A platform for the optimisation of metabolic pathways for glycosylation to achieve a narrow and targeted glycoform distribution

Lead Research Organisation: Imperial College London
Department Name: Life Sciences


Recently, the development of treatments for new disease has shifted away from traditional chemical compounds and towards protein therapeutics (biopharmaceuticals) like antibodies for the treatment of cancer and hormones for chronic diseases. Nearly 70% of these protein therapeutics have sugar molecules attached to them naturally which affect their function and how long they remain in the body. Because the sugars are so important for the drug function, one of the biggest problems in their manufacture is how to control what sugars are added (glycoform) and to ensure that all the proteins produced have the same sugars on them (homogeneous glycoform profile). Current production methods yield a non-homogeneous mix of glycoforms. Also, different glycoforms interact with the immune system in different ways, so it would be of benefit to be able to produce certain glycoforms over others depending on what the drug is and how it is meant to function. Our goal is to develop technology to rapidly determine the effects of different production methods on which glycoforms are produced and how homogeneous the glycoform profile is. To do this we will develop proteins which are produced inside the cells that are also producing the biopharmaceutical that report the concentrations of nutrients that are already known to influence glycoforms. Alongside, we will develop a computer model of the metabolism of the cells which can predict which glycoforms are produced. Using these two together, we should be able to design new media for the cells to use that result in a more homogeneous glycoform profile which we can change based on what the cells are fed with. We can also suggest genetic changes to the cells that would further help us produce a single, designed glycoform. This could lead to the production of drugs that are safer and require lower doses because they have a single glycoform attached which is the most appropriate for the function of that drug.

Technical Summary

Most biopharmaceuticals are glycoproteins and the composition of the glycans has impacts on their activity, stability, and immunogenicity. The ability to produce glycoproteins of a single, defined glycoform would allow designer drugs with maximum stability and reduced immunogenicity that interact with the immune system in the most appropriate way for the disease which they are designed to treat. However, current production methods, most of which involved mammalian cell culture, produce a heterogeneous mix of glycoforms and result in the need for higher doses to ensure efficacy. In theory, both medium design and cellular engineering could be used to manipulate the glycoform profile of a biopharmaceutical on a case-by-case basis. However, this would require a high throughput screening method in which various combinations of media and production strains could be tested to determine which produces the most narrow and appropriate glycoform. The goal of this project is to utilise a combination of genetically-encoded FRET biosensors for primary metabolites whose concentrations are known to affect glycoform profile and a full mathematical model of the production of nucleotide sugar donors, protein secretion and glycosylation in order to accurately predict glycoform profile in a miniaturised assay. In the long term, the platform could be transformed into a fully automated system that takes real-time readings of the fluorescent signal from a culture plate with various cell lines or media to be screened and transmits the results to the model, which calculates (a) the cell growth and metabolic profiles and (b) the expected glycomic profiles for each cell line or condition.

Planned Impact

The proposed work will develop a platform for glycosylation engineering that will allow for testing of conditions to produce glycoproteins with a narrow profile in high throughput. We expect that the work will have economic and societal impacts as well as benefiting the academic community. Biopharmaceutical production has been expanding rapidly in recent years and this trend is expected to continue. Attracting further investment from industrial biotechnology companies has been highlighted as a priority for the government with the formation of the Bioscience Innovation and Growth Team (BIGT). We will engage with the BRIC industrial members at the annual dissemination meeting to ensure that we pursue the correct model systems and targets to ensure that they get the most benefit out of the results. In addition, the platform could equally be applied to other expression systems, particularly microbial ones which are gaining prominence. This could be the target of future applications to the research councils, CASE studentships, or industrial partnership awards using the contacts which are gained through our involvement in BRIC. The proposed work also has significant potential for training researchers for future employment in the industrial biotechnology industries, another goal of the BIGT. We expect that the two PDRAs employed on the project, as well as students associated with satellite projects will gain a significant number marketable skills. The interdisciplinary environment, supported by biannual dissemination meetings within the project investigation group as well as contact with other researchers working on BRIC supported grants will also further enhance their training. For instance, in BRIC Phase 1, several of the PDRAs formed networks which met outside of the dissemination meetings to discuss research techniques. We will ensure that the PDRAs gain valuable transferable skills in scientific writing through preparation of publications and patents and in public speaking through conference presentations. We have included funds for conference travel for the PDRA in our budget. The ability to produce glycoproteins of a single, designed glycoform would greatly increase the utility of these as medications by increasing their stability within the patient, decreasing their immunogenicity, and allowing us to stimulate specifically the part of the immune system that would be most useful in treating the disease. This would result in the need for less glycoprotein in each formulated dose, which in turn could ease problems in downstream purification and storage. All of these would combine to decrease the cost of production and therefore the cost to the consumer and/or the NHS. This will, in turn, increase the accessibility of these types of medications. We will ensure that our results are disseminated widely through active public engagement and through engagement with policy makers and patient lobby groups. For more information on ensuring impact in the scientific community please see academic beneficiaries (above) and the dissemination and exploitation section of the case for support.


