Aggregation and stability problems are one of the major causes of attrition in the development of biopharmaceuticals with a large impact in safety and manufacturing costs. Current methodologies do not offer detailed information on the specific interactions that mediate misfolding and aggregation, particularly in complex systems such as antibodies. The proposed project will combine state-of-the-art computational approaches with in vitro expression and stability data to gain insight into the early events that trigger protein aggregation. It will focus in the use of long (millisecond) time scale molecular dynamics simulations using monoclonal antibodies as an experimental model. Ultimately it is expected that the use of computational tools to predict the behaviour of biological drugs, particularly aggregation and stability, will have a major impact in streamlining the development of safer and better biopharmaceuticals. The proposed project will concentrate on optimising force-fields in order to run meso-scale simulations of antibody fragments with known stability problems. The ultimate goal would be to attempt a simulation on inter-chain interactions (using multiple monomers within a defined unit cell) over a millisecond timescale. This would require access to high performance computational facilities such as HPCx. Professor Mark Sansom's group at Oxford has access to such facilities as well as three powerful (~128 processor) in house clusters. Lonza Biologics plc will provide antibody models to be used in the simulations as well as experimental validation of results. The simulations will allow for elucidating a model of simulated antibody aggregate precursors in the right variant. In addition, the simulations will allow for dynamic comparison in terms of structural and energetic behaviour between different antibody variants. Current Molecular Dynamics simulations in the literature have reported a 30 ns trajectory of a whole antibody using the NAMD code (1). However, the timescale of 30 ns is a relatively short timescale when considering that typically folding events occur on a millisecond timescale, whereas misfolding and aggregation could take significantly longer. Our proposed approach will be more focused on the variable domains of antibodies, as opposed to running simulations of large antibody systems containing the conserved heavy chains. The motivation for this is that variable domains are not only responsible for the binding to antigens, but also seem to be the main contributor to aggregation and stability problems observed in antibodies. Futhermore, by reducing the complexity of the system the timescale of the simulations could be extend into the millisecond scale. Our initial simulation data show a close correlation with experimental observations made internally by Lonza. Aggregation can occur in a variety of flavours (from unfolded, surface interactions, edge interactions) it's also likely that local re-organisation could trigger or be the consequence of aggregation. Gadnell and Gunnarsson have shown cryo-EM structures revealing an Ab with a 'smashed' Fab when aggregated (2). Our proposed simulations will allow for atomic and/or coarse grained level studies of such phenomenon in silico. Another interesting approach will be the investigation of antibodies with surfaces such as membranes. Moreover there is a link between inward rectifying potassium (Kir) channels and immunoglobulins (3) . MSPS' group has a strong international reputation in modelling and simulation of Kir channels,which will allow for the potential PhD candidate to gain greater exposure to the research strengths of the group. References: 1. Chennamsetty N., et al.., J. Mol. Biol., 391, 2, 404-413 (2009). 2. Gadnell M. and Gunnarsson K., Nature Methods 2, 523-524 (2005), 3. Fallen K, et al., Channels (Austin). 3(1): 57-68. (2009)

We have developed a defence mechanism to respond to infection when our body recognises a foreign 'invader'. We have a type of cell known as a B cell in our body, which in response to infection changes into an antibody-producing cell. This is necessary to allow us to fight off infections that would otherwise be very harmful. The antibodies work by attacking the foreign invader and destroying it, thereby clearing the infection and removing the foreign agent. In order for the B cells to change into antibody producing cells (a process known as differentiation), they must adjust the organization of themselves so that they have the required pieces of cellular machinery to make large amounts of antibody. As antibodies are the bodies natural defence against disease, many new antibody type drugs are being developed to help treat human diseases such as cancer and AIDS. However, in order to produce these next generation antibody-based therapeutic 'drugs' we must use mammalian cells to make them. The types of cells we use to make these drugs are not as efficient at producing antibodies as the modified B cell and as a result we are not able to produce enough of these drugs and the cost and demand for them is therefore high. It is thought that this will become even more of a problem as more antibody based drugs are developed. The research proposed here will examine whether we can find mammalian cells with more of the required machinery to produce high-levels of antibodies, or alternatively, if we can manipulate these mammalian cells to produce more of this machinery so that higher yields or amounts of these drugs can be produced more quickly at less cost. At present it is unknown if this is possible, and the process is poorly understood in the mammalian cells used to produce these antibodies. Advanced technology known as proteomics and inducible expression technology will be used to study the differences in the levels of the proteins known to be important for antibody production in differentiated B cells and compare the levels of these proteins in the mammalian cells used for commercial antibody production. We will look for proteins that become either more or less abundant (by altered gene expression, protein synthesis and/or protein degradation) and for subtle molecular modifications to pre-existing proteins known to be able to modify their function (e.g., switch them on or off). Information from current genomics projects will be mined and used in combination with our protein data to identify ways of improving the amount of therapeutic protein 'drug' we can manufacture using these mammalian cells. As stated above, this is extremely important as it is expected that with an increasing number of protein 'drugs' being developed we will lack the capability of producing large enough amounts to meet the required demand for these new drugs.

