Model-assisted approaches for continuous chromatography: Current situation and challenges

Using the reactive/transport dispersive models to simulate a monolithic anion exchanger: Experimental parameter determination, simultaneous model evaluation, and validation

Selecting an adequate model to represent the mass transfer mechanisms occurring in a chromatographic process is generally complicated, which is one of the reasons why monolithic chromatography is scarcely simulated. In this study, the chromatographic separation of model proteins bovine serum albumin (BSA), β-lactoglobulin-A, and β-lactoglobulin-B on an anion exchange monolith was simulated based on experimental parameter determination, simultaneous model testing, and validation under three statistical criteria: retention time, dispersion accuracies, and Pearson correlation coefficient. Experimental characterization of morphologic, physicochemical, and kinetic parameters was performed through volume balances, pressure drop analysis, breakthrough curve analysis, and batch adsorptions. Free Gibbs energy indicated a spontaneous adsorption process for proteins and counterions. Dimensionless numbers were estimated based on height equivalent to a theoretical plate analysis, finding that pore diffusion controlled β-lactoglobulin separation, whereas adsorption/desorption kinetics was the dominant mechanism for BSA. The elution profiles were modeled using the transport dispersive model and the reactive dispersive model coupled with steric mass action (SMA) isotherms because these models allowed to consider most of the mass transport mechanisms that have been described. RDM-SMA presented the most accurate simulations at pH 6.0 and at low (250 mM) and high (400 mM) NaCl concentrations. This simulation will be used as reference to forecast the purification of these proteins from bovine whey waste and to extrapolate this methodology to other monolith-based separations using these three statistical criteria that have not been used previously for this purpose.

Using artificial neural networks to accelerate flowsheet optimization for downstream process development

An optimal purification process for biopharmaceutical products is important to meet strict safety regulations, and for economic benefits. To find the global optimum, it is desirable to screen the overall design space. Advanced model-based approaches enable to screen a broad range of the design-space, in contrast to traditional statistical or heuristic-based approaches. Though, chromatographic mechanistic modeling (MM), one of the advanced model-based approaches, can be speed-limiting for flowsheet optimization, which evaluates every purification possibility (e.g., type and order of purification techniques, and their operating conditions). Therefore, we propose to use artificial neural networks (ANNs) during global optimization to select the most optimal flowsheets. So, the number of flowsheets for final local optimization is reduced and consequently the overall optimization time. Employing ANNs during global optimization proved to reduce the number of flowsheets from 15 to only 3. From these three, one flowsheet was optimized locally and similar final results were found when using the global outcome of either the ANN or MM as starting condition. Moreover, the overall flowsheet optimization time was reduced by 50% when using ANNs during global optimization. This approach accelerates the early purification process design; moreover, it is generic, flexible, and regardless of sample material's type.

Residence time distribution in continuous virus filtration

Regulatory authorities recommend using residence time distribution (RTD) to address material traceability in continuous manufacturing. Continuous virus filtration is an essential but poorly understood step in biologics manufacturing in respect to fluid dynamics and scale-up. Here we describe a model that considers nonideal mixing and film resistance for RTD prediction in continuous virus filtration, and its experimental validation using the inert tracer NaNO3. The model was successfully calibrated through pulse injection experiments, yielding good agreement between model prediction and experiment (R 2 > ${R}^{2}\gt $  0.90). The model enabled the prediction of RTD with variations-for example, in injection volumes, flow rates, tracer concentrations, and filter surface areas-and was validated using stepwise experiments and combined stepwise and pulse injection experiments. All validation experiments achievedR 2 > ${R}^{2}\gt $  0.97. Notably, if the process includes a porous material-such as a porous chromatography material, ultrafilter, or virus filter-it must be considered whether the molecule size affects the RTD, as tracers with different sizes may penetrate the pore space differently. Calibration of the model with NaNO3 enabled extrapolation to RTD of recombinant antibodies, which will promote significant savings in antibody consumption. This RTD model is ready for further application in end-to-end integrated continuous downstream processes, such as addressing material traceability during continuous virus filtration processes.

