Epitope mapping of monoclonal antibodies: a comprehensive comparison of different technologies

Monoclonal antibodies have become an important class of therapeutics in the last 30 years. Because the mechanism of action of therapeutic antibodies is intimately linked to their binding epitopes, identification of the epitope of an antibody to the antigen plays a central role during antibody drug development. The gold standard of epitope mapping, X-ray crystallography, requires a high degree of proficiency with no guarantee of success. Here, we evaluated six widely used alternative methods for epitope identification (peptide array, alanine scan, domain exchange, hydrogen-deuterium exchange, chemical cross-linking, and hydroxyl radical footprinting) in five antibody-antigen combinations (pembrolizumab+PD1, nivolumab+PD1, ipilimumab+CTLA4, tremelimumab+CTLA4, and MK-5890+CD27). The advantages and disadvantages of each technique are demonstrated by our data and practical advice on when and how to apply specific epitope mapping techniques during the drug development process is provided. Our results suggest chemical cross-linking most accurately identifies the epitope as defined by crystallography.

Preclinical characterization and clinical translation of pharmacodynamic markers for MK-5890: a human CD27 activating antibody for cancer immunotherapy

Background

Immune checkpoint inhibitors (ICI) have radically changed cancer therapy, but most patients with cancer are unresponsive or relapse after treatment. MK-5890 is a CD27 agonist antibody intended to complement ICI therapy. CD27 is a member of the tumor necrosis factor receptor superfamily that plays a critical role in promoting responses of T cells, B cells and NK cells.

Methods

Anti-CD27 antibodies were generated and selected for agonist activity using NF-кB luciferase reporter assays. Antibodies were humanized and characterized for agonism using in vitro T-cell proliferation assays. The epitope recognized on CD27 by MK-5890 was established by X-ray crystallography. Anti-tumor activity was evaluated in a human CD27 knock-in mouse. Preclinical safety was tested in rhesus monkeys. Pharmacodynamic properties were examined in mouse, rhesus monkeys and a phase 1 dose escalation clinical study in patients with cancer.

Results

Humanized anti-CD27 antibody MK-5890 (hIgG1) was shown to bind human CD27 on the cell surface with sub-nanomolar potency and to partially block binding to its ligand, CD70. Crystallization studies revealed that MK-5890 binds to a unique epitope in the cysteine-rich domain 1 (CRD1). MK-5890 activated CD27 expressed on 293T NF-κB luciferase reporter cells and, conditional on CD3 stimulation, in purified CD8+ T cells without the requirement of crosslinking. Functional Fc-receptor interaction was required to activate CD8+ T cells in an ex vivo tumor explant system and to induce antitumor efficacy in syngeneic murine subcutaneous tumor models. MK-5890 had monotherapy efficacy in these models and enhanced efficacy of PD-1 blockade. MK-5890 reduced in an isotype-dependent and dose-dependent manner circulating, but not tumor-infiltrating T-cell numbers in these mouse models. In rhesus monkey and human patients, reduction in circulating T cells was transient and less pronounced than in mouse. MK-5890 induced transient elevation of chemokines MCP-1, MIP-1α, and MIP-1β in the serum of mice, rhesus monkeys and patients with cancer. MK-5890 was well tolerated in rhesus monkeys and systemic exposure to MK-5890 was associated with CD27 occupancy at all doses.

Conclusions

MK-5890 is a novel CD27 agonistic antibody with the potential to complement the activity of PD-1 checkpoint inhibition in cancer immunotherapy and is currently undergoing clinical evaluation.

BioPhi: A platform for antibody design, humanization, and humanness evaluation based on natural antibody repertoires and deep learning

