A C-type single-domain antibody with protective efficacy against H1N1 via respiratory administration

Abstract

Influenza A viruses, particularly H1N1, pose a significant public health threat, highlighting the continued need for novel prophylactics and therapeutics. Monoclonal antibodies are promising, but conventional IgGs face challenges in respiratory delivery due to their large size. This study presents A1-B7, a C-type single-domain antibody (C-sdAb) based on a scaffold derived from a human IgG1 CH2 domain, targeting the hemagglutinin (HA) of H1N1 influenza virus. A1-B7 potently neutralizes multiple H1N1 subtypes in vitro. Both prophylactic intranasal administration and therapeutic aerosol inhalation of A1-B7 confer complete protection and reduce live virus in the lungs to undetectable levels. Structural prediction and epitope mapping reveal that A1-B7 binds a conserved stem-region epitope, engaging the HA2 A-helix and a hydrophobic groove. A1-B7 can block low-pH induced conformational change of HA. These findings establish A1-B7 as a promising candidate for further development as a convenient, respiratory tract delivered countermeasure against influenza.

Highlights

• •A C-type single-domain antibody A1-B7 neutralizes H1N1 subtypes of influenza A virus.
• •A1-B7 confers in vivo protection through respiratory administration.
• •A1-B7 recognizes the HA stem region and inhibits conformational change.

Introduction

Influenza A viruses, including seasonal influenza (H1N1, H3N2) and avian influenza (H5N1, H7N9) pose significant and ongoing threats to global public health. The H1N1 subtype plays a critical role in modern influenza history (e.g., 1918 “Spanish Flu” and 2009 H1N1 Pandemic), being one of the primary subtypes responsible for seasonal influenza outbreaks, while also remaining under constant global surveillance due to its pandemic potential (Wooliscroft, 2008; Erbelding et al., 2018). The clinical impact of H1N1 infections ranges from self-limiting respiratory illness to severe pneumonia, acute respiratory distress syndrome (ARDS), and death, especially in high-risk groups such as the young, elderly, and immunocompromised (Iuliano et al., 2018).

Hemagglutinin (HA) is a class I viral fusion glycoprotein on the surface of influenza viruses, playing essential roles in viral attachment, membrane fusion, and host cell entry (Skehel and Wiley, 2000). HA is initially synthesized as a precursor protein, HA0, which is subsequently cleaved by host proteases into two functional subunits—HA1 and HA2. These subunits mediate distinct steps in viral entry: HA1 is responsible for receptor binding, while HA2 facilitates membrane fusion. Due to its essential role in the viral life cycle, HA is a critical target for the development of antiviral drugs and vaccines (Boonstra et al., 2018). Virtually all licensed influenza vaccines indeed exert their protective effects primarily by inducing the human body to produce antibodies targeting the HA (Wei et al., 2020). Although seasonal vaccines serve as a critical countermeasure, vaccines require annual reformulation and have variable efficacy (Osterholm et al., 2012), underscoring the critical need for novel, broad-spectrum therapeutic agents. Monoclonal antibodies (mAbs) targeting various epitopes on HA, have been identified and emerged as promising candidates for both prophylaxis and therapy (Corti et al., 2011; Dreyfus et al., 2012; Laursen et al., 2018; Guthmiller et al., 2022).

However, conventional immunoglobulin G (IgG) antibodies face inherent challenges for respiratory virus intervention. Their large size (∼150 kDa) may impede efficient distribution across the respiratory mucosal epithelium, the primary site of influenza virus replication and entry. Compared to full length, single domain antibodies (sdAbs) derived from the antibody heavy chain variable domain (V-based single domain antibody, V-sdAb, also termed nanobody) are an attractive direction in the development of the next generation of antibody-based therapeutics (Holliger and Hudson, 2005). With V-sdAb technique platform, a plenty of V-sdAbs have been identified and some of them (e.g., Nb21) (Xiang et al., 2020; Nambulli et al., 2021) show clinical potentials. It has been proposed that the CH2 domain (12∼15 kDa) could be engineered as a scaffold for construction of libraries for selection of C-based single domain antibodies (C-sdAbs, C-type nanobodies or C type nanoantibodies) against specific antigens as therapeutic candidates. CH2 contains three flexible loops (Loop BC or Loop 1, Loop DE or Loop 2, and Loop FG or Loop 3) that correspond to three complementary determining regions (CDR1, CDR2 and CDR3) in the antibody variable domain and could be used for the introduction of mutations for library construction (Gong et al., 2012b). Previously, we optimized an IgG1 CH2 domain for improvement of in vivo application potentials (Gong et al., 2013; Cao et al., 2020). Here, we used our C-sdAb technique platform for the selection of C-sdAbs against the hemagglutinin A (HA) of Influenza A H1N1, which resulted in the identification of a C-sdAb termed A1-B7. A1-B7 binds to diverse HA subtypes of H1N1 and neutralizes multiple H1N1 strains. Through AlphaFold-3-predicted structural analysis, epitope competition assays, alanine scanning, we identified that A1-B7 binds to the A-helix (also referred to as the short helix) in the HA2 domain, partial fusion peptide, and the outward-facing β-sheet of the antiparallel beta strand within the stem region. Prophylactic intranasal administration and therapeutic aerosol inhalation of A1-B7 demonstrated 100% protective efficacy in a mouse model.

