Biologics Developability: Optimization of Engineered mAbs With Multi-Parameter Stability Characterization

Stability optimization of engineered mAbs Fahimeh Raoufi1, Marc Bailly1, Laurence Fayadat-Dilman1, Brett Thurlow2 , Stefanie Kall2 1Merck & Co, Discovery Biologics, Protein Sciences, South San Francisco, CA USA 2NanoTemper Technologies , Germany Abstract During biologics development, it is critical to ensure stability of a monoclonal antibody (mAb) with the ultimate goal of reaching the clinic. Biologics discovery often involves huge libraries of candidates with varying biophysical characteristics, which need to be evaluated and optimized for greater developability and downstream success. Understanding how candidate sequence attributes alter biophysical parameters is necessary for improved rational design and delivery of biological candidates. Examining how specific mutations alter the biophysical profile of a mAb is an important first step in the candidate selection and developability workflow. Here, the Protein Sciences Department within Biologics Discovery at Merck used the Prometheus Panta and the parameters obtained from nano-differential scanning fluorimetry (nanoDSF), backreflection (turbidity), and dynamic light scattering (DLS) to characterize a selection of monoclonal antibodies with sequence diversity. Introduction Researchers involved in the discovery phase of therapeutic biologics, particularly monoclonal antibodies, require many parameters to assess the characteristics of their candidates to ensure their streamlined development and long-term success in the clinic. Early phase discovery involves vast libraries of candidates with sequence and epitope diversity. Establishing a better understanding of how mutations in the sequence of the antibodies change biophysical parameters of these diverse candidates helps drive sequence selection.APPLICATION NOTE With Prometheus Panta, it is now possible to do dynamic light scattering (DLS) experiments in tandem with nano-differential scanning fluorimetry (nanoDSF) and backreflection along an entire thermal ramp in order to gain conformational, thermal stability, and turbidity information about your formulation1. Prometheus’s nanoDSF and turbidity measurements obtained from the NT.Plex instrument have been validated as industry-standard, but with the addition of DLS capabilities, it is crucial to ensure the same quality and reliability can be found in the Panta instrumentation. Merck’s Protein Sciences group recently used the Prometheus Panta to evaluate the properties of a series of monoclonal antibody mutants with the objective of supplying data to machine learning programs to improve the candidate selection process2. DLS, nanoDSF, and backreflection were used to evaluate sequence mutations and determine how they affected molecule stability. It was crucial not only to evaluate these candidates with the added DLS ability, but also to determine whether the Panta could be integrated into their workflow without a loss in measurement accuracy from the nanoDSF and turbidity measurements they are typically performing. Once the data were collected in the Panta, they compared data for the same molecules measured previously on the Prometheus NT.Plex to ensure repoducibility. Methods Prometheus Panta measurements Antibody samples were purified to 1 mg/ml in 20 mM sodium acetate pH 5.5 buffer. Single replicates of each sample were run in the Early Access Program (EAP) Prometheus Panta with the 48 individual capillary adapter in place. UV LED excitation was 12% for nanoDSF and turbidity measurements; excitation was set to 100% for DLS experiments. Isothermal measurements for DLS were collected at 25oC using the high-sensitivity mode prior to the thermal denaturation experiments. Temperature denaturation with nanoDSF, backreflection, and DLS acquisition was run from 2595oC at 1oC/min. Data was collected using a beta version of Panta.Control software, and after converted to a format to allow analysis using Panta.Analysis software v1.1. Prometheus NT.Plex measurements Single replicates of each sample were run in the Prometheus NT.Plex. UV LED excitation was 40% for nanoDSF and turbidity measurements. Temperature denaturation with nanoDSF and backreflection acquisition was run from 25-95oC at 1oC/min. Data was collected using version 2.2 of Therm.Control software. ©2021 NanoTemper Technologies, Inc. South San Francisco, CA, USA. All Rights Reserved. 