A Novel Neutralizing Antibody Specific to the DE Loop of VP1 Can Inhibit EV-D68 Infection in Mice Free

Enterovirus D68 (EV-D68) is a major causative agent of mild to severe upper and lower respiratory illnesses and is prevalent in North America, Europe, and Asia in recent years (1–3). Patients infected with EV-D68 exhibit fever, sneezing, rhinorrhea, cough, sore throat, pneumonia, asthma exacerbation, and acute flaccid myelitis, with the possibility of death (2, 4–7). No prophylactic vaccine or therapeutic drug against EV-D68 infection is yet available. Potent Abs represent a new generation of antiviral agents for the prophylaxis and treatment of viral infection (8, 9). Thus, the induction of an Ab response capable of neutralizing EV-D68 isolates is urgently needed to address critical clinical conditions.

Many new mAbs have been successfully isolated and characterized by the microneutralization screening of single B cell sorting from infected patients (10, 11). As a widely used type of animal model, nonhuman primates such as Chinese rhesus macaques infected by a novel pathogen provide valuable information regarding the immune response to infection (12–16). The identification of the functional antisera elicited by infection, including binding capabilities and neutralizing titers, is vitally important for the study of novel pathogens (17, 18). Moreover, comprehensive assessment focusing on the quality, features, and antiviral activity of neutralizing Abs might be conducted to fully define the mAb candidates induced by an infection (19, 20).

EV-D68 is a nonenveloped, single positive-strand RNA virus with an icosahedron structure. There is a “canyon” region on the EV-D68 surface that contains the sialic acid receptor binding sites (21, 22). In addition, the VP1 protein in this canyon is an important structural protein for EV-D68 virus–dominant Ags that has also been used to classify virus genotypes. Despite the variety of strains of EV-D68, the strains detected recently have similar VP1 sequences as long as they belong to the same genetic lineage.

In this study, we report the isolation of a class of VP1-specific mAbs from an EV-D68–infected rhesus macaque and their neutralization activities. A6-1, the most potent of the isolated Abs, exhibited many promising characteristics including a specific Ag–binding capacity, affinity, and specific reactivity. In addition, biochemical binding studies and the characterization of the epitopes recognized by A6-1 demonstrate that its potency is mediated by its unique mode of recognition of a conserved site of vulnerability within the VP1-DE loop located in the canyon region by sequentially blocking the canyon region from binding with α2,6-linked sialic acids on the cellular surface through steric hindrance.

A6-1 protected suckling mice against intranasal infection by the EV-D68 virus. Because A6-1 binds to the DE loop of the EV-D68 VP1 protein, epitopes in this region may be promising candidates for vaccine design and may be the potential targets for inhibiting viral infection.

The EV-D68 Kunming (KM) strain (GenBank accession number: MG991260, https://www.ncbi.nlm.nih.gov/) and the EV-D68 Fermon strain (GenBank accession number: AY426531.1, https://www.ncbi.nlm.nih.gov/) are preserved by the Institute of Medical Biology (IMB), Chinese Academy of Medical Sciences (CAMS). The CA16 G20 strain and the EV71 FY23 strain are also preserved by the IMB, CAMS. The virus titer was 6.5 log cell culture infectious doses (CCID₅₀) per milliliter and was diluted to a suitable concentration for the subsequent experiments. Vero and HeLa cells were acquired from the American Type Culture Collection.

A series of 60 peptides spanning the entire EV-D68 KM VP1 region of EV-D68 were synthesized. Each peptide contained 15 aa residues and had 10 of the same amino acid residues as each adjacent peptide. Other 15-aa peptides of the Fermon strain, which is different from the KM strain in the VP1 gene sequence, were synthesized. In addition, nine peptides covering the 126–135 aa of the VP1 protein were synthesized, each with a single alanine mutation replacement.

