Current state of implementation of in silico tools in the biopharmaceutical industry-Proceedings of the 5th modeling workshop.

The fifth modeling workshop (5MW) was held in June 2023 at Favrholm, Denmark and sponsored by Recovery of Biological Products Conference Series. The goal of the workshop was to assemble modeling practitioners to review and discuss the current state, progress since the last fourth mini modeling workshop (4MMW), gaps and opportunities for development, deployment and maintenance of models in bioprocess applications. Areas of focus were four categories: biophysics and molecular modeling, mechanistic modeling, computational fluid dynamics (CFD) and plant modeling. Highlights of the workshop included significant advancements in biophysical/molecular modeling to novel protein constructs, mechanistic models for filtration and initial forays into modeling of multiphase systems using CFD for a bioreactor and mapped strategically to cell line selection/facility fit. A significant impediment to more fully quantitative and calibrated models for biophysics is the lack of large, anonymized datasets. A potential solution would be the use of specific descriptors in a database that would allow for detailed analyzes without sharing proprietary information. Another gap identified was the lack of a consistent framework for use of models that are included or support a regulatory filing beyond the high-level guidance in ICH Q8-Q11. One perspective is that modeling can be viewed as a component or precursor of machine learning (ML) and artificial intelligence (AI). Another outcome was alignment on a key definition for "mechanistic modeling." Feedback from participants was that there was progression in all of the fields of modeling within scope of the conference. Some areas (e.g., biophysics and molecular modeling) have opportunities for significant research investment to realize full impact. However, the need for ongoing research and development for all model types does not preclude the application to support process development, manufacturing and use in regulatory filings. Analogous to ML and AI, given the current state of the four modeling types, a prospective investment in educating inter-disciplinary subject matter experts (e.g., data science, chromatography) is essential to advancing the modeling community.

Directed evolution of conformation-specific antibodies for sensitive detection of polypeptide aggregates in therapeutic drug formulations.

Biologics such as peptides and proteins possess a number of attractive attributes that make them particularly valuable as therapeutics, including their high activity, high specificity, and low toxicity. However, one of the key challenges associated with this class of drugs is their propensity to aggregate. Given the safety and immunogenicity concerns related to polypeptide aggregates, it is particularly important to sensitively detect aggregates in therapeutic drug formulations as part of the quality control process. Here, we report the development of conformation-specific antibodies that recognize polypeptide aggregates composed of a GLP-1 receptor agonist (liraglutide) and their integration into a sensitive immunoassay for detecting liraglutide amyloid fibrils. We sorted single-chain antibody libraries against liraglutide fibrils using yeast surface display and magnetic-activated cell sorting, and identified several antibodies with high conformational specificity. Interestingly, these antibodies cross-react with amyloid fibrils formed by several other polypeptides, revealing that they recognize molecular features common to different types of fibrils. Moreover, we find that our immunoassay using these antibodies is >50-fold more sensitive than the conventional method for detecting liraglutide aggregation (Thioflavin T fluorescence). We expect that our systematic approach for generating a sensitive, aggregate-specific immunoassay can be readily extended to other biologics to improve the quality and safety of formulated drug products.

Toward in silico CMC: An industrial collaborative approach to model-based process development.

The Third Modeling Workshop focusing on bioprocess modeling was held in Kenilworth, NJ in May 2019. A summary of these Workshop proceedings is captured in this manuscript. Modeling is an active area of research within the biotechnology community, and there is a critical need to assess the current state and opportunities for continued investment to realize the full potential of models, including resource and time savings. Beyond individual presentations and topics of novel interest, a substantial portion of the Workshop was devoted toward group discussions of current states and future directions in modeling fields. All scales of modeling, from biophysical models at the molecular level and up through large scale facility and plant modeling, were considered in these discussions and are summarized in the manuscript. Model life cycle management from model development to implementation and sustainment are also considered for different stages of clinical development and commercial production. The manuscript provides a comprehensive overview of bioprocess modeling while suggesting an ideal future state with standardized approaches aligned across the industry.

Influence of Production Process and Scale on Quality of Polypeptide Drugs: a Case Study on GLP-1 Analogs.

