Insulin purification-Innovation continuum via synthesis of fundamentals, technology, and modeling.

Advancement in all disciplines (art, science, education, and engineering) requires a careful balance of disruption and advancement of classical techniques. Often technologies are created with a limited understanding of fundamental principles and are prematurely abandoned. Over time, knowledge improves, new opportunities are identified, and technology is reassessed in a different light leading to a renaissance. Recovery of biological products is currently experiencing such a renaissance. Crystallization is one example of an elegant and ancient technology that has been applied in many fields and was employed to purify insulins from naturally occurring sources. Crystallization can also be utilized to determine protein structures. However, a multitude of parameters can impact protein crystallization and the "hit rate" for identifying protein crystals is relatively low, so much so that the development of a crystallization process is often viewed as a combination of art and science even today. Supplying the worldwide requirement for insulin (and associated variants) requires significant advances in process intensification to support scale of production and to minimize the overall cost to enable broader access. Expanding beyond insulin, the increasing complexity and diversity of biologics agents challenge the current purification methodologies. To harness the full potential of biologics, there is a need to fully explore a broader range of purification technologies, including nonchromatographic approaches. This impetus requires one to challenge and revisit the classical techniques including crystallization, chromatography, and filtration from a different vantage point and with a new set of tools, including molecular modeling. Fortunately, computational biophysics tools now exist to provide insights into mechanisms of protein/ligand interactions and molecular assembly processes (including crystallization) that can be used to support de novo process development. For example, specific regions or motifs of insulins and ligands can be identified and used as targets to support crystallization or purification development. Although the modeling tools have been developed and validated for insulin systems, the same tools can be applied to more complex modalities and to other areas including formulation, where the issue of aggregation and concentration-dependent oligomerization could be mechanistically modeled. This paper will illustrate a case study juxtaposing historical approaches to insulin downstream processes to a recent production process highlighting the application and evolution of technologies. Insulin production from Escherichia coli via inclusion bodies is an elegant example since it incorporates virtually all the unit operations associated with protein production-recovery of cells, lysis, solubilization, refolding, purification, and crystallization. The case study will include an example of an innovative application of existing membrane technology to combine three-unit operations into one, significantly reducing solids handling and buffer consumption. Ironically, a new separations technology was developed over the course of the case study that could further simplify and intensify the downstream process, emphasizing and highlighting the ever-accelerating pace of innovation in downstream processing. Molecular biophysics modeling was also employed to enhance the mechanistic understanding of the crystallization and purification processes.

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.

Highland games: A benchmarking exercise in predicting biophysical and drug properties of monoclonal antibodies from amino acid sequences.

Biopharmaceutical product and process development do not yet take advantage of predictive computational modeling to nearly the degree seen in industries based on smaller molecules. To assess and advance progress in this area, spirited coopetition (mutually beneficial collaboration between competitors) was successfully used to motivate industrial scientists to develop, share, and compare data and methods which would normally have remained confidential. The first "Highland Games" competition was held in conjunction with the October 2018 Recovery of Biological Products Conference in Ashville, NC, with the goal of benchmarking and assessment of the ability to predict development-related properties of six antibodies from their amino acid sequences alone. Predictions included purification-influencing properties such as isoelectric point and protein A elution pH, and biophysical properties such as stability and viscosity at very high concentrations. Essential contributions were made by a large variety of individuals, including companies which consented to provide antibody amino acid sequences and test materials, volunteers who undertook the preparation and experimental characterization of these materials, and prediction teams who attempted to predict antibody properties from sequence alone. Best practices were identified and shared, and areas in which the community excels at making predictions were identified, as well as areas presenting opportunities for considerable improvement. Predictions of isoelectric point and protein A elution pH were especially good with all-prediction average errors of 0.2 and 1.6 pH unit, respectively, while predictions of some other properties were notably less good. This manuscript presents the events, methods, and results of the competition, and can serve as a tutorial and as a reference for in-house benchmarking by others. Organizations vary in their policies concerning disclosure of methods, but most managements were very cooperative with the Highland Games exercise, and considerable insight into common and best practices is available from the contributed methods. The accumulated data set will serve as a benchmarking tool for further development of in silico prediction tools.

