Improved prediction of antibody VL-VH orientation

AlphaFold2 in biomedical research: facilitating the development of diagnostic strategies for disease

Proteins, as the primary executors of physiological activity, serve as a key factor in disease diagnosis and treatment. Research into their structures, functions, and interactions is essential to better understand disease mechanisms and potential therapies. DeepMind's AlphaFold2, a deep-learning protein structure prediction model, has proven to be remarkably accurate, and it is widely employed in various aspects of diagnostic research, such as the study of disease biomarkers, microorganism pathogenicity, antigen-antibody structures, and missense mutations. Thus, AlphaFold2 serves as an exceptional tool to bridge fundamental protein research with breakthroughs in disease diagnosis, developments in diagnostic strategies, and the design of novel therapeutic approaches and enhancements in precision medicine. This review outlines the architecture, highlights, and limitations of AlphaFold2, placing particular emphasis on its applications within diagnostic research grounded in disciplines such as immunology, biochemistry, molecular biology, and microbiology.

Protein loop structure prediction by community-based deep learning and its application to antibody CDR H3 loop modeling

As of now, more than 60 years have passed since the first determination of protein structures through crystallography, and a significant portion of protein structures can be predicted by computers. This is due to the groundbreaking enhancement in protein structure prediction achieved through neural network training utilizing extensive sequence and structure data. However, substantial challenges persist in structure prediction due to limited data availability, with antibody structure prediction standing as one such challenge. In this paper, we propose a novel neural network architecture that effectively enables structure prediction by reflecting the inherent combinatorial nature involved in protein structure formation. The core idea of this neural network architecture is not solely to track and generate a single structure but rather to form a community of multiple structures and pursue accurate structure prediction by exchanging information among community members. Applying this concept to antibody CDR H3 loop structure prediction resulted in improved structure sampling. Such an approach could be applied in the structural and functional studies of proteins, particularly in exploring various physiological processes mediated by loops. Moreover, it holds potential in addressing various other types of combinatorial structure prediction and design problems.

Accurate prediction of antibody function and structure using bio-inspired antibody language model

In recent decades, antibodies have emerged as indispensable therapeutics for combating diseases, particularly viral infections. However, their development has been hindered by limited structural information and labor-intensive engineering processes. Fortunately, significant advancements in deep learning methods have facilitated the precise prediction of protein structure and function by leveraging co-evolution information from homologous proteins. Despite these advances, predicting the conformation of antibodies remains challenging due to their unique evolution and the high flexibility of their antigen-binding regions. Here, to address this challenge, we present the Bio-inspired Antibody Language Model (BALM). This model is trained on a vast dataset comprising 336 million 40% nonredundant unlabeled antibody sequences, capturing both unique and conserved properties specific to antibodies. Notably, BALM showcases exceptional performance across four antigen-binding prediction tasks. Moreover, we introduce BALMFold, an end-to-end method derived from BALM, capable of swiftly predicting full atomic antibody structures from individual sequences. Remarkably, BALMFold outperforms those well-established methods like AlphaFold2, IgFold, ESMFold and OmegaFold in the antibody benchmark, demonstrating significant potential to advance innovative engineering and streamline therapeutic antibody development by reducing the need for unnecessary trials. The BALMFold structure prediction server is freely available at https://beamlab-sh.com/models/BALMFold.

A novel Fc-enhanced humanized monoclonal antibody targeting B7-H3 suppresses the growth of ESCC

Esophageal squamous cell carcinoma (ESCC) is a prevalent malignant tumor of the digestive tract with a low 5-year survival rate due to the lack of effective treatment methods. Although therapeutic monoclonal antibodies (mAbs) now play an important role in cancer therapy, effective targeted mAbs are still lacking for ESCC. B7-H3 is highly expressed in a variety of tumors and has emerged as a promising therapeutic target. Several mAbs against B7-H3 have advanced to clinical trials, but their development has not yet been pursued for ESCC. Here, we developed a humanized and Fc-engineered anti-B7H3 mAb 24F-Hu-mut2 and systematically evaluated its anti-tumor activity in vitro and in vivo. The 24F-Hu-mut2 was humanized and modified in Fc fragment to obtain stronger antibody-dependent cell-mediated cytotoxicity(ADCC) activity and nanomolar affinity. Furthermore, both of ESCC cell-derived xenograft (CDX) and patient-derived xenograft (PDX) mice models indicated that 24F-Hu-mut2 displayed potent in vivo anti-tumor activity. In addition, a computational docking model showed that the mAb bound to IgC1 and IgC2 domain of B7-H3, which is closer to the cell membrane. Consistently, our ELISA results verified the binding of 24F-Hu-WT and IgC1 and IgC2. Our results indicate that 24F-Hu-mut2 has significant anti-ESCC activity both in vitro and in vivo, and this monoclonal antibody may be a promising antibody against ESCC and other B7-H3 overexpressing tumors.

