Design of experiments applications in bioprocessing: Chromatography process development using split design of experiments

Development of a chromatographic step in a time and resource efficient manner remains a serious bottleneck in protein purification. Chromatographic performance typically depends on raw material attributes, feed material attributes, process factors, and their interactions. Design of experiments (DOE) based process development is often chosen for this purpose. A challenge is, however, in performing a DOE with such a large number of process factors. A split DOE approach based on process knowledge in order to reduce the number of experiments is proposed. The first DOE targets optimizing factors that are likely to significantly impact the process and their effect on process performance is unknown. The second DOE aims to fine-tune another set of interacting process factors, impact of whom on process performance is known from process understanding. Furthermore, modeling of a large set of output response variables has been achieved by fitting the output responses to an empirical equation and then using the parametric constants of the equation as output response variables for regression modeling. Two case studies involving hydrophobic interaction chromatography for removal of aggregates and cation exchange chromatography for separation of charge variants and aggregates have been utilized to illustrate the proposed approach. Proposed methodology reduced total number of experiments by 25% and 72% compared to a single DOE based on central composite design and full factorial design, respectively. The proposed approach is likely to result in a significant reduction in resources required as well as time taken during process development. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 35: e2730, 2019     

Optimal designs for a multiresponse Emax model and efficient parameter estimation

The aim of dose finding studies is sometimes to estimate parameters in a fitted model. The precision of the parameter estimates should be as high as possible. This can be obtained by increasing the number of subjects in the study, N, choosing a good and efficient estimation approach, and by designing the dose finding study in an optimal way. Increasing the number of subjects is not always feasible because of increasing cost, time limitations, etc. In this paper, we assume fixed N and consider estimation approaches and study designs for multiresponse dose finding studies. We work with diabetes dose–response data and compare a system estimation approach that fits a multiresponse Emax model to the data to equation‐by‐equation estimation that fits uniresponse Emax models to the data. We then derive some optimal designs for estimating the parameters in the multi‐ and uniresponse Emax model and study the efficiency of these designs.     

Application of quality by design principles to the development and technology transfer of a major process improvement for the manufacture of a recombinant protein

This study describes the application of quality by design (QbD) principles to the development and implementation of a major manufacturing process improvement for a commercially distributed therapeutic protein produced in Chinese hamster ovary cell culture. The intent of this article is to focus on QbD concepts, and provide guidance and understanding on how the various components combine together to deliver a robust process in keeping with the principles of QbD. A fed-batch production culture and a virus inactivation step are described as representative examples of upstream and downstream unit operations that were characterized. A systematic approach incorporating QbD principles was applied to both unit operations, involving risk assessment of potential process failure points, small-scale model qualification, design and execution of experiments, definition of operating parameter ranges and process validation acceptance criteria followed by manufacturing-scale implementation and process validation. Statistical experimental designs were applied to the execution of process characterization studies evaluating the impact of operating parameters on product quality attributes and process performance parameters. Data from process characterization experiments were used to define the proven acceptable range and classification of operating parameters for each unit operation. Analysis of variance and Monte Carlo simulation methods were used to assess the appropriateness of process design spaces. Successful implementation and validation of the process in the manufacturing facility and the subsequent manufacture of hundreds of batches of this therapeutic protein verifies the approaches taken as a suitable model for the development, scale-up and operation of any biopharmaceutical manufacturing process.     

