ABSTRACT Ricin, a ribosome-inactivating lectin from Ricinus communis seeds, has been used as a bioterrorism agent in multiple cases. While passive immunotherapy with anti-ricin antibodies shows promise in preclinical studies, no approved countermeasure exists. Developing effective monoclonal antibodies (mAbs) is challenging, requiring epitope targeting that ensures neutralization of the two most dominant natural ricin isoforms (D and E). Moreover, high-affinity binding does not always correlate with toxin neutralization, highlighting the importance of epitope specificity in driving protection. Here, we characterized a panel of 17 anti-ricin antibodies, including VHH and IgG mAbs, to determine their affinities, selectivity, and epitopes. Using surface plasmon resonance (SPR) and biolayer interferometry (BLI), we evaluated antibody affinities for the two ricin isoforms (D and E), as well as for ricin agglutinin, a related lectin with markedly lower toxicity. Epitope determination was performed using (1) SPR-based epitope binning, enhanced by network analysis for streamlined bin visualization, and (2) deep mutational scanning with yeast surface display to identify key epitope residues. BLI effectively distinguished low- and high-affinity interactions, while SPR provided superior resolution for determining the highest affinities and lowest dissociation rates. Both epitope-mapping strategies yielded highly consistent results, allowing the identification of critical epitopes associated with potent neutralization and cross-reactivity between ricin isoforms. This study advances our understanding of ricin neutralization by this panel of antibodies, providing key insights into their affinity, epitope specificity, and cross-reactivity. These findings contribute to the rational design of antibody-based therapeutics for ricin intoxication.

ABSTRACT Monoclonal antibodies (mAbs) and fusion proteins for immune applications (FPIA) play a crucial role in treating autoimmune diseases and cancers by targeting cell-surface proteins and triggering multiple immune mechanisms. These functions are mediated by the crystallizable fragment (Fc) region of mAbs and fusion proteins, whose interaction with Fc gamma receptors (FcγRs) can be modulated through Fc amino acid (AA) engineering. To aid research in this area, we developed the IMGT/FcVariantsExplorer tool (https://www.imgt.org/fcvariantsexplorer/) to identify engineered AA changes or variants within the Fc region in mAb and fusion proteins sequences from IMGT/2Dstructure-DB, the AA sequence database of IMGT®, the international ImMunoGeneTics information system®. We used the IMGT® nomenclature of engineered Fc variants involved in antibody effector properties and formats, applying a standardized classification in five categories: ‘Effector,’ ‘Half-life,’ ‘Physicochemical properties,’ ‘Structure,’ and ‘Hybrid.’ We analyzed sequences from 1,107 mAbs and fusion proteins, identifying 483 entries with Fc AA changes, resulting in 211 unique Fc variants in the dataset. We also used web scraping to retrieve associated biological data from literature. All data have been integrated into IMGT/mAb-DB, with links to sequences in IMGT/2Dstructure-DB, enabling users to query Fc variants by their ‘Category’ or ‘Effect.’ This curated dataset reveals key trends in antibody engineering.

ABSTRACT Nanobodies are small, single-domain antibody fragments derived from heavy chain-only antibodies. They combine high binding affinity with advantages such as compact size, stability, solubility, and flexible epitope recognition, making them attractive tools in molecular biology and therapeutic applications. In this study, we engineered and optimized nanobodies for controlled activation of synthetic cytokine receptors, aiming to expand options for receptor customization. Specifically, we used nanobodies as extracellular domains of the gp130 receptor to induce dimerization upon antigen binding. To enable receptor activity, we introduced framework mutations that promote the formation of an i-shaped nanobody (iBody) dimer, adapted from i-shaped antibodies. These mutations enhanced dimerization and enabled low-level ligand-independent receptor activation. AlphaFold modeling identified the key amino acids responsible for forming the iBody interface. Additional modifications reduced intermolecular affinity, thereby minimizing background activation while preserving the structural features necessary for ligand-induced stimulation. This approach effectively broadened the receptor’s activation range. Importantly, these framework mutations were not limited to the gp130-specific nanobody GP11 but were also functional in AIP3, an anti-idiotypic nanobody targeting palivizumab. Here, the modified nanobody fusion receptor could be activated by palivizumab, overcoming prior steric hindrance.