Results

Identification of a C-sdAb against HA

We performed panning of the C-sdAb phage display library against the IFA H1N1 HA (Fig. 1A), which resulted in the identification of several candidate clones with specific but relatively modest binding activity to the HA (Fig. 1B). Among these clones, G6 and F6 demonstrated potential neutralizing activity (Fig. 1C). To increase their efficacy and spectrum, random mutagenesis was introduced into these candidate clones by random error-prone PCR for the construction of libraries with various mutants. Following several rounds of affinity maturation panning with HA, we identified a C-sdAb, designated A1-B7, derived from the F6 clone and harboring two mutations based on F6 (K53 M and K112I). A1-B7 could efficiently bind to the HA from H1N1 subtypes with a half maximal effective concentration (EC₅₀) from 0.48 nM to 7.73 nM based on an enzyme-linked immunosorbent assay (ELISA) (Fig. 1D). We further evaluated the in vitro neutralization activity of A1-B7 against H1N1 A/Puerto Rico/8/34, H1N1 A/human/Hubei/1/2009, and H1N1 clinical isolates, ranged from 0.37 μg/mL to 1.5 μg/mL. We also tested the neutralization activity of A1-B7 against two H3N2 clinical isolations. However, the neutralizing activity against H3N2 was very weak, with IC₅₀ values of 73.2 μg/mL and 37.2 μg/mL, respectively (Fig. 1E, Supplementary Fig. S1A and S1B).

Fig. 1.

Identification of a C-sdAb against HA. A Schematic of panning of a C-sdAb phage-display library against the Influenza A HA and affinity maturation. Created in BioRender. Bihao, W. (2026) https://BioRender.com/kwpyd5r. B Binding activity of original clones to HA from H1N1 A/Michigan/19/21 tested by ELISA. C Neutralizing activity of original clones against H1N1 PR8. The data in (B) to (C) are the means of technical duplicates from a representative result of three independent experiments. D Binding avtivity of A1-B7 to HA from H1N1 subtypes tested by ELISA. E Neutralizing activity of A1-B7 against Influenza A virus subtypes. The EC₅₀ and IC₅₀ values in (D) and (E) are mean ± SD of three independent experiments.

A1-B7 exhibited in vivo efficacy by intranasal administration

A1-B7 exhibited in vitro neutralizing activity. To investigate whether it can exert prophylactic and therapeutic effects via convenient respiratory administration, specifically intranasal delivery in a mouse model, we conducted an in vivo protection study via intranasal administration. In the prophylaxis group, mice were intranasally administered 25 mg/kg of A1-B7 6 h prior to viral challenge. In the treatment group, mice received intranasal A1-B7 at 2 h post-infection and again on the following two days, totaling three doses (Fig. 2A). The results demonstrated that the prophylaxis group achieved 100% protection and significantly prevented weight loss, while the therapeutic group exhibited a 57% protection rate (Fig. 2B, C).

Fig. 2.

The protective effect of A1-B7 via nasal administration. A Diagram of the prophylactic and therapeutic experiments via nasal administration. BALB/c mice were challenge with H1N1 A/Puerto Rico/8/34 (10 × LD₅₀) at day 0 (D0). In the prophylactic group (n = 12), mice were administered A1-B7 intranasally in 25 mg/kg 6 h before viral challenge. In the therapeutic group (n = 12), mice were administered A1-B7 intranasally in 25 mg/kg 2 h after viral challenge and another two doses (2 h, D1 and D2). In the virus group, mice were administered an equal volume of PBS at (2 h, D1 and D2). Created in BioRender. Bihao, W. (2026) https://BioRender.com/kwpyd5r. B Kaplan–Meier survival curve monitoring of infected mice. Statistical significance of the survival curves was calculated by log rank Mantel–Cox test. (∗∗∗P < 0.0002). C Body weight monitoring of infected mice. Data are represented as mean ± SEM (n = 7 per group). Statistical analysis was performed using one-way ANOVA (∗∗P < 0.01, ns not significant). D The viral titers in the lungs and nasal tissue from different groups were determined at 3 dpi. The dashed line represents the detection limit of detection (LOD). The average values are presented. E The copies of viral RNA in the lungs and nasal tissue from different groups were determined at 3 days post-infection by qRT-PCR. The dashed line represents the LOD. The average values are presented. Statistical significance was determined using nonparametric t-test. ∗∗P < 0.01. F Representative lung histopathology (H&E staining) from the prophylactic, therapeutic, and virus groups at 3 days post-infection, as well as normal lung tissue from the uninfected mock control group.

On the fourth day post-infection, five mice from each group were euthanized to collect lung lavage fluid and nasal turbinate tissue homogenates for quantification of infectious viral titers and RNA copy numbers. The findings revealed that both the prophylaxis and treatment groups showed significantly reduced viral loads in both the lungs and nasal turbinate, with no detectable live virus in the prophylaxis group (Fig. 2D). A similar trend was observed in RNA copy numbers, further corroborating the antiviral efficacy of intranasally delivered A1-B7 (Fig. 2E). Lung pathology analysis further confirmed the protective effect of A1-B7. Untreated mice exhibited obvious lung lesions and granulocyte infiltration by day 3 post-infection, mice treated with A1-B7, either prophylactically or therapeutically, showed normal lung tissue architecture with no visible signs of immune cell infiltration (Fig. 2F). A1-B7 demonstrated highly effective preventive efficacy in mice. Although its therapeutic effect was relatively limited, it still significantly alleviated pathological lung damage. For the intranasal administration, the prophylactic and therapeutic effects differ markedly, which may be significantly related to viral clearance in the lungs. In the prophylaxis group, live virus in the lungs was completely cleared by day 4 post-challenge. In contrast, for the therapeutic group, 2 out of 5 mice still had residual virus. Although the virus was suppressed during the first few days after administration and did not cause obvious lung pathological damage, it may have re-amplified substantially later, leading to death starting on day 10. This proportion is roughly consistent with the 57% protection rate.

