As part of the first anniversary issue of Nature Chemical Engineering, we present a collection of opinions from 40 researchers within the field on what they think are the most exciting opportunities that lie ahead for their respective topics.

Therapeutic oligonucleotides (ONs) are typically manufactured via solid-phase synthesis, characterized by limited scalability and huge environmental footprint, limiting their availability. Biomanufactured ONs have the potential to reduce the immunogenic side-effects, and to improve the sustainability of their chemical counterparts. Rhodovulum sulfidophilum was demonstrated a valuable host for the extracellular production of recombinant ONs. However, low viable cell densities and product titer were reported so far. In this work, perfusion cell cultures were established for the intensification of ON biomanufacturing. First, the perfusion conditions were simulated in 50 mL spin tubes, selected as a scale-down model of the process, with the aim of optimizing the medium composition and process parameters. This optimization stage led to an increase in the cell density by 44 % compared to the reference medium formulation. In addition, tests at increasing perfusion rates were conducted until achieving the maximum viable cell density (VCDmax), allowing the determination of the minimum cell-specific perfusion rate (CSPRmin) required to sustain the cell culture. Intriguingly, we discovered in this system also a maximum CSPR, above which growth inhibition starts. By leveraging this process optimization, we show for the first time the conduction of perfusion cultures of R. sulfidophilum in bench-scale bioreactors. This process development pipeline allowed stable cultures for more than 20 days and the continuous biomanufacturing of ONs, testifying the great potential of perfusion processes.

Process models are mathematical formulations (essentially a set of equations) that try to represent the real system/process in a digital or virtual form. These are derived either based on fundamental physical laws often combined with empirical assumptions or learned based on data. The former has been existing for several decades in chemical and process engineering while the latter has recently received a lot of attention with the emergence of several artificial intelligence/machine learning techniques. Hybrid modeling is an emerging modeling paradigm that explores the synergy between existing these two paradigms, taking advantage of the existing process knowledge (or engineering know-how) and information disseminated by the collected data. Such an approach is especially suitable for systems and industries where data generation is significantly resource intensive while at the same time fundamentally not completely deciphered such as the processes involved in the biopharmaceutical pipeline. This technology could, in fact, be the enabler to meeting the demands and goals of several initiatives such as Quality by design, Process Analytical tools, and Pharma 4.0. In addition, it can aid in different process applications throughout process development and Chemistry, Manufacturing, and Control (CMC) to make it more strategic and efficient. This article focuses on providing a step-by-step guide to the different considerations to be made to develop a reliable and applicable hybrid model. In addition, the article aims at highlighting the need for such tools in the biopharmaceutical industry and summarizes the works that advocate its implications. Subsequently, the key qualities of hybrid modeling that make it a key enabler in the biopharmaceutical industry are elaborated with reference to the literature demonstrating such qualities.

Polyhydroxyalkanoates (PHA) are among the most promising bio-based alternatives to conventional petroleum-based plastics. These biodegradable polyesters can in fact be produced by fermentation from bacteria like Cupriavidus necator, thus reducing the environmental footprint of the manufacturing process. However, ensuring consistent product quality attributes is a major challenge of biomanufacturing. To address this issue, the implementation of real-time monitoring tools is essential to increase process understanding, enable a prompt response to possible process deviations and realize on-line process optimization. In this work, a soft sensor based on in situ Raman spectroscopy was developed and applied to the in-line monitoring of PHA biomanufacturing. This strategy allows the collection of quantitative information directly from the culture broth, without the need for sampling, and at high frequency. In fact, through an optimized multivariate data analysis pipeline, this soft sensor allows monitoring cell dry weight, as well as carbon and nitrogen source concentrations with root mean squared errors (RMSE) equal to 3.71, 7 and 0.03 g/L, respectively. In addition, this tool allows the in-line monitoring of intracellular PHA accumulation, with an RMSE of 14 gPHA/gCells. For the first time, also the number and weight average molecular weights of the polymer produced could be monitored, with RMSE of 8.7E4 and 11.6E4 g/mol, respectively. Overall, this work demonstrates the potential of Raman spectroscopy in the in-line monitoring of biotechnology processes, leading to the simultaneous measurement of several process variables in real time without the need of sampling and labor-intensive sample preparations.

