Enhancing antibody-dependent cellular cytotoxicity (ADCC) is a central goal for immunotherapy, yet current strategies face a persistent engineering trade-off between potency, stability, and manufacturability. Here, we report Fc₉₅, an IgG1 Fc variant developed using Microenvironment-Tuned Interface Sculpting (MTIS), a design philosophy that systematically co-optimizes these competing parameters. This rational design process led to an unexpected and powerful discovery: a dual-action mechanism where four synergistic mutations not only directly remodel the FcγRIIIa binding interface for hyper-affinity, but also, through profound local physicochemical alterations, instruct the host cell's machinery to achieve near-complete afucosylation (<15% fucose). This effectively translates a complex cell-line engineering challenge into an elegant problem of protein sequence design. Across five molecular formats targeting oncology and infectious disease, Fc₉₅ consistently surpassed clinical afucosylated antibodies and the Fc3A variant, leading to potency gains of several orders of magnitude in primary cell assays and demonstrating superiority against both high- and low-affinity FcγRIIIa allotypes. In vivo, Fc₉₅ conferred complete, durable protection in a disseminated lymphoma model and mediated robust tumor control and regression in solid tumor settings. Mechanistically, Fc₉₅ acts as a pivotal regulator, inducing comprehensive reprogramming of the tumor microenvironment—driving NK and CD8⁺ T cell infiltration, M1 macrophage polarization, and regulatory T cell depletion—more effectively than benchmark controls. By unifying intentional interface engineering with microenvironment-directed afucosylation, Fc₉₅ emerges as a robust, manufacturable platform that resolves the long-standing efficacy-manufacturability trade-off, paving the way for next-generation immunotherapies.

Abstract Motivation Computational antibody engineering requires reliable prediction of antibody variable-fragment structures, antigen–antibody complexes, and binding interfaces. However, publicly available tools for these tasks have rarely been compared across the complete workflow under a controlled and statistically grounded design. Results We evaluated ImmuneBuilder, IgFold, AlphaFold3, GRAMM, and dyMEAN on 50 non-redundant humanized antibody–antigen complexes using multiple retained predictions and paired statistical testing. All three antibody structure predictors were accurate, with AlphaFold3 performing best overall and for the third complementarity-determining region of the heavy chain. AlphaFold3 also substantially outperformed GRAMM and dyMEAN in complex prediction, producing medium- or high-quality binding interfaces for 46% of the complexes, although overall interface accuracy remained limited. When docking was reliable, AlphaFold3 accurately recovered epitope and paratope residues, salt bridges, and non-bonded contacts, but reproduced hydrogen bonds and fine-grained contact strengths less consistently. These findings provide practical guidance for selecting tools across antibody-modeling workflows and identify persistent limitations in fine-grained interface prediction. Availability and implementation Data, structural predictions, evaluation results, and analysis code are available from Zenodo under record 20710876.

[This corrects the article DOI: 10.3892/ol.2019.10255.].

Monoclonal antibodies (mAbs) have transformed the treatment of cancer and immune disorders, but their single-target nature limits efficacy against heterogeneous tumors and mutating pathogens. Recombinant polyclonal antibodies (RPABs)-defined as mixtures of typically 2-25 defined mAbs produced as a single drug substance from one mixed master cell bank-were proposed to combine the epitope breadth of polyclonal antibodies with the manufacturing consistency of mAbs. This review critically analyzes the four RPABs candidates that have entered clinical trials to date (Sym001, Sym004, Sym013, and Sym015), all developed by Symphogen using its proprietary Sympress™ platform. We identify seven interrelated barriers that collectively explain why no RPABs product has yet received regulatory approval: modest efficacy restricted to biomarker-selected subgroups, significant toxicity, pharmacokinetic mismatch among components, instability of mixed cell banks, lack of standardized quality control methods, regulatory uncertainty, and commercial deprioritization after acquisition. Critically, we distinguish between scientific failure (Sym013, discontinued after early termination of its Phase I trial due to tolerability concerns and pharmacokinetic mismatches) and strategic discontinuation (Sym001 and Sym015, which showed clinical signals but were deprioritized for commercial reasons). All clinical and manufacturing data analyzed in this review are derived exclusively from Symphogen's proprietary Sympress™ platform, as no other RPABs candidate from independent developers has entered clinical trials. We conclude that without independent validation of manufacturing consistency, pharmacokinetic-based component ratio design, and a dedicated regulatory pathway, RPABs face an uncertain future. Recommendations for future development are provided.