The resistance ability of A1-B7 against nebulization

A1-B7 can exert protective effects in vivo via intranasal administration. Given that nebulized ALX-0171 has already been applied in clinical settings (Cunningham et al., 2021), we aimed to test whether A1-B7 can withstand the mechanical stress of nebulization to evaluate its potential for inhalation delivery via nebulization. After being processed by a nebulizer, like another stable C-nanobody previously obtained from our laboratory (Li et al., 2025) (Supplementary Fig. S2A), neither aggregation nor degradation after nebulization was detected by size-exclusion chromatography (SEC) (Fig. 3A–C). Its binding and neutralizing activity also remained unchanged (Fig. 3D, E). However, the initial clone H6 exhibited slight aggregation and decreased activity (Supplementary Fig. S2B). We also examined the temperature sensitivity of A1-B7 and found that its activity remained largely unchanged after treatment at 53 °C or below for 30 min (Fig. 3F). The pH of the human nasal cavity is typically weakly acidic. During inflammation, this pH may change (Zajac et al., 2021), and in conditions such as cystic fibrosis, the nasal environment may become even more acidic (Pezzulo et al., 2012). Therefore, we tested the tolerance of A1-B7 to different pH levels. Under the most extreme conditions selected treatment at pH 1 and 14 for 5 min, the binding activity of A1-B7 still retained at nanomolar level (Fig. 3G).

Fig. 3.

Resistance ability of A1-B7. A The schematic diagrams of nebulization treatment of A1-B7. Created in BioRender. Bihao, W. (2026) https://BioRender.com/mz4je0l. B–E sustainability of A1-B7 to nebulization. Aggregation and degradation were detected by SEC in (B and C), the binding activity reduction was measured by ELISA in (D), and the neutralizing activity reduction was measured in (E). F and G the stability of A1-B7 at different temperatures and extreme pH conditions tested by ELISA. EC₅₀ values are shown as mean ± SD from three independent experiments.

A1-B7 exhibited therapeutic efficacy via aerosol inhalation

Based on preliminary results from intranasal administration, A1-B7 provided only a 57% therapeutic protection rate, and residual live virus was still detected in the lungs by viral titer measurement. We hypothesized that intranasal delivery does not allow sufficient antibody activity in the lung. Given the favorable anti-nebulization properties of A1-B7, we subsequently adopted pulmonary nebulization for therapeutic evaluation at the same dose and time points as the intranasal treatment group. CR9114 was used as a control antibody and administered via intraperitoneal injection as previously reported (Dreyfus et al., 2012) (Fig. 4A).

Fig. 4.

The therapeutic effect of A1-B7 via pulmonary nebulization. A Diagram of the therapeutic experiments via pulmonary nebulization. BALB/c mice were challenge with H1N1 A/Puerto Rico/8/34 (10 × LD₅₀) at day 0 (D0). In the A1-B7 group, mice (n = 12) were administered A1-B7 intranasally in 25 mg/kg 2 h after viral challenge and another two doses (2 h, D1 and D2). In the A1-B7 group, mice (n = 12) were administered CR9114 via intraperitoneal injection in 25 mg/kg 2 h after viral challenge. In the virus group, mice were administered an equal volume of PBS at (2 h, D1 and D2). Created in BioRender. Bihao, W. (2026) https://BioRender.com/kwpyd5r. B Kaplan–Meier survival curve monitoring of infected mice. C Body weight monitoring of infected mice. Data are represented as mean ± SEM (n = 7 per group). Statistical analysis was performed using one-way ANOVA (∗∗P < 0.01, ns not significant). D The viral titers in the lungs and nasal tissue from different groups were determined at 3 dpi. The dashed line represents the detection limit of detection (LOD). The average values are presented. E The copies of viral RNA in the lungs and nasal tissue from different groups were determined at 3 days post-infection by qRT-PCR. The dashed line represents LOD. The average values are presented. Statistical significance was determined using nonparametric t-test. ∗∗P < 0.01, ∗P < 0.05, no significance.

Compared with intranasal administration, pulmonary nebulization achieved complete protection (Fig. 4B) and maintained stable body weight in mice (Fig. 4C). On day 3 post-infection, viral titers measured in bronchoalveolar lavage fluid (BALF) revealed that A1-B7 treatment completely cleared live virus from the lungs (Fig. 4D), demonstrating a significant advantage over intranasal administration. This difference may be the key factor contributing to full protection. Regarding the nasal turbinate, virus was cleared in 4 out of 5 mice; however, one mouse still exhibited a high titer of live virus in the turbinate, and thus no statistically significant difference was observed (Fig. 4D). Furthermore, viral RNA copies in the lungs were nearly undetectable in 4 out of 5 mice, also indicating a marked reduction in viral load (Fig. 4E).

A1-B7 recognized the stem region of HA

To further elucidate the structural basis of this interaction, we employed AlphaFold3 to predict the complex structure of A1-B7 in complex with H1N1 A/California/07/09 HA. The prediction revealed that A1-B7 binds to the stem region and shares a highly similar epitope with CR9114, a representative stem-targeting antibody (Fig. 5A, Supplementary Fig. S3A and S3B). The strong competition between A1-B7 and CR9114 tested by competitive ELISA assays further confirmed a high degree of epitope overlap between the two antibodies (Fig. 5B).

Fig. 5.