A comparison of techniques for studying oogenesis in the European eel Anguilla anguilla

A multi‐technique approach was used to study the changes occurring in European eel Anguilla anguilla ovaries during hormonally‐induced vitellogenesis. Aside from classic techniques used to monitor the vitellogenic process, such as ovary histology, fat content analysis, sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and vitellogenin enzyme linked immunosorbent assay (ELISA), a new technique, Fourier‐transform infrared (FT‐IR) microspectroscopy, was used to analyse A. anguilla ovaries. The results from the different techniques provided different ways of approaching the same process. Although it is considered a time consuming approach, of all the employed techniques, histology provided the most direct evidences about vitellogenesis. SDS–PAGE and ELISA were also useful for studying vitellogenesis, whereas fat analysis cannot be used for this purpose. The FT‐IR analysis provided a representative IR spectrum for each ovarian stage (previtellogenic stage, early vitellogenic stage, mid‐vitellogenic stage and late vitellogenic stage), demonstrating that it is a valid method able to illustrate the distribution of the oocytes within the ovary slices. The chemical maps obtained confirmed changes in lipid concentrations and revealed their distribution within the oocytes at different maturational stages. When the results and the accuracy of the FT‐IR analysis were compared with those of the traditional techniques commonly used to establish the vitellogenic stage, it became evident that FT‐IR is a useful and reliable tool, with many advantages, including the fact that it requires little biological material, the costs involved are low, analysis times are short and last but not least, the fact that it offers the possibility of simultaneously analysing various biocomponents of the same oocyte.     

Characterization of a Monoclonal Antibody Cell Culture Production Process Using a Quality by Design Approach

The goal of quality by design (QbD) in cell culture manufacturing is to develop manufacturing processes which deliver products with consistent critical quality attributes (CQAs). QbD approaches can lead to better process understanding through the use of process parameter risk ranking and statistical design of experiments (DOE). The QbD process starts with an analysis of process parameter risk with respect to CQAs and key performance indicators (KPIs). Initial DOE study designs and their factor test ranges are based on the outcomes of the process parameter risk ranking exercises. Initial DOE studies screen factors for significant influences on CQAs as well as characterize responses for process KPIs. In the case study provided here, multifactor process characterization studies using a scale-down model resulted in significant variation in charge heterogeneity of a monoclonal antibody (MAb) as measured by ion-exchange chromatography (IEC). Iterative DOE studies, using both screening and response surface designs, were used to narrow the operating parameter ranges so that charge heterogeneity could be controlled to an acceptable level. The data from the DOE studies were used to predict worst-case conditions, which were then verified by testing at those conditions. Using the approach described here, multivariate process parameter ranges were identified that yield acceptable CQA levels and that still provide operational flexibility for manufacturing.     

Characterizing the effect of multiple Fc glycan attributes on the effector functions and FcγRIIIa receptor binding activity of an IgG1 antibody

N‐linked Fc glycosylation of IgG1 monoclonal antibody therapeutics can directly influence their mechanism of action by impacting IgG effector functions such as antibody‐dependent cell‐mediated cytotoxicity (ADCC) and complement‐dependent cytotoxicity (CDC). Therefore, identification and detailed characterization of Fc glycan critical quality attributes (CQAs) provides important information for process design and control. A two‐step approach was used to identify and characterize the Fc glycan CQAs for an IgG1 Mab with effector function. First, single factor experiments were performed to identify glycan critical quality attributes that influence ADCC and CDC activities. Next, a full‐factorial design of experiment (DOE) to characterize the possible interactions and relative effect of these three glycan species on ADCC, CDC, and FcγRIIIa binding was employed. Additionally, the DOE data were used to develop models to predict ADCC, CDC, and FcγRIIIa binding of a given configuration of the three glycan species for this IgG1 molecule. The results demonstrate that for ADCC, afuco mono/bi has the largest effect, followed by HM and β‐gal, while FcγRIIIa binding is affected by afuco mono/bi and β‐gal. CDC, in contrast, is affected by β‐gal only. This type of glycan characterization and modeling can provide valuable information for development, manufacturing support and process improvements for IgG products that require effector function for efficacy. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:1181–1192, 2016     

Rapid optimization of protein freeze‐drying formulations using ultra scale‐down and factorial design of experiment in microplates