2APPLICATION NOTE Results Evaluation of two unique families of mAbs IgG-based mAb candidates had mutations introduced in their non-binding regions to profile how single- and double- site mutations altered the stability profiles of these molecules. The aim was to better understand how the genetic and therefore proteomic diversity of early phase antibodies were affected by small changes in the structural core. Each antibody and its mutants were evaluated using nanoDSF, turbidity, and DLS along a thermal ramp to characterize their biophysical characteristics and determine how the mutations affected their stability. When examining the thermal denaturation profiles of the mAbs, it is easy to see their varying behavior. In Figure 1A, four antibodies “Group 1” exhibit different melting temperatures (Tm ) via nanoDSF. Additionally, two antibodies (Protein A and Protein D) exhibit significant turbidity accumulation, while Proteins B and C only have modest turbidity signal increase; this Nomenclature is reflected in the cumulant radius data, which also shows an increased cumulant radius at the onset of the experiment for Protein D. For Group 2, the six antibodies show pronounced differences in their nanoDSF unfolding profiles in Figure 1B. Likewise, there are significant differences in their turbidity and cumulant radii changes. Notably, only Protein G shows no major turbidity changes, which is also reflected in its cumulant radius. Detailed parameter measurements can be found in Table 1. Taken together, these results demonstrate the value of simultaneous measurements of DSF, nanoDSF, and backreflection along a thermal gradient. The results demonstrate the structural variation among IgG-derived mAbs, even within related groups. The values for the antibodies can be quantified to evaluate their fitness for development, and to determine how mutations in the genetic sequence can alter biophysical parameters. Tm1 First melting inflection temperature, at which 50% of the least thermalstable domain is unfolded Ton Onset of unfolding, temperature at which deviation from linear baseline exceeds 0.5% Tagg Onset of aggregation, temperature at which scattering deviates >1.1x of initial value (also Tsize , Tscattering ) Tturb Onset of turbidity, temperature at which backreflection signal deviates from linear baseline by > 0.5% (Referred to as Tagg in Prometheus PR.Plex instruments) PDI Polydispersity Index, a measure of the heterogeneity of particles in a sample; lower values indicate a more homogenous mixture ©2021 NanoTemper Technologies, Inc. South San Francisco, CA, USA. All Rights Reserved. 3APPLICATION NOTE Tm1 Protein A Protein B Protein C Protein D Protein E Protein F Protein G Protein H Protein I 64.99 74.42 70.13 70.43 63.91 69.06 71.47 Ton 60.48 65.23 63.47 63.47 54.65 57.30 Tturb 64.33 83.78 78.48 82.91 Tagg 63.42 75.71 PDI (25ᵒC) 0.1 0.03 79.16 80.84 59.95 79.39 65.26 69.70 69.69 Protein J 69.69 Protein A Protein B 63.52 62.61 62.62 Protein C 82.97 69.68 76.77 67.93 60.41 68.28 82.34 65.47 70.71 67.10 A B Protein D Ratio 350 nm / 330 nm Turbidity [mAU] Cumulant Radius [nm] Temperature [ºC] Ratio 350 nm / 330 nm Turbidity [mAU] Cumulant Radius [nm] 0.01 1.14 0.33 1.17 0.03 0.01 0.04 0.01 Protein E Protein F Table 2: Summary data of first unfolding temperature, unfolding onset, turbidity onset, scattering onset, and PDI for each parent in two groups. A PDI <0.1 is considered highly monodisperse; anything above 0.25 is considered highly polydisperse -- high PDI values highlighted in red. Protein G Protein H Protein I Protein J Temperature [ºC] Figure 1: Profiles of parents from Group 1 (A) and Group 2 (B) antibody groups have unique thermal profiles from intrinsic nanoDSF (top), turbidity (center), and DLS cumulant radius measurement (bottom). ©2021 NanoTemper Technologies, Inc. South San Francisco, CA, USA. All Rights Reserved. 4APPLICATION NOTE Evaluation of mutations on antibodies to determine how they alter multiple stability parameters With the establishment of initial Tm , Ton , rH , and PDI values for the non-mutated antibodies, it is possible to determine whether the mutant mAbs have significant deviations even with only modest structural alterations. Here we have selected two antibody groups and their mutants, representative from each group of mAbs. In Figure 2A, we see Protein A and its mutants A1-4. Note that overall, the changes the mutations made to profiles was minimal compared to the differences in the initial groups, as seen in Figure 1A and 1B. In Figure 2B, we see similar results for Protein G, where the nanoDSF and turbidity data have broadly similar shapes to the non-mutated molecule; note that the non-mutated molecule’s cumulant radius does not indicate a large change in size until a much higher temperature than the mutant molecules. When examining Table 2, it becomes evident that the mutant mAbs did not exhibit significantly increased stability, either thermal or conformational, compared to the non-mutated. However, it also exemplifies the differences that structural changes from single or double mutations can introduce to a molecule. Small changes in the protein sequence can have significant and measurable effects on the biophysical characteristics of a mAb. ©2021 NanoTemper Technologies, Inc. South San Francisco, CA, USA. All Rights Reserved. 5