Assessment of the Protein-Protein Interactions in a Highly Concentrated Antibody Solution by Using Raman Spectroscopy

Convolutional Neural Networks Guided Raman Spectroscopy as a Process Analytical Technology (PAT) Tool for Monitoring and Simultaneous Prediction of Monoclonal Antibody Charge Variants

BACKGROUND

Charge related heterogeneities of monoclonal antibody (mAb) based therapeutic products are increasingly being considered as a critical quality attribute (CQA). They are typically estimated using analytical cation exchange chromatography (CEX), which is time consuming and not suitable for real time control. Raman spectroscopy coupled with artificial intelligence (AI) tools offers an opportunity for real time monitoring and control of charge variants.

OBJECTIVE

We present a process analytical technology (PAT) tool for on-line and real-time charge variant determination during process scale CEX based on Raman spectroscopy employing machine learning techniques.

METHOD

Raman spectra are collected from a reference library of samples with distribution of acidic, main, and basic species from 0-100% in a mAb concentration range of 0-20 g/L generated from process-scale CEX. The performance of different machine learning techniques for spectral processing is compared for predicting different charge variant species.

RESULT

A convolutional neural network (CNN) based model was successfully calibrated for quantification of acidic species, main species, basic species, and total protein concentration with R2 values of 0.94, 0.99, 0.96 and 0.99, respectively, and the Root Mean Squared Error (RMSE) of 0.1846, 0.1627, and 0.1029 g/L, respectively, and 0.2483 g/L for the total protein concentration.

CONCLUSION

We demonstrate that Raman spectroscopy combined with AI-ML frameworks can deliver rapid and accurate determination of product related impurities. This approach can be used for real time CEX pooling decisions in mAb production processes, thus enabling consistent charge variant profiles to be achieved.

Induced Circular Dichroism Analysis of Thermally Induced Conformational Changes on Protein Binding Sites Under a Crowding Environment

Protein-ligand interactions in crowded cellular environments play a crucial role in biological functions. The crowded environment can perturb the overall protein structure and local conformation, thereby influencing the binding pathway of protein-ligand reactions within the cellular milieu. Therefore, a detailed understanding of the local conformation is crucial for elucidating the intricacies of protein-ligand interactions in crowded cellular environments. In this study, we investigated the feasibility of induced circular dichroism (ICD) using 8-anilinonaphthalene-1-sulfonic acid (ANS) for local conformational analysis at the binding site in a crowding environment. Bovine serum albumin (BSA) concentration-dependent measurements were performed to assess the feasibility of ANS-ICD for analyzing protein interior binding sites. The results showed distinct changes in the ANS-ICD spectra of BSA solutions, indicating their potential for analyzing the internal conformation of proteins. Moreover, temperature-dependent measurements were performed in dilute and crowding environments, revealing distinct denaturation pathways of BSA binding sites. Principal component analysis of ANS-ICD spectral changes revealed lower temperature pre-denaturation in the crowded solution than that in the diluted solution, suggesting destabilization of binding sites owing to self-crowding repulsive interactions. The established ANS-ICD method can provide valuable conformational insights into protein-ligand interactions in crowded cellular environments.

Feasibility and performance of cross-clone Raman calibration models in CHO cultivation

Raman spectroscopy is widely used in monitoring and controlling cell cultivations for biopharmaceutical drug manufacturing. However, its implementation for culture monitoring in the cell line development stage has received little attention. Therefore, the impact of clonal differences, such as productivity and growth, on the prediction accuracy and transferability of Raman calibration models is not yet well described. Raman OPLS models were developed for predicting titer, glucose and lactate using eleven CHO clones from a single cell line. These clones exhibited diverse productivity and growth rates. The calibration models were evaluated for clone-related biases using clone-wise linear regression analysis on cross validated predictions. The results revealed that clonal differences did not affect the prediction of glucose and lactate, but titer models showed a significant clone-related bias, which remained even after applying variable selection methods. The bias was associated with clonal productivity and lead to increased prediction errors when titer models were transferred to cultivations with productivity levels outside the range of their training data. The findings demonstrate the feasibility of Raman-based monitoring of glucose and lactate in cell line development with high accuracy. However, accurate titer prediction requires careful consideration of clonal characteristics during model development.