Retaining biopharmaceutical proteins in a stable form is critical to their safety and efficacy, and is a major factor for optimizing the final product. Freeze‐dried formulations offer one route for improved stability. Currently the optimization of formulations for freeze‐drying is an empirical process that requires many time‐consuming experiments and also uses large quantities of product material. Here we describe a generic framework for the rapid identification and optimization of formulation excipients to prevent loss of protein activity during a lyophilization process. Using factorial design of experiment (DOE) methods combined with lyophilization in microplates a range of optimum formulations were rapidly identified that stabilized lactose dehydrogenase (derived from Lactobacillus leichmanii) during freeze‐drying. The procedure outlined herein involves two rounds of factorially designed experiments—an initial screen to identify key excipients and potential interactions followed by a central composite face designed optimization experiment. Polyethylene glycol (PEG) and lactose were shown to have significant effects on maintaining protein stability at the screening stage and optimization resulted in an accurate model that was used to plot a window of operation. The variation of freezing temperatures and rates of sublimation that occur across a microplate during freeze‐drying have been characterized also. The optimum formulation was then freeze‐dried in stoppered vials to verify that the microscale data was relevant to the effects observed at larger pilot scales. This work provides a generic approach to biopharmaceutical formulation screening where possible excipients can be screened for single and interactive effects thereby increasing throughput while reducing costs in terms of time and materials. Biotechnol. Bioeng. 2009; 104: 957–964. © 2009 Wiley Periodicals, Inc.     

Modeling the Residence Time Distribution of Integrated Continuous Bioprocesses

In response to the biopharmaceutical industry advancing from traditional batch operation to continuous operation, the Food and Drug Administration (FDA) has published a draft for continuous integrated biomanufacturing. This draft outlines the most important rules for establishing continuous integration. One of these rules is a thorough understanding of mass flows in the process. A computer simulation framework is developed for modeling the residence time distribution (RTD) of integrated continuous downstream processes based on a unit‐by‐unit modeling approach in which unit operations are simulated one‐by‐one across the entire processing time, and then combined into an integrated RTD model. The framework allows for easy addition or replacement of new unit operations, as well as quick adjustment of process parameters during evaluation of the RTD model. With this RTD model, the start‐up phase to reach steady state can be accelerated, the effects of process disturbances at any stage of the process can be calculated, and virtual tracking of a section of the inlet material throughout the process is possible. A hypothetical biomanufacturing process for an antibody was chosen for showcasing the RTD modeling approach.     

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. Biotechnol. Bioeng. 2013; 110: 1354–1365. © 2012 Wiley Periodicals, Inc.     

A Frailty‐Model‐Based Approach to Estimating the Age‐Dependent Penetrance Function of Candidate Genes Using Population‐Based Case‐Control Study Designs: An Application to Data on the BRCA1 Gene

Summary The population‐based case–control study design is perhaps one of, if not the most, commonly used designs for investigating the genetic and environmental contributions to disease risk in epidemiological studies. Ages at onset and disease status of family members are routinely and systematically collected from the participants in this design. Considering age at onset in relatives as an outcome, this article is focused on using the family history information to obtain the hazard function, i.e., age‐dependent penetrance function, of candidate genes from case–control studies. A frailty‐model‐based approach is proposed to accommodate the shared risk among family members that is not accounted for by observed risk factors. This approach is further extended to accommodate missing genotypes in family members and a two‐phase case–control sampling design. Simulation results show that the proposed method performs well in realistic settings. Finally, a population‐based two‐phase case–control breast cancer study of the BRCA1 gene is used to illustrate the method.     

Defining process design space for biotech products: case study of Pichia pastoris fermentation

The concept of "design space" has been proposed in the ICH Q8 guideline and is gaining momentum in its application in the biotech industry. It has been defined as "the multidimensional combination and interaction of input variables (e.g., material attributes) and process parameters that have been demonstrated to provide assurance of quality." This paper presents a stepwise approach for defining process design space for a biologic product. A case study, involving P. pastoris fermentation, is presented to facilitate this. First, risk analysis via Failure Modes and Effects Analysis (FMEA) is performed to identify parameters for process characterization. Second, small-scale models are created and qualified prior to their use in these experimental studies. Third, studies are designed using Design of Experiments (DOE) in order for the data to be amenable for use in defining the process design space. Fourth, the studies are executed and the results analyzed for decisions on the criticality of the parameters as well as on establishing process design space. For the application under consideration, it is shown that the fermentation unit operation is very robust with a wide design space and no critical operating parameters. The approach presented here is not specific to the illustrated case study. It can be extended to other biotech unit operations and processes that can be scaled down and characterized at small scale.     

