The optimization of upstream and downstream processes for production of recombinant adeno-associated virus (rAAV) with consistent quality depends on the ability to rapidly characterize critical quality attributes (CQAs). In the context of rAAV production, the virus titer, capsid content, and aggregation are identified as potential CQAs, affecting the potency, purity, and safety of rAAV-mediated gene therapy products. Analytical methods to measure these attributes commonly suffer from long turnaround times or low throughput for process development, although rapid, high-throughput methods are beginning to be developed and commercialized. These methods are not yet well established in academic or industrial practice, and supportive data are scarce. Here, we review both established and upcoming analytical methods for the quantification of rAAV quality attributes. In assessing each method, we highlight the progress toward rapid, at-line characterization of rAAV. Furthermore, we identify that a key challenge for transitioning from traditional to newer methods is the scarcity of academic and industrial experience with the latter. This literature review serves as a guide for the selection of analytical methods targeting quality attributes for rapid, high-throughput process characterization during process development of rAAV-mediated gene therapies.

Manufacturing of recombinant adeno-associated virus (rAAV) viral vectors remains challenging, with low yields and low full:empty capsid ratios in the harvest. To elucidate the dynamics of recombinant viral production, we develop a mechanistic model for the synthesis of rAAV viral vectors by triple plasmid transfection based on the underlying biological processes derived from wild-type AAV. The model covers major steps starting from exogenous DNA delivery to the reaction cascade that forms viral proteins and DNA, which subsequently result in filled capsids, and the complex functions of the Rep protein as a regulator of the packaging plasmid gene expression and a catalyst for viral DNA packaging. We estimate kinetic parameters using dynamic data from literature and in-house triple transient transfection experiments. Model predictions of productivity changes as a result of the varied input plasmid ratio are benchmarked against transfection data from the literature. Sensitivity analysis suggests that (1) the poorly coordinated timeline of capsid synthesis and viral DNA replication results in a low ratio of full virions in harvest, and (2) repressive function of the Rep protein could be impeding capsid production at a later phase. The analyses from the mathematical model provide testable hypotheses for evaluation and reveal potential process bottlenecks that can be investigated.

No abstract

This article describes the process characterization and development of models to inform a process control strategy to prepare (R,R)-epoxy ketone 2, an intermediate in the manufacture of carfilzomib. Model calibration for relevant unit operations and the development of a dynamic integrated flowsheet-level model in gPROMS FormulatedProducts software enabled investigation of the impact of process disturbances and model uncertainties on the critical quality attributes (CQAs) and identification of critical process disturbances and failure modes to guide a process control strategy. The model development was similar to that described in the previous parts of this series, but with the added complexity of comparing two distinct kinetic formulations for the epoxidation reaction. The main CQAs for this process were (1) the conversion of enone 1 (target ≥99.0 mol % conversion) and (2) the purity target for solids prior to cake wash (target ≥97.5% purity by weight). Conversion of enone was not always achieved with the expected disturbances: whereas 99.5% conversion was expected for normal operating conditions, 97.2% conversion was predicted for the worst-case combination of disturbances. The chiral purity of crystalline (R,R)-epoxy ketone 2 was not always achieved with the expected disturbances: 98.2% purity was expected for normal operating conditions, and 96.7% purity was expected for the worst-case combination of disturbances. These analyses allowed for rank ordering of critical process parameters that impact conversion and suitable manipulated variables to develop a robust process control strategy for the manufacturing scheme.

This article details efforts to characterize and develop a process control strategy for the manufacture of enone 2, a carfilzomib drug substance intermediate obtained through a Barbier-type Grignard reaction of morpholine amide 1. This includes the development of a novel mechanistic model for the heterogeneous Barbier-type Grignard reaction. After the model was characterized with laboratory-scale batch experiments, its performance was compared with experimental data collected under continuous operating conditions. Under nominal operating conditions, the experimentally measured conversion of morpholine amide varied from 94.3% to 96.7%, a range that was encompassed by the model. With a mechanistic model validated under continuous operating conditions, relationships between the magnesium charging interval and the variability in conversion of morpholine amide 1 to enone 2 were determined to further explore the experimental design space. The remaining unit operations were subsequently characterized, and the models developed for the individual operations were integrated into a flowsheet-level dynamic process model implemented in the gPROMS FormulatedProducts software. The impact of various process disturbances and model uncertainties on the critical quality attributes were then investigated, and critical process parameters, failure modes, and control strategies to address these disturbances were identified. The process was found to be most sensitive to operational disturbances in the supplied reactants: morpholine amide 1 and 2-bromopropene (2-BP). As 1 is manufactured upstream by the process described in Part 1 of this series, in silico analysis of potential process control strategies focused on manipulation of the 2-BP concentration and flow rate into the primary reactor. Overall, this work highlights the benefits of using mathematical modeling to deepen the understanding of pharmaceutical manufacturing processes and enable integrated unit operations in a continuous manufacturing setting.

