D structure, refolding conditions and known aggregation propensity. This and other model proteins will be experimentally investigated systematically with the plans to : (i) Determine the sensitivity to aggregation based on variations in pH, ionic strength, counter-ions, concentration, additives, surfactants and temperature (ii) Identify common mechanisms which drive aggregation phenomena including the importance of misfolding and partial unfolding. (iii) Apply well established colloidal aggregation (DVLO) and flocculation theories to problems in protein aggregation phenomena. (iv) Consider solution hydration/dehydration behaviour of proteins, exploiting the known like behaviour of water soluble polymers including the complex competition between proteins, counter ions, stabilisers for the water molecules. (v) Identify the role played by surfaces as well as contaminant particles in aggregation phenomena (vi) Propose methods for predicting protein stability to aggregation and well as strategies to mitigate the risks of aggregation.

The specific productivity of the majority of recombinant proteins secreted by Pichia pastoris is relatively poor. Identification of high producing recombinant strains is largely an empirical process with little understanding of the characteristics that lead to high productivity on scale up of culture conditions from shake flask scale to high cell density industrial fermentations. Current practice when expressing proteins using Pichia pastoris is to produce recombinants using integrating vectors leading to position independant expression. The most widely used system is based on the alcohol oxidase promoter (AOX) which requires derepression and then induction by methanol. Copy number of the target gene is an important variable but does not necessarily correlate with titre. Selection of the 'best' clone for scale up and manufacturing is complicated by the fact that clones achieving high secreted protein titre in shake flasks do not always achieve high titre in high cell density fermentations. This is also significantly influenced by the impact of methanol utilisation phenotype (mut+, mutS or mut-) of the host. Studies (limited to laboratory processes) have shown that poor expression in Pichia pastoris can in some cases be correlated with folding problems in the ER, which induces an unfolded protein response (UPR). BRIC project 06/18 (David Leak, Imperial College (IC)) is (1) developing metabolomic methods and applying these with transcriptomics to understand the underlying physiology of UPR stressed, non-stressed and non-expressing cultures and, (2) developing easily monitored reporters of UPR to allow ideal expression conditions to be determined on a small scale. This project will use industrially relevant proteins (induced with methanol and non-methanol) and protein models (low and high secretion titres) with an industrial fermentation platform to characterise and better understand the inter-play between the parameters described above and induction of the UPR (UPR tools developed by BRIC 06/18). Metabolomic information (based on extracellular (footprint) metabolites) will also be collected. The student will spend a significant amount of time working at Avecia: 2-3 model proteins, AOX recombinants in range of mut hosts, establishing copy number and titre in shake flask culture, establishing research cell banks of clones for further work (matrix of mut/copy number/titre), establishing fermentation performance of 4-6 clones in 15L fed-batch fermentations - including analysis for footprint metabolites and/or UPR linked GFP expression. The above work is envisaged to take 15 mo with the student spending an equal amount of time at IC (see below) and Avecia. The student will initially work at IC (ca.4 mo) becoming familiarised with the UPR reporter systems being developed and will making constructs (e.g. GFP in various mut backgrounds) necessary for integration into the work at Avecia. The first half (15-18 mo) will define a pattern/decision matrix. Depending on the outcome of this we will either: (1) Expand work with GFP, which provides us with a ready made system for the sorting of variants, making it ammenable for high-throughput analysis to get a more fundamental understanding of the causes of UPR. This coupled with the transcriptomic analysis being done at IC (BRIC 06/18) should point to some host features that can be developed further to improve secretion and protein titre. Or, (2) Investigate the possibility of online control of UPR, which would require taking one/more of the poorly expressing constructs and dissecting the expression conditions to establish the conditions at which specific productivity plateaus/declines and then then investigating whether we can develop a feedback control regime which keeps at this level. The second phase of the work will be done mainly at IC, but decisions as to which route to follow will involve all parties taking into account the timescale and academic needs of the student.