Protein PEGylation, i.e. functionalization with poly(ethylene glycol) chains, has been demonstrated an ef-ficient way to improve the therapeutic index of these biopharmaceuticals. We demonstrated that Multi -column Countercurrent Solvent Gradient Purification (MCSGP) is an efficient process for the separation of PEGylated proteins (Kim et al., Ind. and Eng. Chem. Res. 2021, 60, 29, 10764-10776), thanks to the internal recycling of product-containing side fractions. This recycling phase plays a critical role in the economy of MCSGP as it avoids wasting valuable product, but at the same time impacts its productivity extending the overall process duration. In this study, our aim is to elucidate the role of the gradient slope within this recycling stage on the yield and productivity of MCSGP for two case-studies: PEGylated lysozyme and an industrially relevant PEGylated protein. While all the examples of MCSGP in the literature refer to a single gradient slope in the elution phase, for the first time we systematically investigate three differ-ent gradient configurations: i) a single gradient slope throughout the entire elution, ii) recycling with an increased gradient slope, to shed light on the competition between volume of the recycled fraction and required inline dilution and iii) an isocratic elution during the recycling phase. The dual gradient elution proved to be a valuable solution for boosting the recovery of high-value products, with the potential for alleviating the pressure on the upstream processing.(c) 2023 Elsevier B.V. All rights reserved.

Oligonucleotides (ONs) are breaking through in the biopharmaceutical industry as a promising class of biotherapeutics. The main success of these molecules is due to their peculiar way of acting in the cellular process, regulating the gene expression and hence influencing the protein synthesis at a pretranslational level. Although the Food and Drug Administration (FDA) already approved a few ON-based therapeutics, their production cost strongly limits large-scale manufacturing: a situation that can be alleviated through process intensification. In this study, we address this problem by developing an efficient and continuous chromatographic purification process for ONs. In particular, we considered the chromatographic purification of an ON crude prepared by chemical synthesis using anion exchange resins. We demonstrate that in this system the competitive adsorption of the various species on the same sites of the resin leads to the displacement of the more weakly adsorbing species by the more strongly adsorbing ones. This phenomenon affects the behavior of the chromatographic units and it has been investigated in detail. Then, we developed a continuous countercurrent solvent gradient purification (MCSGP) process, which can significantly improve the productivity and buffer consumption compared to a classical single-column, batch chromatographic process.

Conjugation of biopharmaceuticals to polyethylene glycol chains, known as PEGylation, is nowadays an efficient and widely exploited strategy to improve critical properties of the active molecule, including stability, biodistribution profile, and reduced clearance. A crucial step in the manufacturing of PEGylated drugs is the purification. The reference process in industrial settings is single-column chromatography, which can meet the stringent purity requisites only at the expenses of poor product recoveries. A valuable solution to this trade-off is the Multicolumn Countercurrent Solvent Gradient Purification (MCSGP), which allows the internal and automated recycling of product-containing side fractions that are typically discarded in the batch processes. In this study, an ad hoc design procedure was applied to the single-column batch purification of an industrially relevant PEGylated protein, with the aim of defining optimal collection window, elution duration and elution buffer ionic strength to be then transferred to the MCSGP. This significantly alleviates the design of the continuous operation, subjected to manifold process parameters. The MCSGP designed by directly transferring the optimal parameters allowed to improve the yield and productivity by 8.2% and 17.8%, respectively, when compared to the corresponding optimized batch process, ensuring a purity specification of 98.0%. Once the efficacy of MCSGP was demonstrated, a detailed analysis of its cost of goods was performed and compared to the case of single-column purification. To the best of our knowledge, this is the first example of a detailed economic investigation of the MCSGP across different manufacturing scenarios and process cadences of industrial relevance, which demonstrated not only the viability of this continuous technology but also its flexibility.

