Comparison of data science workflows for root cause analysis of bioprocesses.

Root cause analysis (RCA) is one of the most prominent tools used to comprehensively evaluate a biopharmaceutical production process. Despite of its widespread use in industry, the Food and Drug Administration has observed a lot of unsuitable approaches for RCAs within the last years. The reasons for those unsuitable approaches are the use of incorrect variables during the analysis and the lack in process understanding, which impede correct model interpretation. Two major approaches to perform RCAs are currently dominating the chemical and pharmaceutical industry: raw data analysis and feature-based approach. Both techniques are shown to be able to identify the significant variables causing the variance of the response. Although they are different in data unfolding, the same tools as principal component analysis and partial least square regression are used in both concepts. Within this article we demonstrate the strength and weaknesses of both approaches. We proved that a fusion of both results in a comprehensive and effective workflow, which not only increases better process understanding. We demonstrate this workflow along with an example. Hence, the presented workflow allows to save analysis time and to reduce the effort of data mining by easy detection of the most important variables within the given dataset. Subsequently, the final obtained process knowledge can be translated into new hypotheses, which can be tested experimentally and thereby lead to effectively improving process robustness.

Mathematical model of adherent Vero cell growth and poliovirus production in animal component free medium.

Sabin-IPV (or sIPV, inactivated polio vaccine based on attenuated Sabin strains) is anticipated to replace the oral polio vaccine for the endgame in polio eradication. Optimization of sIPV production will lead to a better economically feasible vaccine. To assist process optimization, we studied Sabin type 1 poliovirus (PV) infection kinetics on Vero cells in controlled bioreactor vessels. The aim of our study was to develop a descriptive mathematical model able to capture the dynamics of adherent Vero cell growth and PV infection kinetics in animal component free medium. The model predicts the cell density, metabolites profiles, and viral yields in time. We found that the multiplicity of infection (MOI) and the time of infection (TOI) within the investigated range did not affect maximal PV yields, but they did affect the process time. The latter may be reduced by selecting a low TOI and a high MOI. Additionally, we present a correlation between viral titers and D-antigen, a measure for immunogenicity, of Sabin type 1 PV. The developed model is adequate for further studies of the cell metabolism and infection kinetics and may be used to identify control strategies to increase viral productivity. Increased viral yields reduce costs of polio vaccines with large implications on public health.

Isoelectric point determination of live polioviruses by capillary isoelectric focusing with whole column imaging detection.

Using a capillary isoelectric focusing-whole column imaging detection (CIEF-WCID) method, the isoelectric points (pI) of complete intact polioviruses were determined. The polioviruses that were analyzed are the commonly used viruses for the production of inactivated polio vaccines (IPV)-Mahoney (type 1), MEF (type 2), and Saukett (type 3)-as well as for attenuated oral polio vaccines (OPV) and Sabin types 1, 2, and 3. A method for analyzing biological hazardous components (biological safety level 2) was set up for the CIEF-WCID analyzer used. This method is based on closed circuits. The determined pI's were 6.2 for Mahoney, 6.7 for MEF-1, and 5.8 for Saukett. The pI's of Sabin types 1, 2, and 3 viruses were 7.4, 7.2, and 6.3, respectively. Resolution of the virus peaks was shown to be reproducible. Using this adjusted CIEF-WCID technique, the pI of biologically hazardous components like toxins or viruses can be determined, which is beneficial for the development of vaccine production methods among others.

Scale-down of the inactivated polio vaccine production process.

The anticipated increase in the demand for inactivated polio vaccines resulting from the success in the polio eradication program requires an increase in production capacity and cost price reduction of the current inactivated polio vaccine production processes. Improvement of existing production processes is necessary as the initial process development has been done decades ago. An up-to-date lab-scale version encompassing the legacy inactivated polio vaccine production process was set-up. This lab-scale version should be representative of the large scale, meaning a scale-down model, to allow experiments for process optimization that can be readily applied. Initially the separate unit operations were scaled-down at setpoint. Subsequently, the unit operations were applied successively in a comparative manner to large-scale manufacturing. This allows the assessment of the effects of changes in one unit operation to the consecutive units at small-scale. Challenges in translating large-scale operations to lab-scale are discussed, and the concessions that needed to be made are described. The current scale-down model for cell and virus culture (2.3-L) presents a feasible model with its production scale counterpart (750-L) when operated at setpoint. Also, the current scale-down models for the DSP unit operations clarification, concentration, size exclusion chromatography, ion exchange chromatography, and inactivation are in agreement with the manufacturing scale. The small-scale units can be used separately, as well as sequentially, to study variations and critical product quality attributes in the production process. Finally, it is shown that the scale-down unit operations can be used consecutively to prepare trivalent vaccine at lab-scale with comparable characteristics to the product produced at manufacturing scale.