Polyreactivity of antibodies from different B-cell subpopulations is determined by distinct sequence patterns of variable region

An antibody molecule that can bind to multiple distinct antigens is defined as polyreactive. In the present study, we performed statistical analyses to assess sequence correlates of polyreactivity of >600 antibodies cloned from different B-cell types of healthy humans. The data revealed several sequence patterns of variable regions of heavy and light immunoglobulin chains that determine polyreactivity. The most prominent identified patterns were increased number of basic amino acid residues, reduced frequency of acidic residues, increased number of aromatic and hydrophobic residues, and longer length of CDR L1. Importantly, our study revealed that antibodies isolated from different B-cell populations used distinct sequence patterns (or combinations of them) for polyreactive antigen binding. Furthermore, we combined the data from sequence analyses with molecular modeling of selected polyreactive antibodies and demonstrated that human antibodies can use multiple pathways for achieving antigen-binding promiscuity. These data reconcile some contradictions in the literature regarding the determinants of antibody polyreactivity. Moreover, our study demonstrates that the mechanism of polyreactivity of antibodies evolves during immune response and might be tailored to specific functional properties of different B-cell compartments. Finally, these data can be of use for efforts in the development and engineering of therapeutic antibodies.

Structure and Dynamics Guiding Design of Antibody Therapeutics and Vaccines

Antibodies and other new antibody-like formats have emerged as one of the most rapidly growing classes of biotherapeutic proteins. Understanding the structural features that drive antibody function and, consequently, their molecular recognition is critical for engineering antibodies. Here, we present the structural architecture of conventional IgG antibodies alongside other formats. We emphasize the importance of considering antibodies as conformational ensembles in solution instead of focusing on single-static structures because their functions and properties are strongly governed by their dynamic nature. Thus, in this review, we provide an overview of the unique structural and dynamic characteristics of antibodies with respect to their antigen recognition, biophysical properties, and effector functions. We highlight the numerous technical advances in antibody structure prediction and design, enabled by the vast number of experimentally determined high-quality structures recorded with cryo-EM, NMR, and X-ray crystallography. Lastly, we assess antibody and vaccine design strategies in the context of structure and dynamics.

Structural modeling of antibody variable regions using deep learning-progress and perspectives on drug discovery

AlphaFold2 has hallmarked a generational improvement in protein structure prediction. In particular, advances in antibody structure prediction have provided a highly translatable impact on drug discovery. Though AlphaFold2 laid the groundwork for all proteins, antibody-specific applications require adjustments tailored to these molecules, which has resulted in a handful of deep learning antibody structure predictors. Herein, we review the recent advances in antibody structure prediction and relate them to their role in advancing biologics discovery.

Fast, accurate antibody structure prediction from deep learning on massive set of natural antibodies

Antibodies have the capacity to bind a diverse set of antigens, and they have become critical therapeutics and diagnostic molecules. The binding of antibodies is facilitated by a set of six hypervariable loops that are diversified through genetic recombination and mutation. Even with recent advances, accurate structural prediction of these loops remains a challenge. Here, we present IgFold, a fast deep learning method for antibody structure prediction. IgFold consists of a pre-trained language model trained on 558 million natural antibody sequences followed by graph networks that directly predict backbone atom coordinates. IgFold predicts structures of similar or better quality than alternative methods (including AlphaFold) in significantly less time (under 25 s). Accurate structure prediction on this timescale makes possible avenues of investigation that were previously infeasible. As a demonstration of IgFold's capabilities, we predicted structures for 1.4 million paired antibody sequences, providing structural insights to 500-fold more antibodies than have experimentally determined structures.

Structural insights into the broad protection against H1 influenza viruses by a computationally optimized hemagglutinin vaccine

Influenza virus poses an ongoing human health threat with pandemic potential. Due to mutations in circulating strains, formulating effective vaccines remains a challenge. The use of computationally optimized broadly reactive antigen (COBRA) hemagglutinin (HA) proteins is a promising vaccine strategy to protect against a wide range of current and future influenza viruses. Though effective in preclinical studies, the mechanistic basis driving the broad reactivity of COBRA proteins remains to be elucidated. Here, we report the crystal structure of the COBRA HA termed P1 and identify antigenic and glycosylation properties that contribute to its immunogenicity. We further report the cryo-EM structure of the P1-elicited broadly neutralizing antibody 1F8 bound to COBRA P1, revealing 1F8 to recognize an atypical receptor binding site epitope via an unexpected mode of binding.

Prediction of the structural interface between fibroblast growth factor23 and Burosumab using alanine scanning and molecular docking

Burosumab, an FGF23 targeting monoclonal antibody, was approved by the FDA in 2018 for use in children and adults with X-linked hypophosphatemia (or XLH). While several clinical studies have demonstrated the long-term safety and efficacy of Burosumab, the molecular basis of FGF23-Burosumab interaction which underpins its mechanism of action remains unknown. In this study, we employed molecular docking combined with alanine scanning of epitope and paratope to predict a model of FGF23-Burosumab interaction. Then, we used the model to understand the species-species cross-reactivity of Burosumab and to reverse engineer mouse FGF23 with 'back to human' mutations to bind Burosumab. Finally, we redesigned the CDRs with two mutations to engineer an affinity enhanced variant of the antibody. Our study provides insights into the FGF23-Burosumab interaction and demonstrates that alanine-scanning coupled with molecular docking can be used to optimize antibody candidates (e.g., structure-guided affinity maturation) for therapeutic use.