The biopharmaceutical market has experienced a tremendous growth in the last years. However, this growth should be balanced considering the difficulty in bioproduct development and the associated high manufacturing costs. These limitations are pushing towards process intensification, stimulated by the Quality by Design (QbD) initiative. Integrated continuous biomanufacturing has emerged as a promising approach towards high throughputs and reliable product quality attributes. At the same time, to face the increased management complexity and compensate fluctuations at the level of the different unit operations, model-based digitalization stands as a fundamental element. In this review, we discuss the state of the art in integrated biomanufacturing and digitalization to highlight their potential towards process intensification. The continuous technologies adopted in the upstream and downstream processing are first reviewed, with a focus on perfusion bioreactors and continuous chromatography. Then, model-based digitalization and its potential in the monitoring and control of integrated bioprocesses are discussed.

In this work, we aim to introduce the concept of the degree of hybridization for cell culture process modeling. We propose that a family of hybrid models can be created with varying fractions of process knowledge explicitly encoded in the model, defined as the degree of hybridization, with the two extremes being fully data-driven (0%) and fully mechanistic (100%) models. Subsequently, the aim is to compare the different models based on different metrics: model accuracy, the experimental effort for model development, extrapolation capability, the capability of generating new process understanding, and ease of utilization in practice, and to demonstrate that this could provide an additional degree of freedom for model selection. We could quantitatively demonstrate that for the cell culture process, either extreme has limitations. The major drawback of the data-driven model is the poor performance at low data availability as well as poor extrapolation capability, inability to provide process understanding, and subsequently inefficient practical application. On the other hand, the mechanistic model has poor accuracy due to the addition of excessive knowledge that then biases the models. Moving from data-driven to mechanistic models, the performance of the models improves progressively, as long as the knowledge added is not too biased. We show that the choice of the hybrid models to be used is based on the goal of model development. For instance, hybrid models including mass balances on each species show better performance in transferring models across different modes of operation. On the other hand, models with a higher degree of hybridization allow for more process interpretation possibilities. For modeling accuracy, amount of training data, extrapolation, and practical applications, the Hybrid Rate (HR) model is found to have the optimal degree of hybridization. This is likely due to the compromise between adding process knowledge and increasing the model parameters achieved by the HR model. The HR model features the incorporation of mass balance and channels the data-driven modeling to cell-specific rates, and thus these two pieces of information appear to be the most crucial ones. Finally, we believe that the concept will be instrumental in progressively developing and testing hypotheses about complex processes such as cell cultures.

The biopharmaceutical industries are continuously faced with the pressure to reduce the development costs and accelerate development time scales. The traditional approach of heuristic-based or platform process-based optimization is soon getting obsolete, and more generalized tools for process development and optimization are required to keep pace with the emerging trends. Thus, advanced model-based methods that can reduce the can ensure accelerated development of robust processes with minimal experiments are necessary. Though mechanistic models for chromatography are quite popular, their success is limited by the need to have accurate knowledge of adsorption isotherms and mass transfer kinetics. As an alternative, in this work, a hybrid modeling approach is proposed. Thereby, the chromatographic unit behavior is learned by a combination of neural network and mechanistic model while fitting suitable experimental breakthrough curves. Since this approach does not require identifying suitable mechanistic assumptions for all the phenomena, it can be developed with lower effort. Thus, allowing the scientists to concentrate their focus on process development. The performance of the hybrid model is compared with the mechanistic Lumped kinetic Model for in-silico data and experiments conducted on a system of industrial relevance. The flexibility of the hybrid modeling approach results in about three times higher accuracies compared to Lumped Kinetic Model. This is validated for five different isotherm models used to simulate data, with the hybrid model showing about two to three times lower prediction errors in all the cases. Not only in prediction, but we could also show that the hybrid model is more robust in extrapolating across process conditions with about three times lower error than the LKM. Additionally, it could be demonstrated that an appropriately tailored formulation of the hybrid model can be used to generate representations for the underlying principles such as adsorption equilibria and mass transfer kinetics.

