Multi-modal fluorescence spectroscopy for online analysis of proteins in bioprocesses

Biopharmaceutical proteins are typically purified from clarified fermentation broths using multiple chromatographic steps where each separates the proteins based on one or more physico-chemical feature such as net charge, size, hydrophobicity and biological affinity. The elution peak containing the native protein product at each step is often contaminated by host cell proteins, or with slightly modified versions of the product which can be almost indistinguishable from the native form and therefore very challenging to remove. The precise product profile at each step is therefore very sensitive to small changes in upstream variability, buffer composition and pH, and the gradual fouling of chromatographic resins that affects performance through repeated re-use. It is therefore imperative to be able to monitor the product profile, preferably in-line or on-line, to be able to adjust the process parameters in real time, or to make a responsive decision as to when to start and stop collecting the product fraction within the elution peak. Real time, low cost and low volume analysis of proteins and protein mixtures suitable for online monitoring of chromatography, is generally limited to simple absorbance, refractive index and conductivity measurements which only provide basic peak detection and no detailed characterisation of the protein product profile. Multi-angle laser light scattering (MALLS) has some potential for online characterisation of approximate molecular masses, whereas accurate online mass-spectrometry is too expensive and technically demanding. We will take advantage of our recent advances in both the biophysical analysis and rapid laser-induced temperature perturbations of unlabelled proteins in microcapillary flow, as well as state-of-the-art optical components, to establish a low volume flow-detector for use in chromatography, that can evaluate the heterogeneity of the protein product profile in real time. A single set of optics for the intrinsic fluorescence of proteins will be set up to measure fluorescence intensity (FLI), time-resolved fluorescence (TRF) and fluorescence correlation spectroscopy (FCS) and so simultaneously characterise orthogonal features of the protein product profile. These will measure protein quantity (peak detection), and detect underlying variability in solution conformation, oligomeric state, and particle sizes, including the soluble aggregates. FCS is more sensitive and quantitative for relative particle concentrations than DLS or MALLS which are disproportionately sensitive to larger particles. To further resolve product heterogeneity, the flowing sample (continuously split from the main elution stream) will also be subjected to a rapid (12 ms) temperature jump of up to 70degC using our recently demonstrated microfluidic IR-induced heating technology. This will induce partial structural unfolding of protein domains, and the dissociation of soluble oligomers, with kinetics and amplitudes that are characteristic to each different protein species in the sample, and thus provide further online resolution of the sample complexity relative to known reference standards or previous elution profiles. As an additional bonus, the detector will also be suitable for standalone sample analysis, such as for the profiling of dosage formulations and their viscosities. The benefits of each new spectroscopic mode will be demonstrated for a wide range of relevant proteins from our lab and from other BRIC members, including IgG, Fab, GCSF. To test the range of heterogeneity that can be detected, these will be syringe-pumped into the detector and analysed in partially and fully purified forms, and also after deliberate modification by partial proteolysis, partial unfolding and aggregation at low pH, oxidation, and shear damage. Online application to chromatography will then be demonstrated using a flow splitter to give a continuous flow into the detector in parallel to fraction collection.

Chromatographic elution peaks of native proteins often contain host cell proteins, or modified proteins that are almost indistinguishable from the native form. The precise product profile is very sensitive to upstream variability, buffer composition and pH, and the gradual fouling of chromatographic resins through repeated re-use. It is therefore imperative to be able to monitor the product profile in real time, to be able to adjust the process parameters or make an informed decision to start and stop collecting the product fraction within the elution peak. Real time, low cost and low volume online monitoring of chromatography is generally limited to absorbance, refractive index and conductivity, providing only basic peak detection with no detailed characterisation of the protein species profile. We will build upon our recent advances in intrinsic fluorescence analysis and rapid laser-induced temperature perturbations of unlabelled proteins in microcapillary flow, as well as state-of-the-art optical components, to establish a low volume flow-detector, that can evaluate protein peak heterogeneity in real time. Fluorescence intensity, time-resolved fluorescence and fluorescence correlation spectroscopy will be established in the device for orthogonal measurements of protein quantity, and variability in solution conformation, oligomeric state, and particle size. A sample split from the main elution stream will also be subjected to a rapid +70C temperature jump using our microfluidic IR-induced heating technology. This will induce partial structural unfolding of protein domains and the dissociation of soluble oligomers, with kinetics and amplitudes, characteristic to each protein species in the sample, to provide further online resolution of the sample complexity. As an additional bonus, the detector will also be suitable for standalone sample analysis, such as for the profiling of dosage formulations and their viscosities.

"WHO WILL BENEFIT FROM THE RESEARCH?"

UK-based companies within the BRIC community will benefit from research into robust and effective analytics for bioprocessing. Improved online monitoring of downstream bioprocesses will minimise offline analyses, speed up downstream bioprocess optimisation, and reduce the regulatory risks associated with product integrity and reproducibility through better control of product quality over chromatographic resin lifetimes. It will allow them to more effectively monitor the outputs of bioprocess operations such as chromatography and protein refolding during bioprocess development, manufacturing scale up, and during the final manufacturing stage. The project will establish a new measurement technology for low-cost on-line detection of protein chromatography, and also for off-line analysis of multiple properties for complex protein mixtures. The new instrumentation also fits the bioprocessing research challenges for protein products priority of BRIC, via robust, quantitative and low cost measurement of protein sample complexity, quality and soluble protein aggregate content. By spiking very low concentrations of inert fluorescent particle standards the device will, as a secondary output, also measure viscosity in nL to <1uL samples of dosage formulations at 10-200 mg/ml protein, to provide early injectability assessments. Current 100ul to mL scale rheometers require significant quantities of protein. The project also fits the High-throughput bioprocess development priority, as a new detector for ultra-scale down chromatography optimisations in low volume Hi-Trap FPLC or uHPLC systems.

Potential patients will benefit because the research will significantly improve our characterisation of biopharmaceutical protein products and bioprocess performance at an early stage, aiding reduction in their development times, which is particularly crucial for those addressing previously unmet clinical needs. Benefits to the NHS relate to the possibility of constraining costs. Proteins are innately complex and labile so that bioprocess development times and hence costs tend to be high. The capacity to treat conditions such as rheumatoid arthritis much more effectively in ageing populations is vital but it still poses a problem with respect to stretched NHS budgets.

"HOW WILL THEY BENEFIT FROM THE RESEARCH?"
The research will generate novel and simple to operate online bioprocess monitoring techniques to complement existing ones and therefore better characterise bioprocess performance in real time. This will reveal heterogeneities in protein products much earlier in development than currently possible, by depending less on time-consuming protein analytical services elsewhere, which are typically prioritised for final product characterisations. Real-time characterisation will also enable continuous process control to optimise performance. This will lead to more rapid bioprocess development at pilot scale, and also better product quality control at larger scales. These will enable the UK biotechnology industry to design more efficient bioprocesses that minimise the presence of contaminants with similar properties to the fully native protein product, such as misfolded, truncated, disulphide scrambled or alternatively liganded species.

The UK economy will benefit because academic research will complement the country's strength in bioscience discovery. Collaboration between bioprocess engineers and protein biophysicists on industrially relevant therapeutic proteins will ensure effective knowledge and skills transfer between the science and engineering base and UK industry. This will expand their position in the global healthcare market and attract further R&D investment from global business which recognises the UK as a good place to conduct these activities. Such retention of expertise, know-how and intellectual property will aid the capacity to remain internationally competitive.