10 25 50
Description The aim of the award was to gain a better understanding of how what we feed mammalian cells in industrial cell culture impacts their efficiency at making the protein of interest and the final quality of that product. This was achieved through a combination of mathematical modelling and experimentation. In order to be able to fit the parameters of the mathematical model, we also developed a high throughput way of measuring the response of cells under different conditions. This award led to the development of a modelling framework that links mathematical models at multiple scales to predict the final sugar structures on a monoclonal antibody as an example protein. The model connects the extracellular concentrations of key nutrients to cell growth, the intracellular concentrations of metabolites needed for protein production and sugar addition (glycosylation), and the glycosylation reactions in the secretory pathway. The model was additionally used to predict the outcome of various feeding strategies and to design a feeding strategy that would allow for increased addition of galactose on the protein product (a property that is associated with increased efficacy). The model predictions and optimisation were validated in experiments. With respect to high throughput measurements, we developed fluorescent biosensors based on Forster Resonance Energy Transfer between fluorescent proteins. We showed that they could be used to measure the intracellular concentrations of glucose and glutamine in living cells and that the response of these cells to feeding could be quantitatively measured, which facilitates mathematical modelling. We also developed a novel chromatography assay for measuring the concentrations of nucleotide sugar donors, which are key precursors for the glycosylation reaction. This method was transferred to an industrial collaborator (Symphogen) and has also been used in additional projects within the research group.

Overall, the project employed three PDRAs who were trained at the interface of computational and experimental bioprocessing. Two of these went on to be employed in industry (one in a computational environment, the other experimental) and one started an independent academic position at University College Dublin. The project also led to two successive industrial CASE studentships with MedImmune (now AstraZeneca) working on further developing our understanding of the link between process conditions and product quality.

Significant additional collaborative projects were developed with GSK and with Symphogen, who specialise in therapeutic antibody production. In the collaboration with GSK, we examined how the mathematical model could be extended to different cell lines and products given a new set of experimental data. This led to a new framework for how such models can be adjusted, which is currently being explored in an additional PhD project funded in the Department of Chemical Engineering
Exploitation Route This project was funded through the Bioprocess Research Industry Club (BRIC). Because of the industrial involvement in the club the research was shaped as the grant progressed according to the needs and desires of the biopharmaceutical industry. Therefore we developed technologies that they ultimately had a heavy interest in using as exemplified by the PhD studentships and further research projects with MedImmune (now AstraZeneca), GSK, and Symphogen.

Additionally, other research groups have adopted the Forster Resonance Energy Transfer biosensors, the chromatography method, and mathematical models developed by the work. We have sent the FRET sensor plasmids to three other research laboratories (in the US and Europe). The chromatography method has been adopted by industry and has also been used for other academic work both in the study of glycosylation and to measure the concentrations of nucleotides in cell-free reactions. Finally, the modelling framework has been widely adopted (the original research paper has been cited 100 times and many of the follow up papers have more than 50 citations at this time).
Sectors Healthcare,Pharmaceuticals and Medical Biotechnology

