Rational affinity maturation of anti-amyloid antibodies with high conformational and sequence specificity

Alec A. Desai, Matthew D. Smith, Yulei Zhang, Emily K. Makowski, Julia E. Gerson, Edward Ionescu, Charles G. Starr, Jennifer M. Zupancic, Shannon J. Moore, Alexandra Sutter, Magdalena I. Ivanova, Geoffrey G. Murphy, Henry L. Paulson, Peter M. TessierJournal of Biological Chemistry 2021 · DOI

TL;DR

Researchers have developed a novel directed evolution strategy to create highly specific antibodies against amyloid-beta (Aβ) aggregates, which are implicated in neurodegenerative diseases like Alzheimer's. These new antibodies exhibit improved affinity and conformational specificity compared to existing clinical-stage treatments, while also demonstrating significantly reduced off-target binding.

What Problem Does This Paper Address?

Amyloidogenic proteins, such as amyloid-beta (Aβ), aggregate into toxic structures that are central to neurodegenerative diseases like Alzheimer's and Parkinson's. Antibodies that can specifically target these harmful aggregates are crucial for developing diagnostics, therapeutics, and research tools. However, generating such antibodies is challenging due to the complex, heterogeneous, and often insoluble nature of protein aggregates. Traditional methods like immunization can lead to antibodies that lack specificity or bind to unintended targets. Existing directed evolution techniques, while powerful, often struggle with the insolubility of aggregate antigens and can lead to a loss of conformational specificity during affinity maturation. This paper aims to address these limitations by developing a systematic and efficient method for creating antibodies with high affinity, precise conformational recognition, and minimal off-target binding.

Methods

The core of this work is a directed evolution strategy for affinity maturation of a lead antibody (AF1) that already shows some specificity for Aβ fibrils. The process involves:

  1. Rational Library Design: Instead of random mutagenesis, the researchers designed a library of antibody variants by strategically sampling amino acid combinations in the antibody's antigen-binding site (complementarity-determining regions or CDRs). This design was guided by natural antibody diversity and constraints to favor specificity and avoid problematic residues (like positively charged ones that can increase non-specific binding).
  2. Yeast Surface Display: The designed antibody library was displayed on the surface of yeast cells. This is a standard technique for screening large antibody libraries.
  3. Iterative Sorting (MACS): The yeast library was subjected to multiple rounds of magnetic-activated cell sorting (MACS). Crucially, this involved:
  4. Negative Selection: Repeatedly removing yeast cells that bound to disaggregated Aβ (to ensure conformational specificity) and islet amyloid polypeptide (IAPP) fibrils (to ensure sequence specificity).
  5. Positive Selection: Selecting yeast cells that bound to aggregated Aβ fibrils.
  6. Deep Sequencing: After sorting, the enriched antibody libraries were analyzed using deep sequencing to identify mutations that were strongly correlated with improved binding to Aβ aggregates.
  7. Characterization: Promising antibody variants were then expressed and rigorously tested for their affinity, conformational specificity (binding to aggregated vs. disaggregated Aβ), and non-specific binding to other proteins.

Datasets

  • Human Antibody CDR Data: Used to inform the rational design of the antibody library based on natural diversity.
  • Alzheimer's Aβ42 Fibrils and Disaggregated Aβ42: Used as antigens for selection and specificity testing.
  • Islet Amyloid Polypeptide (IAPP) Fibrils: Used as a negative selection target to improve sequence specificity.
  • Clinical-stage Aβ Antibodies (Aducanumab and Crenezumab): Used as benchmarks for comparison of affinity, specificity, and biophysical properties.
  • Mouse Brain Samples (5xFAD transgenic and wild-type): Used to evaluate antibody performance in vivo.
  • Human Brain Tissue Lysates (Alzheimer's patients and controls): Used for ex vivo testing.
  • Polyspecificity Reagent (PSR): A mixture of soluble membrane proteins from CHO cells, used to assess non-specific binding.

Key Insights and Ideas

  • Rational Design is Key: Moving beyond random mutagenesis, a targeted approach to library design based on natural antibody diversity and biophysical principles significantly improved the efficiency of finding high-quality antibodies.
  • Strategic Negative Selection is Crucial: The iterative use of negative selections against disaggregated Aβ and other protein aggregates was vital for ensuring the final antibodies were conformationally specific and avoided cross-reactivity.
  • Avoiding Positively Charged Residues: Eliminating positively charged residues in the CDRs, which are often selected for in affinity maturation due to avidity effects, proved effective in reducing non-specific binding and potentially mitigating strong avidity effects during sorting.
  • Near-Neutral Fv Charge for Specificity: The study suggests that antibodies with near-neutral or weakly positively charged variable fragment (Fv) regions are optimal for achieving high antibody specificity, a finding that contrasts with some other Aβ antibodies that have highly charged Fvs.
  • In Vivo Relevance: The developed antibodies not only performed well in vitro but also demonstrated specific recognition of Aβ aggregates in mouse models and human brain tissue, validating their potential for in vivo applications.

