Antibody design using deep learning: from sequence and structure design to affinity maturation.

Deep learning has achieved impressive results in various fields such as computer vision and natural language processing, making it a powerful tool in biology. Its applications now encompass cellular image classification, genomic studies and drug discovery. While drug development traditionally focused deep learning applications on small molecules, recent innovations have incorporated it in the discovery and development of biological molecules, particularly antibodies. Researchers have devised novel techniques to streamline antibody development, combining in vitro and in silico methods. In particular, computational power expedites lead candidate generation, scaling and potential antibody development against complex antigens. This survey highlights significant advancements in protein design and optimization, specifically focusing on antibodies. This includes various aspects such as design, folding, antibody-antigen interactions docking and affinity maturation.

Development and experimental validation of computational methods for human antibody affinity enhancement.

High affinity is crucial for the efficacy and specificity of antibody. Due to involving high-throughput screens, biological experiments for antibody affinity maturation are time-consuming and have a low success rate. Precise computational-assisted antibody design promises to accelerate this process, but there is still a lack of effective computational methods capable of pinpointing beneficial mutations within the complementarity-determining region (CDR) of antibodies. Moreover, random mutations often lead to challenges in antibody expression and immunogenicity. In this study, to enhance the affinity of a human antibody against avian influenza virus, a CDR library was constructed and evolutionary information was acquired through sequence alignment to restrict the mutation positions and types. Concurrently, a statistical potential methodology was developed based on amino acid interactions between antibodies and antigens to calculate potential affinity-enhanced antibodies, which were further subjected to molecular dynamics simulations. Subsequently, experimental validation confirmed that a point mutation enhancing 2.5-fold affinity was obtained from 10 designs, resulting in the antibody affinity of 2 nM. A predictive model for antibody-antigen interactions based on the binding interface was also developed, achieving an Area Under the Curve (AUC) of 0.83 and a precision of 0.89 on the test set. Lastly, a novel approach involving combinations of affinity-enhancing mutations and an iterative mutation optimization scheme similar to the Monte Carlo method were proposed. This study presents computational methods that rapidly and accurately enhance antibody affinity, addressing issues related to antibody expression and immunogenicity.

AI-based IsAb2.0 for antibody design.

Therapeutic antibody design has garnered widespread attention, highlighting its interdisciplinary importance. Advancements in technology emphasize the critical role of designing nanobodies and humanized antibodies in antibody engineering. However, current experimental methods are costly and time-consuming. Computational approaches, while progressing, faced limitations due to insufficient structural data and the absence of a standardized protocol. To tackle these challenges, our lab previously developed IsAb1.0, an in silico antibody design protocol. Yet, IsAb1.0 lacked accuracy, had a complex procedure, and required extensive antibody bioinformation. Moreover, it overlooked nanobody and humanized antibody design, hindering therapeutic antibody development. Building upon IsAb1.0, we enhanced our design protocol with artificial intelligence methods to create IsAb2.0. IsAb2.0 utilized AlphaFold-Multimer (2.3/3.0) for accurate modeling and complex construction without templates and employed the precise FlexddG method for in silico antibody optimization. Validated through optimization of a humanized nanobody J3 (HuJ3) targeting HIV-1 gp120, IsAb2.0 predicted five mutations that can improve HuJ3-gp120 binding affinity. These predictions were confirmed by commercial software and validated through binding and neutralization assays. IsAb2.0 streamlined antibody design, offering insights into future techniques to accelerate immunotherapy development.

Accelerating antibody discovery and design with artificial intelligence: Recent advances and prospects.

Therapeutic antibodies are the largest class of biotherapeutics and have been successful in treating human diseases. However, the design and discovery of antibody drugs remains challenging and time-consuming. Recently, artificial intelligence technology has had an incredible impact on antibody design and discovery, resulting in significant advances in antibody discovery, optimization, and developability. This review summarizes major machine learning (ML) methods and their applications for computational predictors of antibody structure and antigen interface/interaction, as well as the evaluation of antibody developability. Additionally, this review addresses the current status of ML-based therapeutic antibodies under preclinical and clinical phases. While many challenges remain, ML may offer a new therapeutic option for the future direction of fully computational antibody design.