Development of an animal component free production process for Sabin inactivated polio vaccine.

Inactivated polio vaccine production using attenuated Sabin strains (sIPV) instead of wild type polio viruses (cIPV) is an initiative encouraged by the World Health Organization. This use of attenuated viruses is preferred as it reduces risks related to potential outbreaks during IPV production. Previously, an sIPV production process was set up based on the cIPV production process. Optimizing this process while using only animal component free (ACF) substances allows reduction of operational costs and mitigates risks of adverse effects related with animal derived compounds. Here, development of a process for production of sIPV using only ACF compounds, is described. The upstream process required a change in cell growth medium from serum-containing medium to ACF medium, while virus production media remained the same as the already used M199 medium was free of animal components. In the downstream process multiple modifications in existing unit operations were made including addition of a diafiltration step prior to inactivation. After optimizing each unit operation, robustness of the whole process was demonstrated using design of experiments (DoE) methodology. By using DoE we were able to vary different process parameters across unit operations to assess the impact on our quality attributes. The developed process was robust as the observed variation for quality attributes due to differences in process parameters remained within specification. The resulting pilot process showed not only to be robust, but also to have a considerable higher product yield when compared to the serum containing sIPV process. Product yields are now comparable to the cIPV process based on using wild type polio viruses. Moreover, the potency of the produced vaccine was comparable that of cIPV vaccine. The developed ACF sIPV process can be transferred to vaccine manufacturers at the end-of pre-clinical development phase, at lab- or pilot scale, before production of clinical trial material.

Multivariate data analysis on historical IPV production data for better process understanding and future improvements.

Historical manufacturing data can potentially harbor a wealth of information for process optimization and enhancement of efficiency and robustness. To extract useful data multivariate data analysis (MVDA) using projection methods is often applied. In this contribution, the results obtained from applying MVDA on data from inactivated polio vaccine (IPV) production runs are described. Data from over 50 batches at two different production scales (700-L and 1,500-L) were available. The explorative analysis performed on single unit operations indicated consistent manufacturing. Known outliers (e.g., rejected batches) were identified using principal component analysis (PCA). The source of operational variation was pinpointed to variation of input such as media. Other relevant process parameters were in control and, using this manufacturing data, could not be correlated to product quality attributes. The gained knowledge of the IPV production process, not only from the MVDA, but also from digitalizing the available historical data, has proven to be useful for troubleshooting, understanding limitations of available data and seeing the opportunity for improvements.

Quantitative CPP Evaluation from Risk Assessment Using Integrated Process Modeling.

Risk assessments (RAs) are frequently conducted to assess the potential effect of process parameters (PPs) on product quality attributes (e.g., a critical quality attribute (CQA)). To evaluate the PPs criticality the risk priority number (RPN) for each PP is often calculated. This number is generated by the multiplication of three factors: severity, occurrence, and detectability. This mathematical operation may result in some potential errors due to the multiplication of ordinal scaled values and the assumption that the factors contribute equally to the PPs criticality. To avoid these misinterpretations and to assess the out of specification (OOS) probability of the drug substance, we present a novel and straightforward mathematical algorithm. This algorithm quantitatively describes the PPs effect on each CQA assessed within the RA. The transcription of severity and occurrence to model effect sizes and parameters distribution are the key elements of the herein developed approach. This approach can be applied to any conventional RA within the biopharmaceutical industry. We demonstrate that severity and occurrence contribute differently to the PP criticality and compare these results with the RPN number. Detectability is used in a final step to precisely sort the contribution of each factor. To illustrate, we show the misinterpretation risk of the PP critically by using the conventional RPN approach.

Accelerating bioprocess development by analysis of all available data: A USP case study.