PURPOSE: Manufacturing processes for polypeptide/protein drugs are designed to ensure robust quality, efficacy and safety. Process differences introduced by follow-on manufacturers may result in changes in quality and clinical outcomes. This study investigated the impact of production methods on the stability and impurities of liraglutide and semaglutide drug substances/products, and the potential impact on drug quality, efficacy and safety. METHODS: State-of-the-art analytical methods were used to compare physical and chemical stability, and impurity profiles of drug substances/products from different suppliers. Identified polypeptide-related impurities were evaluated for immunogenicity potential by in silico T cell epitope prediction. Semaglutide immunogenicity in clinical trials (SUSTAIN) was evaluated using a tiered antibody analysis. RESULTS: Manufacturing scale and process strongly impacted the physical stability of the products. Trace metals increased high-molecular-weight protein formation for liraglutide and semaglutide. Synthetic and recombinant liraglutide produced by five suppliers had distinct impurity profiles compared with the originator. In silico evaluation suggested that new impurities could be immunogenic. Immunogenicity of semaglutide in clinical trials was lower than for liraglutide. CONCLUSIONS: Differences in manufacturing processes affect chemical/physical stability and impurity profile, and may impact immunogenicity. Follow-on versions of liraglutide and semaglutide, and possibly other polypeptides, should be clinically evaluated for efficacy and safety.

Sensitive detection of glucagon aggregation using amyloid fibril-specific antibodies.

Sensitive detection of protein aggregates is important for evaluating the quality of biopharmaceuticals and detecting misfolded proteins in several neurodegenerative diseases. However, it is challenging to detect extremely low concentrations (<10 ppm) of aggregated protein in the presence of high concentrations of soluble protein. Glucagon, a peptide hormone used in the treatment of extreme hypoglycemia, is aggregation-prone and forms amyloid fibrils. Detection of glucagon fibrils using conformation-specific antibodies is an attractive approach for identifying such aggregates during process and formulation development. Therefore, we have used yeast surface display and magnetic-activated cell sorting to sort single-chain antibody libraries to identify antibody variants with high conformational specificity for glucagon fibrils. Notably, we find several high-affinity antibodies that display excellent selectivity for glucagon fibrils, and we have integrated these antibodies into a sensitive immunoassay. Surprisingly, the sensitivity of our assay-which involves direct (nonantibody mediated) glucagon immobilization in microtiter plates-can be significantly enhanced by pretreating the microtiter plates with various types of globular proteins before glucagon immobilization. Moreover, increased total concentrations of glucagon peptide also significantly improve the sensitivity of our assay, which appears to be due to the strong seeding activity of immobilized fibrils at high glucagon concentrations. Our final assay is highly sensitive (fibril detection limit of ~0.5-1 ppm) and is >20 times more sensitive than detection using a conventional, amyloid-specific fluorescent dye (Thioflavin T). We expect that this type of sensitive immunoassay can be readily integrated into the drug development process to improve the generation of safe and potent peptide therapeutics.

High-throughput downstream process development for cell-based products using aqueous two-phase systems (ATPS) - A case study.

The availability of preparative-scale downstream processing strategies for cell-based products presents a critical juncture between fundamental research and clinical development. Aqueous two-phase systems (ATPS) present a gentle, scalable, label-free, and cost-effective method for cell purification, and are thus a promising tool for downstream processing of cell-based therapeutics. Here, the application of a previously developed robotic screening platform that enables high-throughput cell partitioning analysis in ATPS is reported. In the present case study a purification strategy for two model cell lines based on high-throughput screening (HTS)-data and countercurrent distribution (CCD)-modeling, and validated the CCD-model experimentally is designed. The obtained data are shown an excellent congruence between CCD-model and experimental data, indicating that CCD-models in combination with HTS-data are a powerful tool in downstream process development. Finally, the authors are shown that while cell cycle phase significantly influences cell partitioning, cell type specific differences in surface properties are the main driving force in charge-dependent separation of HL-60 and L929 cells. In order to design a highly robust purification process it is, however, advisable to maintain constant growth conditions.

High-throughput downstream process development for cell-based products using aqueous two-phase systems.

As the clinical development of cell-based therapeutics has evolved immensely within the past years, downstream processing strategies become more relevant than ever. Aqueous two-phase systems (ATPS) enable the label-free, scalable, and cost-effective separation of cells, making them a promising tool for downstream processing of cell-based therapeutics. Here, we report the development of an automated robotic screening that enables high-throughput cell partitioning analysis in ATPS. We demonstrate that this setup enables fast and systematic investigation of factors influencing cell partitioning. Moreover, we examined and optimized separation conditions for the differentiable promyelocytic cell line HL-60 and used a counter-current distribution-model to investigate optimal separation conditions for a multi-stage purification process. Finally, we show that the separation of CD11b-positive and CD11b-negative HL-60 cells is possible after partial DMSO-mediated differentiation towards the granulocytic lineage. The modeling data indicate that complete peak separation is possible with 30 transfers, and >93% of CD11b-positive HL-60 cells can be recovered with >99% purity. The here described screening platform facilitates faster, cheaper, and more directed downstream process development for cell-based therapeutics and presents a powerful tool for translational research.