Despite recent advances in transgenic animal models and display technologies, humanization of mouse sequences remains one of the main routes for therapeutic antibody development. Traditionally, humanization is manual, laborious, and requires expert knowledge. Although automation efforts are advancing, existing methods are either demonstrated on a small scale or are entirely proprietary. To predict the immunogenicity risk, the human-likeness of sequences can be evaluated using existing humanness scores, but these lack diversity, granularity or interpretability. Meanwhile, immune repertoire sequencing has generated rich antibody libraries such as the Observed Antibody Space (OAS) that offer augmented diversity not yet exploited for antibody engineering. Here we present BioPhi, an open-source platform featuring novel methods for humanization (Sapiens) and humanness evaluation (OASis). Sapiens is a deep learning humanization method trained on the OAS using language modeling. Based on an in silico humanization benchmark of 177 antibodies, Sapiens produced sequences at scale while achieving results comparable to that of human experts. OASis is a granular, interpretable and diverse humanness score based on 9-mer peptide search in the OAS. OASis separated human and non-human sequences with high accuracy, and correlated with clinical immunogenicity. BioPhi thus offers an antibody design interface with automated methods that capture the richness of natural antibody repertoires to produce therapeutics with desired properties and accelerate antibody discovery campaigns. The BioPhi platform is accessible at https://biophi.dichlab.org and https://github.com/Merck/BioPhi.

Differences in human IgG1 and IgG4 S228P monoclonal antibodies viscosity and self-interactions: Experimental assessment and computational predictions of domain interactions

Human/humanized IgG4 antibodies have reduced effector function relative to IgG1 antibodies, which is desirable for certain therapeutic purposes. However, the developability and biophysical properties for IgG4 antibodies are not well understood. This work focuses on the head-to-head comparison of key biophysical properties, such as self-interaction and viscosity, for 14 human/humanized, and chimeric IgG1 and IgG4 S228P monoclonal antibody pairs that contain the identical variable regions. Experimental measurements showed that the IgG4 S228P antibodies have similar or higher self-interaction and viscosity than that of IgG1 antibodies in 20 mM sodium acetate, pH 5.5. We report sequence and structural drivers for the increased viscosity and self-interaction detected in IgG4 S228P antibodies through a combination of experimental data and computational models. Further, we applied and extended a previously established computational model for IgG1 antibodies to predict the self-interaction and viscosity behavior for each antibody pair, providing insight into the structural characteristics and differences of these two isotypes. Interestingly, we observed that the IgG4 S228P swapped variants, where the CH3 domain was swapped for that of an IgG1, showed reduced self-interaction behavior. These domain swapped IgG4 S228P molecules also showed reduced viscosity from experiment and coarse-grained simulations. We also observed that experimental diffusion interaction parameter (kD) values have a high correlation with computational diffusivity prediction for both IgG1 and IgG4 S228P isotypes.

Abbreviations: AHc, constant region Hamaker constant; AHv, variable region Hamaker constant; CDRs, Complementarity-determining regions; CG, Coarse-grained model; CH1, Constant heavy chain 1; CH2 Constant heavy chain 2; CH3 Constant heavy chain 3; chgCH3 Effective charge on the CH3 region; CL Constant light chain; cP, Centipoise; DLS, Dynamic light scattering; Fab, Fragment antigen-binding; Fc, Fragment crystallizable; Fv, Variable domaing; (r) Radial distribution function; H1 CDR1 of Heavy Chain; H2 CDR2 of Heavy Chain; H3 CDR3 of Heavy Chain; HVI, High viscosity index; IgG1 human immunoglobulin of IgG1 subclass; IgG4 human immunoglobulin of IgG4 subclass; kD, Diffusion interaction parameter; L1 CDR1 of Light Chain; L2 CDR2 of Light Chain; L3 CDR3 of Light Chain; mAb, Monoclonal antibody; MD, Molecular dynamics; PPI Protein–protein interactions; SCM, Spatial charge map; UP-SEC, Ultra-high-performance size-exclusion chromatography; VH, Variable domain of Heavy Chain; VL, Variable domain of Light Chain

Cryo-EM structures of inhibitory antibodies complexed with arginase 1 provide insight into mechanism of action

Human Arginase 1 (hArg1) is a metalloenzyme that catalyzes the hydrolysis of L-arginine to L-ornithine and urea, and modulates T-cell-mediated immune response. Arginase-targeted therapies have been pursued across several disease areas including immunology, oncology, nervous system dysfunction, and cardiovascular dysfunction and diseases. Currently, all published hArg1 inhibitors are small molecules usually less than 350 Da in size. Here we report the cryo-electron microscopy structures of potent and inhibitory anti-hArg antibodies bound to hArg1 which form distinct macromolecular complexes that are greater than 650 kDa. With local resolutions of 3.5 Å or better we unambiguously mapped epitopes and paratopes for all five antibodies and determined that the antibodies act through orthosteric and allosteric mechanisms. These hArg1:antibody complexes present an alternative mechanism to inhibit hArg1 activity and highlight the ability to utilize antibodies as probes in the discovery and development of peptide and small molecule inhibitors for enzymes in general.