Reviewing the process intensification landscape through the introduction of a novel, multitiered classification for downstream processing

A demand for process intensification in biomanufacturing has increased over the past decade due to the ever-expanding market for biopharmaceuticals. This is largely driven by factors such as a surge in biosimilars as patents expire, an aging population, and a rise in chronic diseases. With these market demands, pressure upon biomanufacturers to produce quality products with rapid turnaround escalates proportionally. Process intensification in biomanufacturing has been well received and accepted across industry based on the demonstration of its benefits of improved productivity and efficiency, while also reducing the cost of goods. However, while these benefits have been shown empirically, the challenges of adopting process intensification into industry remain, from smaller independent start-up to big pharma. Traditionally, moving from batch to a process intensification scheme has been viewed as an "all or nothing" approach involving continuous bioprocessing, in which the factors of complexity and significant capital costs hinder its adoption. In addition, the literature is crowded with a variety of terms used to describe process intensification (continuous, periodic counter-current, connected, intensified, steady-state, etc.). Often, these terms are used inappropriately or as synonyms, which generates confusion in the field. Through a detailed review of current state-of-the-art systems, consumables, and process intensification case studies, we herein propose a defined approach in the implementation of downstream process intensification through a standardized nomenclature and viewing it as distinct independent levels. These can function separately as intensified single-unit operations or be built upon by integration with other process steps allowing for simple, incremental, cost-effective implementation of process intensification in the manufacturing of biopharmaceuticals.

Current challenges and recent advances on the path towards continuous biomanufacturing

Continuous biopharmaceutical manufacturing is currently a field of intense research due to its potential to make the entire production process more optimal for the modern, ever-evolving biopharmaceutical market. Compared to traditional batch manufacturing, continuous bioprocessing is more efficient, adjustable, and sustainable and has reduced capital costs. However, despite its clear advantages, continuous bioprocessing is yet to be widely adopted in commercial manufacturing. This article provides an overview of the technological roadblocks for extensive adoptions and points out the recent advances that could help overcome them. In total, three key areas for improvement are identified: Quality by Design (QbD) implementation, integration of upstream and downstream technologies, and data and knowledge management. First, the challenges to QbD implementation are explored. Specifically, process control, process analytical technology (PAT), critical process parameter (CPP) identification, and mathematical models for bioprocess control and design are recognized as crucial for successful QbD realizations. Next, the difficulties of end-to-end process integration are examined, with a particular emphasis on downstream processing. Finally, the problem of data and knowledge management and its potential solutions are outlined where ontologies and data standards are pointed out as key drivers of progress.

Physics-informed neural networks to solve lumped kinetic model for chromatography process

Numerical method is widely used for solving the mechanistic models of chromatography process, but it is time-consuming and hard to response in real-time. Physics-informed neural network (PINN) as an emerging technology combines the structure of neural network with physics laws, and is getting noticed for solving physics problems with a balanced accuracy and calculation speed. In this research, a proof-of-concept study was carried out to apply PINN to chromatography process simulation. The PINN model structure was designed for the lumped kinetic model (LKM) with all LKM parameters. The PINN structure, training data and model complexity were optimized, and an optimal mode was obtained by adopting an in-series structure with a nonuniform training data set focusing on the breakthrough transition region. A PINN for LKM (LKM-PINN) consisting of four neural networks, 12 layers and 606 neurons was then used for the simulation of breakthrough curves of chromatography processes. The LKM parameters were estimated with two breakthrough curves and used to infer the breakthrough curves at different residence times, loading concentrations and column sizes. The results were comparable to that obtained with numerical methods. With the same raw data and constraints, the average fitting error for LKM-PINN model was 0.075, which was 0.081 for numerical method. With the same initial guess, the LKM-PINN model took 160 s to complete the fitting, while the numerical method took 7 to 72 min, depending on the fitting settings. The fitting speed of LKM-PINN model was further improved to 30 s with random initial guess. Thus, the LKM-PINN model developed in this study is capable to be applied to real-time simulation for digital twin.