Description The model of nucleotide sugar donor metabolism and glycosylation developed in this project has attracted interest from the industrial members of the BRIC community and internationally. This resulted in a collaboration with Symphogen that resulted in two publications as well as a project funded by GSK. In the GSK project, the researcher used in house data to re-parameterise the model for industrial cell lines so that it could be used to predict the effects of changes in the bioprocess on glycoform.
First Year Of Impact 2016
Sector Pharmaceuticals and Medical Biotechnology
Impact Types Economic

Description BBSRC CASE Studentship
Amount £103,042 (GBP)
Funding ID BB/M016315/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 10/2015 
End 09/2019
Description BBSRC Impact Award
Amount £15,000 (GBP)
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 11/2015 
End 03/2016
Description BRIC CASE Studentship
Amount £103,932 (GBP)
Funding ID BB/J003808/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 10/2011 
End 10/2015
Description Industrial sponsor
Amount £66,000 (GBP)
Organisation GlaxoSmithKline (GSK) 
Sector Private
Country Global
Start 03/2016 
End 09/2016
Description GSK 
Organisation GlaxoSmithKline (GSK)
Country Global 
Sector Private 
PI Contribution We are training the model developed in the project to the data collected by the partner.
Collaborator Contribution They have provided data and funding for postdoctoral researchers (see also further funding).
Impact None yet
Start Year 2016
Description Johns Hopkins Protein Glycosylation Modelling 
Organisation Johns Hopkins University
Department Department of Chemical and Biomolecular Engineering
Country United States 
Sector Academic/University 
PI Contribution Collaboration on protein glycosylation modelling
Collaborator Contribution Academic research exchange
Impact No outputs to date
Start Year 2017
Description Symphogen 
Organisation Symphogen A/S
Country Denmark 
Sector Private 
PI Contribution We have analysed samples from industrial CHO cultures and adapted the model to a different cell line and different growing conditions
Collaborator Contribution The partner contributed samples to the project
Impact Two publications resulted: 1. Fan Y, Del Val IJ, Mueller C, Sen JW, Rasmussen SK, Kontoravdi C, Weilguny D, Andersen MRet al., 2015, Amino Acid and Glucose Metabolism in Fed-Batch CHO Cell Culture Affects Antibody Production and Glycosylation, BIOTECHNOLOGY AND BIOENGINEERING, Vol: 112, Pages: 521-535 2. Fan Y, Del Val IJ, Muller C, Lund AM, Sen JW, Rasmussen SK, Kontoravdi C, Baycin-Hizal D, Betenbaugh MJ, Weilguny D, Andersen MRet al., 2015, A multi-pronged investigation into the effect of glucose starvation and culture duration on fed-batch CHO cell culture, BIOTECHNOLOGY AND BIOENGINEERING, Vol: 112, Pages: 2172-2184,
Start Year 2012
Description Imperial Fringe 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? Yes
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact A public engagement activity. In 2013 this was followed by a lecture by Greg Winter. The general concepts of biosensors were discussed as well as their application in manufacturing. Demonstration on pH biosensor using red cabbage and several household items of differing pH (washing up liquid, vinegar, cola)

no actual impacts realised to date
Year(s) Of Engagement Activity 2013,2014
Description Launch of Imperial West Campus (March 6, 2013) 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact We were invited to have a stand at the Launch of the Imperial West campus which was attended by Boris Johnson. The purpose of the event was to showcase research aimed towards industrial translation. The biosensors were highlighted as a potential cutting-edge manufacturing technology for bioprocesses in the UK. We produced visual demonstrations (a small bioreactor and some agar plates) as well as posters and a banner

No impacts to date
Year(s) Of Engagement Activity 2013
Description Science Museum Lates 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Public/other audiences
Results and Impact The Science Museum Late in BioRevolution was a public activity at the Science Museum in London. An estimated 200 visitors were invited to see a series of demonstrations. We were hosting 'The Icing on the Cake' which is all about the intricate patterns of sugars which coat modern therapeutic drugs.

The event stimulated awareness among the general public on the role of glycosylation in the modulation of the efficacy of therapeutics.
Year(s) Of Engagement Activity 2014