Sequence-based design and construction of synthetic nanobody library.

Nanobody (Nb), the smallest antibody fragments known to bind antigens, is now widely applied to various studies, including protein structure analysis, bioassay, diagnosis, and biomedicine. The traditional approach to generating specific nanobodies involves animal immunization which is time-consuming and expensive. As the understanding of the antibody repertoire accumulation, the synthetic library, which is devoid of animals, has attracted attention widely in recent years. Here, we describe a synthetic phage display library (S-Library), designed based on the systematic analysis of the next-generation sequencing (NGS) of nanobody repertoire. The library consists of a single highly conserved scaffold (IGHV3S65*01-IGHJ4*01) and complementary determining regions of constrained diversity. The S-Library containing 2.19 × 108 independent clones was constructed by the one-step assembly and rapid electro-transformation. The S-Library was screened against various targets (Nb G8, fusion protein of Nb G8 and green fluorescent protein, bovine serum albumin, ovalbumin, and acetylcholinesterase). In comparison, a naïve library (N-Library) from the source of 13 healthy animals was constructed and screened against the same targets as the S-Library. Binders were isolated from both S-Library and N-Library. The dynamic affinity was evaluated by the biolayer interferometry. The data confirms that the feature of the Nb repertoire is conducive to reducing the complexity of library design, thus allowing the S-Library to be built on conventional reagents and primers.

AI-driven antibody design with generative diffusion models: current insights and future directions.

Therapeutic antibodies are at the forefront of biotherapeutics, valued for their high target specificity and binding affinity. Despite their potential, optimizing antibodies for superior efficacy presents significant challenges in both monetary and time costs. Recent strides in computational and artificial intelligence (AI), especially generative diffusion models, have begun to address these challenges, offering novel approaches for antibody design. This review delves into specific diffusion-based generative methodologies tailored for antibody design tasks, de novo antibody design, and optimization of complementarity-determining region (CDR) loops, along with their evaluation metrics. We aim to provide an exhaustive overview of this burgeoning field, making it an essential resource for leveraging diffusion-based generative models in antibody design endeavors.

Prediction of Paratope-Epitope Pairs Using Convolutional Neural Networks.

Antibodies play a central role in the adaptive immune response of vertebrates through the specific recognition of exogenous or endogenous antigens. The rational design of antibodies has a wide range of biotechnological and medical applications, such as in disease diagnosis and treatment. However, there are currently no reliable methods for predicting the antibodies that recognize a specific antigen region (or epitope) and, conversely, epitopes that recognize the binding region of a given antibody (or paratope). To fill this gap, we developed ImaPEp, a machine learning-based tool for predicting the binding probability of paratope-epitope pairs, where the epitope and paratope patches were simplified into interacting two-dimensional patches, which were colored according to the values of selected features, and pixelated. The specific recognition of an epitope image by a paratope image was achieved by using a convolutional neural network-based model, which was trained on a set of two-dimensional paratope-epitope images derived from experimental structures of antibody-antigen complexes. Our method achieves good performances in terms of cross-validation with a balanced accuracy of 0.8. Finally, we showcase examples of application of ImaPep, including extensive screening of large libraries to identify paratope candidates that bind to a selected epitope, and rescoring and refining antibody-antigen docking poses.

Re-Engineering Therapeutic Anti-Aß Monoclonal Antibody to Target Amyloid Light Chain.