Bioprocess development generates extensive datasets from different unit operations and sources (e.g. time series, quality measurements). The development of such processes can be accelerated by evaluating all data generated during the experimental design. This can only be achieved by having a clearly defined data logging and analysis strategy. The latter is described in this manuscript. It consists in a combination of a feature based approach along with principal component analysis and partial least square regression. Application of this combined strategy is illustrated by applying it in an upstream processing (USP) case study. Data from the development and optimization of an animal component free USP of Sabin inactivated poliovirus vaccine (sIPV) was evaluated. During process development, 26 bioreactor runs at scales ranging from 2.3 to 16 L were performed. Several operational parameters were varied, and data was routinely analyzed following a design of experiments (DoE) methodology. With the strategy described here, it became possible to scrutinize all data from the 26 runs in a single data study. This included the DoE response parameters, all data generated by the bioreactor control systems, all offline data, and its derived calculations. This resulted in a more detailed, reliable and exact view on the most important parameters affecting bioreactor performance. In this case study, the strategy was applied for the analysis of previously produced data. Further development will use this data analysis methodology for continuous enhancing and accelerating process development, intensified DoE and integrated process modelling.

Kinetic model for adherent Vero cell growth and poliovirus production in batch bioreactors.

The production of poliovirus vaccines in adherent Vero cells in batch bioreactors usually consists of a two-step upstream process: (1) Vero cell cultivation on microcarriers and (2) poliovirus proliferation. In this study we developed a mathematical model to describe this two-step process. We introduced the calculation of the oxygen uptake rate (OUR) and a correction of measurement for the sampling effect in order to ensure the high quality data sets. Besides the data of the OUR, we selected glucose concentration, Vero cell concentration and the virus titer for daily in process control to evaluate the progress of the process. With the selected data sets, the described model can accurately describe poliovirus production by Vero cells. Several other regular in process control samples (e.g. lactate concentration, ammonia concentration, and amino acids concentration) were excluded from the model, simplifying the process control analysis and minimizing labor.

Specific production rate of VHH antibody fragments by Saccharomyces cerevisiae is correlated with growth rate, independent of nutrient limitation.

Saccharomyces cerevisiae carrying a multicopy integrated expression vector containing the gene encoding a Llama antibody fragment, has been cultivated in continuous cultures both under carbon and nitrogen limiting conditions with galactose as the sole carbon source. VHH-R2 expression was under control of the inducible GAL7 promoter. Induction however, was independent of the galactose consumption rate and maximal at all growth rates. VHH-R2 was secreted with 70% efficiency at all growth rates and under both limitations. The specific production rate increased linear with increasing growth rate in a growth-associated manner. However, when grown under nitrogen limitation at growth rates above 0.09 h(-1), the extracellular VHH-R2 was less active or part of the VHH-R2 was in an inactive form. From our results we conclude that to obtain a maximal amount of VHH per kilogram biomass per hour, VHH production should be done in carbon limited continuous cultures at high specific growth rates.

sIPV process development for costs reduction.

Polio is expected to be eradicated within only a few years from now. Upon polio eradication, the use of oral polio vaccines, which can cause circulating and virulent vaccine derived polio viruses, will be stopped. From this moment onwards, inactivated polio vaccines (IPV) will be used for worldwide vaccination against polio. An increased demand for IPV is thus anticipated. As a result, process development studies regarding the IPV production process, developed in the 1960s, have intensified. Studies on yield optimization aiming at costs reduction as well as the use of alternative polio viruses, which are more biosafe for manufacturing, are actual. Here our strategy to setup a new IPV production process using attenuated Sabin polio virus strains is presented. Moreover, aspects on reduction of the costs of goods and the impact of process optimization on sIPV costs are reviewed.

Assessment of mass transfer and mixing in rigid lab-scale disposable bioreactors at low power input levels.

Mass transfer, mixing times and power consumption were measured in rigid disposable stirred tank bioreactors and compared to those of a traditional glass bioreactor. The volumetric mass transfer coefficient and mixing times are usually determined at high agitation speeds in combination with sparged aeration as used for single cell suspension and most bacterial cultures. In contrast, here low agitation speeds combined with headspace aeration were applied. These settings are generally used for cultivation of mammalian cells growing adherent to microcarriers. The rigid disposable vessels showed similar engineering characteristics compared to a traditional glass bioreactor. On the basis of the presented results appropriate settings for adherent cell culture, normally operated at a maximum power input level of 5 W m(-3) , can be selected. Depending on the disposable bioreactor used, a stirrer speed ranging from 38 to 147 rpm will result in such a power input of 5 W m(-3) . This power input will mix the fluid to a degree of 95% in 22 ± 1 s and produce a volumetric mass transfer coefficient of 0.46 ± 0.07 h(-1) .