As the predominant antibody isotype at human mucosal surfaces, immunoglobulin A (IgA) exerts critical functions in mucosal defense and systemic immune homeostasis. The two IgA subclasses, IgA1 and IgA2, possess distinct structural and functional features and assemble into dimeric IgA (dIgA) and secretory IgA (sIgA) under physiological conditions. With advances in cryo-electron microscopy, the molecular architecture and assembly of sIgA have been clearly characterized. While IgA protects hosts from pathogenic infection and sustains immune balance, aberrantly altered IgA contributes to the pathogenesis of IgA nephropathy and autoimmune encephalitis. To date, IgA-based therapeutics have displayed promising therapeutic efficacy against respiratory disorders, inflammatory bowel diseases and oral fungal infections. Through standardized retrieval and categorical sorting of the published literature covering IgA structure, physiological functions, disease pathogenic mechanisms and therapeutic research, this review systematically organizes relevant research advances, analyzes the bottlenecks and translational prospects of IgA-targeted therapeutics, and provides objective references to accelerate the clinical translation of IgA antibody agents.

ABSTRACT Antibody‐antibiotic conjugates (AACs) provide a transformative platform for eradicating intracellular methicillin‐resistant Staphylococcus aureus (MRSA). However, current AACs largely follow antibody‐drug conjugates (ADCs) design principles and rely on host‐lysosomal proteases for activation, rendering them ineffective against extracellular (planktonic) bacteria. Furthermore, we demonstrate that high bacterial burden induces host‐lysosomal dysfunction, fundamentally compromising the efficacy of host‐dependent AACs against intracellular reservoirs. To address these limitations, we developed a pathogen‐centric AAC (AZO‐AAC) featuring an azobenzene‐based linker that is triggered by bacterial‐secreted azoreductase. This strategy shifts the activation mechanism from host‐cell machinery to the pathogen itself, enabling antibiotic release independent of host‐cell integrity. Surprisingly, our designed AZO‐AAC achieves nanomolar‐scale eradication of both planktonic and intracellular MRSA, reaching levels below the limit of detection in scenarios where traditional AACs fail. In murine models of peritonitis and septicemia, a single dose (60 mg/kg) of AZO‐AAC resulted in a six‐log reduction in bacterial burden and the preservation of normal tissue architecture. By decoupling activation from host‐cell status, this pathogen‐responsive platform provides a robust strategy for the targeted elimination of complex, multi‐niche MRSA infections.

Single-domain properties and ease of gene manipulation of the nanobody (Nb) make it suitable for affinity maturation in vitro. In this study, the anti-ochratoxin A (OTA) Nb3G was selected as a model antibody to explore feasible approach of affinity enhancement. Artificial intelligence (AI)-augmented molecular modeling tools were used to analyze the interaction between OTA and Nb3G, and nine key residues (Thr28, Tyr32, Asp54, Thr57, Trp99, Phe103, Glu106, Trp110, Tyr112) were selected to construct a site-directed mutagenesis library. After biological screening and identification, the affinity-enhanced mutants 1H and 11B were obtained, exhibiting half maximal inhibitory concentration (IC50) values of 0.31 ng/mL and 0.25 ng/mL, corresponding to 1.26-fold and 1.56-fold sensitivity improvements compared with the parent Nb (IC50 0.39 ng/mL). Enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance (SPR) further confirmed 2-fold and 4-fold increases in titer, and 6.27-fold and 11.49-fold improvements in affinity, respectively. The mutants also demonstrated superior stability under diverse conditions and improved competitive inhibition performance with lower variability. Structural analysis indicated that deeper antigen-binding cavities may facilitate more efficient OTA encapsulation, enhancing detection sensitivity. Recoveries in spiked oat and coffee samples ranged from 94.20 % to 113.85 %. Overall, the synthetic nanobody library strategy developed in this work provides an effective method for the preparation of high-affinity nanobodies against small molecule antigens.