Predicting Antibody Developability Profiles Through Early Stage Discovery Screening

Monoclonal antibodies play an increasingly important role for the development of new drugs across multiple therapy areas. The term 'developability' encompasses the feasibility of molecules to successfully progress from discovery to development via evaluation of their physicochemical properties. These properties include the tendency for self-interaction and aggregation, thermal stability, colloidal stability, and optimization of their properties through sequence engineering. Selection of the best antibody molecule based on biological function, efficacy, safety, and developability allows for a streamlined and successful CMC phase. An efficient and practical high-throughput developability workflow (100 s-1,000 s of molecules) implemented during early antibody generation and screening is crucial to select the best lead candidates. This involves careful assessment of critical developability parameters, combined with binding affinity and biological properties evaluation using small amounts of purified material (<1 mg), as well as an efficient data management and database system. Herein, a panel of 152 various human or humanized monoclonal antibodies was analyzed in biophysical property assays. Correlations between assays for different sets of properties were established. We demonstrated in two case studies that physicochemical properties and key assay endpoints correlate with key downstream process parameters. The workflow allows the elimination of antibodies with suboptimal properties and a rank ordering of molecules for further evaluation early in the candidate selection process. This enables any further engineering for problematic sequence attributes without affecting program timelines.

Pharmacokinetic Properties of Humanized IgG1 and IgG4 Antibodies in Preclinical Species: Translational Evaluation

Assessment of the factors that regulate antibody exposure-response relationships in the relevant animal models is critical for the design of successful translational strategies from discovery to the clinic. Depending on the specific clinical indication, preclinical development paradigms may require that the efficacy or dosing-related attributes for the existing antibody be assessed in various species when cross-reactivity of the lead antibody to the intended species is justified. Additionally, with the success of monoclonal antibodies for management of various human conditions, a parallel interest in therapeutic use of these novel modalities in various veterinary species has followed. The protective role of neonatal Fc receptor (FcRn) in regulation of IgG homeostasis and clearance is now well recognized and the "nonspecific clearance" of antibodies through bone marrow-derived phagocytic and vascular endothelial cells (via lysosomal processes) is modulated by interactions with FcRn receptors. In this study, we have attempted to examine the PK properties of human IgG antibodies in dog and monkey. These studies establish a translational framework for evaluation of IgG antibody PK properties across species.

Deep mutational scanning of an antibody against epidermal growth factor receptor using mammalian cell display and massively parallel pyrosequencing

We developed a method for deep mutational scanning of antibody complementarity-determining regions (CDRs) that can determine in parallel the effect of every possible single amino acid CDR substitution on antigen binding. The method uses libraries of full length IgGs containing more than 1000 CDR point mutations displayed on mammalian cells, sorted by flow cytometry into subpopulations based on antigen affinity and analyzed by massively parallel pyrosequencing. Higher, lower and neutral affinity mutations are identified by their enrichment or depletion in the FACS subpopulations. We applied this method to a humanized version of the anti-epidermal growth factor receptor antibody cetuximab, generated a near comprehensive data set for 1060 point mutations that recapitulates previously determined structural and mutational data for these CDRs and identified 67 point mutations that increase affinity. The large-scale, comprehensive sequence-function data sets generated by this method should have broad utility for engineering properties such as antibody affinity and specificity and may advance theoretical understanding of antibody-antigen recognition.

Affinity and cross-reactivity engineering of CTLA4-Ig to modulate T cell costimulation