Structural analysis of A1-B7. A Superimposition of the AlphaFold3-predicted structure of A1-B7 bound to H1 HA and the structure of CR9114 bound to H5 HA (PDB: 9DWE). HA is shown in gray, A1-B7 in red, and the VH domain of CR9114 in green. B Epitope competition between A1-B7 and CR9114 assessed by ELISA. C Detailed view of the A1-B7 and HA interaction interface. HA1 subunit is colored blue, HA2 orange, and A1-B7 red. Key interaction sites including the A helix, the antiparallel β-sheet, and the critical residue W348 on the fusion peptide (FP), are labeled in figure. D A1-B7 predicted epitope interaction amino acid alignment. The sequence logo was constructed in WebLogo 3. E–G The binding activities of CR9114 and A1-B7 to the corresponding mutant HA proteins were compared after substituting the beta sheet motif_28-32. The data are the means of technical duplicates from a representative result of three independent experiments. H Comparison of the antibody binding area under the curve (AUC) following alanine substitution at the corresponding sites on HA. The average values of three independent experiments are presented. I The binding activity comparison of A1-B7 containing mutation F32I. The data are the means of technical duplicates from a representative result of three independent experiments.

As indicated by the predicted model, A1-B7 resembles other stem-binding antibodies in its mode of interaction. Specifically, its complementarity-determining region Loop3 primarily engages with the A-helix of HA2, a relatively conserved structural element. Concurrently, its Loop 1 interacts with a part of fusion peptide which includes a hydrophobic groove formed by residues Trp³⁴⁸ and spatially adjacent Ile³⁷², which constitutes a key component of the stem supersite. Notably, Phe³¹ in A1-B7 plays a critical role in this interaction—a feature commonly observed in other antibodies targeting this epitope, such as CR9114 (Dreyfus et al., 2012), MEDI8852 (Kallewaard et al., 2016), and FI6v3 (Corti et al., 2011). The highly conserved (345VDGW348, 345IDGW348 in H1N1 A/Puerto Rico/8/34) is a part of fusion peptide. Additionally, the Loop 2 of A1-B7 also contacts the antiparallel β-sheet epitope in the HA1 subunit (Fig. 5C). Amino acid alignment of the A1-B7 footprint revealed that its binding site within the fusion peptide region and A-helix of the HA2 subunit is highly conserved. However, the beta-sheet region of HA1 involved in binding is poorly conserved, which may explain its relatively lower activity against H3N2 (Fig. 5D).

To experimentally validate the predicted epitope, we performed epitope-substitution and alanine-scanning mutagenesis on HA. Following the replacement of “HSVNL” on the antiparallel β-sheet with “NATEL” derived from H3 or “HAKDI” derived from H5 located in HA1, we found that CR9114 retained effective binding, whereas the binding activity of A1-B7 was completely abolished (Fig. 5E–G). Meanwhile, reverse mutating “HAKDI” on H5 back to “HSVNL” restores strong binding of A1-B7. This indicates that this β-sheet constitutes one of the primary binding sites for A1-B7. Single point mutations within the partial fusion peptide and key residues on the A-helix all significantly reduced or completely abolished binding to A1-B7 (Fig. 5H, Supplementary Fig. S3D and S3E). In contrast, CR9114 is not sensitive to these mutations although its binding sites are similar. Likewise, after substituting the critical Phe³¹ on A1-B7 Loop 1, the binding to the antigen was also completely abolished (Fig. 5I), confirming the reliability of these structural predictions.

A1-B7 blocks low-pH induced conformational change and membrane fusion

The above results indicate that A1-B7 binds to the stem region and therefore does not directly block the interaction between HA and sialic acid receptors. Based on epitope studies, antibodies targeting the stem region typically exert neutralizing effects by inhibiting HA proteolytic activation and the membrane fusion step (Dreyfus et al., 2012; Kallewaard et al., 2016; Laursen et al., 2018). We observed that A1-B7 inhibits the low pH-induced membrane fusion by HA at the cellular level (Fig. 6A and B, Supplementary Fig. S4), found that A1-B7 does not inhibit proteolytic cleavage (Fig. 6C, D) precisely because its epitope does not cover the cleavage site. Therefore, we conducted an HA conformational change inhibition assay (Fig. 6E) and found that, similar to CR9114 (Laursen et al., 2018), A1-B7 can inhibit the low pH-induced conformational change of HA and the subsequent dissociation of the HA1 subunit (Fig. 6F and G). These findings indicate that A1-B7 exerts its neutralizing effect through inhibiting low-pH conformational changes of HA.

Fig. 6.

Neutralizing mechanism of A1-B7. A A1-B7 inhibition of cell membrane fusion assay. GFP images are shown. Cells without HA transfection (empty vector) and cells not subjected to low pH treatment serve as blank controls. CR9114 serves as positive control. C-Nb_bac serves as a negative control. B The inhibition rate measured by the renilla luciferase activity. Data is shown as mean ± SD of technical duplicates. C Schematic diagram of pyrolytic inhibition (Laursen et al., 2018). D Lane 1: HA0 without trypsin. Lane 2: HA digested with trypsin. Lane 3: HA digested with trypsin in the presence of A1-B7. E Schematic diagram of low pH induced conformational inhibition. Created in BioRender. Bihao, W. (2026) https://BioRender.com/kwpyd5r. F Detect the HA1 eluted by DTT, residual HA1 from the resin eluted by BXT buffer, and HA2 eluted by BXT buffer after low-PH induced conformational change, respectively. Lanes 1, 2, and 3 correspond to: without antibody, with A1-B7, and with CR9114, respectively. The data shown are representative of three independent experiments. G The inhibition rate of the antibody was calculated based on the efficiency of residual HA1 on resin normalized to HA2. The data are shown as mean ± SD. Statistical analysis was performed using one-way ANOVA (∗∗P < 0.01, ∗P < 0.05).