Design and Analysis of Crossover Trials for Absorbing Binary Endpoints

Summary The crossover is a popular and efficient trial design used in the context of patient heterogeneity to assess the effect of treatments that act relatively quickly and whose benefit disappears with discontinuation. Each patient can serve as her own control as within‐individual treatment and placebo responses are compared. Conventional wisdom is that these designs are not appropriate for absorbing binary endpoints, such as death or HIV infection. We explore the use of crossover designs in the context of these absorbing binary endpoints and show that they can be more efficient than the standard parallel group design when there is heterogeneity in individuals' risks. We also introduce a new two‐period design where first period “survivors” are rerandomized for the second period. This design combines the crossover design with the parallel design and achieves some of the efficiency advantages of the crossover design while ensuring that the second period groups are comparable by randomization. We discuss the validity of the new designs and evaluate both a mixture model and a modified Mantel–Haenszel test for inference. The mixture model assumes no carryover or period effects while the Mantel–Haenszel approach conditions out period effects. Simulations are used to compare the different designs and an example is provided to explore practical issues in implementation.     

Statistical vs. Stochastic experimental design: An experimental comparison on the example of protein refolding

Optimization of experimental problems is a challenging task in both engineering and science. In principle, two different design of experiments (DOE) strategies exist: statistical and stochastic methods. Both aim to efficiently and precisely identify optimal solutions inside the problem‐specific search space. Here, we evaluate and compare both strategies on the same experimental problem, the optimization of the refolding conditions of the lipase from Thermomyces lanuginosus with 26 variables under study. Protein refolding is one of the main bottlenecks in the process development for recombinant proteins. Despite intensive effort, the prediction of refolding from sequence information alone is still not applicable today. Instead, suitable refolding conditions are typically derived empirically in large screening experiments. Thus, protein refolding should constitute a good performance test for DOE strategies. We compared an iterative stochastic optimization applying a genetic algorithm and a standard statistical design consisting of a D‐optimal screening step followed by an optimization via response surface methodology. Our results revealed that only the stochastic optimization was able to identify optimal refolding conditions (～1.400 U g^?1 refolded activity), which were 3.4‐fold higher than the standard. Additionally, the stochastic optimization proved quite robust, as three independent optimizations performed similar. In contrast, the statistical DOE resulted in a suboptimal solution and failed to identify comparable activities. Interactions between process variables proved to be pivotal for this optimization. Hence, the linear screening model was not able to identify the most important process variables correctly. Thereby, this study highlighted the limits of the classic two‐step statistical DOE. © 2012 American Institute of Chemical Engineers Biotechnol. Prog., 2012     

Large‐scale identification of genetic design strategies using local search

In the past decade, computational methods have been shown to be well suited to unraveling the complex web of metabolic reactions in biological systems. Methods based on flux–balance analysis (FBA) and bi‐level optimization have been used to great effect in aiding metabolic engineering. These methods predict the result of genetic manipulations and allow for the best set of manipulations to be found computationally. Bi‐level FBA is, however, limited in applicability because the required computational time and resources scale poorly as the size of the metabolic system and the number of genetic manipulations increase. To overcome these limitations, we have developed Genetic Design through Local Search (GDLS), a scalable, heuristic, algorithmic method that employs an approach based on local search with multiple search paths, which results in effective, low‐complexity search of the space of genetic manipulations. Thus, GDLS is able to find genetic designs with greater in silico production of desired metabolites than can feasibly be found using a globally optimal search and performs favorably in comparison with heuristic searches based on evolutionary algorithms and simulated annealing.     