Amyloidosis involves the deposition of misfolded proteins. Even though it is caused by different pathogenic mechanisms, in aggregate, it shares similar features. Here, we tested and confirmed a hypothesis that an amyloid antibody can be engineered by a few mutations to target a different species. Amyloid light chain (AL) and ß-amyloid peptide (Aß) are two therapeutic targets that are implicated in amyloid light chain amyloidosis and Alzheimer's disease, respectively. Though crenezumab, an anti-Aß antibody, is currently unsuccessful, we chose it as a model to computationally design and prepare crenezumab variants, aiming to discover a novel antibody with high affinity to AL fibrils and to establish a technology platform for repurposing amyloid monoclonal antibodies. We successfully re-engineered crenezumab to bind both Aß42 oligomers and AL fibrils with high binding affinities. It is capable of reversing Aß42-oligomers-induced cytotoxicity, decreasing the formation of AL fibrils, and alleviating AL-fibrils-induced cytotoxicity in vitro. Our research demonstrated that an amyloid antibody could be engineered by a few mutations to bind new amyloid sequences, providing an efficient way to reposition a therapeutic antibody to target different amyloid diseases.

De novo design and Rosetta-based assessment of high-affinity antibody variable regions (Fv) against the SARS-CoV-2 spike receptor binding domain (RBD).

The continued emergence of new SARS-CoV-2 variants has accentuated the growing need for fast and reliable methods for the design of potentially neutralizing antibodies (Abs) to counter immune evasion by the virus. Here, we report on the de novo computational design of high-affinity Ab variable regions (Fv) through the recombination of VDJ genes targeting the most solvent-exposed hACE2-binding residues of the SARS-CoV-2 spike receptor binding domain (RBD) protein using the software tool OptMAVEn-2.0. Subsequently, we carried out computational affinity maturation of the designed variable regions through amino acid substitutions for improved binding with the target epitope. Immunogenicity of designs was restricted by preferring designs that match sequences from a 9-mer library of "human Abs" based on a human string content score. We generated 106 different antibody designs and reported in detail on the top five that trade-off the greatest computational binding affinity for the RBD with human string content scores. We further describe computational evaluation of the top five designs produced by OptMAVEn-2.0 using a Rosetta-based approach. We used Rosetta SnugDock for local docking of the designs to evaluate their potential to bind the spike RBD and performed "forward folding" with DeepAb to assess their potential to fold into the designed structures. Ultimately, our results identified one designed Ab variable region, P1.D1, as a particularly promising candidate for experimental testing. This effort puts forth a computational workflow for the de novo design and evaluation of Abs that can quickly be adapted to target spike epitopes of emerging SARS-CoV-2 variants or other antigenic targets.

IsAb: a computational protocol for antibody design.

The design of therapeutic antibodies has attracted a large amount of attention over the years. Antibodies are widely used to treat many diseases due to their high efficiency and low risk of adverse events. However, the experimental methods of antibody design are time-consuming and expensive. Although computational antibody design techniques have had significant advances in the past years, there are still some challenges that need to be solved, such as the flexibility of antigen structure, the lack of antibody structural data and the absence of standard antibody design protocol. In the present work, we elaborated on an in silico antibody design protocol for users to easily perform computer-aided antibody design. First, the Rosetta web server will be applied to generate the 3D structure of query antibodies if there is no structural information available. Then, two-step docking will be used to identify the binding pose of an antibody-antigen complex when the binding information is unknown. ClusPro is the first method to be used to conduct the global docking, and SnugDock is applied for the local docking. Sequentially, based on the predicted binding poses, in silico alanine scanning will be used to predict the potential hotspots (or key residues). Finally, computational affinity maturation protocol will be used to modify the structure of antibodies to theoretically increase their affinity and stability, which will be further validated by the bioassays in the future. As a proof of concept, we redesigned antibody D44.1 and compared it with previously reported data in order to validate IsAb protocol. To further illustrate our proposed protocol, we used cemiplimab antibody, a PD-1 checkpoint inhibitor, as an example to showcase a step-by-step tutorial.

Structural mechanism of Fab domain dissociation as a measure of interface stability.