Discussion

Influenza treatment has long relied on neuraminidase inhibitors (e.g., oseltamivir) and more recently, the cap-dependent endonuclease inhibitor baloxavir marboxil. However, the emergence of resistant influenza strains poses a growing clinical concern. For oseltamivir, mutations such as H275Y, R293K, and N295S in the neuraminidase have been documented to reduce drug susceptibility (Kiso et al., 2004; Moscona, 2009; Bloom et al., 2010; Hurt et al., 2012). A recent study indicates that even the K130 N/E mutation occurring on HA can also confer resistance to oseltamivir (Zhang et al., 2025). Similarly, baloxavir resistance-associated mutations (e.g., I38 T/F/M in the PA subunit) have been detected in treated patients, particularly among children and in circulating strains, raising concerns about reduced therapeutic efficacy and potential transmission of resistant variant (Hayden et al., 2018; Omoto et al., 2018).

The mAbs have been demonstrated as alternative antiviral strategies. Several anti-influenza mAbs have entered clinical development. For example, broadly neutralizing antibodies targeting the hemagglutinin (HA) protein, such as MHAA4549A (Lim et al., 2021) and VIS410 (Sloan et al., 2020), have advanced to Phase II trials. In these trials, they have demonstrated a favorable safety profile and significant virologic activity in both human challenge models and naturally infected patients. Additionally, CT-P27 (a triple antibody combination) has shown prophylactic and therapeutic potential in early-phase studies (Sun et al., 2024). These antibodies generally target functionally critical and conserved regions, notably the HA stem, highlighting their potential as antiviral therapeutics. SdAbs constitute a versatile and innovative therapeutic platform, distinguished by a set of key attributes: small size, high stability, excellent tissue penetration, suitability for direct respiratory administration via intranasal or aerosolized inhalation, and amenability to multivalent engineering. When administered via the respiratory route, sdAbs offer distinct advantages for combating respiratory viruses like influenza virus: (1) directly targeting the primary site of infection (the respiratory tract), enabling high local concentrations; (2) ensuring a rapid onset of action; and (3) minimizing systemic exposure and associated side effects. A nanobody targeting RSV, ALX-0171, which is administered via nasal delivery, has entered clinical trials, demonstrating the potential of nanobodies in combating respiratory viruses (Larios Mora et al., 2018; Cunningham et al., 2021). Consistently, we recently characterized the pharmacokinetics of C-sdAbs in vivo (Li et al., 2025). Here, we isolated a C-sdAb, designated A1-B7, which not only exhibits potent and broad neutralizing activity against multiple H1N1 influenza strains in vitro, but also demonstrates robust prophylactic efficacy in vivo when administered intranasally. Although the therapeutic effect of intranasal administration is poor, pulmonary nebulization can achieve complete protection, indicating that different application scenarios require different routes of administration. A1-B7 recognize the stem region of HA. While AlphaFold 3 provides valuable structural insights, the predicted models are static snapshots and may have limited accuracy in flexible loops or non-conserved regions. Therefore, our epitope and mechanistic conclusions are primarily drawn from experimental data (alanine scanning, mutagenesis, and binding assays), with the computational model used as a guide. These results provide a proof-of-concept for the prophylactic and therapeutic potential of C-sdAbs against respiratory viruses. Future efforts will focus on the optimization of A1-B7 to enhance its neutralizing activity and expand its antiviral spectrum toward clinical translation.

Conclusions

In conclusion, this study identifies a potent C-sdAb against influenza H1N1 and provides compelling evidence for the feasibility of respiratory mucosal delivery of sdAbs as an effective prophylactic and therapeutic strategy. A1-B7 represents a promising candidate for further development as a convenient intranasal countermeasure capable of rapid action at the infection site.

Materials and methods

Animals

Female BALB/c mice (6–8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The animal center at Wuhan Institute of Virology (WIV), Chinese Academy of Sciences (CAS), provided housing and routine maintenance for these mice. According to the animal care and use regulations approved by the WIV institutional review board (ethics approval No. WIVA34202504), the infection experiments were conducted within the Animal Biosafety Level 2 (ABSL-2) Laboratory at WIV, CAS.

Cell and virus

Madin-Darby canine kidney (MDCK) cells (obtained from the National Virus Resource Center, WIV, CAS) were cultured in DMEM (Gibco) with 10% FBS (Gibco, Grand Island, NY, USA), 100 units/mL penicillin, and 100 μg/mL streptomycin (Gibco) at 37 °C and 5% CO₂. HEK293F cells (purchased from Invitrogen, Carlsbad, CA, USA) were cultured using Freestyle 293 expression medium (Invitrogen) in an incubator at 37 °C and 8% CO₂ with shaking at 120 rpm/min.

The H1N1 A/Puerto Rico/8/34, H1N1 A/human/Hubei/1/2009 (An et al., 2014) were provided from the virus collection at Wuhan Institute of Virology, CAS, China. The H1N1 and H3N2 clinical isolates were provided by Suizhou center for disease control and prevention. They were propagated in 10 day-old embryonated chicken eggs at 37 °C for 48 h. After virus growth, eggs were cooled overnight at 4 °C. The allantoic fluid recovered from the chilled eggs was centrifugated (4000g for 10 min at 4 °C) to remove cellular debris. The viruses were then divided into aliquots and stored at −80 °C.

