Analytical Strategies for Developability Assessment of Therapeutic Proteins

by Aaron Beach, Paul Wassmann, Thorsten Lorenz
Novartis Institute for BioMedical Research Inc.
Novartis Biologics Center, Integrated Biologics Profiling (IBP)

Introduction

Over the recent decades, the pharmaceutical industry has shifted focus to new modalities for disease treatment. The so-called “Biologics”, involves administration of proteins cells and nucleic acids derived drugs,¹ with therapeutic proteins (TP) currently representing the major portion in the biologics field. To reduce the risk of potential failure during technical development due to unexpected liabilities, protein candidates require a careful assessment of their developability profile as early as possible in the process. The integrated biologics profiling concept provides such an assessment for molecules transitioning from research to development, holistically evaluating manufacturability, biophysical characteristics, stability and in vivo fitness.^2,3 Until today most biopharmaceutical treatments have been antibody-based products, but the next generation of biopharmaceuticals is growing out of a higher diversity of molecular formats.^4-6 In order to provide an adequate developability assessment for the rapidly changing portfolio, new approaches and technologies must be employed, spanning all areas.

Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special offers
from American Pharmaceutical Review – all delivered right to your inbox! Sign up now!

To address these challenges in an efficient and targeted manner, we apply project stage-dependent work packages into three distinct phases with appropriately targeted analytical strategies. The first phase, termed Screening, allows for evaluation of an extremely targeted and project specific set of molecule parameters or liabilities early in a project to overcome or eliminate critical developability issues. Factors analyzed and optimized include but are not limited to expression vectors and hosts, titer, and molecule integrity in order to identify a suitable set of candidates, conditions and process attributes. The analytical support during this stage is driven by higher throughput and primarily generic methodologies. The second phase, termed Profiling, in most cases requires analytics optimization such that assay performance can distinguish between candidates on the basis of initial production methods (both expression and purification) and a preliminary quality attribute assessment for relative ranking of the candidates. The final phase, termed Lead development, utilizes detailed analytical characterization to provide a comprehensive developability assessment and risk/effort assessment of the final candidate(s), as well as final cell line, an initial scalable purification process, and preformulation data including accelerated stability. In this article, we will provide examples of the latest analytical approaches applied during the developability assessment process to enable characterization of higher complexity therapeutic proteins (TPs) across all three of these phases.

Molecular Liabilities of Therapeutics Proteins

Biologics development in light of new protein formats introduces a paradigm shift, in that a platform approach which is highly successful for monoclonal antibody (mAb) production for preclinical studies needs to be adapted for other protein classes. The challenges from this change do not only apply to expression and purification, but to analytical characterization as well. The entire strategy of analyzing these new entities requires rethinking to address new types of liabilities and higher investment in new analytical techniques.

Since mAbs reveal a significant homology in sequence and higher order structure, potential liabilities and properties can be assessed with a generic analytical strategy. Monitoring identity, biophysical attributes, high and low molecular weight species (HMWs/LMWs), and post translational modifications (PTMs) often does not require major changes to previously established assays. TPs, which are highly heterogeneous, can span a much wider range of biophysical attributes such as strong hydrophobicity, e.g. due to payloads or extensions and isoelectric point shifts due to hyper sialyation of glycans.⁷ TPs have high diversity with regard to glycosylation attributes, such as a high number of sites for glycosylation per molecule, multibranched glycans, or O-glycosylation motifs. Some formats are less amenable to high level recombinant production requiring expression improvement guided by appropriate analytics.⁸ TPs with oligomeric structure mediated by non-covalent interactions or engineered covalent disulfide bonds introduce a need for monitoring chain misassembly. Additional modifications such as phosphorylation⁹ and sulfation can be often found in different TP functional domains, thus the analytical strategy should include respective assays to characterize these modifications. Traditional degradation pathways may also not fit to new formats due to instabilities inherent to a TP (lower melting temperatures, non-covalent interactions, sensitivity to proteolytic damage). Sample preparation amenable to mAbs often is not suitable, as the physico-chemical properties of new molecules may be more sensitive to certain conditions (high concentration, extreme pH, surfactant compatibility, etc). All of these potential liabilities have a chance to further convolute analytical evaluation when their cumulative effect obfuscates proper readout.

With the variety of molecules and potential modifications, standard techniques might often not be appropriate for understanding or detecting certain liabilities. This necessitates the development of more tailor made approaches, as well as changing reference analytical methods to avoid artefactual findings. The utilization of modern approaches to analytical method development in light of these new challenges are of high value and examples are provided in the subsequent paragraphs.