Raman Spectroscopic Analysis of Highly-Concentrated Antibodies under the Acid-Treated Conditions

PURPOSE

Antibody drugs are usually formulated as highly-concentrated solutions, which would easily generate aggregates, resulting in loss of efficacy. Although low pH increases the colloidal dispersion of antibodies, acid denaturation can be an issue. Therefore, knowing the physical properties at low pH under high concentration conditions is important.

METHODS

Raman spectroscopy was used to investigate pH-induced conformational changes of antibodies at 50 mg/ml. Experiments in pH 3 to 7 were performed for human serum IgG and recombinant rituximab.

RESULTS

We detected the evident changes at pH 3 in Tyr and Trp bands, which are the sensitive markers of intermolecular interactions. Thermal transition analysis over the pH range demonstrated that the thermal transition temperature (Tm) was highest at pH 3. Acid-treated and neutralized one showed higher Tm than that of pH 7, indicating that their extent of intermolecular interactions correlated with the Tm values. Onset temperature was clearly different between concentrated and diluted samples. Colloidal analyses confirmed the findings of the Raman analysis.

CONCLUSION

Our studies demonstrated the positive correlation between Raman analysis and colloidal information, validating as a method for evaluating antibody conformation associated with aggregation propensities.

An Intercompany Perspective on Practical Experiences of Predicting, Optimizing and Analyzing High Concentration Biologic Therapeutic Formulations

Developing high-dose biologic drugs for subcutaneous injection often requires high-concentration formulations and optimizing viscosity, solubility, and stability while overcoming analytical, manufacturing, and administration challenges. To understand industry approaches for developing high-concentration formulations, the Formulation Workstream of the BioPhorum Development Group, an industry-wide consortium, conducted an inter-company collaborative exercise which included several surveys. This collaboration provided an industry perspective, experience, and insight into the practicalities for developing high-concentration biologics. To understand solubility and viscosity, companies desire predictive tools, but experience indicates that these are not reliable and experimental strategies are best. Similarly, most companies prefer accelerated and stress stability studies to in-silico or biophysical-based prediction methods to assess aggregation. In addition, optimization of primary container-closure and devices are pursued to mitigate challenges associated with high viscosity of the formulation. Formulation strategies including excipient selection and application of studies at low concentration to high-concentration formulations are reported. Finally, analytical approaches to high concentration formulations are presented. The survey suggests that although prediction of viscosity, solubility, and long-term stability is desirable, the outcome can be inconsistent and molecule dependent. Significant experimental studies are required to confirm robust product definition as modeling at low protein concentrations will not necessarily extrapolate to high concentration formulations.

Raman Marker Bands for Secondary Structure Changes of Frozen Therapeutic Monoclonal Antibody Formulations During Thawing

In this work we use Raman spectroscopy for protein characterization in the frozen state. We investigate the behavior of frozen therapeutic monoclonal antibody IgG1 formulation upon thawing by Raman spectroscopy. Secondary and tertiary structure of the protein in three different mab formulations in the frozen state are followed through observation of marker bands for α-helix, β-sheet and random coil. We identify the tyrosine intensity ratio I₈₅₆/I₈₃₀ as a marker for mab aggregation. Upon fast cooling (40 °C/min) to -80 °C we observe a significant increase of random coil and α -helical structures, while this is not the case for slower cooling (20 °C/min) to -80 °C. Most changes in the protein's secondary structure are observed in the course of thawing in the range up to -20 °C, when passing through the glass transitions and cold-crystallization of the two types of freeze-concentrated solutions formed through macro- and microcryoconcentration. An increase of protein concentration and the addition of mannitol suppress secondary structural changes but do no impact on aggregation.

Dipolar Interactions and Protein Hydration in Highly Concentrated Antibody Formulations