ABSTRACT Experimental screening for biopharmaceutical developability properties typically relies on resource-intensive, and time-consuming assays such as size exclusion chromatography (SEC). This study highlights the potential of in silico models to accelerate the screening process by exploring sequence and structure-based machine learning techniques. Specifically, we compared surrogate models based on pre-computed features extracted from sequence and predicted structure with sequence-based approaches using protein language models (PLMs) like ESM-2. In addition to different end-to-end fine-tuning strategies for PLM, we have also investigated the integration of the structural information of the antibodies into the prediction pipeline through graph neural networks (GNN). We applied these different methods for predicting protein aggregation propensity using a dataset of approximately 1200 Immunoglobulin G (IgG1) molecules. Through this empirical evaluation, our study identifies the most effective in silico approach for predicting developability properties for SEC assays, thereby adding insights to existing screening efforts for accelerating the antibody development process.

ABSTRACT Acute myeloid leukemia (AML) is a heterogeneous malignancy with poor clinical outcome. Aberrant expression of CD7 in AML patients is linked to shorter overall survival and lack of response to standard of care therapy. CD33/CD7 co-expression on leukemic blasts occurs in approximately one-third of AML patients and is known to be absent in normal myeloid cells. We propose that CD33+CD7+ AML constitutes an aggressive subgroup characterized by poorer prognosis and enrichment in stem-cell associated gene signatures. To address the substantial unmet need in this patient cohort, we developed the antibody–drug conjugate BVX001, a CD33xCD7-targeted bispecific antibody-binding fragment linked to an auristatin payload. Importantly, BVX001 relies on simultaneous binding to CD33 and CD7 in cis through an ‘AND-gated’ design, for optimal delivery of its cytotoxic payload. Consequently, BVX001 did not affect healthy myeloid progenitors or T cells at concentrations at which its monospecific counterparts showed toxicity. BVX001 induced significant tumor regression in AML cell line and patient-derived xenografts and increased overall survival. Finally, BVX001 showed significant blast ablation and reduced leukemic stem cell frequency in AML patient samples with both high and low target co-expression. Together, our findings support BVX001 as a new and promising approach for the treatment of this aggressive CD33+CD7+ AML subtype, currently lacking targeted therapeutic options.

ABSTRACT Since the approval of OKT3® in 1986, monoclonal antibodies (mAbs) have become a cornerstone of modern therapeutics. However, their complex physicochemical properties pose challenges, particularly for high-concentration formulation and subcutaneous administration. Excipient selection is crucial for maintaining mAb stability and efficacy, yet existing studies often lack systematic, cross-source analyses. This study integrates data from marketed products and patents to investigate formulation trends and excipient preferences. Data were retrieved from the Drugs@FDA databases, CAS Formulations Database and Derwent Innovation as of December 31, 2024. Extracted information included target, indication, dosage form, route of administration, and formulation composition. The associations between formulation-related factors (e.g., antibody concentration, route of administration) and excipient selection were evaluated using proportion tests. A total of 6,119 patent records and 108 marketed mAb products (covering 388 patented and 141 marketed formulations) were analyzed. Proportion tests revealed significant associations between antibody concentration and the use of histidine (marketed p = 0.0017) and citric acid (marketed p = 0.0047). The route of administration also influenced excipient choice, notably for hyaluronidase (marketed p = 0.0167; patent p = 0.0056). In addition, lyophilized formulations accounted for a relatively small fraction of both marketed (14.18%) and patented (14.69%) products, with sucrose emerging as the predominant lyoprotectant. This study analyzed excipient usage in marketed and patent formulations by time, concentration, and administration route. High-concentration products more frequently included histidine, arginine, and hyaluronidase, while low-concentration ones used citric/phosphoric acid, trehalose, and NaCl. Intravenous formulations commonly used phosphate/citric buffers, while histidine, arginine, hyaluronidase, and methionine buffers were favored for subcutaneous administration. Lyophilized formulations consistently contained sucrose as the main excipient to mitigate freeze-drying stresses. Additionally, surfactants were essential across formulations to prevent surface-induced aggregation. Patent data could provide early indications of emerging formulation strategies, though further validation is needed to confirm their predictive value.