NanoNet: Rapid and accurate end-to-end nanobody modeling by deep learning

Antibodies are a rapidly growing class of therapeutics. Recently, single domain camelid VHH antibodies, and their recognition nanobody domain (Nb) appeared as a cost-effective highly stable alternative to full-length antibodies. There is a growing need for high-throughput epitope mapping based on accurate structural modeling of the variable domains that share a common fold and differ in the Complementarity Determining Regions (CDRs). We develop a deep learning end-to-end model, NanoNet, that given a sequence directly produces the 3D coordinates of the backbone and Cβ atoms of the entire VH domain. For the Nb test set, NanoNet achieves 3.16Å average RMSD for the most variable CDR3 loops and 2.65Å, 1.73Å for the CDR1, CDR2 loops, respectively. The accuracy for antibody VH domains is even higher: 2.38Å RMSD for CDR3 and 0.89Å, 0.96Å for the CDR1, CDR2 loops, respectively. NanoNet run times allow generation of ∼1M nanobody structures in less than 4 hours on a standard CPU computer enabling high-throughput structure modeling. NanoNet is available at GitHub: https://github.com/dina-lab3D/NanoNet

Antibody structure prediction using interpretable deep learning

Therapeutic antibodies make up a rapidly growing segment of the biologics market. However, rational design of antibodies is hindered by reliance on experimental methods for determining antibody structures. Here, we present DeepAb, a deep learning method for predicting accurate antibody F_V structures from sequence. We evaluate DeepAb on a set of structurally diverse, therapeutically relevant antibodies and find that our method consistently outperforms the leading alternatives. Previous deep learning methods have operated as “black boxes” and offered few insights into their predictions. By introducing a directly interpretable attention mechanism, we show our network attends to physically important residue pairs (e.g., proximal aromatics and key hydrogen bonding interactions). Finally, we present a novel mutant scoring metric derived from network confidence and show that for a particular antibody, all eight of the top-ranked mutations improve binding affinity. This model will be useful for a broad range of antibody prediction and design tasks.

•

DeepAb, a deep learning method for antibody structure, is presented

•

Structures from DeepAb are more accurate than alternatives

•

Outputs of DeepAb provide interpretable insights into structure predictions

•

DeepAb predictions should facilitate design of novel antibody therapeutics

Accurate structure models are critical for understanding the properties of potential therapeutic antibodies. Conventional methods for protein structure determination require significant investments of time and resources and may fail. Although greatly improved, methods for general protein structure prediction still cannot consistently provide the accuracy necessary to understand or design antibodies. We present a deep learning method for antibody structure prediction and demonstrate improvement over alternatives on diverse, therapeutically relevant benchmarks. In addition to its improved accuracy, our method reveals interpretable outputs about specific amino acids and residue interactions that should facilitate design of novel therapeutic antibodies.

Spike-Dependent Opsonization Indicates Both Dose-Dependent Inhibition of Phagocytosis and That Non-Neutralizing Antibodies Can Confer Protection to SARS-CoV-2

Spike-specific antibodies are central to effective COVID19 immunity. Research efforts have focused on antibodies that neutralize the ACE2-Spike interaction but not on non-neutralizing antibodies. Antibody-dependent phagocytosis is an immune mechanism enhanced by opsonization, where typically, more bound antibodies trigger a stronger phagocyte response. Here, we show that Spike-specific antibodies, dependent on concentration, can either enhance or reduce Spike-bead phagocytosis by monocytes independently of the antibody neutralization potential. Surprisingly, we find that both convalescent patient plasma and patient-derived monoclonal antibodies lead to maximum opsonization already at low levels of bound antibodies and is reduced as antibody binding to Spike protein increases. Moreover, we show that this Spike-dependent modulation of opsonization correlate with the outcome in an experimental SARS-CoV-2 infection model. These results suggest that the levels of anti-Spike antibodies could influence monocyte-mediated immune functions and propose that non-neutralizing antibodies could confer protection to SARS-CoV-2 infection by mediating phagocytosis.