Molecular interaction mechanisms in high-concentrated protein systems are of fundamental importance for the rational development of biopharmaceuticals such as monoclonal antibody (mAb) formulations. In such high-concentrated protein systems, the intermolecular distances between mAb molecules are reduced to the size of the protein diameter (approx. 10 nm). Thus, protein-protein interactions are more pronounced at high concentrations; so a direct extrapolation of physicochemical properties obtained from measurements at a low protein concentration of the corresponding properties at a high protein concentration is highly questionable. Besides the charge-charge interaction, the effects of molecular crowding, dipolar interaction, changes in protein hydration, and self-assembling tendency become more relevant. Here, protein hydration, protein dipole moment, and protein-protein interactions were studied in protein concentrations up to 200 mg/mL (= 1.3 mM) in different formulations for selected mAbs using dielectric relaxation spectroscopy (DRS). These data are correlated with the second virial coefficient, A₂, the diffusion interaction parameter, k_D, the elastic shear modulus, G', and the dynamic viscosity, η. When large contributions of dipolar protein-protein interactions were observed, the tendency of self-assembling and an increase in solution viscosity were detected. These effects were examined using specific buffer conditions. Furthermore, different types of protein-water interactions were identified via DRS, whereby the effect of high protein concentration on protein hydration was investigated for different high-concentrated liquid formulations (HCLFs).

Multi-attribute Raman spectroscopy (MARS) for monitoring product quality attributes in formulated monoclonal antibody therapeutics

Rapid release of biopharmaceutical products enables a more efficient drug manufacturing process. Multi-attribute methods that target several product quality attributes (PQAs) at one time are an essential pillar of the rapid-release strategy. The novel, high-throughput, and nondestructive multi-attribute Raman spectroscopy (MARS) method combines Raman spectroscopy, design of experiments, and multivariate data analysis (MVDA). MARS allows the measurement of multiple PQAs for formulated protein therapeutics without sample preparation from a single spectroscopic scan. Variable importance in projection analysis is used to associate the chemical and spectral basis of targeted PQAs, which assists in model interpretation and selection. This study shows the feasibility of MARS for the measurement of both protein purity-related and formulation-related PQAs; measurements of protein concentration, osmolality, and some formulation additives were achieved by a generic multiproduct model for various protein products containing the same formulation components. MARS demonstrates the potential to be a powerful methodology to improve the efficiency of biopharmaceutical development and manufacturing, as it features fast turnaround time, good robustness, less human intervention, and potential for automation.

Effects of Viscosity and Refractive Index on the Emission and Diffusion Properties of Alexa Fluor 405 Using Fluorescence Correlation and Lifetime Spectroscopies

Fluorescence Correlation Spectroscopy (FCS) studies of the interaction of polymers or proteins in solution are strongly affected by the viscosity and refractive index of the medium, and the effects are likely to be more significant with the use of short wavelength excitation (e.g., 405 nm diode lasers). Failing to account for these issues can lead to incorrect measurement of average size, conformational changes, and dynamic behaviour of polymers and proteins. Steady-state, time-resolved, and FCS measurements of Alexa 405 in glycerol:water mixtures were performed to determine its suitability for FCS measurements with 405 nm excitation. The effects of the refractive index and viscosity on the diffusion coefficient and photophysical parameters (lifetime and relative quantum yield) of the fluorophore were determined. Alexa 405 lifetime decreased from 3.55 ns in water to 3.25 ns in a 50:50 glycerol:water mixture, while its diffusion coefficient dropped from 333 ± 16 to 44 ± 1 µm²s^- 1. Lifetime data collected from micromolar solutions of Alexa 405 did however also suggest that as solvent polarity decreased, aggregates (excimers) were formed as evidenced by the appearance of a rising edge in the decay plots. The interdependence between lifetime, refractive index, and diffusion coefficient could be accurately fitted by a simple polynomial function indicating that the probe is well behaved and predictable in the glycerol:water model system. Overall, Alexa 405 is a most promising and reliable probe for FCS measurement using violet laser diode excitation sources.

Revisiting the Rate-Limiting Step of the ANS-Protein Binding at the Protein Surface and Inside the Hydrophobic Cavity

8-Anilino-1-naphthalenesulfonic acid (ANS) is used as a hydrophobic fluorescence probe due to its high intensity in hydrophobic environments, and also as a microenvironment probe because of its unique ability to exhibit peak shift and intensity change depending on the surrounding solvent environment. The difference in fluorescence can not only be caused by the microenvironment but can also be affected by the binding affinity, which is represented by the binding constant (K). However, the overall binding process considering the binding constant is not fully understood, which requires the ANS fluorescence binding mechanism to be examined. In this study, to reveal the rate-limiting step of the ANS-protein binding process, protein concentration-dependent measurements of the ANS fluorescence of lysozyme and bovine serum albumin were performed, and the binding constants were analyzed. The results suggest that the main factor of the binding process is the microenvironment at the binding site, which restricts the attached ANS molecule, rather than the attractive diffusion-limited association. The molecular mechanism of ANS-protein binding will help us to interpret the molecular motions of ANS molecules at the binding site in detail, especially with respect to an equilibrium perspective.