How can measurement, monitoring, modeling and control advance cell culture in industrial biotechnology?

This report highlights the potential of measurement, monitoring, modeling and control (M(3) C) methodologies in animal and human cell culture technology. In particular, state-of-the-art of M(3) C technologies and their industrial relevance of existing technology are addressed. It is a summary of an expert panel discussion between biotechnologists and biochemical engineers with both academic and industrial backgrounds. The latest ascents in M(3) C are discussed from a cell culture perspective for industrial process development and production needs. The report concludes with a set of recommendations for targeting M(3) C research toward the industrial interests. These include issues of importance for biotherapeutics production, miniaturization of measurement techniques and modeling methods.

Novel peptide ligand with high binding capacity for antibody purification.

Small synthetic ligands for protein purification have become increasingly interesting with the growing need for cheap chromatographic materials for protein purification and especially for the purification of monoclonal antibodies (mAbs). Today, Protein A-based chromatographic resins are the most commonly used capture step in mAb down stream processing; however, the use of Protein A chromatography is less attractive due to toxic ligand leakage as well as high cost. Whether used as an alternative to the Protein A chromatographic media or as a subsequent polishing step, small synthetic peptide ligands have an advantage over biological ligands; they are cheaper to produce, ligand leakage by enzymatic degradation is either eliminated or significantly reduced, and they can in general better withstand cleaning in place (CIP) conditions such as 0.1M NaOH. Here, we present a novel synthetic peptide ligand for purification of human IgG. Immobilized on WorkBeads, an agarose-based base matrix from Bio-Works, the ligand has a dynamic binding capacity of up to 48 mg/mL and purifies IgG from harvest cell culture fluid with purities and recovery of >93%. The binding affinity is ∼105 M⁻¹ and the interaction is favorable and entropy-driven with an enthalpy penalty. Our results show that the binding of the Fc fragment of IgG is mediated by hydrophobic interactions and that elution at low pH is most likely due to electrostatic repulsion. Furthermore, we have separated aggregated IgG from non-aggregated IgG, indicating that the ligand could be used both as a primary purification step of IgG as well as a subsequent polishing step.

Exploring variation in binding of Protein A and Protein G to immunoglobulin type G by isothermal titration calorimetry.

Bacterial Protein A (PrtA) and Protein G (PrtG) are widely used for affinity purification of antibodies. An understanding of how PrtA and PrtG bind to different isotypes of immunoglobulin type G (IgG) and to their corresponding Fc fragments is essential for the development of PrtA and PrtG mimetic ligands and for the establishment of generic processes for the purification of various antibodies. In this paper, the interactions between the two IgG-binding proteins and IgG of two different subclasses, IgG1 and IgG4, as well as their analogous Fc fragments have been studied by isothermal titration calorimetry. The results indicate that both protein ligands bind IgG and Fc fragments strongly with Ka values in the range of 10(7) -10(8) M(-1) and for both ligands, the interaction with both IgG isotypes is enthalpically driven though entropically unfavorable. Moreover, variation in the standard entropic and standard enthalpic contribution to binding between the two isotypes as well as between IgG and Fc fragment implies that the specific interaction with PrtA varies according to IgG isotype. In contrast to PrtA, PrtG bound to F(ab')(2) fragment with a Ka value of 5.1 × 10(5) M(-1) ; thus underscoring the usefulness of PrtA as a preferred ligand for generic antibody purification processes.

A label-free methodology for selective protein quantification by means of absorption measurements.

The application of high throughput experimentation (HTE) in protein purification process development has created an analytical bottleneck. Using a new label-free and non-invasive methodology for analyzing multicomponent protein mixtures by means of spectral measurements, we show that the analytical throughput for selective protein quantification can be increased significantly. An analytical assay based on this new methodology was shown to generate very precise results. Further, the assay was successfully applied as analytics for a resin screening performed in HTE mode. The increase in analytical throughput was obtained without decreasing the level of information when compared to analytical chromatography. This proves its potential as a valuable analytical tool in conjugation with high throughput process development (HTPD). Further, fast selective protein quantification can enhance process control in a commercial production environment and, hence, minimize the need for off-line release analysis.