Loss of T cell antigen recognition arising from changes in peptide and major histocompatibility complex protein flexibility: implications for vaccine design.

Modification of the primary anchor positions of antigenic peptides to improve binding to major histocompatibility complex (MHC) proteins is a commonly used strategy for engineering peptide-based vaccine candidates. However, such peptide modifications do not always improve antigenicity, complicating efforts to design effective vaccines for cancer and infectious disease. Here we investigated the MART-1(27-35) tumor antigen, for which anchor modification (replacement of the position two alanine with leucine) dramatically reduces or ablates antigenicity with a wide range of T cell clones despite significantly improving peptide binding to MHC. We found that anchor modification in the MART-1(27-35) antigen enhances the flexibility of both the peptide and the HLA-A*0201 molecule. Although the resulting entropic effects contribute to the improved binding of the peptide to MHC, they also negatively impact T cell receptor binding to the peptide·MHC complex. These results help explain how the "anchor-fixing" strategy fails to improve antigenicity in this case, and more generally, may be relevant for understanding the high specificity characteristic of the T cell repertoire. In addition to impacting vaccine design, modulation of peptide and MHC flexibility through changes to antigenic peptides may present an evolutionary strategy for the escape of pathogens from immune destruction.

Cation exchange as a single polishing step for conjugated peptides.

Purification of peptides typically includes expensive reverse phase (RP) processes, which utilize high pressure and large volumes of solvent. For two conjugated peptides, chromatography process development targeted a low-pressure aqueous process that could achieve target product purities of ≥95%, comparable to purities seen with traditional RP. A high throughput screening approach of different modalities was used to identify binding and elution conditions on a cation exchange resin and small-scale columns were used in order to assess impurity removal and process yield. The parameters for load and gradient elution were optimized to increase product purity and process productivity with a wide operating window identified where high purity and productivity are achieved. Computational modeling was then used to validate experimental chromatography results and to gain insight on the effect of the chemical modifications on the surface properties of the two peptides. Both modeling and experimental data showed that with optimization, cation exchange could be utilized as a single polishing step for conjugated peptides. Similar purities were achieved as those seen with RP with up to double the productivity.

A general and efficient approach for NMR studies of peptide dynamics in class I MHC peptide binding grooves.

T-Cell receptor recognition of peptides bound by major histocompatibility complex (MHC) proteins initiates a cellular immune response. Dynamics of peptides within MHC binding grooves can influence TCR recognition, yet NMR studies which could address this rigorously have been hindered by the expense of isotopically labeled peptides and the large size of peptide-MHC complexes. Here we describe a methodology for characterizing peptide dynamics within MHC binding grooves via NMR, using a biosynthetic approach for producing labeled peptide. With the Tax(11-19) peptide bound to the human class I MHC HLA-A*0201, we demonstrate that peptide generated in this manner can be well characterized in MHC binding grooves by NMR, providing opportunities to more precisely study the role of peptide dynamics in TCR recognition. Demonstrating the utility of such studies, the data with the Tax(11-19) peptide indicate the presence of slow conformational exchange in the peptide, supporting an "induced-fit" style TCR binding mechanism.

Structures of MART-126/27-35 Peptide/HLA-A2 complexes reveal a remarkable disconnect between antigen structural homology and T cell recognition.