Combining capture–recapture data and pedigree information to assess heritability of demographic parameters in the wild

Quantitative genetic analyses have been increasingly used to estimate the genetic basis of life‐history traits in natural populations. Imperfect detection of individuals is inherent to studies that monitor populations in the wild, yet it is seldom accounted for by quantitative genetic studies, perhaps leading to flawed inference. To facilitate the inclusion of imperfect detection of individuals in such studies, we develop a method to estimate additive genetic variance and assess heritability for binary traits such as survival, using capture–recapture (CR) data. Our approach combines mixed‐effects CR models with a threshold model to incorporate discrete data in a standard ‘animal model’ approach. We employ Markov chain Monte Carlo sampling in a Bayesian framework to estimate model parameters. We illustrate our approach using data from a wild population of blue tits (Cyanistes caeruleus) and present the first estimate of heritability of adult survival in the wild. In agreement with the prediction that selection should deplete additive genetic variance in fitness, we found that survival had low heritability. Because the detection process is incorporated, capture–recapture animal models (CRAM) provide unbiased quantitative genetics analyses of longitudinal data collected in the wild.     

Optimal culture conditions for keratinase production by a novel thermophilic Myceliophthora thermophila strain GZUIFR‐H49‐1

Aims: To investigate the effect of medium compositions and culture conditions on keratinase production by a novel thermophilic fungus Myceliophthora thermophila (Apinis) Oorschot strain GZUIFR‐H49‐1. Methods and Results: The thermophilic strain GZUIFR‐H49‐1 with keratinolytic ability was characterized and identified as a strain of M. thermophila on the basis of its morphological characters and molecular analysis of ITS1‐5.8S‐ITS2 rDNA sequence. Among the medium compositions tested, the soluble starch (SS), urea, sodium thiosulfate and CaCl₂ were the most effective C‐source, N‐source, S‐source and mineral ion, respectively, by employing the single‐factor experiment. The urea and pH value were the significant factors (P < 0·05) for the keratinase production in this experiment condition using Plackett–Burman factorial design. The conditions of keratinase production were further optimized by Box–Behnken design. Consequently, there was a 6·4‐fold increase (5100 U l^?1) in the keratinase activity than the initial value (800 U l^?1) by this optimal process. Conclusions: This study indicated that the optimization design proved a useful and powerful tool for the development of optimal medium compositions and culture conditions. Myceliophthora thermophila strain GZUIFR‐H49‐1 was a promising fungus strain for keratinase production. Significance and Impact of the Study: This study characterized a novel thermophilic M. thermophila strain GZUIFR‐H49‐1 with potential applications for keratinase production. These conditions of keratinase production obtained by means of optimization design will be accumulated as potential information for exploration and utilization to the new fungus isolate.     

Multivariate statistical monitoring as applied to clean‐in‐place (CIP) and steam‐in‐place (SIP) operations in biopharmaceutical manufacturing

Multivariate statistical process monitoring (MSPM) is becoming increasingly utilized to further enhance process monitoring in the biopharmaceutical industry. MSPM can play a critical role when there are many measurements and these measurements are highly correlated, as is typical for many biopharmaceutical operations. Specifically, for processes such as cleaning‐in‐place (CIP) and steaming‐in‐place (SIP, also known as sterilization‐in‐place), control systems typically oversee the execution of the cycles, and verification of the outcome is based on offline assays. These offline assays add to delays and corrective actions may require additional setup times. Moreover, this conventional approach does not take interactive effects of process variables into account and cycle optimization opportunities as well as salient trends in the process may be missed. Therefore, more proactive and holistic online continued verification approaches are desirable. This article demonstrates the application of real‐time MSPM to processes such as CIP and SIP with industrial examples. The proposed approach has significant potential for facilitating enhanced continuous verification, improved process understanding, abnormal situation detection, and predictive monitoring, as applied to CIP and SIP operations. © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 30:505–515, 2014     