Biopharmaceuticals are subjected to very strict purity requirements to be marketed. At the same time, peptides and other biomolecules are industrially synthesized through techniques (e.g., solid-phase synthesis) often leading to the formation of many impurities with molecular characteristics very similar to the target product. Therefore, the purification of these mixtures via preparative chromatography can be very challenging. This typically involves ternary or central-cut separations, characterized by chromatograms where the central peak, corresponding to the target product, exhibits significant overlapping on both sides with impurities slightly more or less adsorbable. In single-column (batch) preparative chromatography, this leads to a typical yield-purity tradeoff, meaning that high purity can be obtained at the cost of low yield and vice versa, with obvious consequences on the overall production costs. This study demonstrates how this limitation can be alleviated using the continuous countercurrent operating mode, conducted on a multicolumn system, as a tool for process intensification. In particular, the Multicolumn Countercurrent Solvent Gradient Purification (MCSGP) process has been applied to the purification of an industrial crude mixture of icatibant, which is a peptidomimetic antagonist of bradykinin B2-receptor that has been recently also considered for the treatment of patients affected by COVID-19 disease. It is shown that MCSGP allows conjugating process simplicity (using only two columns) with a significant improvement in process performance, compared to the corresponding batch process. This includes all process performance parameters: yield, productivity, and buffer consumption for a given purity specification of icatibant.

In process engineering, two paradigms of modeling approaches exist: the mechanistic and the data-driven approaches with the former being completely based on knowledge while the latter completely based on data. In our previous work, we highlighted the advantages of using hybrid models that explores the synergy between mechanistic and data-driven models. Here we introduce the concept of developing a series of hybrid models constituted by a progressively increasing extent of process knowledge. Thus, aligning the models on the "degrees of hybridization" axis with data-driven model being 0% hybridized and mechanistic model being 100% hybridized. In this work, the proposed concept is demonstrated for the application of a chromatographic capture step where the models are evaluated based on (i) prediction accuracy, (ii) extrapolation ability, (iii) providing process understanding, and (iv) practical application. We show the limitations of both model variant extremes. On one hand, the performance of the mechanistic model is compromised due to an excessive imposition of knowledge, thus affecting its predictive capabilities and efficiency in practical utility. On the other hand, the data-driven model inherently is not suitable for application such as multicolumn chromatography or to gain process understanding. In contrast, a series of hybrid models could be developed with better and versatile performance in term of prediction, extrapolation, process understanding, and practical utility. We show that for general process applications the different hybrid model variants and their ensembles have comparable performance. We illustrate the criteria for selection of a particular hybrid model variant based on different considerations such as complexity of training or model development, acquired understanding, and data requirement.

The rapid growth and diffusion of biopharmaceuticals, including polyethylene glycol (PEG)-protein conjugates, required to set up stringent regulatory requirements, in terms of purity and clinical safety. In order to produce proteins that meet these requirements, the development of robust purification methods has been one of the main goals in the field of downstream processing in the last years. Most of these methods rely on single or sequential batch chromatographic separations, enabling one to reach the purity specification but often at the expense of low yield. An appealing alternative is the Multicolumn Countercurrent Solvent Gradient Purification (MCSGP). Thanks to the internal recycling of the fractions where product and impurities coelute, this process can provide a significant improvement in the process yield, preserving the purity specification. The drawback lies in the increased number of process parameters, compared to the batch, which complicates the design and optimization of this unit. In this work, we propose an ad hoc design procedure for the optimization of central-cut separations at a target purity. Using PEGylated lysozyme as a model system, we illustrate the use of this procedure to identify the load, the elution gradient as well as the collection intervals for optimal yield and productivity, at fixed purity specifications, for a batch (single) column system. The obtained optimal process parameters are subsequently transferred to the MCSGP unit. The so-designed MCSGP process exhibited superior performances compared to the corresponding optimal batch process by increasing the yield and productivity by 17.0% and 2.1%, respectively, at the purity specification of 80%.