Spectroscopic Evidence of the Salt-Induced Conformational Change around the Localized Electric Charges on the Protein Surface of Fibronectin Type III

The effect of salt on the electrostatic interaction of a protein is an important issue, because addition of salt affects protein stability and association/aggregation. Although adding salt is a generally recognized strategy to improve protein stability, this improvement does not necessarily occur. The lack of an effect upon the addition of salt was previously confirmed for the tenth fibronectin type III domain from human fibronectin (FN3) by thermal stability analysis. However, the detailed molecular mechanism is unknown. In the present study, by employing the negatively charged carboxyl triad on the surface of FN3 as a case study, the molecular mechanism of the inefficient NaCl effect on protein stability was experimentally addressed using spectroscopic methods. Complementary analysis using Raman spectroscopy and 8-anilino-1-naphthalenesulfonic acid fluorescence revealed the three-phase behavior of the salt-protein interaction between NaCl and FN3 over a wide salt concentration range from 100 mM to 4.0 M, suggesting that the Na⁺-specific binding to the negatively charged carboxyl triad causes a local conformational change around the binding site with an accompanying structural change in the overall protein, which contributes to the protein's structural destabilization. This spectroscopic evidence clarifies the molecular understanding of the inefficiency of salt to improve protein stability. The findings will inform the optimization of formulation conditions.

Science and art of protein formulation development

Protein pharmaceuticals have become a significant class of marketed drug products and are expected to grow steadily over the next decade. Development of a commercial protein product is, however, a rather complex process. A critical step in this process is formulation development, enabling the final product configuration. A number of challenges still exist in the formulation development process. This review is intended to discuss these challenges, to illustrate the basic formulation development processes, and to compare the options and strategies in practical formulation development.

Spectroscopic Analysis of Protein-Crowded Environments Using the Charge-Transfer Fluorescence Probe 8-Anilino-1-Naphthalenesulfonic Acid

The molecular behaviors of proteins under crowding conditions are crucial for understanding the protein actions in intracellular environments. Under a crowded environment, the distance between protein molecules is almost the same size as the molecular level, thus, both the excluded volume effect and short ranged soft chemical interaction on protein surface could induce the complicated influence on the protein behavior cooperatively. Recently, various kinds of analytical approaches from macroscopic to microscopic aspects have been made to evaluate the crowding effect. The method, however, has not been established to evaluate the surface specific interactions on protein surface. In this study, the analytical method to evaluate the crowding effect has been suggested by using a charge-transfer fluorescence probe, ANS. By employing the unique property of ANS attaching to charged residues on the surface of lysozyme, the crowding effect was focused, while the case was compared as a reference, in which ANS is confined in hydrophobic pockets of BSA. Consequently, the surface specific changes of fluorescence spectra were readily observed under the crowded environment, whereas the fluorescence spectra of ANS in protein inside did not change. This result suggests the fluorescence spectra of ANS binding to protein surface have the capability to estimate the crowding effect of proteins.

Spectroscopic methods for assessing the molecular origins of macroscopic solution properties of highly concentrated liquid protein solutions

In cases of subcutaneous injection of therapeutic monoclonal antibodies, high protein concentrations (>50 mg/ml) are often required. During the development of these high concentration liquid formulations (HCLF), challenges such as aggregation, gelation, opalescence, phase separation, and high solution viscosities are more prone compared to low concentrated protein formulations. These properties can impair manufacturing processes, as well as protein stability and shelf life. To avoid such unfavourable solution properties, a detailed understanding about the nature of these properties and their driving forces are required. However, the fundamental mechanisms that lead to macroscopic solution properties, as above mentioned, are complex and not fully understood, yet. Established analytical methods for assessing the colloidal stability, i.e. the ability of a native protein to remain dispersed in solution, are restricted to dilute conditions and provide parameters such as the second osmotic virial coefficient, B₂₂, and the diffusion interaction coefficient, k_D. These parameters are routinely applied for qualitative estimations and identifications of proteins with challenging solution behaviours, such as high viscosities and aggregation, although the assays are prepared for low protein concentration conditions, typically between 0.1 and 20 mg/ml ("ideal" solution conditions). Quantitative analysis of samples of high protein concentration is difficult and it is hard to obtain information about the driving forces of such solution properties and corresponding protein-protein self-interactions. An advantage of using specific spectroscopic methods is the potential of directly analysing highly concentrated protein solutions at different solution conditions. This allows for collecting/gaining valuable information about the fundamental mechanisms of solution properties of the high protein concentration regime. In addition, the derived parameters might be more predictive as compared to the parameters originating from assays which are optimized for the low protein concentration range. The provided information includes structural data, molecular dynamics at various timescales and protein-solvent interactions, which can be obtained at molecular resolution. Herein, we provide an overview about spectroscopic techniques for analysing the origins of macroscopic solution behaviours in general, with a specific focus on pharmaceutically relevant high protein concentration and formulation conditions.