Construction and expression of H1N1 HA ectodomain and CR9114

The codon-optimized gene encoding the HA from H1N1 A/Puerto Rico/8/34, A/New Caledonia/20/99, A/California/07/09, and A/Michigan/19/21(Laursen et al., 2018) were commercially synthesized and cloned into the mammalian expression vector pCAGGS with a C-terminal T4 fibritin trimerization motif, Avi tag, and His tag. For the conformational change assay, a flag tag was introduced at the N-terminal of HA A/California/07/09. Plasmids were transiently transfected into Freestyle 293 cells (293F) with polyethylenimine. After five days, the supernatants were harvested, filtered and purified by Ni-Excel according to a standard protocol. All the purified proteins were concentrated by a 10-kDa centrifugal filter device (Millipore, Burlington, MA, USA). Proteins were stored at −80 °C for further experiments.

The gene encoding CR9114 (IgG format) was synthesized and cloned into the vector pVITRO2-neo-mcs (InvivoGen, San Diego, CA, USA). Plasmids were transiently transfected into Freestyle 293 cells (293F) with polyethylenimine. After five days, the supernatants were harvested, filtered and loaded onto Protein A resin (GE HealthCare, Chicago, IL, USA) for purification according to a standard protocol.

Selection of candidate C-sdAbs

To select candidate C-sdAbs, we used a C-sdAb phage display library(Wang et al., 2021) for panning against HA. After four rounds of panning, polyclonal phage ELISA was applied to analyze the enrichment. Then, approximately 200 single colonies were picked for monoclonal phage ELISA and sequencing. Several parental clones were selected for the binding test.

Expression and purification of C-sdAbs

The candidate C-sdAbs were expressed according to a reported protocol (Gong et al., 2012a). In brief, the phagemid was transformed into E. coli HB2151 for expression under isopropyl β-d-1-thiogalactopyranoside (IPTG) induction for 14 h at 37 °C. Bacteria were harvested and resuspended in phosphate-buffered saline (PBS) supplemented with polymyxin B (Sangon Biotech, Shanghai, China) and oscillated for 45 min at 30 °C. After separating cell debris by centrifugation at 13,000g for 30 min, 0.45 μM-filtered supernatants were loaded onto Ni-NTA resin for purification according to a standard protocol. Fractions were pooled, concentrated, and buffer exchanged by a 3-kDa centrifugal filter device (Millipore).

Affinity maturation

To increase the neutralization activity of candidate clones, affinity maturation was performed according to a previously reported protocol (Zhang et al., 2004). Error-prone PCR was conducted for random mutagenesis in candidate C-sdAbs. A phage-display library with DNA fragments containing mutated genes of C-sdAbs was constructed. For sequential panning, different HAs were used, and the amounts of antigens were also reduced. The maturation procedures were repeated several times. One affinity-matured clone termed A1-B7 was finally identified after comparison with the parental clones.

Antigen binding measured by enzyme-linked immunosorbent assay (ELISA)

Protein antigens were coated in 96-well half-area microplates (Corning #3690) overnight at 4 °C. The plates were washed with PBST (PBS with 0.05% Tween 20 (Diamond Biotechnology, China) and then blocked with 3% milk (Bio-Rad) in PBS for 1 h at 37 °C. Fifty microliters of serially diluted candidate C-sdAbs in 1% milk were added and incubated for 1.5 h at 37 °C and detected by HRP-conjugated anti-Flag (Sigma-Aldrich, St. Louis, MO, USA) antibody. The absorbance at 405 nm was measured after incubation with ABTS substrate (Invitrogen). The data were plotted using GraphPad Prism, and the EC₅₀ (half maximal effective concentration) was calculated.

Authentic virus neutralization assay

MDCK cells were seeded in 96-well cell culture plates at 1 × 10⁵ cells per well and cultured overnight in an incubator at 37 °C and 5% CO₂. Final concentrations of antibody starting at 100 μg/mL and threefold-fold serially diluted were mixed with an equal volume of virus, and the mixture was incubated at 37 °C for 1 h. This mixture contains no serum components and is supplemented with TPCK-trypsin (Sigma) at a final concentration of 1 μg/mL. After removing the supernatant of MDCK cells followed PBS washing for twice, the antibody-virus mixture was added for infection. Cells were cultured at 37 °C and 5% CO₂ for 24 h. Observe the cytopathic effect (CPE) induced by the drug under a microscope. Take 40 μL of the supernatant and mix it with 20 μL of MES (2-Morpholinoethanesulphonic acid, MedChemExpress, Monmouth Junction, NJ, USA) buffer containing 2 μg/mL MUNANA (4-methylumbelliferyl-α-N-acetyl-neuraminate, MedChemExpress). Incubate the mixture at 37 °C for 45 min. Terminate the reaction by adding 100 μL of 0.14 μM NaOH per well as the stop solution. Measure the fluorescence value using a multifunctional microplate reader with an excitation wavelength of 355 nm and an emission wavelength of 485 nm. In the experiment, the wells with only maintenance medium added were cell-only control wells, and the wells with maintenance medium and virus dilution were virus control wells. Fluorescence values of each well were determined and plotted by GraphPad Prism. The IC₅₀ (half maximal inhibitory concentration) was calculated using nonlinear regression. All the infection experiments were performed in a biosafety level-2 (BSL-2) laboratory at WIV, CAS.