This article details process characterization efforts and the development of corresponding process models to inform a process control strategy to produce a carfilzomib drug substance intermediate, morpholine amide 3. Model calibration for relevant unit operations and development of a dynamic integrated flowsheet-level model in gPROMS FormulatedProducts software allowed an investigation of the impact of process disturbances and model uncertainties on critical quality attributes (CQAs) and identification of critical process disturbances and failure modes to guide the process control strategy. The main CQA for this step was the conversion of Boc-D-leucine monohydrate (≥95%). The model was used to ensure that a state of control would be maintained in the presence of disturbances to target process parameters. The process was found to be robust against the analyzed disturbance scenarios, including worst-case combined disturbances. One case study highlights the dynamics of flow blockage for a key reagent, N,N′-carbonyldiimidazole (CDI), resulting from line clogging and the relationship between blockage duration and reaction conversion. The blockage was studied in silico, and the model demonstrated that the acceptance criteria for reaction conversion were met even with a flow rate reduction of 40%. The detrimental impact on the product concentration in the downstream process, however, required modification of the final distillation operation. The revised distillation column operation was demonstrated to address this concern and tolerated variable concentrations of morpholine amide while achieving the target water specification (<0.25 wt %). The results of this in silico analysis were verified with a production-scale demonstration of morpholine amide synthesis at a throughput of 12 kg/day to experimentally evaluate the impact of disturbances on the control strategy and overall performance of the system.

Metal-oxide nanoparticles with high surface area, controllable functionality and thermal and mechanical stability provide high affinity for enzymes when the next generation of biosensor applications are being considered. We report on the synthesis of metal-oxide-based nanoparticles (with different physical and chemical properties) using hydrothermal processing, photo-deposition and silane functionalization. Physical and chemical properties of the user-synthesized nanoparticles were investigated using scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDX), and Raman scattering, respectively. Thus, characterized metal-oxide-based nanoparticles served as nanosupports for the immobilization of soybean peroxidase enzyme (a model enzyme) through physical binding. The enzyme–nanosupport interface was evaluated to assess the optimum nanosupport characteristics that preserve enzyme functionality and its catalytic behavior. Our results showed that both the nanosupport geometry and its charge influence the functionality and catalytic behavior of the bio-metal-oxide hybrid system.

The methylotrophic yeast Pichia pastoris is widely used as a microbial host for recombinant protein production. Bioreactor models for P. pastoris can inform understanding of cellular metabolism and can be used to optimize bioreactor operation. This article constructs an extensive macroscopic bioreactor model for P. pastoris which describes substrates, biomass, total protein, other medium components, and off‐gas components. Species and elemental balances are introduced to describe uptake and evolution rates for medium components and off‐gas components. Additionally, a pH model is constructed using an overall charge balance, acid/base equilibria, and activity coefficients to describe production of recombinant protein and precipitation of medium components. The extent of run‐to‐run variability is modeled by distributions of a subset of the model parameters, which are estimated using the maximum likelihood method. Model prediction from the extensive macroscopic bioreactor model well describes experimental data with different operating conditions. The probability distributions of the model predictions quantified from the parameter distribution are quantifiably consistent with the run‐to‐run variability observed in the experimental data. The uncertainty description in this macroscopic bioreactor model identifies the model parameters that have large variability and provides guidance as to which aspects of cellular metabolism should be the focus of additional experimental studies. The model for medium components with pH and precipitation can be used for improving chemically defined medium by minimizing the amount of components needed while meeting cellular requirements.