Improved poliovirus D-antigen yields by application of different Vero cell cultivation methods.

Vero cells were grown adherent to microcarriers (Cytodex 1; 3 g L(-1)) using animal component free media in stirred-tank type bioreactors. Different strategies for media refreshment, daily media replacement (semi-batch), continuous media replacement (perfusion) and recirculation of media, were compared with batch cultivation. Cell densities increased using a feed strategy from 1×10(6) cells mL(-1) during batch cultivation to 1.8, 2.7 and 5.0×10(6) cells mL(-1) during semi-batch, perfusion and recirculation, respectively. The effects of these different cell culture strategies on subsequent poliovirus production were investigated. Increased cell densities allowed up to 3 times higher D-antigen levels when compared with that obtained from batch-wise Vero cell culture. However, the cell specific D-antigen production was lower when cells were infected at higher cell densities. This cell density effect is in good agreement with observations for different cell lines and virus types. From the evaluated alternative culture methods, application of a semi-batch mode of operations allowed the highest cell specific D-antigen production. The increased product yields that can easily be reached using these higher cell density cultivation methods, showed the possibility for better use of bioreactor capacity for the manufacturing of polio vaccines to ultimately reduce vaccine cost per dose. Further, the use of animal-component-free cell- and virus culture media shows opportunities for modernization of human viral vaccine manufacturing.

Next generation inactivated polio vaccine manufacturing to support post polio-eradication biosafety goals.

Worldwide efforts to eradicate polio caused a tipping point in polio vaccination strategies. A switch from the oral polio vaccine, which can cause circulating and virulent vaccine derived polioviruses, to inactivated polio vaccines (IPV) is scheduled. Moreover, a manufacturing process, using attenuated virus strains instead of wild-type polioviruses, is demanded to enhance worldwide production of IPV, especially in low- and middle income countries. Therefore, development of an IPV from attenuated (Sabin) poliovirus strains (sIPV) was pursued. Starting from the current IPV production process based on wild type Salk strains, adaptations, such as lower virus cultivation temperature, were implemented. sIPV was produced at industrial scale followed by formulation of both plain and aluminium adjuvanted sIPV. The final products met the quality criteria, were immunogenic in rats, showed no toxicity in rabbits and could be released for testing in the clinic. Concluding, sIPV was developed to manufacturing scale. The technology can be transferred worldwide to support post polio-eradication biosafety goals.

Inactivated polio vaccine development for technology transfer using attenuated Sabin poliovirus strains to shift from Salk-IPV to Sabin-IPV.

Industrial-scale inactivated polio vaccine (IPV) production dates back to the 1960s when at the Rijks Instituut voor de Volksgezondheid (RIV) in Bilthoven a process was developed based on micro-carrier technology and primary monkey kidney cells. This technology was freely shared with several pharmaceutical companies and institutes worldwide. In this contribution, the history of one of the first cell-culture based large-scale biological production processes is summarized. Also, recent developments and the anticipated upcoming shift from regular IPV to Sabin-IPV are presented. Responding to a call by the World Health Organization (WHO) for new polio vaccines, the development of Sabin-IPV was continued, after demonstrating proof of principle in the 1990s, at the Netherlands Vaccine Institute (NVI). Development of Sabin-IPV plays an important role in the WHO polio eradication strategy as biocontainment will be critical in the post-OPV cessation period. The use of attenuated Sabin strains instead of wild-type Salk polio strains will provide additional safety during vaccine production. Initially, the Sabin-IPV production process will be based on the scale-down model of the current, and well-established, Salk-IPV process. In parallel to clinical trial material production, process development, optimization and formulation research is being carried out to further optimize the process and reduce cost per dose. Also, results will be shown from large-scale (to prepare for future technology transfer) generation of Master- and Working virus seedlots, and clinical trial material (for phase I studies) production. Finally, the planned technology transfer to vaccine manufacturers in low and middle-income countries is discussed.