An affinity threshold for maximum efficacy in anti-PD-1 immunotherapy

Monoclonal antibodies targeting the programmed cell death protein 1 (PD-1) remain the most prevalent cancer immunotherapy both as a monotherapy and in combination with additional therapies. Despite the extensive success of anti-PD-1 monoclonal antibodies in the clinic, the experimental relationship between binding affinity and functional potency for anti-PD-1 antibodies in vivo has not been reported. Anti-PD-1 antibodies with higher and lower affinity than nivolumab or pembrolizumab are entering the clinic and show varied preclinical efficacy. Here, we explore the role of broad-ranging affinity variation within a single lineage in a syngeneic immunocompetent mouse model. By developing a panel of murine anti-PD-1 antibodies with varying affinity (ranging from KD = 20 pM - 15 nM), we find that there is a threshold affinity required for maximum efficacy at a given dose in the treatment of the MC38 adenocarcinoma model with anti-PD-1 immunotherapy. Physiologically based pharmacokinetic modeling complements interpretation of the experimental results and highlights the direct relationship between dose, affinity, and PD-1 target saturation in the tumor.

A high-affinity human TCR-like antibody detects celiac disease gluten peptide-MHC complexes and inhibits T cell activation

Antibodies specific for peptides bound to human leukocyte antigen (HLA) molecules are valuable tools for studies of antigen presentation and may have therapeutic potential. Here, we generated human T cell receptor (TCR)-like antibodies toward the immunodominant signature gluten epitope DQ2.5-glia-α2 in celiac disease (CeD). Phage display selection combined with secondary targeted engineering was used to obtain highly specific antibodies with picomolar affinity. The crystal structure of a Fab fragment of the lead antibody 3.C11 in complex with HLA-DQ2.5:DQ2.5-glia-α2 revealed a binding geometry and interaction mode highly similar to prototypic TCRs specific for the same complex. Assessment of CeD biopsy material confirmed disease specificity and reinforced the notion that abundant plasma cells present antigen in the inflamed CeD gut. Furthermore, 3.C11 specifically inhibited activation and proliferation of gluten-specific CD4⁺ T cells in vitro and in HLA-DQ2.5 humanized mice, suggesting a potential for targeted intervention without compromising systemic immunity.

Robustification of RosettaAntibody and Rosetta SnugDock

In recent years, the observed antibody sequence space has grown exponentially due to advances in high-throughput sequencing of immune receptors. The rise in sequences has not been mirrored by a rise in structures, as experimental structure determination techniques have remained low-throughput. Computational modeling, however, has the potential to close the sequence–structure gap. To achieve this goal, computational methods must be robust, fast, easy to use, and accurate. Here we report on the latest advances made in RosettaAntibody and Rosetta SnugDock—methods for antibody structure prediction and antibody–antigen docking. We simplified the user interface, expanded and automated the template database, generalized the kinematics of antibody–antigen docking (which enabled modeling of single-domain antibodies) and incorporated new loop modeling techniques. To evaluate the effects of our updates on modeling accuracy, we developed rigorous tests under a new scientific benchmarking framework within Rosetta. Benchmarking revealed that more structurally similar templates could be identified in the updated database and that SnugDock broadened its applicability without losing accuracy. However, there are further advances to be made, including increasing the accuracy and speed of CDR-H3 loop modeling, before computational approaches can accurately model any antibody.

Modeling Immunity with Rosetta: Methods for Antibody and Antigen Design

Structure-based antibody and antigen design has advanced greatly in recent years, due not only to the increasing availability of experimentally determined structures but also to improved computational methods for both prediction and design. Constant improvements in performance within the Rosetta software suite for biomolecular modeling have given rise to a greater breadth of structure prediction, including docking and design application cases for antibody and antigen modeling. Here, we present an overview of current protocols for antibody and antigen modeling using Rosetta and exemplify those by detailed tutorials originally developed for a Rosetta workshop at Vanderbilt University. These tutorials cover antibody structure prediction, docking, and design and antigen design strategies, including the addition of glycans in Rosetta. We expect that these materials will allow novice users to apply Rosetta in their own projects for modeling antibodies and antigens.

Interaction of clinical-stage antibodies with heme predicts their physiochemical and binding qualities

Immunoglobulin repertoires contain a fraction of antibodies that recognize low molecular weight compounds, including some enzymes' cofactors, such as heme. Here, by using a set of 113 samples with variable region sequences matching clinical-stage antibodies, we demonstrated that a considerable number of these antibodies interact with heme. Antibodies that interact with heme possess specific sequence traits of their antigen-binding regions. Moreover they manifest particular physicochemical and functional qualities i.e. increased hydrophobicity, higher propensity of self-binding, higher intrinsic polyreactivity and reduced expression yields. Thus, interaction with heme is a strong predictor of different molecular and functional qualities of antibodies. Notably, these qualities are of high importance for therapeutic antibodies, as their presence was associated with failure of drug candidates to reach clinic. Our study reveled an important facet of information about relationship sequence-function in antibodies. It also offers a convenient tool for detection of liabilities of therapeutic antibodies.