ABSTRACT The feline McDonough sarcoma-like tyrosine kinase 3 (FLT3)/FLT3 ligand (FLT3L) signaling pathway regulates the development and activity of dendritic cells (DCs) and other myeloid cells, including monocytes. FLT3L, DCs, and monocytes have been implicated in several autoimmune diseases. Here, we describe the development and characterization of a human immunoglobulin G1λ monoclonal antibody (AMG 329; formerly MEDI1116/VIB-1116/HZN-1116) targeting human FLT3L. AMG 329 was derived from a large human combined antibody display library; it was optimized to enhance affinity for FLT3L and reduce antibody dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity. Binding affinity was determined by surface plasmon resonance interaction analysis. Specificity of FLT3L was measured using cell-based flow cytometry and an in vitro functional neutralization assay. ADCC activity was measured using an in vitro cell culture system. Toxicity and toxicokinetics were evaluated in cynomolgus monkeys during AMG 329 dosing (5–100 mg/kg; ≤ 27 weeks) and recovery (≤32 weeks). The AMG 329 antigen-binding region selectively bound to human and cynomolgus monkey FLT3L with affinities of 170 and 63 pM, respectively. AMG 329 specifically bound to and neutralized soluble and cell-bound human FLT3L and did not induce ADCC. AMG 329 administration generally reduced circulating plasmacytoid, conventional DC, and classical monocyte relative proportions in cynomolgus monkeys in a non–dose-dependent manner. Disruption of the FLT3/FLT3L signaling pathway presents a new potential therapeutic approach to treat autoimmune and inflammatory diseases. AMG 329 is a selective human monoclonal antibody antagonist of FLT3L that is currently being investigated in clinical studies.

ABSTRACT Developability studies provide essential data to identify monoclonal antibodies (mAbs) with optimal drug-like properties, which are indicative of a molecule’s suitability for large-scale manufacturing, long-term storage, and ease of administration. Hydrophobicity is a critical molecular attribute that affects solubility, aggregation, and stability at high protein concentrations and is routinely assessed in these studies. Although traditional analytical hydrophobic interaction chromatography (aHIC) is considered the benchmark for measuring hydrophobicity, its application in early developability studies is limited because the process requires serial sample injections, which is time-intensive and impractical for the evaluation of hundreds of molecules. To overcome this limitation, we developed an alternative aHIC method that uses a plate-based assay format, enabling rapid screening of large sample sets. Compatible with automation platforms, this surrogate aHIC method demonstrates excellent accuracy in distinguishing between low- and high-risk molecules, proving to be an efficient tool for preliminary developability assessments. This innovative assay provides a robust, timesaving, and sample-efficient means of evaluating hydrophobicity that readily supports early phase biotherapeutic antibody discovery through selection of mAbs with favorable drug-like properties. Furthermore, the potential for adaptation of this method to various molecular formats suggests its broad applicability in biotherapeutic discovery.

ABSTRACT Bispecific T cell engager (TCE) therapies have demonstrated transformative clinical success in the treatment of hematological cancers, but the lack of antigens that are sufficiently selective for malignant cells has hampered the success of TCEs in the solid-tumor space. To overcome the on-target, off-tumor toxicities that result from the expression of even low levels of tumor-associated antigens in healthy tissues, we sought to identify a TCE target with highly tumor-restricted expression patterns. Here, we characterize cancer-testes antigen Preferentially Expressed Antigen in Melanoma (PRAME) as a highly selective tumor antigen and identify a proteasomal degradation peptide PRAME425–433 (PRAME425) presented in the context of major histocompatibility complex I (MHCI) as an attractive TCE target. We designed a TCR-mimic (TCRm) antibody screening cascade that prioritizes screening anti-PRAME pMHC binders in off-target T cell dependent cellular cytotoxicity assays in a potent TCE format, rather than relying solely on traditional pMHC binding assays, to determine specificity. Using this screening cascade, we discovered antibodies that selectively bind PRAME425 pMHC without over-recognition of off-target peptides or MHCI via a TCR-like binding geometry. We further solved the first structure of an anti-PRAME425 pMHC TCRm antibody in complex with PRAME425/HLA-A *02:01 using cryo electron microscopy to confirm the TCRm antibody binds in a TCR-like binding geometry and specifically recognizes the PRAME425 peptide. By formatting these novel TCRm antibodies into potent TCEs, we demonstrate PRAME425 pMHC-specific killing of tumor cells, representing a new class of anti-PRAME pMHC biologics.

ABSTRACT Variable heavy (VH) and variable light (VL) chain pairing is a critical determinant of antibody diversity, stability, and antigen-binding specificity. Identifying productive VH – VL combinations experimentally is labor-intensive and costly, motivating the development of computational methods that can more efficiently predict compatible heavy – light chain pairs. In this work, we present a comprehensive framework that includes a new benchmark dataset and three deep learning models, each trained with a different negative sampling strategy: random pairing, V-gene mismatching, and full V(D)J germline mismatching. Our dataset includes natural pairs and these three types of synthetic negatives to simulate increasingly realistic biological constraints. Furthermore, we present a lightweight yet highly effective BERT-based model that achieves over 90% accuracy in discriminating natural from synthetic VH – VL pairs. Through extensive evaluation, we demonstrate that V(D)J-informed negative sampling significantly improves model generalization and biological interpretability. By providing reproducible baselines and a biologically grounded benchmark, this work lays the foundation for future development of efficient computational tools in antibody engineering.