Key Results

  • Improved Affinity: The affinity-matured antibodies showed a significant increase in binding affinity for Aβ fibrils, with EC50 values in the 4-13 nM range, comparable to or better than clinical-stage antibodies like crenezumab (9 ± 1 nM) and aducanumab (3 ± 0.2 nM).
  • High Conformational Specificity: The new antibodies retained high specificity for Aβ aggregates, similar to aducanumab and their parental antibody (AF1), and unlike crenezumab, which binds to both aggregated and disaggregated forms.
  • Extremely Low Non-Specific Binding: The affinity-matured antibodies exhibited very low levels of non-specific interactions, comparable to or better than crenezumab, and significantly lower than aducanumab.
  • In Vivo Recognition: The antibodies successfully detected Aβ aggregates in transgenic mouse brains and human Alzheimer's patient brain tissue.
  • Favorable Biophysical Properties: The antibodies maintained high monomeric protein content and favorable thermal stability, comparable to clinical-stage antibodies.

Limitations and Future Work

  • Avidity Effects: While the strategy mitigated avidity effects, they remain a challenge in selecting antibodies against multivalent aggregates. The authors note that the selection of mutations was based on sets of six, but other residues at non-mutated sites also contributed to binding, suggesting a more complex interplay than initially modeled.
  • Deep Sequencing Analysis: The deep sequencing analysis was a "scratch on the surface," and errors in initial evaluation meant the reported variants might not represent the absolute best possible mutations. This suggests further optimization is possible.
  • Exposure Time in Immunoblots: The authors caution that conformational specificity observed in immunoblots can be dependent on exposure time, making interpretation complex.
  • Future Directions: The authors propose applying this methodology to further increase affinities, evolve conformational specificity for other aggregate types (like prefibrillar oligomers), and potentially combine these antibodies with strategies to enhance blood-brain barrier transport for improved therapeutic efficacy and safety. They are also testing the generality of these methods for antibodies against other amyloidogenic proteins.

Summary updated

Abstract

The aggregation of amyloidogenic polypeptides is strongly linked to several neurodegenerative disorders, including Alzheimer's and Parkinson's diseases. Conformational antibodies that selectively recognize protein aggregates are leading therapeutic agents for selectively neutralizing toxic aggregates, diagnostic and imaging agents for detecting disease, and biomedical reagents for elucidating disease mechanisms. Despite their importance, it is challenging to generate high-quality conformational antibodies in a systematic and site-specific manner due to the properties of protein aggregates (hydrophobic, multivalent, and heterogeneous) and limitations of immunization (uncontrolled antigen presentation and immunodominant epitopes). Toward addressing these challenges, we have developed a systematic directed evolution procedure for affinity maturing antibodies against Alzheimer's Aβ fibrils and selecting variants with strict conformational and sequence specificity. We first designed a library based on a lead conformational antibody by sampling combinations of amino acids in the antigen-binding site predicted to mediate high antibody specificity. Next, we displayed this library on the surface of yeast, sorted it against Aβ42 aggregates, and identified promising clones using deep sequencing. The resulting antibodies displayed similar or higher affinities than clinical-stage Aβ antibodies (aducanumab and crenezumab). Moreover, the affinity-matured antibodies retained high conformational specificity for Aβ aggregates, as observed for aducanumab and unlike crenezumab. Notably, the affinity-maturated antibodies displayed extremely low levels of nonspecific interactions, as observed for crenezumab and unlike aducanumab. We expect that our systematic methods for generating antibodies with unique combinations of desirable properties will improve the generation of high-quality conformational antibodies specific for diverse types of aggregated conformers.
BibTeX
@article{desai2021rational,
  author = {Alec A. Desai and Matthew D. Smith and Yulei Zhang and Emily K. Makowski and Julia E. Gerson and Edward Ionescu and Charles G. Starr and Jennifer M. Zupancic and Shannon J. Moore and Alexandra Sutter and Magdalena I. Ivanova and Geoffrey G. Murphy and Henry L. Paulson and Peter M. Tessier},
  title = {{Rational affinity maturation of anti-amyloid antibodies with high conformational and sequence specificity}},
  journal = {Journal of Biological Chemistry},
  year = {2021},
  doi = {10.1016/j.jbc.2021.100508},
}
Figure