State-of-the-Art Tools for TP Characterization and Efficient Analytical Method Development

Early high-throughput candidate and attribute assessment

The analytical strategy is motivated with the same philosophy outlined earlier, with a phased and risk-based approach. During the screening phase, developing project specific methods would be inefficient and resource intense to screen a large number of candidates or combinations of candidates and expression systems. The initial use of generic, high throughput methods is geared to obtain a qualitative or semi-quantitative assessment and to differentiate between candidates on specific parameters. Typical assays suitable for such high-throughput readouts are dynamic light scattering for aggregation propensity, differential scanning fluorimetry for conformational stability or plate-based UV readouts for monitoring precipitation. The wide range of biophysical characteristics of TPs can benefit from higher throughput assessment of such features in an early screening phase. For example, in certain cases it might be beneficial to understand the conformational stability of complex TPs with regard to environmental conditions (i.e. pH, buffer ionic strength) so that biological readout, early process development, and further analytical characterization can be executed in appropriate buffering conditions. While this applicability for screening eff orts occurs earlier in the process on intermediate material, the final highly purified drug substance would also profit from this biophysical screening approach, narrowing conditions for conformational and colloidal stability with minimal sample consumption and analysis time.¹⁰

In cases where only a small number of candidates enters the screening phase, often the focus is put on technical enablers to overcome expression challenges, or issues with molecular integrity. In these cases, assays for determining titer and protein degradation with adequate performance and throughput are applied to screen a broad set of expression conditions.

Efficient Method Development Using Design of Experiment (DoE) Approach

With the rise of complex protein formats, method development requires a change from the one-dimensional approach. This approach is often time and resource intense and adequate to improve methods to very high performance and sensitivity for exceptional challenges with antibody projects. Such a stepwise approach for liquid chromatography (LC) often tests a “standard method” for suitability, and then explores a variety of available method adjustments (resins, mobile phase, operating conditions) one by one to optimize assay performance (Figure 1). The combination of individual improvements during stepwise method development provides a set of conditions for the final optimized method. As the number of parameters to screen increase with complex TPs, successful early method development is less feasible with available resources and the limitations of a stepwise approach.

A more advanced approach leverages the concepts of DoE. The design space approach allows multiple method parameters to be defined before experimentation and analyzed in a matrix, testing many factors simultaneously while minimizing unnecessary efforts. Increasing sample size helps in the interpretation of factors relationship to the output, however in most cases it is not necessary to perform a full factorial testing, meaning the evaluation of all controlled factors. Setting an optimality criteria before starting experimentation can help to narrow the number of conditions to be analyzed in a given design space.¹¹ These controlled factors change from experiment to experiment, while the output in combination with uncontrolled factors allow for significance of each factor to be examined. Further extending this concept, and applying an automated method development platform,¹² can significantly streamline liquid chromatography method development through automatic control, method generation and multivariate data analysis.

Design space approaches are not only applicable for LC method development, but have been exemplified in capillary electrophoresissodium dodecyl sulfate (CE-SDS) method optimization for mAbs.¹³ Being able to assess interactions of various factors (denaturation temperature and time, SDS percentage, sample buffer pH) in tandem for novel protein formats significantly reduces resource investment in method development compared to the traditional stepwise approaches. When the design space approach is implemented, a combinatorial factor analysis with a defined performance goal space can provide the best resolving and most robust method (Figure 2).

The opportunity to apply these concepts for robustness is particularly valuable when applied earlier during developability assessment. We have successfully applied this concept to reverse phase method development, where an increase in resolution of two species was monitored by testing and interpreting over 15 factors was achieved in two days. As a result, a powerful in- process control method could be provided to support our early downstream process development. Deploying robust analytical methods early avoids potential delays and large resource investments.

State-of-the-Art and Combinational Analytical Technologies for Tackling Complexity of TPs

Newer concepts such as two dimensional liquid chromatography (2DLC) are coming into popularity¹⁴ for modern biologics characterization, and represents an opportunity for enhancing nonstandardized protein format analysis. Automation with robotics also provides an opportunity to simplify sample preparation, efficiently creating large sample sets across a broad set of conditions.

Complex biologics often require the combination of results from two independent methods or may require laborious analytical fraction collection and re-analysis of these fractions with orthogonal methodology to gain insight into unanticipated properties. New advances in UHPLC such as 2DLC with multiple heart cutting¹⁵ simplify this process by enabling on-line, automatic second dimensional analysis of individual peaks from the first dimension, without sacrificing the complete first dimension analysis. Pairing orthogonal LC methods with 2DLC also minimizes resident times in less favorable elution conditions, which avoids false positive readouts that re-chromatography could introduce due to storage condition instabilities post-elution. With regard to developability assessment, we have applied this technology to provide insight into molecular liabilities that may not be intuitive or conclusively related without direct analysis of impurities, e.g. co-eluting species in size exclusion chromatography. Such a platform could be implemented in a variety of cases, such as high throughput screening,¹⁶ but we see highest utility when leveraged for deep characterization of a smaller sample set¹⁷ containing otherwise co-eluting species in a sample under one dimensional analysis (Figure 3).