Small structural changes in peptides presented by major histocompatibility complex (MHC) molecules often result in large changes in immunogenicity, supporting the notion that T cell receptors are exquisitely sensitive to antigen structure. Yet there are striking examples of TCR recognition of structurally dissimilar ligands. The resulting unpredictability of how T cells will respond to different or modified antigens impacts both our understanding of the physical bases for TCR specificity as well as efforts to engineer peptides for immunomodulation. In cancer immunotherapy, epitopes and variants derived from the MART-1/Melan-A protein are widely used as clinical vaccines. Two overlapping epitopes spanning amino acid residues 26 through 35 are of particular interest: numerous clinical studies have been performed using variants of the MART-1 26-35 decamer, although only the 27-35 nonamer has been found on the surface of targeted melanoma cells. Here, we show that the 26-35 and 27-35 peptides adopt strikingly different conformations when bound to HLA-A2. Nevertheless, clonally distinct MART-1(26/27-35)-reactive T cells show broad cross-reactivity towards these ligands. Simultaneously, however, many of the cross-reactive T cells remain unable to recognize anchor-modified variants with very subtle structural differences. These dichotomous observations challenge our thinking about how structural information on unligated peptide/MHC complexes should be best used when addressing questions of TCR specificity. Our findings also indicate that caution is warranted in the design of immunotherapeutics based on the MART-1 26/27-35 epitopes, as neither cross-reactivity nor selectivity is predictable based on the analysis of the structures alone.

Characterization of a novel modification of a CHO-produced mAb: Evidence for the presence of tyrosine sulfation.

Herein we describe the investigation of a Chinese hamster ovary (CHO)-expressed human mAb molecule found partially modified by a +80 Da adduct. This mass difference, suggestive of a single sulfation or phosphorylation addition, was observed by mass analysis of the intact and reduced molecule by mass spectrometry (MS). The modification was located on tyrosine 31 (Y31) of the light chain in the complementarity-determining region 1 by liquid chromatography (LC)-MS peptide mapping and electron transfer dissociation fragmentation. The complete loss of the 80 Da modification moiety during collision induced dissociation fragmentation suggested this modification could not be a tyrosine phosphorylation. Treatment of the mAb with alkaline phosphatase confirmed our hypothesis. Western blot experiment using anti-tyrosine sulfation antibody and LC retention time correlation with corresponding synthetic sulfated peptides further confirmed the identification of tyrosine sulfation on the light chain. The unique sequence motif with neighboring acidic amino acids and local secondary structure might play a role to make Y31 a substrate residue for sulfation. This type of modification, to our knowledge, has not been previously reported for CHO-produced human IgG antibodies.

Targeted purification development enabled by computational biophysical modeling.

Chromatographic and non-chromatographic purification of biopharmaceuticals depend on the interactions between protein molecules and a solid-liquid interface. These interactions are dominated by the protein-surface properties, which are a function of protein sequence, structure, and dynamics. In addition, protein-surface properties are critical for in vivo recognition and activation, thus, purification strategies should strive to preserve structural integrity and retain desired pharmacological efficacy. Other factors such as surface diffusion, pore diffusion, and film mass transfer can impact chromatographic separation and resin design. The key factors that impact non-chromatographic separations (e.g., solubility, ligand affinity, charges and hydrophobic clusters, and molecular dynamics) are readily amenable to computational modeling and can enhance the understanding of protein chromatographic. Previously published studies have used computational methods such as quantitative structure-activity relationship (QSAR) or quantitative structure-property relationship (QSPR) to identify and rank order affinity ligands based on their potential to effectively bind and separate a desired biopharmaceutical from host cell protein (HCP) and other impurities. The challenge in the application of such an approach is to discern key yet subtle differences in ligands and proteins that influence biologics purification. Using a relatively small molecular weight protein (insulin), this research overcame limitations of previous modeling efforts by utilizing atomic level detail for the modeling of protein-ligand interactions, effectively leveraging and extending previous research on drug target discovery. These principles were applied to the purification of different commercially available insulin variants. The ability of these computational models to correlate directionally with empirical observation is demonstrated for several insulin systems over a range of purification challenges including resolution of subtle product variants (amino acid misincorporations). Broader application of this methodology in bioprocess development may enhance and speed the development of a robust purification platform.