Zaire Ebola virus (EBOV) is one of the main pathogens causing Ebola hemorrhagic fever. The whole human monoclonal antibody Mab114, which can specifically recognize EBOV, has good functions of neutralizing the virus and improving the survival rate of infected animals. The longer in vivo half-life of antibody drugs is mainly achieved through pH dependent binding of the receptor FcRn. This study used computer simulation methods to construct the spatial conformation of human Fc and human FcRn complexes based on molecular docking and mechanical optimization. The key structural domains and important amino acid sites for the binding of human Fc and human FcRn were determined. Based on this, two Fc mutants were designed, and two Fc mutant antibodies Mab114-M5 and Mab114-M6 were constructed using whole human monoclonal antibody Mab114 as the model antibody. The affinity, in vivo and in vitro neutralization function, in vivo serum metabolism of human FcRn transgenic mice, and in vivo antiviral activity were verified through experiments. The results showed that the mutant antibody had a significantly longer retention time in mice compared to Mab114, and also exhibited similar activity in neutralizing EBOV as the maternal antibody Mab114. Fc mutants have the potential to have a long half-life in vivo, thereby reducing medication frequency and providing new ideas for the development of long-acting antibody therapy drugs.

A core complex with transcription factors (TFs) TAL1/TCF3/GATA2 and adaptors LMO2/LDB1 lies at the top of the hematopoietic transcriptional hierarchy. The mechanism(s) underlying the expression of these components remain elusive. Adaptor RACK1 interacts with multiple TFs and modulates their activation and/or stability. However, a role of RACK1 in the transcriptional control of hematopoietic stem cell (HSC) fates hasn’t been disclosed. Here, we report that RACK1 is expressed across various hematopoietic cell types. Adulthood Rack1 deletion in type I interferon- (IFN-I)-responsive cells leads to rapid and profound hematopoietic failure and HSC loss. HSC exhaustion upon adulthood Rack1 deletion results from cell-intrinsic defects with massive apoptosis. Single-cell RNA sequencing indicates that adulthood Rack1 deletion in IFN-I-responsive cells leads to aberrant lineage-geneset-scores of transcriptional HSCs and the emergence of stressed HSCs. Furthermore, prenatal deletion of Rack1 in hematopoietic cells results in reduced and defective HSCs in the fetal liver. Mechanistically, RACK1 prevents HSC loss through maintaining the protein level of LDB1. The direct interaction between RACK1 and LDB1 suppressing its ubiquitination and subsequent degradation, thereby stabilizes LDB1. Therefore, RACK1 maintains adult and fetal mouse HSCs through, at least partially, directly binding to and stabilizing LDB1.

Background Filoviruses (Ebola and Marburg) cause severe high-mortality infectious diseases. Neutralizing antibodies represent promising antiviral candidates, but broadly reactive antibodies against multiple filoviruses are rare. Antibody-ligand fusion strategies that combine conserved glycoprotein-targeting antibodies with IGF2R ligands can mediate endosomal antibody delivery. Previously, we showed that an NPC2-fused AF-03 (an antibody to the GP1 core region) exhibited pan-filovirus neutralization. However, this NPC2 fusion faced quality control challenges and its application is limited by heterogeneous glycosylation, batch inconsistency, and poor CMC feasibility. Methods Human IGF2 or its Y27L variant was fused to the N-terminus of the AF-03 light chain via a peptide linker. Cellular attachment was measured by flow cytometry and ELISA. Internalization was assessed by flow cytometry, western blotting and confocal imaging. Neutralizing activity against pseudotyped filoviruses was determined using a luciferase reporter system in vitro. Results The IGF2-fused antibody (termed AF03-IL) exhibited broad-spectrum neutralizing activity against multiple pseudotyped filovirus in vitro. However, its potency, including internalization and neutralization, was much weaker than that of the NPC2-fused antibody (termed AF03-NL). The reduced activity was caused by IGF2 binding to IGF1R on the cell surface. To overcome this limitation, we introduced a single point mutation (Y27L) into the IGF2 domain. This new construct, named AF03-ILm1, showed superior internalization and neutralization activity compared to the parental AF03-IL. Conclusions The anti-filovirus activity of IGF2-fused AF-03 is impaired by cross-engagement with IGF1R that is rescued by Y27L introduction. The work also might have implications into the design of IGF2-based protein degradation molecules.