Small molecule drugs (e.g. antibiotics) have traditionally been the mainstay of treatments and therapies in man, however in the last 10-20 years protein based drugs (e.g. herceptin, which is often used to treat breast cancer) have developed to such a point that these now constitute a significant section of the pharmaceutical market. There are several categories of protein based drugs, one of which, monoclonal antibodies, constitutes the largest number of protein molecules in a class either in use or in clinical trials. Many protein based drugs are challenging to produce because they (a) require particular helper proteins to fold and assemble into their final active state and (b) are decorated on their surfaces by sugars and other molecules that are essential to their bioactivity. Due to the high precision required to produce biotherapeutics, such protein based drugs for the treatment of diseases are usually produced by cells kept in culture under defined conditions. One problem with this is that the cells we use to make proteins for therapeutic uses are not as efficient as we would like them to be and the cells respond to small changes in the environment in which they are grown. This can affect the consistency and quality of the final drug-substance or protein drug. As a consequence, we may not be able to produce enough of these drugs and/or the cost of producing them is too high. This proposal sets out to address a key area that underpins recombinant protein synthesis yields from mammalian cells in culture, the role of trace metals (e.g. magnesium, manganese, iron, zinc, copper, nickel, colbalt) in, and their influence upon, mammalian cell growth and therapeutic recombinant protein (rP) production. The concentrations of such trace metals in the solution in which cells are grown can impact upon the therapeutic protein drug quality (particularly how these impact upon safety and efficacy of the drug substance and batch-to-batch variation/reproducibility of the process used to manufacture it) and heterogeneity. During this project we will build upon the synergistic expertise of the applicants to develop and deliver new understanding of key metal biology related to the cellular processes that ultimately determine recombinant protein heterogeneity and yield from Chinese hamster ovary (CHO) cells. CHO cells are the current gold standard mammalian cell line used in industry to produce therapeutic recombinant proteins. The studies will, for the first time, investigate the role of metal biology extra- and intra-cellularly (both total metal ion concentrations and free/buffered when the metal is bound to proteins) in underpinning the phenotype of recombinant CHO cell lines and determine how metal concentrations, cellular flux, and metal transporters may be manipulated to provide culture processes with better process control (e.g. which metal ions to monitor when screening raw materials). This will lead to more consistent drug substance production, improved safety, efficacy and reduced costs/improved security of the supply chain and longer term with cell lines with enhanced industrial phenotypes e.g. increased and prolonged growth, reduced rP heterogeneity, improved glycosylation profiles. Without improved process control and expression systems the biotechnology/pharmaceutical industries will lack the capability to produce large enough amounts of these valuable and effective drugs to meet the demand at a price that is affordable for health care providers.

The biopharmaceutical industry relies upon screening and adaptation procedures to identify stable, productive transfectants able to grow in defined media to produce a recombinant protein (rP). This procedure is not ideal early in the drug development pipeline when it is necessary to screen and evaluate potential rP drug targets and/or when the more rapid generation of recombinant material would allow more molecules into first-in-human studies faster. Transient gene expression therefore provides a means for the generation of rapid amounts of rP for such studies. Transient expression technology for the production of biotherapeutic relevant rPs has mostly utilised suspension adapted human embryonic kidney cells (HEK293 cells) and calcium phosphate precipitation although recently Chinese hamster ovary (CHO) cells have been used. Indeed, a number of companies are investigating transient expression technology to produce hundreds of mgs of rPs in a matter of days using non-viral DNA delivery systems and cells cultured in bioreactors. Recent reports also demonstrate the potential to improve titres in transient processes using a systematic approach including optimisation of the vector and transfection conditions, use of cell cycle regulators (p18, p21) and fibroblast growth factor, and addition of valproic acid to the media. The optimisation of the ratio of IgG heavy and light chain genes and the use of fed-batch cultures when generating IgG material using CHO transient monoclonal antibody (mAb) expression have been reported to improve yields whilst gene optimisation of heavy and light chain transcripts also increases mAb production levels during transient expression in CHO cells. The project Lonza Biologics has a CHO suspension adapted host cell line (CHOK1SV) which is able to achieve high cell concentrations and maintain viability for long periods of time compared to other suspension adapted CHO lines. Lonza also has its own proprietary vector expression system, the GS System. In this project the student will develop a novel CHOK1SV and GS vector based (DNA) transient expression system for the rapid production of 100's of mg's of rPs. This