Animal experiment with intranasal administration

Mice were randomly divided into three groups: prophylaxis (n = 12), treatment (n = 12), and virus control (n = 12). Prophylaxis group: Mice were anesthetized with isoflurane and administered the antibody (25 mg/kg, each nostril received 10 μL of A1-B7 at a concentration of 25 mg/mL) intranasally 6 h before viral challenge. Treatment group: Mice received intranasal antibody (25 mg/kg, each nostril received 10 μL of A1-B7 at a concentration of 25 mg/mL) at 2 h, 24 h, and 48 h post-infection. Virus control group: Mice were administered PBS intranasally at the same time points. All mice were anesthetized with isoflurane and inoculated intranasally with 20 μL of influenza virus suspension (10 × LD₅₀ = 91 TCID₅₀). Anesthesia was maintained via intraperitoneal injection of tribromoethanol (200 μL/mouse). On day 4 post-infection, five mice from each group were euthanized. Lungs were lavaged with 1.5 mL of PBS containing 0.1% BSA for bronchoalveolar lavage fluid (BALF) collection. The left lung was fixed in 4% paraformaldehyde (PFA) for histopathological analysis (performed by Servicebio, Wuhan). Nasal turbinates were collected and homogenized for viral load analysis. The remaining mice (n = 7 per group) were monitored daily for body weight changes and survival for up to 14 days. For live virus titer measurement, the supernatants of BALF and Nasal turbinate homogenate after centrifugation were diluted from 1:10 to 1:100,000 and added to a monolayer of MDCK cells 100 μL per well as described in authentic virus neutralization. The limit of detection (LOD) was calculated as the reciprocal of the starting dilution divided by the inoculation volume in milliliters: LOD = 10/0.1 = 100 TCID₅₀/mL. To determine the viral RNA copy number, 140 μL of tissue homogenate supernatant was extracted according to the manual of the QIAamp® Viral RNA Mini Kit (Qiagen, Hilden, Germany). Using the HiScript® II One Step qRT-PCR SYBR Green Kit (Vazyme Biotech, Nanjing, China), amplification primers (F: CAGAGGTCTATTTGGAGCCA, R: GTGTGCTTTTTTGATCCGCT) were designed with the HA fragment to perform real-time quantitative reverse transcription PCR (real-time qRT-PCR) to determine the RNA copy number. The LOD was defined as the mean copy number calculated from the standard curve for the negative control wells.

Stability and nebulizing resistance testing

To examine the nebulization resistant capacity of A1-B7, A1-B7 was nebulized using a microsprayer nebulizer (Beijing Huironghe Technology Co., Ltd.) and applied to size exclusion chromatography (SEC) and ELISA to examine the monodispersity and binding activity. The SEC was conducted over a Superdex75 10/300 GL column (GE Healthcare) in PBS. We have included a stable C-type nanobody C-Nb_bac (Li et al., 2025) as a positive control as well as an initial clone H6 with insufficient stability, serving as a negative control. The expression and SEC of C-Nb_bac is consistent with A1-B7. To test the stability under the temperature, 0.22 μm filtered A1-B7 was incubated at 37 °C, 45 °C, 53 °C, 61 °C, and 69 °C for 30 min. To test the resistance of A1-B7 to extreme pH conditions, we incubated A1-B7 in buffers at pH = 1 and pH = 14 for 5 min, followed by replacement with PBS. The binding activity of each treated A1-B7 was tested by ELISA as described above.

Animal experiment with aerosol inhalation

Mice were randomly divided into three groups: A1-B7 treatment (n = 12), CR9114 treatment (n = 12), and virus control (n = 12). All mice were anesthetized with isoflurane and inoculated intranasally with 20 μL of influenza virus suspension (10 × LD₅₀ = 91 TCID₅₀). A1-B7 group: Mice were anesthetized with isoflurane and administered the antibody (25 mg/kg, 50 μL of A1-B7 at a concentration of 10 mg/mL) by aerosol inhalation at 2 h, 24 h, and 48 h post-infection. CR9114 group: Mice were anesthetized with isoflurane and administered the antibody (25 mg/kg, 100 μL of CR9114 at a concentration of 5 mg/mL) by intraperitoneal injection at 2 h. Virus control group: Mice were administered PBS by aerosol inhalation at 2 h, 24 h, and 48 h post-infection. Subsequent body weight monitoring, tissue collection, and viral load measurement were performed as described above.

Structural prediction and analysis

The complex structure of the A1-B7 in complex with influenza hemagglutinin (HA) (sequence of A/California/07/09 was used) was predicted using AlphaFold3 (Jumper et al., 2021; Abramson et al., 2024). The complex structure was predicted (AlphaFold Server, https://alphafoldserver.com) in multimer mode with five random seeds and full genetic search. The output with the highest predicted confidence was selected for modeling. Structural visualization and figure generation were performed with UCSF ChimeraX (Pettersen et al., 2021). Protein–protein interaction interfaces, including buried surface area and hydrogen bonds, were analyzed using the PDBePISA server (Krissinel and Henrick, 2007).

Competitive ELISA

The HA was coated on the plate overnight at 4 °C. A1-B7 2 nM was mixed with the competitors CR9114 at threefold serially diluted concentrations from 500 nM to 0.69 nM. The bound A1-B7was detected by HRP-conjugated anti-Flag antibody (Sigma-Aldrich). The ELISA procedures were performed as described above.

Binding assay of epitope substitution and alanine scanning by ELISA

To further verify the footprints involved in the interaction, the corresponding sequence of each epitope from different groups was introduced into H1N1 A/California/07/09 HA by overlap extension PCR. The “HSVNL” to “NATEL” and “HAKDI” substitution was based on H1N1 A/California/07/09 HA. The “NATEL” to “HSVNL” substitution was based on H3N2 A/Hong Kong/45/2018 HA. The “HAKDI” to “HSVNL” substitution was based on H5N6 A/Guangzhou/39715/2014 HA. Recombinant HA was expressed and purified. Binding to mutated HA was measured by ELISA as described above. Alanine scanning was performed based on preliminary information from the predicted structure to further determine the detailed residues significant to the interaction. A1-B7 bearing the mutation F31I was constructed and purified as described above, the binding activity to HA was measured by ELISA.