BACKGROUND AND AIMS:Tumor necrosis factor-α (TNF-α) is a well-characterized causal mediator in the pathogenesis of acute liver failure (ALF). Despite the efficacy of TNF-α inhibitors (e.g., antibodies and Fc-fused receptor proteins) in animal models, clinic translation has not been successful so far, largely due to sustained repression of TNF-α signaling and Fc-mediated deleterious immune responses. To address these obstacles, we aim to develop a new modality targeting TNF-α by employing lysosome-targeting protein degradation technology. METHODS:Murine Niemann-Pick type C 2 (NPC2) or a mutated form of human insulin-like growth factor 2 (IGF2) (Y27L, IGF2m) was fused to the C-terminus of a TNF-α-targeting nanobody (Nb) via a peptide linker. The ability of Nb-NPC2 and Nb-IGF2m to deliver TNF-α into the lysosomal compartment for degradation, thereby inhibiting TNF-α-induced apoptosis, was examined in IGF2R-expressing cell lines. ALF models induced by co-injection of D-Galactosamine (GalN) with human TNF-α or LPS were used to evaluate the protective efficacy of the fusion proteins. RESULTS:Nb-NPC2 potently suppressed TNF-α-induced cellular apoptosis via the interaction between mannose-6-phosphate (M6P)-bearing NPC2 and IGF2R. Mechanistically, the fusion protein actively hijacked TNF-α to be transported into the lysosome for degradation. Nb-NPC2 administration protected mice from ALF-induced lethality and tissue damage by rapidly clearing TNF-α in circulation. Furthermore, a non-glycosylated version (i.e., Nb-IGF2m) exhibited comparable effects in vitro and in vivo. CONCLUSIONS:Nb-NPC2 (IGF2m) represents a novel avenue with therapeutic potential for ALF, primarily via lysosome-targeting degradation of TNF-α. IMPACT AND IMPLICATIONS:Acute liver failure is a life-threatening inflammatory disease in which TNF-α serves as a central mediator to trigger hepatocyte death, thereby causing vascular dysfunction and liver injury. However, attempts to block TNF-α using antibodies or Fc-fusion proteins in ALF patients have provided no survival benefit to date, largely attributable to sustained suppression of TNF-α as well as Fc-mediated deleterious immune reactions. Our study proposed a novel modality by employing a lysosome-targeting degradation strategy that rapidly cleared TNF-α in circulation and consequently exhibited protective efficacy in ALF mouse models. Given the huge unmet medical need for ALF therapy, the therapeutic potential of this avenue deserves further validation.

Ba ckground: Filoviruses (Ebola and Marburg) cause severe, high-mortality diseases. Neutralizing antibodies are ideal therapeutics, but broad-spectrum ones are rare. A strategy is to couple an antibody targeting conserved filovirus glycoprotein epitopes with ligands for IGF2R to enable endosomal delivery. Previously, we have shown that NPC2-fused AF-03 (an antibody to the GP1 core region) exhibited pan-filovirus neutralization but faced quality control challenges due to NPC2’s mannose-6-phosphate motif.M ethods: human IGF2 or the variant (Y27L) fused to the N-terminus of the light chain of AF-03 via a peptide linker. The capacity of cellular attachment and internalization was examined by flow cytometry, ELISA, western blotting and confocal imaging respectively. The potency to neutralize pseudotypic filovirus species in vitro was determined using a luciferase reporter system.R esults: IGF2-fused AF-03 (termed AF03-IL) exhibited pan-filovirus neutralizing activity in vitro. However, its potency, including the internalization and neutralization activity, was much weaker than NPC2-fused AF-03 (termed AF03-NL), which was due to the perturbation by ligation of IGF2 to IGF1R on the cell surface. Accordingly, AF03-ILm1 was generated, in which one point mutation (Y27L) was introduced in IGF2 to evade the ligation to IGF1R. Compared with the parental cargo, AF03-ILm1 exhibited superior internalization and neutralization activity.C onclusions: AF03-ILm1 represents a promising therapeutic cargo for filovirus-caused diseases. The data here also provide new insights into the design of IGF2-based protein degradation molecules.