Identification and engineering of human variable regions that allow expression of stable single-chain T cell receptors.

Single-chain antibody fragments (scFv), consisting of two linked variable regions (V(H) and V(L)), are a versatile format for engineering and as potential antigen-specific therapeutics. Although the analogous format for T cell receptors (TCRs), consisting of two linked V regions (Vα and Vß; referred to here as scTv), could provide similar opportunities, all wild-type scTv proteins examined to date are unstable. This obstacle has prevented scTv fragments from being widely used for engineering or therapeutics. To further explore whether some stable human scTv fragments could be expressed, we used a yeast system in which display of properly folded domains correlates with ability to express the folded scTv in soluble form. We discovered that, unexpectedly, scTv fragments that contained the human Vα2 region (IMGT: TRAV12 family) were displayed and properly associated with different Vß regions. Furthermore, a single polymorphic residue (Ser(α49)) in the framework region conferred additional thermal stability. These stabilized Vα2-containing scTv fragments could be expressed at high levels in Escherichia coli, and used to stain target cells that expressed the specific pep-HLA-A2 complexes. Thus, the scTv fragments can serve as a platform for engineering TCRs with diverse specificities, and possibly for therapeutic or diagnostic applications.

Thermodynamics of T-cell receptor-peptide/MHC interactions: progress and opportunities.

alphabeta T-cell receptors (TCRs) recognize peptide antigens presented by class I or class II major histocompatibility complex molecules (pMHC). Here we review the use of thermodynamic measurements in the study of TCR-pMHC interactions, with attention to the diversity in binding thermodynamics and how this is related to the variation in TCR-pMHC interfaces. We show that there is no enthalpic or entropic signature for TCR binding; rather, enthalpy and entropy changes vary in a compensatory manner that reflects a narrow free energy window for the interactions that have been characterized. Binding enthalpy and entropy changes do not correlate with structural features such as buried surface area or the number of hydrogen bonds within TCR-pMHC interfaces, possibly reflecting the myriad of contributors to binding thermodynamics, but likely also reflecting a reliance on van't Hoff over calorimetric measurements and the unaccounted influence of equilibria linked to binding. TCR-pMHC binding heat capacity changes likewise vary considerably. In some cases, the heat capacity changes are consistent with conformational differences between bound and free receptors, but there is little data indicating these conformational differences represent the need to organize disordered CDR loops. In this regard, we discuss how thermodynamics may provide additional insight into conformational changes occurring upon TCR binding. Finally, we highlight opportunities for the further use of thermodynamic measurements in the study of TCR-pMHC interactions, not only for understanding TCR binding in general, but also for understanding specifics of individual interactions and the engineering of TCRs with desired molecular recognition properties.

T cell receptor binding transition states and recognition of peptide/MHC.

T cell receptor recognition of peptide/MHC has been described as proceeding through a "two-step" process in which the TCR first contacts the MHC molecule prior to formation of the binding transition state using the germline-encoded CDR1 and CDR2 loops. The receptor then contacts the peptide using the hypervariable CDR3 loops as the transition state decays to the bound state. The model subdivides TCR binding into peptide-independent and peptide-dependent steps, demarcated at the binding transition state. Investigating the two-step model, here we show that two TCRs that recognize the same peptide/MHC bury very similar amounts of solvent-accessible surface area in their transition states. However, 1300-1500 A2 of surface area is buried in each, a significant amount suggestive of participation of peptide and associated CDR3 surface. Consistent with this interpretation, analysis of peptide and TCR variants indicates that stabilizing contacts to the peptide are formed within both transition states. These data are incompatible with the original two-step model, as are transition state models built using the principle of minimal frustration commonly employed in the investigation of protein folding and binding transition states. These findings will be useful in further explorations of the nature of TCR binding transition states, as well as ongoing efforts to understand the mechanisms by which T cell receptors recognize the composite peptide/MHC surface.