The increasing demand for efficient and robust processes in the purification of monoclonal antibodies (mAbs) has recently brought frontal chromatography to the forefront. Applied during the polishing step, it enables the removal of high molecular weight aggregates from the target product, achieving high purities. Typically, this process is operated in batch using a single column, which makes it intrinsically subjected to a purity-yield tradeoff. This means that high purities can only be achieved at the cost of lowering the product yield and vice versa. Recently, a two-column continuous implementation of frontal chromatography, referred to as Flow2, was developed. Despite being able of alleviating the purity-yield tradeoff typical of batch operations, the increase in the number of process parameters complicates its optimal design, with the risk of not exploiting its full potential. In this study, we developed an ad hoc design procedure (DP) suitable for the optimization of both batch frontal chromatography and Flow2 in terms of purity, yield, and productivity. This procedure provided similar results as a multiobjective optimization based on genetic algorithm but with lower computational effort. Then, batch and Flow2 operated at their optimal conditions were compared. Besides showing a more favorable Pareto front of yield and productivity at a specified purity, the Flow2 process demonstrated improved robustness compared to the batch process with respect to modifications in the loading linear velocity, washing buffer ionic strength and loading time, thus providing an appealing operation for integrated processes.

This chapter introduces the necessary concepts to develop mammalian cell perfusion cultures for the expression of therapeutic proteins at lab scale. We highlight the operation of the orbitally shaken tubes and of a classical glass vessel reactor system coupled to an external alternating tangential flow (ATF) device. Two different experiments can be performed in the shake-tube system: (1) the VCDmax experiment exploring the maximum achievable viable cell density at a given medium exchange rate and (2) the VCDSS experiment for the prediction of process performance at constant viable cell density and a given medium exchange rate for the design of the benchtop bioreactor process. In addition, the operation of the benchtop system is discussed containing start-up and control procedures for a long-term production run.

In this chapter, we give an overview of the challenges and objectives in operating mammalian cell perfusion cultures and provide guidelines for the design and set-up of lab-scale bioreactor systems. Next, the control structure needed to maintain long-term stable and viable cultures is illustrated, followed by a section on the media design. In the last section, we discuss steady-state operation, reactor, and process dynamics, including product quality considerations. Comparisons with current technologies, particularly with respect to product quality, are also discussed.

This chapter provides a general introduction for the application of mechanistic and statistical models in bioreactor process development and optimisation. We first introduce the equations governing the behaviour of batch and continuous stirred tank reactors (CSTR) for a given chemical system. After presenting the criteria to generalise these concepts to biological systems and bioprocesses, we discuss the implementation of mechanistic models for the simulation and control of the bioreactor performance with particular emphasis on product quality attributes, such as N-linked glycosylation. In addition, we describe the implementation of statistical and hybrid models and their application in process development and reactor optimisation. Lastly, we compare the use of the various modelling techniques for process monitoring and control.

Multivariate latent variable methods have become a popular and versatile toolset to analyze bioprocess data in industry and academia. This work spans such applications from the evaluation of the role of the standard process variables and metabolites to the metabolomics level, that is, to the extensive number metabolic compounds detectable in the extracellular and intracellular domains. Given the substantial effort currently required for the measurement of the latter groups, a tailored methodology is presented that is capable of providing valuable process insights as well as predicting the glycosylation profile based on only four experiments measured over 12 cell culture days. An important result of the work is the possibility to accurately predict many of the glycan variables based on the information of three experiments. An additional finding is that such predictive models can be generated from the more accessible process and extracellular information only, that is, without including the more experimentally cumbersome intracellular data. With regards to the incorporation of omics data in the standard process analytics framework in the future, this works provides a comprehensive data analysis pathway which can efficiently support numerous bioprocessing tasks.

Emulsion polymerization of the three-monomer system butyl acrylate-styrene-methacrylic acid was performed in batch using a commercial maltodextrin derived from starch degradation as stabilizer. Stable latexes with narrow particle size distributions were obtained in all examined cases. A method was developed to analyze and quantify the partitioning of the maltodextrin between the continuous phase (supernatant) and the particle phase. Significant differences between the polysaccharides adsorbed onto particles with or without emulsion polymerization reaction were observed. The possible reactions of maltodextrin in presence of a radical initiator were studied in aqueous phase, thus confirming maltodextrin degradation. The formation of copolymers involving the original monomers and the stabilizer according to two different reactive pathways was also confirmed. In terms of adsorbed maltodextrin, two different contributions were observed: maltodextrin physically adsorbed and maltodextrin chemically grafted and/or physically incorporated into the polymer.