Model-assisted process development, characterization and design of continuous chromatography for antibody separation

Continuous manufacturing in monoclonal antibody production has generated increased interest due to its consistent quality, high productivity, high equipment utilization, and low cost. One of the major challenges in realizing continuous biological manufacturing lies in implementing continuous chromatography. Given the complex operation mode and various operation parameters, it is challenging to develop a continuous process. Due to the process parameters being mainly determined by the breakthrough curves and elution behaviors, chromatographic modeling has gradually been used to assist in process development and characterization. Model-assisted approaches could realize multi-parameter interaction investigation and multi-objective optimization by integrating continuous process models. These approaches could reduce time and resource consumption while achieving a comprehensive and systematic understanding of the process. This paper reviews the application of modeling tools in continuous chromatography process development, characterization and design. Model-assisted process development approaches for continuous capture and polishing steps are introduced and summarized. The challenges and potential of model-assisted process characterization are discussed, emphasizing the need for further research on the design space determination strategy and parameter robustness analysis method. Additionally, some model applications for process design were highlighted to promote the establishment of the process optimization and process simulation platform.

Digital twin of a continuous chromatography process for mAb purification: Design and model-based control

Model-based design of integrated continuous train coupled with online process analytical technology (PAT) tool can be a potent facilitator for monitoring and control of Critical Quality Attributes (CQAs) in real time. Charge variants are product related variants and are often regarded as CQAs as they may impact potency and efficacy of drug. Robust pooling decision is required for achieving uniform charge variant composition for mAbs as baseline separation between closely related variants is rarely achieved in process scale chromatography. In this study, we propose a digital twin of a continuous chromatography process, integrated with an online HPLC-PAT tool for delivering real time pooling decisions to achieve uniform charge variant composition. The integrated downstream process comprised continuous multicolumn capture protein A chromatography, viral inactivation in coiled flow inverter reactor (CFIR), and multicolumn CEX polishing step. An online HPLC was connected to the harvest tank before protein A chromatography. Both empirical and mechanistic modeling have been considered. The model states were updated in real time using online HPLC charge variant data for prediction of the initial and final cut point for CEX eluate, according to which the process chromatography was directed to switch from collection to waste to achieve the desired charge variant composition in the CEX pool. Two case studies were carried out to demonstrate this control strategy. In the first case study, the continuous train was run for initially 14 h for harvest of fixed charge variant composition as feed. In the second case study, charge variant composition was dynamically changed by introducing forced perturbation to mimic the deviations that may be encountered during perfusion cell culture. The control strategy was successfully implemented for more than ±5% variability in the acidic variants of the feed with its composition in the range of acidic (13%-17%), main (18%-23%), and basic (59%-68%) variants. Both the case studies yielded CEX pool of uniform distribution of acidic, main and basic profiles in the range of 15 ± 0.8, 31 ± 0.3, and 53 ± 0.5%, respectively, in the case of empirical modeling and 15 ± 0.5, 31 ± 0.3, and 53 ± 0.3%, respectively, in the case of mechanistic modeling. In both cases, process yield for main species was >85% and the use of online HPLC early in the purification train helped in making quicker decision for pooling of CEX eluate. The results thus successfully demonstrate the technical feasibility of creating digital twins of bioprocess operations and their utility for process control.

Potential use of Cypridina luciferin for quantifying alpha 1-acid glycoprotein in human serum

Alpha 1-acid glycoprotein (AGP) is an acute phase protein in mammals, including humans. The amount of AGP in human serum varies in response to certain diseases; thus, many efforts have been made to develop methods for quantifying human AGP. We recently discovered that luminescence occurs merely by mixing Cypridina luciferin with human AGP under human serum-free neutral or basic buffer conditions. In this study, we tested an application of Cypridina luciferin for quantifying AGP contained in human serum. Our luminescence spectrum measurements of Cypridina luciferin with human serum samples showed that the maximum emission wavelength with human serum (480 nm) differed from that with human AGP (464 nm) due to the abundant presence of endogenous human serum albumin (HSA). Furthermore, the luminescence intensities of Cypridina luciferin with human AGP in HSA-depleted human serum were consistent with those in a human serum-free basic buffer, but those in human serum were not. These results indicated that depletion of HSA in human serum was required to use Cypridina luciferin for quantifying AGP in human serum. Additionally, we found that the luminescence intensity of Cypridina luciferin with bovine AGP was approximately tenfold lower than that with human AGP.