CTLA4-Ig is an Fc fusion protein containing the extracellular domain of CTLA-4, a receptor known to deliver a negative signal to T cells. CTLA4-Ig modulates T cell costimulatory signals by blocking the CD80 and CD86 ligands from binding to CD28, which delivers a positive T cell costimulatory signal. To engineer CTLA4-Ig variants with altered binding affinity to CD80 and CD86, we employed a high-throughput protein engineering method to map the ligand binding surface of CTLA-4. The resulting mutagenesis map identified positions critical for the recognition of each ligand on the three CDR-like loops of CTLA-4, consistent with the published site-directed mutagenesis and x-ray crystal structures of the CTLA-4/CD80 and CTLA-4/CD86 complexes. A number of single amino acid substitutions were identified that equally affected the binding affinity of CTLA4-Ig for both ligands as well as those that differentially affected binding. All of the high-affinity variants showed improved off-rates, with the best one being a 17.5-fold improved off-rate over parental CTLA4-Ig binding to CD86. Allostimulation of human CD4(+) T cells showed that improvement of CD80 and CD86 binding activity augmented inhibition of naive and primed T cell activation. In general, increased affinity for CD86 resulted in more potent inhibition of T cell response than did increased affinity for CD80. Optimization of the affinity balance to CD80 and CD86 to particular disease settings may lead to development of a CTLA4-Ig molecule with improved efficacy and safety profiles.

Analysis of the dose rate produced by control rods discharged from a BWR into the irradiated fuel pool

BWR control rods become activated by neutron reactions into the reactor. Therefore, when they are withdrawn from the reactor, they must be stored into the storage pool for irradiated fuel at a certain depth under water. Dose rates on the pool surface and the area surrounding the pool should be lower than limits for workers. The MCNP code based on the Monte Carlo method has been applied to model this situation and to calculate dose rates at points of interest.

Evidence for evolutionarily conserved secondary structure in the H19 tumor suppressor RNA

The molecular basis for function of the mammalian H19 as a tumor suppressor is poorly understood. Large, conserved open reading frames (ORFs) are absent from both the human and mouse cDNAs, suggesting that it may act as an RNA. Contradicting earlier reports, however, recent studies have shown that the H19 transcript exists in polysomal form and is likely translated. To distinguish between possible functional roles for the gene product, we have characterized the sequence requirements for H19-mediated in vitro suppression of tumor cell clonogenicity and analyzed the sequence of the gene cloned from a range of mammals. A cDNA version of the human gene, lacking the unusually short introns characteristic of imprinted genes, is as effective as a genomic copy in blocking anchorage-independent growth by G401 cells. The first 710 nucleotides of the gene can be deleted with no effect on in vitro activity. Further truncations from either the 5'- or 3'-end, however, cause a loss of suppression of clonogenicity. Using conserved sequences within the H19 gene as PCR primers, genomic DNA fragments were amplified from a range of mammalian species that span the functional domain defined by deletion analysis. Sequences from cat, lynx, elephant, gopher and orangutan complement the previous database of sequences from human, mouse, rat and rabbit. Hypothetical translation of the resulting sequences shows an absence of conserved ORFs of any size. Free energy and covariational analysis of the RNA sequences was used to identify potential helical pairings within the H19 transcript. A set of 16 helices are supported by covariation (i.e. conservation of base pairing potential in the absence of primary sequence conservation). The predicted RNA pairings consist largely of local hairpins but also include several long range interactions that bridge the 5'- and 3'-ends of the functional domain. Given the evolutionary conservation of structure at the RNA level and the absence of conservation at the protein level, we presume that the functional product of the H19 gene is a structured RNA.

RNA secondary structure prediction based on free energy and phylogenetic analysis

We describe a computational method for the prediction of RNA secondary structure that uses a combination of free energy and comparative sequence analysis strategies. Using a homology-based sequence alignment as a starting point, all favorable pairings with respect to the Turner energy function are identified. Each potentially paired region within a multiple sequence alignment is scored using a function that combines both predicted free energy and sequence covariation with optimized weightings. High scoring regions are ranked and sequentially incorporated to define a growing secondary structure. Using a single set of optimized parameters, it is possible to accurately predict the foldings of several test RNAs defined previously by extensive phylogenetic and experimental data (including tRNA, 5 S rRNA, SRP RNA, tmRNA, and 16 S rRNA). The algorithm correctly predicts approximately 80% of the secondary structure. A range of parameters have been tested to define the minimal sequence information content required to accurately predict secondary structure and to assess the importance of individual terms in the prediction scheme. This analysis indicates that prediction accuracy most strongly depends upon covariational information and only weakly on the energetic terms. However, relatively few sequences prove sufficient to provide the covariational information required for an accurate prediction. Secondary structures can be accurately defined by alignments with as few as five sequences and predictions improve only moderately with the inclusion of additional sequences.