The C-terminus of the BoNT/A heavy chain (BoNT/AHC) mediates binding to its receptor, SV2, a critical step for toxicity. Antibody inhibition of this interaction enhances neuronal survival. We previously identified a functional anti-BoNT/AHC nanobody, HM. To extend its in vivo half-life, we designed and prepared two Fc-optimized nanoparticles, HM-Fc5 and HM-Fc6. Structural modeling (homology/docking) of the HM Fv-AHC complex predicted that HM engages key AHC residues (Tyr1155, Phe1160, Ile1161, Val1184, Asn1188, Lys1189, Glu1190), which overlap with the SV2 binding site. This suggests HM’s protective mechanism involves blocking toxin-receptor binding and cellular entry. HM-Fc5 and HM-Fc6 retained the stability and function of the parental HM antibody while exhibiting prolonged in vivo half-life. These optimized nanobodies offer economical candidates potentially enabling longer dosing intervals, beneficial for prophylaxis or chronic disease treatment. Significance Statement: The purpose of the study is to design and prepare two Fc optimized nanoparticles, HM-Fc5 and HM-Fc6, and predict the key residues involved in the interaction between HMs and AHC. The experimental results showed that HM-Fc5 and HM-Fc6 have the same stability as the parent HM antibody but have a longer half-life in vivo. The key residues Tyr1155, Phe1160, Ile1161, Val1184, Asn1188, Lys1189, and Glu1190 overlap with the SV2 binding site. Our experimental results indicate that these nanobody candidates are not only more economical and convenient, but may also have longer dosing intervals, providing strong evidence and reference for prolonging the in vivo half-life of nanomaterials.

Programmed cell death-ligand 1 (PD-L1)/PD-1 axis is crucial for maintenance of immune homeostasis and its impairment partially accounts for the pathogenesis of inflammatory diseases. Hence, augmenting PD-L1/PD-1 signals represents a novel strategy to prevent destructive inflammation and induce immune tolerance. Recently, we developed a new cargo by conjugating the ectodomain of PD-L1 with pHLIP, a low pH-responding and membrane-inserting peptide, and demonstrated its potent immune-suppressive activity under weakly acidic (pH6.1) conditions in vitro. Herein, we further showed that PD-L1-pHLIP (termed as PD-L1-pHLIPwt) responded well to weakly acidic buffer, but not in nearly neutral pH (pH6.8) solutions. To overcome this obstacle, pHLIPwt was replaced by a variant harboring two mutations (Asp14Gla and Asp25Aad) and PD-L1 ectodomain was conjugated to the N-terminus of pHLIP variant via sulfo-SMCC linker (termed as PD-L1-pHLIPva). PD-L1-pHLIPva potently inhibited T effector function including proliferation, activation as well as proinflammatory cytokine release in nearly neutral pH buffer through PD-L1/PD-1 interaction. The inhibitory function of PD-L1-pHLIPva was attributed to more amounts of PD-L1 anchored on the surface of several types of immune cells compared with PD-L1-pHLIPwt. Given that the niche in the lesions of inflammation is weakly acidic even nearly neutral pH, PD-L1-pHLIPva represents a new arsenal to potentially dampen excessive inflammatory reactions.

Botulinum neurotoxin (BoNT) is a highly lethal toxin produced by the anaerobic bacterium Clostridium botulinum, which leads to nerve paralysis following poisoning. At present, there is no specific drug officially approved. Antibodies, particularly single-domain antibodies, represent safe and effective candidates for specific drugs against BoNT. In this study, the receptor-binding domain of botulinum toxin (BoNT/AHCC) was utilized to immunize Bactrian camels, resulting in the generation of a nanobody phage library. From this library, a high-affinity binding antibody, designated A1, and a neutralizing antibody, named HM, were successfully obtained through SPR-based screening. The affinity constant of HM for botulinum toxin is 1.08E-11 M. Results from computer simulations indicate that HM binds at the same site as SV2C. Furthermore, experimental findings demonstrate that HM exhibits significant blocking activity at both the in vitro binding level and the cellular level. In mouse toxicity experiments, HM has been shown to offer protection against a 20 LD50 dose of BoNT/A. Consequently, HM mitigates botulinum toxin poisoning in mice by obstructing the binding of AHCC to SV2C.