Therapeutic antibodies should not only recognize antigens specifically, but also need to be free from developability issues, such as poor stability. Thus, the mechanistic understanding and characterization of stability are critical determinants for rational antibody design. In this study, we use molecular dynamics simulations to investigate the melting process of 16 antigen binding fragments (Fabs). We describe the Fab dissociation mechanisms, showing a separation in the VH-VL and in the CH1-CL domains. We found that the depths of the minima in the free energy curve, corresponding to the bound states, correlate with the experimentally determined melting temperatures. Additionally, we provide a detailed structural description of the dissociation mechanism and identify key interactions in the CDR loops and in the CH1-CL interface that contribute to stabilization. The dissociation of the VH-VL or CH1-CL domains can be represented by conformational changes in the bend angles between the domains. Our findings elucidate the melting process of antigen binding fragments and highlight critical residues in both the variable and constant domains, which are also strongly germline dependent. Thus, our proposed mechanisms have broad implications in the development and design of new and more stable antigen binding fragments.

State-of-the-Art Approaches to Heterologous Expression of Bispecific Antibodies Targeting Solid Tumors.

Bispecific antibodies (bsAbs) are some of the most promising biotherapeutics due to the versatility provided by their structure and functional features. bsAbs simultaneously bind two antigens or two epitopes on the same antigen. Moreover, they are capable of directing immune effector cells to cancer cells and delivering various compounds (radionuclides, toxins, and immunologic agents) to the target cells, thus offering a broad spectrum of clinical applications. Current review is focused on the technologies used in bsAb engineering, current progress and prospects of these antibodies, and selection of various heterologous expression systems for bsAb production. We also discuss the platforms development of bsAbs for the therapy of solid tumors.

Prediction of polyreactive and nonspecific single-chain fragment variables through structural biochemical features and protein language-based descriptors.

BACKGROUND: Monoclonal antibodies (mAbs) have been used as therapeutic agents, which must overcome many developability issues after the discovery from in vitro display libraries. Especially, polyreactive mAbs can strongly bind to a specific target and weakly bind to off-target proteins, which leads to poor antibody pharmacokinetics in clinical development. Although early assessment of polyreactive mAbs is important in the early discovery stage, experimental assessments are usually time-consuming and expensive. Therefore, computational approaches for predicting the polyreactivity of single-chain fragment variables (scFvs) in the early discovery stage would be promising for reducing experimental efforts. RESULTS: Here, we made prediction models for the polyreactivity of scFvs with the known polyreactive antibody features and natural language model descriptors. We predicted 19,426 protein structures of scFvs with trRosetta to calculate the polyreactive antibody features and investigated the classifying performance of each factor for polyreactivity. In the known polyreactive features, the net charge of the CDR2 loop, the tryptophan and glycine residues in CDR-H3, and the lengths of the CDR1 and CDR2 loops, importantly contributed to the performance of the models. Additionally, the hydrodynamic features, such as partial specific volume, gyration radius, and isoelectric points of CDR loops and scFvs, were newly added to improve model performance. Finally, we made the prediction model with a robust performance ([Formula: see text]) with an ensemble learning of the top 3 best models. CONCLUSION: The prediction models for polyreactivity would help assess polyreactive scFvs in the early discovery stage and our approaches would be promising to develop machine learning models with quantitative data from high throughput assays for antibody screening.

Multistate design of influenza antibodies improves affinity and breadth against seasonal viruses.

Influenza is a yearly threat to global public health. Rapid changes in influenza surface proteins resulting from antigenic drift and shift events make it difficult to readily identify antibodies with broadly neutralizing activity against different influenza subtypes with high frequency, specifically antibodies targeting the receptor binding domain (RBD) on influenza HA protein. We developed an optimized computational design method that is able to optimize an antibody for recognition of large panels of antigens. To demonstrate the utility of this multistate design method, we used it to redesign an antiinfluenza antibody against a large panel of more than 500 seasonal HA antigens of the H1 subtype. As a proof of concept, we tested this method on a variety of known antiinfluenza antibodies and identified those that could be improved computationally. We generated redesigned variants of antibody C05 to the HA RBD and experimentally characterized variants that exhibited improved breadth and affinity against our panel. C05 mutants exhibited improved affinity for three of the subtypes used in design by stabilizing the CDRH3 loop and creating favorable electrostatic interactions with the antigen. These mutants possess increased breadth and affinity of binding while maintaining high-affinity binding to existing targets, surpassing a major limitation up to this point.