Order‐restricted Scores Test for the Evaluation of Population‐based Case–control Studies when the Genetic Model is Unknown

The Cochran–Armitage (CA) linear trend test for proportions is often used for genotype‐based analysis of candidate gene association. Depending on the underlying genetic mode of inheritance, the use of model‐specific scores maximises the power. Commonly, the underlying genetic model, i.e. additive, dominant or recessive mode of inheritance, is a priori unknown. Association studies are commonly analysed using permutation tests, where both inference and identification of the underlying mode of inheritance are important. Especially interesting are tests for case–control studies, defined by a maximum over a series of standardised CA tests, because such a procedure has power under all three genetic models. We reformulate the test problem and propose a conditional maximum test of scores‐specific linear‐by‐linear association tests. For maximum‐type, sum and quadratic test statistics the asymptotic expectation and covariance can be derived in a closed form and the limiting distribution is known. Both the limiting distribution and approximations of the exact conditional distribution can easily be computed using standard software packages. In addition to these technical advances, we extend the area of application to stratified designs, studies involving more than two groups and the simultaneous analysis of multiple loci by means of multiplicity‐adjusted p‐values for the underlying multiple CA trend tests. The new test is applied to reanalyse a study investigating genetic components of different subtypes of psoriasis. A new and flexible inference tool for association studies is available both theoretically as well as practically since already available software packages can be easily used to implement the suggested test procedures.     

Strategic deployment of CHO expression platforms to deliver Pfizer's Monoclonal Antibody Portfolio

Development of stable cell lines for expression of large‐molecule therapeutics represents a significant portion of the time and effort required to advance a molecule to enabling regulatory toxicology studies and clinical evaluation. Our development strategy employs two different approaches for cell line development based on the needs of a particular project: a random integration approach for projects where high‐level expression is critical, and a site‐specific integration approach for projects in which speed and reduced employee time spend is a necessity. Here we describe both our random integration and site‐specific integration platforms and their applications in support of monoclonal antibody development and production. We also compare product quality attributes of monoclonal antibodies produced with a nonclonal cell pool or clonal cell lines derived from the two platforms. Our data suggests that material source (pools vs. clones) does not significantly alter the examined product quality attributes. Our current practice is to leverage this observation with our site‐specific integration platform, where material generated from cell pools is used for an early molecular assessment of a given candidate to make informed decisions around development strategy. © 2017 American Institute of Chemical Engineers Biotechnol. Prog., 33:1463–1467, 2017     

Interface Engineering of MXene Composite Separator for High‐Performance Li–Se and Na–Se Batteries

Selenium (Se), due to its high electronic conductivity and high energy density, has recently attracted considerable interest as a cathode material for rechargeable Li/Na batteries. However, the poor cycling stability originating from the severe shuttle effect of polyselenides hinders their practical applications. Herein, highly stable Li/Na–Se batteries are developed using ultrathin (≈270 nm, loading of 0.09 mg cm^?2) cetrimonium bromide (CTAB)/carbon nanotube (CNT)/Ti₃C₂T_x MXene hybrid modified polypropylene (PP) (CCNT/MXene/PP) separators. The hybrid separator can immobilize the polyselenides via enhanced Lewis acid–base interactions between CTAB/MXene and polyselenides, which is demonstrated by theoretical calculations and X‐ray photoelectron spectroscopy. The incorporation of CNT helps to improve the electrolyte infiltration and facilitate the ionic transport. In situ permeation experiments are conducted for the first time to visually study the behavior of polyselenides, revealing the prohibited shuttle effect and protected Li anode from corrosion with CCNT/MXene/PP separators. As a result, the Li–Se batteries with CCNT/MXene/PP separators deliver an outstanding cycling performance over 500 cycles at 1C with an extremely low capacity decay of 0.05% per cycle. Moreover, the hybrid separators also perform well in Na–Se batteries. This study develops a preferable separator–electrolyte interface and the concept can be applied in other conversion‐type battery systems.