Geometric potentials from deep learning improve prediction of CDR H3 loop structures

Motivation

Antibody structure is largely conserved, except for a complementarity-determining region featuring six variable loops. Five of these loops adopt canonical folds which can typically be predicted with existing methods, while the remaining loop (CDR H3) remains a challenge due to its highly diverse set of observed conformations. In recent years, deep neural networks have proven to be effective at capturing the complex patterns of protein structure. This work proposes DeepH3, a deep residual neural network that learns to predict inter-residue distances and orientations from antibody heavy and light chain sequence. The output of DeepH3 is a set of probability distributions over distances and orientation angles between pairs of residues. These distributions are converted to geometric potentials and used to discriminate between decoy structures produced by RosettaAntibody and predict new CDR H3 loop structures de novo.

Results

When evaluated on the Rosetta antibody benchmark dataset of 49 targets, DeepH3-predicted potentials identified better, same and worse structures [measured by root-mean-squared distance (RMSD) from the experimental CDR H3 loop structure] than the standard Rosetta energy function for 33, 6 and 10 targets, respectively, and improved the average RMSD of predictions by 32.1% (1.4 Å). Analysis of individual geometric potentials revealed that inter-residue orientations were more effective than inter-residue distances for discriminating near-native CDR H3 loops. When applied to de novo prediction of CDR H3 loop structures, DeepH3 achieves an average RMSD of 2.2 ± 1.1 Å on the Rosetta antibody benchmark.

Availability and Implementation

DeepH3 source code and pre-trained model parameters are freely available at https://github.com/Graylab/deepH3-distances-orientations.

Supplementary information

Supplementary data are available at Bioinformatics online.

Conformational Ensembles of Antibodies Determine Their Hydrophobicity

A major challenge in the development of antibody biotherapeutics is their tendency to aggregate. One root cause for aggregation is exposure of hydrophobic surface regions to the solvent. Many current techniques predict the relative aggregation propensity of antibodies via precalculated scales for the hydrophobicity or aggregation propensity of single amino acids. However, those scales cannot describe the nonadditive effects of a residue’s surrounding on its hydrophobicity. Therefore, they are inherently limited in their ability to describe the impact of subtle differences in molecular structure on the overall hydrophobicity. Here, we introduce a physics-based approach to describe hydrophobicity in terms of the hydration free energy using grid inhomogeneous solvation theory (GIST). We apply this method to assess the effects of starting structures, conformational sampling, and protonation states on the hydrophobicity of antibodies. Our results reveal that high-quality starting structures, i.e., crystal structures, are crucial for the prediction of hydrophobicity and that conformational sampling can compensate errors introduced by the starting structure. On the other hand, sampling of protonation states only leads to good results when combined with high-quality structures, whereas it can even be detrimental otherwise. We conclude by pointing out that a single static homology model may not be adequate for predicting hydrophobicity.

Novel murine mAbs define specific and cross-reactive epitopes on the latex profilin panallergen Hev b 8

The production of specific antibodies able to recognize allergens from different sources or block interactions between allergens and antibodies mediating allergic reactions is crucial for developing successful tools for diagnostics and therapeutics. Panallergens are highly conserved proteins present in widely different species, implicated in relevant cross-reactions. The panallergen latex profilin (Hev b 8) has been associated with the latex-food-pollen syndrome. We generated five monoclonal IgGs and one IgE from murine hybridomas against recombinant Hev b 8 and evaluated their interaction with this allergen using ELISA and biolayer interferometry (BLI). Affinity purified mAbs exhibited high binding affinities towards rHev b 8, with K_D1 values ranging from 10^-10 M to 10^-11 M. Some of these antibodies also recognized the recombinant profilins from maize and tomato (Zea m 12 and Sola l 1), and the ash tree pollen (Fra e 2). Competition ELISA demonstrated that some mAb pairs could bind simultaneously to rHev b 8. Using BLI, we detected competitive, non-competitive, and partial-competition interactions between pairs of mAbs with rHev b 8, suggesting the existence of at least two non-overlapping epitopes on the surface of this allergen. Three-dimensional models of the Fv of 1B4 and 2D10 IgGs and docking simulations of these Fvs with rHev b 8 revealed these epitopes. Furthermore, these two mAbs inhibited the interaction of polyclonal IgE and IgG4 antibodies from profilin-allergic patients with rHev b 8, indicating that the mAbs and the antibodies present in sera from allergic patients bind to overlapping epitopes on the allergen. These mAbs can be useful tools for immune-localization studies, immunoassay development, or standardization of allergenic products.

Computational approaches to therapeutic antibody design: established methods and emerging trends

Antibodies are proteins that recognize the molecular surfaces of potentially noxious molecules to mount an adaptive immune response or, in the case of autoimmune diseases, molecules that are part of healthy cells and tissues. Due to their binding versatility, antibodies are currently the largest class of biotherapeutics, with five monoclonal antibodies ranked in the top 10 blockbuster drugs. Computational advances in protein modelling and design can have a tangible impact on antibody-based therapeutic development. Antibody-specific computational protocols currently benefit from an increasing volume of data provided by next generation sequencing and application to related drug modalities based on traditional antibodies, such as nanobodies. Here we present a structured overview of available databases, methods and emerging trends in computational antibody analysis and contextualize them towards the engineering of candidate antibody therapeutics.