Model-based evaluation and model-free strategy for process development of three-column periodic counter-current chromatography

Multi-column counter-current chromatography is an advanced technology used for continuous capture processes to improve process productivity, resin capacity utilization and product consistency. However, process development is difficult due to process complexity. In this work, some general and convenient guidances for three-column periodic counter-current chromatography (3C-PCC) were developed. Boundaries and distributions of operating windows of 3C-PCC processes were clarified by model-based predictions. Interactive effects of feed concentration (c₀), resin properties (q_max and D_e), recovery and regeneration times (t_RR) were evaluated over a wide range for maximum productivity (P_max). Furthermore, variation of P_max was analyzed considering the constraint factors (capacity utilization target and flow rate limitation). The plateau value of P_max was determined by q_max and t_RR. The operating conditions for P_max were controlled by q_max, t_RR and c₀ interactively, and a critical concentration existed to judge whether the operating conditions of P_max under constraints. Based on the comprehensive understanding on 3C-PCC processes, a model-free strategy was proposed for process development. The optimal operating conditions could be determined based on a set of breakthrough curves, which was used to optimize process performance and screen resins. The approach proposed was validated using monoclonal antibody (mAb) capture with a 3C-PCC system under various mAb and feed concentrations. The results demonstrated that a thorough model-based process understanding on multi-column counter-current chromatography is important and could improve process development and establish a model-free strategy for more convenient applications.

Downstream processing of virus-like particles with aqueous two-phase systems: applications and challenges

The advancement of recombinant virus-like particle-based vaccines has attracted global attention owing to substantially safety and high efficacy in provoking a protective immunity against various chronic and infectious diseases in humans and animals. A robust, low-cost and scalability separation and purification technology is of utmost importance in the downstream processing of recombinant virus-like particles to produce affordable and safe vaccines. Being a relatively simple, environmentally friendly and efficient biomolecules recovery approach, aqueous two-phase systems have received great attention from researchers worldwide. This review aims to highlight the challenges and outlook in addition to the current applications of aqueous two-phase systems in downstream processing of virus-like particles. The efforts will confidently reinforce scholars' knowledge and fill in the valuable research gap in the aspect of concerning recombinant virus-like particle-based vaccines development, particularly related to the virus-like particles downstream production processes. This article is protected by copyright. All rights reserved.

A novel framework of surrogate-based feasibility analysis for establishing design space of twin-column continuous chromatography

Multi-column periodic counter-current chromatography (PCC) has attracted wide attention for the primary capture for the purpose of achieving continuous biomanufacturing. Consequently, determining the design space of the continuous capture process is very important to facilitate process understanding and improving product quality. In this work, we proposed a novel approach to identify the design space of continuous chromatography to balance the computational complexity and model predictions. Specifically, surrogate-based feasibility analysis with adaptive sampling is applied to establish the design space of twin-column CaptureSMB process. The surrogate model is constructed based on the developed mechanistic model for the identification of the design space. The effects of process variables (including interconnected loading time, interconnected flowrate, and batch flowrate) on the design space are comprehensively examined based on an active set strategy. Besides, essential factors like recovery-regeneration time and constraints of column performance parameters (yield, productivity, and capacity utilization) are thoroughly investigated. The impact of design variables such as column length is also studied.

Patterns of protein adsorption in ion-exchange particles and columns: Evolution of protein concentration profiles during load, hold, and wash steps predicted for pore and solid diffusion mechanisms

Elucidation of protein transport mechanism in ion exchanges is essential to model separation performance. In this work we simulate intraparticle adsorption profiles during batch adsorption assuming typical process conditions for pore, solid and parallel diffusion. Artificial confocal laser scanning microscopy images are created to identify apparent differences between the different transport mechanisms. Typical sharp fronts for pore diffusion are characteristic for Langmuir equilibrium constants of K_L ≥1. Only at K_L = 0.1 and lower, the profiles are smooth and practically indistinguishable from a solid diffusion mechanism. During hold and wash steps, at which the interstitial buffer is removed or exchanged, continuation of diffusion of protein molecules is significant for solid diffusion due to the adsorbed phase concentration driving force. For pore diffusion, protein mobility is considerable at low and moderate binding strength. Only when pore diffusion if completely dominant, and the binding strength is very high, protein mobility is low enough to restrict diffusion out of the particles. Simulation of column operation reveals substantial protein loss when operating conditions are not adjusted appropriately.