Effects of urea induced protein conformational changes on ion exchange chromatographic behavior.

Urea is widely employed to facilitate protein separations in ion exchange chromatography at various scales. In this work, five model proteins were used to examine the chromatographic effects of protein conformational changes induced by urea in ion exchange chromatography. Linear gradient experiments were carried out at various urea concentrations and the protein secondary and tertiary structures were evaluated by far UV CD and fluorescence measurements, respectively. The results indicated that chromatographic retention times were well correlated with structural changes and that they were more sensitive to tertiary structural change. Steric Mass Action (SMA) isotherm parameters were also examined and the results indicated that urea induced protein conformational changes could affect both the characteristic charge and equilibrium constants in these systems. Dynamic light scattering analysis of changes in protein size due to urea-induced unfolding indicated that the size of the protein was not correlated with SMA parameter changes. These results indicate that while urea-induced structural changes can have a marked effect on protein chromatographic behavior in IEX, this behavior can be quite complicated and protein specific. These differences in protein behavior may provide insight into how these partially unfolded proteins are interacting with the resin material.

Quality by design--thermodynamic modelling of chromatographic separation of proteins.

A desired goal of the PAT framework is to design and develop well-understood processes that will consistently ensure a predefined quality at the end of the manufacturing process. Such procedures would be consistent with the basic tenet of quality by design and could reduce risks to quality and regulatory concerns while improving efficiency. To support a more in-depth understanding of the design and development of a chromatographic purification process the paper discusses the general thermodynamic principles of ligand-binding and models of multi-component adsorption in ion-exchange and hydrophobic chromatography. The parameters in the models are easy to determine and have a well-defined physical significance. Examples demonstrate how the model parameters can be determined from experimental data and in order to validate the model, simulated chromatograms are compared to the corresponding experimental chromatograms. Finally it is demonstrated how a simulation can be used to explain an aberration.

Comparison of chromatographic ion-exchange resins VI. Weak anion-exchange resins.

A comparative study on weak anion exchangers was performed to investigate the pH dependence, binding strength, particle size distribution, and static and dynamic capacity of the chromatographic resins. The resins tested included: DEAE Sepharose FF, Poros 50 D, Fractogel EMD DEAE (M), MacroPrep DEAE Support, DEAE Ceramic HyperD 20, and Toyopearl DEAE 650 M. Testing was performed with five different model proteins: Anti-FVII mAb (immunoglobulin G), aprotinin, bovine serum albumin (BSA), Lipolase (Novozymes), and myoglobin. Retention showed an expected increasing trend as a function of pH for proteins with low pI. A decrease in retention was observed for some resins at pH 9 likely due to initiation of deprotonation of the weak anion-exchange ligands. Expected particle size distribution was obtained for all resins compared to previous studies. Binding strength to weak anion-exchange resins as a function of ionic strength depends on the specific protein. Binding and elution at low salt concentration may be performed with Toyopearl DEAE 650 M, while binding and elution at high salt concentration may be performed with MacroPrep DEAE Support. Highest binding capacities were generally obtained with Poros 50 D followed by DEAE Ceramic HyperD 20. A general good agreement was obtained between this study and data obtained by the suppliers. Verification of binding strength trends with model proteins was achieved with human growth hormone (hGH) and a hGH variant on the same resins with different elution salts, sodium chloride, sodium hydrogenphosphate, sodium sulphate, and sodium acetate. Static capacity measurements obtained in the traditional experimental set-up were compared with high-throughput screening (HTS) technique experiments with reasonable agreement. Isotherm data obtained from HTS techniques and pulse experiments were successfully combined with mathematical modelling to simulate, develop and optimise the separation process of two model proteins, Lipolase and BSA. The data presented in this paper may be used for selection of resins for testing in process development.

Optimized removal of soluble host cell proteins for the recovery of met-human growth hormone inclusion bodies from Escherichia coli cell lysate using crossflow microfiltration.

Cross-flow membrane microfiltration was used under optimal conditions to recover met-growth hormone inclusion bodies (IBs) from Escherichia coli cell lysate by removal of the host-cell (bacterial) proteins (HCP) under minimal fouling conditions. This is the first step of a two-step process in which the goal was to isolate IBs at high yield from the HCP. These undesired soluble HCP were removed by passing them through the membrane while retaining the insolubles, including the aggregated IBs. Experiments were conducted at constant permeate flux with flat-sheet membranes of different pore sizes and chemistry, with feeds of varying pH and ionic strengths to determine the optimum combination for HCP removal. Diafiltration, the washing away of impurities with protein-free buffer, was then employed to ensure removal of the host cell proteins at the optimum conditions. About 90% removal of the HCP was obtained in about 5 diavolumes, maintaining high protein transmission and low membrane fouling.