VH -VL interdomain dynamics observed by computer simulations and NMR

The relative orientation of the two variable domains, VH and VL , influences the shape of the antigen binding site, that is, the paratope, and is essential to understand antigen specificity. ABangle characterizes the VH -VL orientation by using five angles and a distance and compares it to other known structures. Molecular dynamics simulations of antibody variable domains (Fvs) reveal fluctuations in the relative domain orientations. The observed dynamics between these domains are confirmed by NMR experiments on a single-chain variable fragment antibody (scFv) in complex with IL-1β and an antigen-binding fragment (Fab). The variability of these relative domain orientations can be interpreted as a structural feature of antibodies, which increases the antibody repertoire significantly and can enlarge the number of possible binding partners substantially. The movements of the VH and VL domains are well sampled with molecular dynamics simulations and are in agreement with the NMR ensemble. Fast Fourier transformation of the ABangle metrics allows to assign timescales of 0.1-10 GHz to the fastest collective interdomain movements. The results clearly show the necessity of dynamics to understand and characterize the favorable orientations of the VH and VL domains implying a considerable binding interface flexibility and reveal in all antibody fragments (Fab, scFv, and Fv) very similar VH -VL interdomain variations comparable to the distributions observed for known X-ray structures of antibodies. SIGNIFICANCE STATEMENT: Antibodies have become key players as therapeutic agents. The binding ability of antibodies is determined by the antigen-binding fragment (Fab), in particular the variable fragment region (Fv). Antigen-binding is mediated by the complementarity-determining regions consisting of six loops, each three of the heavy and light chain variable domain VH and VL . The relative orientation of the VH and VL domains influences the shape of the antigen-binding site and is a major objective in antibody design. In agreement with NMR experiments and molecular dynamics simulations, we show a considerable binding site flexibility in the low nanosecond timescale. Thus we suggest that this flexibility and its implications for binding and specificity should be considered when designing and optimizing therapeutic antibodies.

Macromolecular modeling and design in Rosetta: recent methods and frameworks

The Rosetta software for macromolecular modeling, docking and design is extensively used in laboratories worldwide. During two decades of development by a community of laboratories at more than 60 institutions, Rosetta has been continuously refactored and extended. Its advantages are its performance and interoperability between broad modeling capabilities. Here we review tools developed in the last 5 years, including over 80 methods. We discuss improvements to the score function, user interfaces and usability. Rosetta is available at http://www.rosettacommons.org.

Antibody humanization-the Influence of the antibody framework on the CDR-H3 loop ensemble in solution

Antibody engineering of non-human antibodies has focused on reducing immunogenicity by humanization, being a major limitation in developing monoclonal antibodies. We analyzed four series of antibody binding fragments (Fabs) and a variable fragment (Fv) with structural information in different stages of humanization to investigate the influence of the framework, point mutations and specificity on the complementarity determining region (CDR)-H3 loop dynamics. We also studied a Fv without structural information of the anti-idiotypic antibody Ab2/3H6, because it completely lost its binding affinity upon superhumanization, as an example of a failed humanization. Enhanced sampling techniques in combination with molecular dynamics simulations allow to access micro- to milli-second timescales of the CDR-H3 loop dynamics and reveal kinetic and thermodynamic changes involved in the process of humanization. In most cases, we observe a reduced conformational diversity of the CDR-H3 loop when grafted on a human framework and find a conformational shift of the dominant CDR-H3 loop conformation in solution. A shallow side minimum of the conformational CDR-H3 loop ensemble attached to the murine framework becomes the dominant conformation in solution influenced by the human framework. Additionally, we observe in the case of the failed humanization that the potentially binding competent murine CDR-H3 loop ensemble in solution shows nearly no kinetical or structural overlap with the superhumanized variant, thus explaining the loss of binding.

High-Throughput Antibody Structure Modeling and Design Using ABodyBuilder

Antibodies are proteins of the adaptive immune system; they can be designed to bind almost any molecule, and are increasingly being used as biotherapeutics. Experimental antibody design is an expensive and time-consuming process, and computational antibody design methods can now be used to help develop new therapeutics and diagnostics. Within the design pipeline, accurate antibody structure modeling is essential, as it provides the basis for antibody-antigen docking, binding affinity prediction, and estimating thermal stability. Ideally, models should be rapidly generated, allowing the exploration of the breadth of antibody space. This allows methods to replicate the natural processes of antibody diversification (e.g., V(D)J recombination and somatic hypermutation), and cope with large volumes of data that are typical of next-generation sequencing datasets. Here we describe ABodyBuilder and PEARS, algorithms that build and mutate antibody model structures. These methods take ~30 s to generate a model antibody structure.