Efficient generation of single domain antibodies with high affinities and enhanced thermal stabilities

Single domain antibodies (sdAbs), made of natural single variable regions of camelid or cartilaginous fish antibodies, or unpaired variable regions of mouse or human IgGs, are some of the more promising biologic modalities. However, such conventional sdAbs have difficulties of either using unwieldy animals for immunization or having high affinity deficiencies. Herein, we offer a versatile method to generate rabbit variable domain of heavy chain (rVH) derived sdAbs with high affinities (K_D values of single digit nM or less) and enhanced thermal stabilities (equal to or even higher than those of camelid derived sdAbs). It was found that a variety of rVH binders, including those with high affinities, were efficiently acquired using an rVH-displaying phage library produced at a low temperature of 16 °C. By a simple method to introduce an additional disulfide bond, their unfolding temperatures were increased by more than 20 °C without severe loss of binding affinity. Differential scanning calorimetry analysis suggested that this highly efficient thermal stabilization was mainly attributed to the entropic contribution and unique thermodynamic character of the rVHs.

Applications of Raman Spectroscopy in Biopharmaceutical Manufacturing: A Short Review

The production of active pharmaceutical ingredients (APIs) is currently undergoing its biggest transformation in a century. The changes are based on the rapid and dramatic introduction of protein- and macromolecule-based drugs (collectively known as biopharmaceuticals) and can be traced back to the huge investment in biomedical science (in particular in genomics and proteomics) that has been ongoing since the 1970s. Biopharmaceuticals (or biologics) are manufactured using biological-expression systems (such as mammalian, bacterial, insect cells, etc.) and have spawned a large (>€35 billion sales annually in Europe) and growing biopharmaceutical industry (BioPharma). The structural and chemical complexity of biologics, combined with the intricacy of cell-based manufacturing, imposes a huge analytical burden to correctly characterize and quantify both processes (upstream) and products (downstream). In small molecule manufacturing, advances in analytical and computational methods have been extensively exploited to generate process analytical technologies (PAT) that are now used for routine process control, leading to more efficient processes and safer medicines. In the analytical domain, biologic manufacturing is considerably behind and there is both a huge scope and need to produce relevant PAT tools with which to better control processes, and better characterize product macromolecules. Raman spectroscopy, a vibrational spectroscopy with a number of useful properties (nondestructive, non-contact, robustness) has significant potential advantages in BioPharma. Key among them are intrinsically high molecular specificity, the ability to measure in water, the requirement for minimal (or no) sample pre-treatment, the flexibility of sampling configurations, and suitability for automation. Here, we review and discuss a representative selection of the more important Raman applications in BioPharma (with particular emphasis on mammalian cell culture). The review shows that the properties of Raman have been successfully exploited to deliver unique and useful analytical solutions, particularly for online process monitoring. However, it also shows that its inherent susceptibility to fluorescence interference and the weakness of the Raman effect mean that it can never be a panacea. In particular, Raman-based methods are intrinsically limited by the chemical complexity and wide analyte-concentration-profiles of cell culture media/bioprocessing broths which limit their use for quantitative analysis. Nevertheless, with appropriate foreknowledge of these limitations and good experimental design, robust analytical methods can be produced. In addition, new technological developments such as time-resolved detectors, advanced lasers, and plasmonics offer potential of new Raman-based methods to resolve existing limitations and/or provide new analytical insights.