will build upon preliminary work from both the academic and industrial labs that show manipulation of transient systems can lead to increased yields. The test molecules will include a model mAb (gene optimised and non-optimised), tPA and rhEPO. Initially the student will investigate delivery systems and internalisation of the DNA to the nucleus after modification of the vector (e.g. via use of nuclear localisation sequences and partitioning elements). The student will also determine the influence of environmental conditions (e.g. temperature, pH) and media composition and feeding on product yield. The student will also investigate host cell engineering strategies to improve yield and speed of the transient expression process. In particular the student will focus upon cell cycle engineering (e.g. p21) and chaperones and foldases that we have shown in the laboratory in transient studies increases the high-level transient expression of model proteins (e.g. Torsin A, translation initiation factors). Finally, expression of the Vaccinia E3L protein in CHO strains has a positive effect upon transient expression of mAb. E3L acts to block the activation of cellular genes that respond to viral infection, in particular PKR. Work in the lab at Kent has shown that PKR activity can also be modulated by manipulation of p58 involved in preventing or modulating the activity of this kinase. The student will therefore investigate the effect of modulating PKR activity on transient yields. The outcomes: 1. Development of novel vector, CHOK1SV derived host and processing (media/feeds) technology for the expression of 100's mg of rPs 2. An understanding of the limitations upon transient expression of rPs from CHOK1SV cells

This project aims to develop a considerably improved method for the purification of monoclonal antibodies (mAbs) at a preparative, pilot or industrial scale by using a new, emerging technology with a liquid stationary phase. In 2005 the sales of protein therapeutics was estimated to be worth over $50 billion, of which $15 billion was contributed by antibodies. It has been projected that antibodies will show three fold faster growth than other proteins of 21% per annum over the next 5 years to 2010 reaching over $35 billion (Datamonitor 2006). There were 21 protein blockbusters in 2005 versus 6 in 2000 and 7 of the last 11 were antibodies. This is led by the increased interest shown by major pharma companies such as AstraZeneca, GSK and Merck in acquiring new protein drugs. Biologics are predicted to account for 57% of large pharma growth between 2004 and 2010. Since outsourced production is expected to increase from 58% to 69% over the 5 years from 2006, Lonza Biologics, as one of the largest contract manufacturing organisations (CMO) and therefore a predominant force in the process development and production of antibodies to the biologics sector, have a vested interest in production methods for antibodies. It has been noted by governments that the supply of protein therapeutics has a high cost and this has been illustrated by several high profile cases in the British courts where patients have won access to drugs previously denied to them by NICE. As a result there have been agency lead projects to reduce these costs, such as the AIMS project in Europe and the FDA's 21st Century Initiative, which have tried to identify and lower the barriers to entry of new production techniques. This is on top of the industry's efforts to lower such costs internally. At Lonza, the Downstream Processing (DSP) Process Engineering Group has been formed to dedicate resource to investigating such innovative technologies and the use of liquid / liquid extraction by counter current methods is seen as one of the key emerging technologies in this field. MAbs can be produced by cell fermentation and this is the approach adopted by Lonza. The titre in these fermentations is increasing to generally over 5g/L and can be as much as 10g/L. Fermentations can be as large as 20,000L with as much as 200kg mAb coming from a single fermenter. With such masses reaching the purification processes, traditional fixed bed columns can no longer cope with the load placed on them and multiple cycles are required, leading to plant throughput issues. This, together with the limited binding capacity of such columns, means that resin lifespan has been decreased to fewer batches. Combined, these issues have generated an important need to investigate alternatives. This project aims to develop a different method of purification, one that uses a liquid rather than a solid stationary phase. It can therefore operate with the presence of particulate matter direct from the fermentation, and be scaled up to large sizes without pressure problems. The aim is to develop a separation protocol for the preparative-scale purification of monoclonal antibodies using counter-current chromatography (CCC) or centrifugal partition chromatography (CPC). This protocol may use an aqueous two-phase system (ATPS) or an aqueous-organic solvent system. It may also adopt a novel approach such as the use of ionic liquids, the use of affinity ligands within the liquid stationary phase, or the use of a unique continuous counter-current extraction process, approaches that can only be employed when the stationary phase is a fluid. In summary, the goal is to purify mAbs directly from cell fermentations using liquid-liquid technology at a scale up to 200kg per fermenter. To achieve this goal, the research project will be wide ranging, looking at a number of possible options, though always taking full advantage of the fluid nature of the stationary phase.