Bispecific antibodies (BsAbs) targeting PD-1 and LAG-3 offer a promising strategy in cancer immunotherapy by enhancing antitumor immunity and overcoming resistance to PD-1 blockade. Despite the growing interest in PD-1/LAG-3 BsAbs, a systematic comparison of different BsAbs formats remains lacking, leaving a gap in the rational design of optimized therapeutics. In this study, we systematically compared three BsAb formats-YG-003D1 (Ab-ScFv format), YG-003D2 (DVD format), and YG-003D3 (Knob-into-Hole (KIH) format)-to evaluate their structural, functional, and pharmacokinetic properties, providing critical insights into their therapeutic potential. YG-003D1 exhibited the strongest binding and blocking activity due to its tetravalent Ab-ScFv structure, which facilitated dual-target engagement with minimal steric hindrance. However, it had relatively low expression yields and a tendency to form aggregates, which could impact manufacturability and long-term stability. YG-003D2, utilizing a DVD format, exhibited mild steric hindrance in dual-target engagement, leading to a moderate reduction in blocking efficiency, particularly in LAG-3 inhibition. Nonetheless, its bivalency for both PD-1 and LAG-3 may provide advantages in specific therapeutic contexts. In contrast, YG-003D3, with its asymmetric KIH format, demonstrated the most favorable balance of manufacturability, stability, and pharmacokinetics. It had high expression yields, minimal aggregation, and a half-life comparable to IgG, making it the most promising candidate for clinical development. However, its monovalent binding per target resulted in slightly reduced blocking potency compared to YG-003D1. While YG-003D3 demonstrated the best overall balance of properties, alternative formats such as YG-003D1 could be refined through Fc engineering or linker optimization to enhance manufacturability and reduce aggregation. Similarly, YG-003D2's steric hindrance could be mitigated by introducing flexible linkers to improve dual-target engagement. Further modifications to YG-003D3, such as affinity tuning or Fc engineering, could enhance its blocking potency while retaining its favorable pharmacokinetics. These insights not only provide a rational framework for PD-1/LAG-3 bispecific inhibitor design but also serve as a reference for broader applications of BsAbs in immunotherapy.

Antigen peptides that are presented by a major histocompatibility complex (MHC) and recognized by a T cell receptor (TCR) have an essential role in immunotherapy. Although substantial progress has been made in predicting MHC presentation, accurately predicting the binding interactions between antigen peptides, MHCs and TCRs remains a major computational challenge. In this paper, we propose a unified deep framework (called UniPMT) for peptide, MHC and TCR binding prediction to predict the binding between the peptide and the CDR3 of TCR β in general, presented by class I MHCs. UniPMT is comprehensively validated by a series of experiments and achieved state-of-the-art performance in the peptide–MHC–TCR, peptide–MHC and peptide–TCR binding prediction tasks with up to 15

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

SARS-CoV-2 Omicron sublineages escape most preclinical/clinical neutralizing antibodies in development, suggesting that previously employed antibody screening strategies are not well suited to counteract the rapid mutation of SARS-CoV-2. Therefore, there is an urgent need to screen better broad-spectrum neutralizing antibody. In this study, a comprehensive approach to design broad-spectrum inhibitors against both SARS-CoV-1 and SARS-CoV-2 by leveraging the structural diversity of nanobodies is proposed. This includes the de novo design of a fully human nanobody library and the camel immunization-based nanobody library, both targeting conserved epitopes, as well as the development of multivalent nanobodies that bind nonoverlapping epitopes. The results show that trivale B11-E8-F3, three nanobodies joined tandemly in trivalent form, have the broadest spectrum and efficient neutralization activity, which spans from SARS-CoV-1 to SARS-CoV-2 variants. It is also demonstrated that B11-E8-F3 has a very prominent preventive and some therapeutic effect in animal models of three authentic viruses. Therefore, B11-E8-F3 has an outstanding advantage in preventing SARS-CoV-1/SARS-CoV-2 infections, especially in immunocompromised populations or elderly people with high-risk comorbidities.

Marburg virus (MARV) is one of the filovirus species that cause deadly hemorrhagic fever in humans, with mortality rates up to 90%. Neutralizing antibodies represent ideal candidates to prevent or treat virus disease. However, no antibody has been approved for MARV treatment to date. In this study, we identified a novel human antibody named AF-03 that targeted MARV glycoprotein (GP). AF-03 possessed a high binding affinity to MARV GP and showed neutralizing and protective activities against the pseudotyped MARV in vitro and in vivo. Epitope identification, including molecular docking and experiment-based analysis of mutated species, revealed that AF-03 recognized the Niemann-Pick C1 (NPC1) binding domain within GP1. Interestingly, we found the neutralizing activity of AF-03 to pseudotyped Ebola viruses (EBOV, SUDV, and BDBV) harboring cleaved GP instead of full-length GP. Furthermore, NPC2-fused AF-03 exhibited neutralizing activity to several filovirus species and EBOV mutants via binding to CI-MPR. In conclusion, this work demonstrates that AF-03 represents a promising therapeutic cargo for filovirus-caused disease.