Conclusion

While platform approaches have benefits with well-established biologics formats such as antibodies or Fabs, the underlying diversity of TPs requires a new tiered strategy during developability assessment, using generic approaches in early stages, and later increasing investment in format specific methodologies to address the breadth of different molecular properties. Some of the approaches described here have already been implemented for successful candidate characterization and selection of TPs according to the integrated biologics profiling concept. The prospective portions of the analytical strategies described herein represent a field which greatly profits from thoughtful design to accelerate development of therapeutic proteins and potentially expand the knowledge base for predicting characterization assay parameters appropriate for each molecular format.

We believe that a flexible approach and adequate upfront investment in analytical method development for new modalities of biologics in risk based and phase dependent approach will be a key component in transitioning a candidate into development through an understanding of its critical physicochemical and biochemical properties.

Acknowledgements

We would like to thank all current and former members of the integrated biologics profiling (IBP) unit for their contributions to this article. Additionally we would like to thank many colleagues in Research and Development functions for fruitful discussions, supporting a constant improvement of our developability assessment strategy.

A special acknowledgment to our colleagues in the analytics team at IBP, especially Ankita Pandey for her contributions to the design space method development approach.

References

1. Phung, W., Sandoval, W., Kelley, R. F., & Lill, J. R. (2017). Analytical Tools for Biologics Molecular Assessment. Analytical Characterization of Biotherapeutics, 1-3.
2. Lorenz T. et al. 2014. Developability Assessment of Biologics by Integrated Biologics Profiling. American Pharmaceutical Review
3. Jarasch A, et al. 2015 Developability Assessment During the Selection of Novel Therapeutic Antibodies
4. Banta S, Dooley K, Shur O. 2013. Replacing antibodies: Engineering new binding proteins. Annu Rev Biomed Eng 15:93–113.
5. Ueda T. 2014. Next-generation optimized biotherapeutics—A review and preclinical study. Biochem Biophys Acta 1844(11):2053–2057.
6. Wurch T, Lowe P, Caussanel V, Bes C, Beck A, Corvaia N. 2008. Development of novel protein scaffolds as alternatives to whole antibodies for imaging and therapy: Status on discovery research and clinical validation. Curr Pharm Biotechnol 9:502–509
7. Song, E., Mayampurath, A., Yu, C. Y., Tang, H., & Mechref, Y. (2014). Glycoproteomics: Identifying the glycosylation of prostate specifi c antigen at normal and high isoelectricpoints by LC–MS/MS. Journal of proteome research, 13(12), 5570-5580.
8. Karengera, E., Durocher, Y., De Crescenzo, G., & Henry, O. (2017). Combining metabolic and process engineering strategies to improve recombinant glycoprotein production and quality. Applied microbiology and biotechnology, 101(21), 7837-7851.
9. Ketcham, S. A., Ashraf, M., & Madhavarao, C. N. (2017). Direct quantifi cation of protein glycan phosphorylation. BioTechniques, 63(3), 117-123.
10. Kunz, P., Flock, T., Soler, N., Zaiss, M., Vincke, C., Sterckx, Y., & Hoheisel, J. D. (2017). Exploiting sequence and stability information for directing nanobody stability engineering. Biochimica et Biophysica Acta (BBA)-General Subjects, 1861(9), 2196-2205.
11. SAS. JMP, Version 4; SAS Institute Inc.: Cary, NC.
12. Verseput, R., & Turpin, J. A. (2015). Fusion QbD and the ‘Perfect Storm’of Technologies Driving QbD-aligned LC Method Development. Chromatography Today, Aug./Sept, 64-68.
13. Michels, D. A., Parker, M., & Salas‐Solano, O. (2012). Quantitative impurity analysis of monoclonal antibody size heterogeneity by CE‐LIF: Example of development and validation through a quality‐by‐design framework. Electrophoresis, 33(5), 815-826.
14. Maloney, D. R. S. T. D. (2017). Recent Advances in Two-Dimensional Liquid Chromatography for Pharmaceutical and Biopharmaceutical Analysis. Recent Advances in Two-Dimensional Liquid Chromatography for Pharmaceutical and Biopharmaceutical Analysis.
15. Pursch, M., Lewer, P., & Buckenmaier, S. (2017). Resolving Co-Elution Problems of Components in Complex Mixtures by Multiple Heart-Cutting 2D-LC. Chromatographia, 80(1), 31-38.
16. K. Sandra, M. Steenbeke, I. Vandenheede, G. Vanhoenacker, P. Sandra, The versatility of heart-cutting and comprehensive two-dimensional liquid chromatography in monoclonal antibody clone selection, J. Chromatogr. A, 1523 (2017) 283-292.
17. S.H. Yang, J. Wang, K. Zhang. Validation of a two-dimensional liquid chromatography method for quality control testing of pharmaceutical materials, J. Chromatogr. A, 1492 (2017) 89-97.