High-Affinity Antibodies Designing of SARS-CoV-2 Based on Molecular Dynamics Simulations.

SARS-CoV-2 has led to a global pandemic of new crown pneumonia, which has had a tremendous impact on human society. Antibody drug therapy is one of the most effective way of combating SARS-CoV-2. In order to design potential antibody drugs with high affinity, we used antibody S309 from patients with SARS-CoV as the target antibody and RBD of S protein as the target antigen. Systems with RBD glycosylated and non-glycosylated were constructed to study the influence of glycosylation. From the results of molecular dynamics simulations, the steric effects of glycans on the surface of RBD plays a role of "wedge", which makes the L335-E340 region of RBD close to the CDR3 region of the heavy chain of antibody and increases the contact area between antigen and antibody. By mutating the key residues of antibody at the interaction interface, we found that the binding affinities of antibody mutants G103A, P28W and Y100W were all stronger than that of the wild-type, especially for the G103A mutant. G103A significantly reduces the distance between the binding region of L335-K356 in the antigen and P28-Y32 of heavy chain in the antibody through structural transition. Taken together, the antibody design method described in this work can provide theoretical guidance and a time-saving method for antibody drug design.

Antibody side chain conformations are position-dependent.

Side chain prediction is an integral component of computational antibody design and structure prediction. Current antibody modelling tools use backbone-dependent rotamer libraries with conformations taken from general proteins. Here we present our antibody-specific rotamer library, where rotamers are binned according to their immunogenetics (IMGT) position, rather than their local backbone geometry. We find that for some amino acid types at certain positions, only a restricted number of side chain conformations are ever observed. Using this information, we are able to reduce the breadth of the rotamer sampling space. Based on our rotamer library, we built a side chain predictor, position-dependent antibody rotamer swapper (PEARS). On a blind test set of 95 antibody model structures, PEARS had the highest average χ1 and χ1+2 accuracy (78.7% and 64.8%) compared to three leading backbone-dependent side chain predictors. Our use of IMGT position, rather than backbone ϕ/ψ, meant that PEARS was more robust to errors in the backbone of the model structure. PEARS also achieved the lowest number of side chain-side chain clashes. PEARS is freely available as a web application at http://opig.stats.ox.ac.uk/webapps/pears.

Analysis of interleukin 23 and 7G10 interactions for computational design of lead antibodies against immune-mediated inflammatory diseases.

Wealth of structural data on theurapeutic targets in complex with monoclonal antibodies (mAbs) and advances in molecular modeling algorithms present exciting opportunities in the field of novel biologic design. Interleukin 23 (IL23), a well-known drug target for autoimmune diseases, in complex with mAb 7G10 offers prospect to design potent lead antibodies by traversing the complete epitope-paratope interface. Herein, key interactions aiding antibody-based neutralization in IL23-7G10 complex are resolute through PyMOL, LigPlot+, Antibody i-Patch, DiscoTope and FoldX. Six amino acids Ser31, Val33, Asn55, Lys59 in heavy chain and His34, Ser93 in light chain are subjected to in silico mutagenesis with residues Met, Trp, Ile, Leu and Arg. A set of 431 mutant macromolecules are outlined. Binding affinities of these molecules with IL23 are estimated through protein-protein docking by employing ZDOCK, ClusPro and RosettaDock. Subsequently, the macromolecules revealed comparable result with 7G10 are cross validated through binding free-energy calculations by applying Molecular Mechanics/Poisson Boltzman Surface Area method in CHARMM. Thirty nine designed theoretical antibodies showed improved outcome in all evaluations; from these, top 10 molecules showed at least nine unit better binding affinity compared to the known mAb. These molecules have the potential to act as lead antibodies. Subsequent molecular dynamics simulations too favored prospective of best ranked molecule to have therapeutic implications in autoimmune and inflammatory diseases. Abbreviations: IL23: interleukin 23; IL17: interleukin17; Ab: antibody; Ag: antigen; mAbs: monoclonal antibodies; STAT3: signal transducer and activator of transcription 3; STAT4: signal transducer and activator of transcription 4; PDB: protein databank; MM/PBSA: molecular mechanics Poisson-Boltzmann surface area; Ag-Ab: antigen- antibody complex; SPC/E: extended simple point charge; SD: steepest descents; PME: particle mesh ewald; dG: binding free energies; Fv: variable fragment.