Plasma Cells Are the Most Abundant Gluten Peptide MHC-expressing Cells in Inflamed Intestinal Tissues From Patients With Celiac Disease

BACKGROUND & AIMS

Development of celiac disease is believed to involve the transglutaminase-dependent response of CD4⁺ T cells toward deamidated gluten peptides in the intestinal mucosa of individuals with specific HLA-DQ haplotypes. We investigated the antigen presentation process during this mucosal immune response.

METHODS

We generated monoclonal antibodies (mAbs) specific for the peptide-MHC (pMHC) complex of HLA-DQ2.5 and the immunodominant gluten epitope DQ2.5-glia-α1a using phage display. We used these mAbs to assess gluten peptide presentation and phenotypes of presenting cells by flow cytometry and enzyme-linked immune absorbent spot (ELISPOT) in freshly prepared single-cell suspensions from intestinal biopsies from 40 patients with celiac disease (35 untreated and 5 on a gluten-free diet) as well as 18 subjects with confirmed noninflamed gut mucosa (controls, 12 presumed healthy, 5 undergoing pancreatoduodenectomy, and 1 with potential celiac disease).

RESULTS

Using the mAbs, we detected MHC complexes on cells from intestinal biopsies from patients with celiac disease who consume gluten, but not from patients on gluten-free diets. We found B cells and plasma cells to be the most abundant cells that present DQ2.5-glia-α1a in the inflamed mucosa. We identified a subset of plasma cells that expresses B-cell receptors (BCR) specific for gluten peptides or the autoantigen transglutaminase 2 (TG2). Expression of MHC class II (MHCII) was not restricted to these specific plasma cells in patients with celiac disease but was observed in an average 30% of gut plasma cells from patients and controls.

CONCLUSIONS

A population of plasma cells from intestinal biopsies of patients with celiac disease express MHCII; this is the most abundant cell type presenting the immunodominant gluten peptide DQ2.5-glia-α1a in the tissues from these patients. These results indicate that plasma cells in the gut can function as antigen-presenting cells and might promote and maintain intestinal inflammation in patients with celiac disease or other inflammatory disorders.

High-resolution mapping of the polyclonal immune response to the human platelet alloantigen HPA-1a (PlA1)

Antibodies to platelet-specific antigens are responsible for 2 clinically important bleeding disorders: posttransfusion purpura and fetal/neonatal alloimmune thrombocytopenia (FNAIT). The human platelet-specific alloantigen 1a/1b (HPA-1a/1b; also known as PlA1/A2) alloantigen system of human platelet membrane glycoprotein (GP) IIIa is controlled by a Leu33Pro polymorphism and is responsible for ∼80% of the cases of FNAIT. Local residues surrounding polymorphic residue 33 are suspected to have a profound effect on alloantibody binding and subsequent downstream effector events. To define the molecular requirements for HPA-1a alloantibody binding, we generated transgenic mice that expressed murine GPIIIa (muGPIIIa) isoforms harboring select humanized residues within the plexin-semaphorin-integrin (PSI) and epidermal growth factor 1 (EGF1) domains and examined their ability to support the binding of a series of monoclonal and polyclonal HPA-1a-specific antibodies. Humanizing the PSI domain of muGPIIIa was sufficient to recreate the HPA-1a epitope recognized by some HPA-1a-specific antibodies; however, humanizing distinct amino acids within the linearly distant but conformationally close EGF1 domain was required to enable binding of others. These results reveal the previously unsuspected complex heterogeneity of the polyclonal alloimmune response to this clinically important human platelet alloantigen system. High-resolution mapping of this alloimmune response may improve diagnosis of FNAIT and should facilitate the rational design and selection of contemplated prophylactic and therapeutic anti-HPA-1a reagents.

Construction and Characterization of a Humanized Anti-Epstein-Barr Virus gp350 Antibody with Neutralizing Activity in Cell Culture

Acute Epstein-Barr virus (EBV) infection in immunosuppressed transplant patients can give rise to a malignant B-cell proliferation known as post-transplant lymphoproliferative disease (PTLD). The EBV major virion surface glycoprotein (gp)350 is a principal target of naturally occurring neutralizing antibodies and is viewed as the best target to prevent acute infection and PTLD in at-risk transplant recipients. We have constructed a humanized (hu) version of the murine anti-gp350 neutralizing monoclonal antibody 72a1. The hu72a1 IgG1 antibody displayed no significant anti-mouse activity, recognized both gp350 and its splice variant gp220 as well as a gp350 peptide that was shown to constitute the principal EBV gp350 neutralizing epitope when tested in immunoassays. Hu72a1 antibody blocked in vitro EBV infection of B cells at a level which equaled that of a mouse-human chimeric 72a1 antibody construct. This work provides a further structural and immunological understanding of the 72a1 antibody interaction with EBV gp350, and constitutes a launch point for future anti-EBV therapeutic antibodies designed to block EBV infection and prevent PTLD while eliminating the deleterious antigenic murine features of the original 72a1 antibody.