HA induced membrane fusion inhibition assay

To determine whether A1-B7 can inhibit HA-mediated membrane fusion, a membrane fusion assay was performed based on dual split proteins (DSP) (Kondo et al., 2010). Full-length HA and DSP1-7 were co-transfected into 293T cells, while DSP8-11 was transfected into MDCK cells. After 24 h post-transfection, A1-B7 were serially diluted and incubated at 37 °C with 293T cells expressing HA and DSP1-7. Subsequently, this system was mixed with MDCK cells expressing DSP8-11 with 1:1 and seeded onto a 96-well plate together. After 24 h, the pre-warmed low-pH treating buffer (10 mM MES with 150 mM NaCl, pH 5.0) (Su et al., 2008) was added to the co-cultured cells. After 8 min of incubation, it was quickly aspirated. The cells were then washed once, followed by replenishment with fresh culture medium for continued incubation. The GFP and bright field images were recorded using a fluorescence inverted microscope (Leica Microsystems, Wetzlar, Germany). Renilla luciferase detection was performed according to the instructions of the Renilla Luciferase Assay Kit (Beyotime Biotechnology, Shanghai, China) for quantitative analysis.

HA proteolysis assay

Both HA alone and HA with a 6-fold molar excess of A1-B7 were subjected to proteolysis using TPCK-treated trypsin at a final concentration of 2 μg/mL. The digestion was carried out overnight at 4 °C, followed by analysis via SDS-PAGE to assess the cleavage efficiency (Laursen et al., 2018).

HA conformational change assay

First, site-specific biotinylation of the HA protein with N-terminal Flag tag was performed targeting its Avi tag located at the C-terminus of HA2 (Biotin Labeling Kit for Avi-tag Protein with BirA, Beyotime). Proteolysis was then carried out as described above. 20 μg of the digested HA was bound to equal amounts of Strep-TactinXT resin (iba) from the same batch, followed by two washes with PBS. An excess amount of A1-B7 or CR9114 was added and incubated at 37 °C for 30 min, followed by two PBS washes. The resin was then treated with 0.1 M sodium acetate buffer (pH 5.0) at 37 °C for 10 min and washed twice with PBS. Subsequently, 40 μL of 20 mM DTT (BioFroxx, Pfungstadt, Germany) was added, incubated at 37 °C, and the eluate was collected. Finally, the resin was eluted by 40 μL elution BXT buffer (IBA Lifesciences, Göttingen, Germany). The sample of 20 μL was analyzed by SDS-PAGE and Western blotting. The shedding HA₁ was detected by HRP-conjugated anti-Flag antibody (Sigma-Aldrich). HA2 eluted from the resin was detected using an HRP-conjugated anti-His antibody (Protentech, Wuhan, China), and HA1 was detected by HRP-conjugated anti-Flag antibody. The inhibition rate was calculated by comparing the amount of residual HA1 after normalization to the amount of HA2 detected.

Data availability

All the data generated during the current study are included in the manuscript.

Ethic statement

All animal experiments were carried out in the Animal Biosafety Level 2 (ABSL-2) Laboratory at WIV, CAS in accordance with the WIV institutional review board guidelines for animal care and use (ethics number: WIVA34202504).

Author contributions

Bihao Wu: conceptualization, data curation, formal analysis, investigation, methodology, software, visualization, writing-original draft. Haiwei Zhang: data curation, investigation, methodology, validation. Xiangxing He: data curation, investigation, methodology. Zhe Zhang: data curation, investigation, methodology. Xiaohui Liu: conceptualization, project administration, resources, supervision. Rui Gong: conceptualization, funding acquisition, project administration, supervision, validation, writing-review & editing.

Conflict of interests

The authors declare no competing interests.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

Supplementary Figure S1.

In vitro activity of A1-B7. A Binding activity of A1-B7 to HA from H1N1 subtypes tested by ELISA. B Neutralizing activity of A1-B7 against H1N1 and H3N2.

Supplementary Figure S2.

Control antibody resistance to nebulization. A Size exclusion chromatography and binding activity changes of C-Nbbac pre and post nebulization. B Size exclusion chromatography and binding activity changes of H6 pre and post nebulization.

Supplementary Figure S3.

Confidence assessment of the AlphaFold-predicted structure of A1-B7 in complex with HA. A A representative predicted structural model of A1-B7 bound to HA. The pLDDT (predicted Local Distance Difference Test) score, a per-residue confidence metric ranging from 0 to 100, is color-coded on the structure (dark blue: very high confidence > 90; light blue: 70–90; yellow: 50–70; orange/red: < 50). The global pTM (predicted Template Modeling score) and interface ipTM (interface pTM) values are indicated next to the structure. pTM > 0.5 suggests correct global folding, while ipTM > 0.5 indicates a reliable inter-protein interaction interface. B Expected position error of the structure, typically presented as a paired error matrix or per-residue plot. This value (in Å) estimates the positional deviation (RMSD) for each residue relative to the true structure; lower errors (darker colors in a matrix) imply higher confidence in relative positioning, especially across the protein-protein interface. C SDS-PAGE of mutated HA with single Ala substitution. D–E Binding activity change of CR9114 and A1-B7 to mutated HA tested by ELISA.

Supplementary Figure S4.

Cell-cell fusion inhibition of A1-B7. The brightfield and merged images are shown.