PD-L1 Nanobody Competitively Inhibits the Formation of the PD-1/PD-L1 Complex: Comparative Molecular Dynamics Simulations.

The anti-PD-L1 monoclonal antibody (mAb) targeting PD-1/PD-L1 immune checkpoint has achieved outstanding results in clinical application and has become one of the most popular anti-cancer drugs. The mechanism of molecular recognition and inhibition of PD-L1 mAbs is not yet clear, which hinders the subsequent antibody design and modification. In this work, the trajectories of PD-1/PD-L1 and nanobody/PD-L1 complexes were obtained via comparative molecular dynamics simulations. Then, a series of physicochemical parameters including hydrogen bond, dihedral angle distribution, pKa value and binding free energy, and so forth, were all comparatively analyzed to investigate the recognition difference between PD-L1 and PD-1 and nanobody. Both LR113 (the amino acid residues in PD-L1 are represented by the lower left sign of L) and LR125 residues of PD-L1 undergo significant conformational change after association with mAbs, which dominates a strong electrostatic interaction. Solvation effect analysis revealed that solvent-water enhanced molecular recognition between PD-L1 and nanobody. By combining the analyses of the time-dependent root mean squared fluctuation (RMSF), free energy landscape, clustering and energy decomposition, the potential inhibition mechanism was proposed that the nanobody competitively and specifically bound to the ß-sheet groups of PD-L1, reduced the PD-L1's flexibility and finally blocked the formation of PD-1/PD-L1 complex. Based on the simulation results, site-directed mutagenesis of ND99 (the amino acid residues in Nano are displayed by the lower left sign of N) and NQ116 in the nanobody may be beneficial for improving antibody activity. This work offers some structural guidance for the design and modification of anticancer mAbs based on the structure of the PD-1/PD-L1 complex.

Computational de novo design of antibodies binding to a peptide with high affinity.

Antibody drugs play a critical role in infectious diseases, cancer, autoimmune diseases, and inflammation. However, experimental methods for the generation of therapeutic antibodies such as using immunized mice or directed evolution remain time consuming and cannot target a specific antigen epitope. Here, we describe the application of a computational framework called OptMAVEn combined with molecular dynamics to de novo design antibodies. Our reference system is antibody 2D10, a single-chain antibody (scFv) that recognizes the dodecapeptide DVFYPYPYASGS, a peptide mimic of mannose-containing carbohydrates. Five de novo designed scFvs sharing less than 75% sequence similarity to all existing natural antibody sequences were generated using OptMAVEn and their binding to the dodecapeptide was experimentally characterized by biolayer interferometry and isothermal titration calorimetry. Among them, three scFvs show binding affinity to the dodecapeptide at the nM level. Critically, these de novo designed scFvs exhibit considerably diverse modeled binding modes with the dodecapeptide. The results demonstrate the potential of OptMAVEn for the de novo design of thermally and conformationally stable antibodies with high binding affinity to antigens and encourage the targeting of other antigen targets in the future. Biotechnol. Bioeng. 2017;114: 1331-1342. © 2017 Wiley Periodicals, Inc.

Computer-Aided Antibody Design: An Overview.

The use of monoclonal antibody as the next generation protein therapeutics with remarkable success has surged the development of antibody engineering to design molecules for optimizing affinity, better efficacy, greater safety and therapeutic function. Therefore, computational methods have become increasingly important to generate hypotheses, interpret and guide experimental works. In this chapter, we discussed the overall antibody design by computational approches.