How B-Cell Receptor Repertoire Sequencing Can Be Enriched with Structural Antibody Data

Next-generation sequencing of immunoglobulin gene repertoires (Ig-seq) allows the investigation of large-scale antibody dynamics at a sequence level. However, structural information, a crucial descriptor of antibody binding capability, is not collected in Ig-seq protocols. Developing systematic relationships between the antibody sequence information gathered from Ig-seq and low-throughput techniques such as X-ray crystallography could radically improve our understanding of antibodies. The mapping of Ig-seq datasets to known antibody structures can indicate structurally, and perhaps functionally, uncharted areas. Furthermore, contrasting naïve and antigenically challenged datasets using structural antibody descriptors should provide insights into antibody maturation. As the number of antibody structures steadily increases and more and more Ig-seq datasets become available, the opportunities that arise from combining the two types of information increase as well. Here, we review how these data types enrich one another and show potential for advancing our knowledge of the immune system and improving antibody engineering.

Web-accessible molecular modeling with Rosetta: The Rosetta Online Server that Includes Everyone (ROSIE)

The Rosetta molecular modeling software package provides a large number of experimentally validated tools for modeling and designing proteins, nucleic acids, and other biopolymers, with new protocols being added continually. While freely available to academic users, external usage is limited by the need for expertise in the Unix command line environment. To make Rosetta protocols available to a wider audience, we previously created a web server called Rosetta Online Server that Includes Everyone (ROSIE), which provides a common environment for hosting web-accessible Rosetta protocols. Here we describe a simplification of the ROSIE protocol specification format, one that permits easier implementation of Rosetta protocols. Whereas the previous format required creating multiple separate files in different locations, the new format allows specification of the protocol in a single file. This new, simplified protocol specification has more than doubled the number of Rosetta protocols available under ROSIE. These new applications include pK_a determination, lipid accessibility calculation, ribonucleic acid redesign, protein-protein docking, protein-small molecule docking, symmetric docking, antibody docking, cyclic toxin docking, critical binding peptide determination, and mapping small molecule binding sites. ROSIE is freely available to academic users at http://rosie.rosettacommons.org.

Nonspecificity in a nonimmune human scFv repertoire

Efforts to develop effective antibody therapeutics are frequently hampered by issues such as aggregation and nonspecificity, often only detected in late stages of the development process. In this study, we used a high throughput cross-reactivity assay to select nonspecific clones from a naïve human repertoire scFv library displayed on the surface of yeast. Most antibody families were de-enriched; however, the rarely expressed VH6 family was highly enriched among nonspecific clones, representing almost 90% of isolated clones. Mutational analysis of this family reveals a dominant role of CDRH2 in driving nonspecific binding. Homology modeling of a panel of VH6 antibodies shows a constrained β-sheet structure in CDRH2 that is not present in other families, potentially contributing to nonspecificity of the family. These findings confirm the common decision to exclude VH6 from synthetic antibody libraries, and support VH6 polyreactivity as a possible important role for the family in early ontogeny and cause for its overabundance in cases of some forms of autoimmunity.

Modeling and docking of antibody structures with Rosetta

We describe Rosetta-based computational protocols for predicting the 3D structure of an antibody from sequence (RosettaAntibody) and then docking the antibody to protein antigens (SnugDock). Antibody modeling leverages canonical loop conformations to graft large segments from experimentally determined structures, as well as offering (i) energetic calculations to minimize loops, (ii) docking methodology to refine the V_L-V_H relative orientation and (iii) de novo prediction of the elusive complementarity determining region (CDR) H3 loop. To alleviate model uncertainty, antibody-antigen docking resamples CDR loop conformations and can use multiple models to represent an ensemble of conformations for the antibody, the antigen or both. These protocols can be run fully automated via the ROSIE web server (http://rosie.rosettacommons.org/) or manually on a computer with user control of individual steps. For best results, the protocol requires roughly 1,000 CPU-hours for antibody modeling and 250 CPU-hours for antibody-antigen docking. Tasks can be completed in under a day by using public supercomputers.

Accurate Structure Prediction of CDR H3 Loops Enabled by a Novel Structure-Based C-Terminal Constraint

Ab structure prediction has made great strides, but accurately modeling CDR H3 loops remains elusive. Unlike the other five CDR loops, CDR H3 does not adopt canonical conformations and must be modeled de novo. During Antibody Modeling Assessment II, we found that biasing simulations toward kinked conformations enables generating low-root mean square deviation models (Weitzner et al. 2014. Proteins 82: 1611-1623), and since then, we have presented new geometric parameters defining the kink conformation (Weitzner et al. 2015. Structure 23: 302-311). In this study, we use these parameters to develop a new biasing constraint. When applied to a benchmark set of high-quality CDR H3 loops, the average minimum root mean square deviation sampled is 0.93 Å, compared with 1.34 Å without the constraint. We then test the performance of the constrained de novo method for homology modeling and rigid-body docking and present the results for 1) the Antibody Modeling Assessment II targets, 2) the 2009 RosettaAntibody benchmark set, and 3) the high-quality set.