Comparison of chromatographic ion-exchange resins V. Strong and weak cation-exchange resins.

Strong and weak cation-exchangers were compared for a number of chromatographic parameters, i.e. pH dependence, efficiency, binding strength, particle size distribution, static and dynamic capacity, and scanning electron microscopy (SEM) pictures. Chromatographic resins investigated were Fractogel EMD SO3- (M), Fractogel EMD SE Hicap (M), Fractogel EMD COO- (M), MacroPrep 25S, MacroPrep High S, MacroPrep CM, CM HyperZ, and Matrex Cellufine C-500. Testing was done with three proteins: Anti-FVII Mab (IgG), aprotinin, and lysozyme. For lysozyme and aprotinin with pI above experimental pH, dependence of pH on retention was generally low, though some pronounced decrease of retention with increasing pH was observed for CM HyperZ. For Anti-FVII Mab with pI<7.5, binding was observed on several resins at pH 7.5. Efficiency results present the expected trend of increasing dependence of plate height as a function of increasing flow rate, and the highest flow dependence was observed for Fractogel EMD COO-. Particle size distribution was determined by two independent methods, coulter counting and SEM pictures, with fair agreement. Binding strength data of cation-exchange resins as a function of ionic strength depends on the protein, but binding and elution at high salt concentration may in general be performed with MacroPrep resins. Comparison of dynamic capacity data at 10% break-through and static capacity measurements shows that a very diverse utilization of approximately 25-90% of the total available capacity is employed during chromatographic operation. The effect of competitive binding from yeast fermentation components on dynamic binding capacity of aprotinin was studied showing a significant decrease in binding capacity. Sepharose FF, Toyopearl 650 M, and Ceramic HyperD F strong and weak cation-exchange resins were included in this study. Resins with good pure aprotinin capacity also performed well for aprotinin in fermentation broth, but the highest relative capacity was obtained with MacroPrep High S having a fairly low pure component dynamic capacity. Results of this paper may be used in the selection of resins for further testing in biopharmaceutical protein purification process development.

Comparison of chromatographic ion-exchange resins IV. Strong and weak cation-exchange resins and heparin resins.

A comparative study was performed on heparin resins and strong and weak cation exchangers to investigate the pH dependence, efficiency, binding strength, particle size distribution, static and dynamic capacity, and scanning electron microscopy pictures of chromatographic resins. The resins tested include: Heparin Sepharose FF, SP Sepharose FF, CM Sepharose FF, Heparin Toyopearl 650 m, SP Toyopearl 650 m, CM Toyopearl 650 m, Ceramic Heparin HyperD M, Ceramic S HyperD 20, and Ceramic CM HyperD F. Testing was performed with four different proteins: anti-FVII Mab (IgG), aprotinin, lysozyme, and myoglobin. Dependence of pH on retention was generally very low for proteins with high isoelectric point (pI), though some decrease of retention with increasing pH was observed for CM Ceramic HyperD F and S Ceramic HyperD 20. Binding of anti-FVII Mab with pI < 7.5 was observed on several resins at pH 7.5. Efficiency results show the expected trend of increasing dependence of the plate height with increasing flow rate of Ceramic HyperD resins followed by Toyopearl 650 m resins and the highest flow dependence of the Sepharose FF resins corresponding to their pressure resistance. Determination of particle size distribution by two independent methods, coulter counting and SEM, was in good agreement. Binding strength of cation-exchange resins as a function of ionic strength varies depending on the protein. Binding and elution at high salt concentration may be performed with Ceramic HyperD resins, while binding and elution at low salt concentration may be performed with model proteins on heparin resins. Employing proteins with specific affinity for heparin, a much stronger binding is observed, however, some cation exchangers may still be good substitutions for heparin resins. Dynamic capacity at 10% breakthrough compared to static capacity measurements and dynamic capacity displays that approximately 40-80% of the total available capacity is utilized during chromatographic operation depending on flow rate. A general good agreement was obtained between results of this study and data obtained by others. Results of this study may be used in the selection of resins for testing during protein purification process development.