We labeled the EV-D68 VP1 protein or A6-1 mAb according to the Lightning-Link Fluorescein Conjugation Kit instructions (Innova Biosciences, Baraham, Cambridge, U.K.). Before we added purified VP1 protein (1 mg/ml) to the Lightning-Link mixture, we added 1 μl of LL-modifier reagent for each 10 μl of VP1 protein to be labeled; then we mixed gently. We pipetted the VP1 sample (with added LL-modifier) directly onto the Lightning-Link mixture. We resuspended gently by withdrawing and redispensing the liquid once or twice using a pipette. After incubating the mixture at room temperature (20–25°C) in the dark for 3 h (or more), we added 1 μl of LL-quencher FD reagent for every 10 μl of VP1 protein used. After 30 min, the conjugate could then be used.

All animal experiments maintained the replacement, refinement, and reduction ethical principles. The animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the IMB, CAMS. Two healthy 6-mo-old rhesus monkeys (Macaca mulatta; identifier 15247 weight of 1.15 kg is female, identifier 15249 weight of 1.2 kg is male) were administered an EV-D68 virus suspension with 4.5 log CCID₅₀ titer via nasal delivery. Each monkey was kept in a single cage and fed according to IACUC guidelines. The animal keepers recorded the weight and body temperature of the monkeys daily and collected blood, nasal swabs, and fecal specimens from both monkeys at 1, 3, 5, 7, 9, 11, 14, and 28 d postinfection (dpi). The anti–EV-D68 Ab in both monkeys was confirmed as negative before the study was initiated.

Total RNA was extracted from different cell cultures, fresh tissue homogenates, or nasal washes from the experimental animals using the TRNzol-A⁺ Reagent Mini Kit (Tiangen Biotech, Beijing, China) according to the manufacturer’s instructions. The EV-D68 RNA viral loads were determined using a one-step quantitative real-time RT-PCR (qRT-PCR) assay as previously described (23).

For cell surface staining, fresh PBMCs were resuspended in 100 μl of FACS staining buffer (3% BSA in PBS) containing Abs against CD20, CD27, and VP1/FITC-labeled Ags and incubated for 30 min at room temperature in the dark. The stained cells were analyzed and sorted by a BD FACSJazz cell sorter (BD Biosciences) into 96-well plates containing 20 μl of cell lysis buffer according to the gating strategy. The CD27⁺ cells were gated from CD20⁺ populations that excluded CD3⁺ cells. Then, the EV-D68 VP1 Ag–specific memory B cells were defined as CD27⁺VP1⁺. All flow cytometry data were analyzed by FlowJo V10 software (FlowJo, Ashland, OR). Row H of the 96-well plate was used as a blank control and contained only lysis buffer from the Takara One Step cDNA Synthesis Kit (Takara Bio, Japan). The sorted target cells were stored at −80°C.

The sequences of the monkey Ig H chain variable gene (V_H) and L chain variable gene (V_L) chains were amplified by RT-PCR and nested PCR using the methods and primers in Table I (24). In brief, Ig V_H and V_L were amplified in 96-well PCR plates containing sorted single B cells. The first-strand cDNA was synthesized according to the PrimeScript II 1st Strand cDNA Synthesis Kit (Takara Bio), and Ig H chain, Ig lambda chain, and Ig kappa chain were amplified by two rounds of PCR under the following conditions.

The first round of H chain PCR contained 5 μl of the reverse transcription (RT) reaction, 16 μl of PrimeSTAR Max Premix (2×), 1 μl of each forward primer, and 1 μl of each reverse primer. The second round of H chain PCR contained 5 μl of the first-round reaction, 19 μl of PrimeSTAR Max Premix (2×), 1 μl of each forward primer, and 1 μl of each reverse primer. The first-round procedure was as follows: 98°C for 1 min; 50 cycles of 98°C for 10 s, 55°C for 5 s, and 72°C for 5 s; and finally, 72°C for 5 min. The second-round procedure was as follows: 98°C for 1 min; 50 cycles of 98°C for 10 s, 60°C for 5 s, and 72°C for 5 s; and finally, 72°C for 5 min.

The first round of L chain λ PCR contained 6.5 μl of the RT reaction, 15 μl of PrimeSTAR Max Premix (2×), and 0.8 μl of each forward primer. The second round of L chain PCR contained 6.5 μl of the first-round reaction, 15 μl of PrimeSTAR Max Premix (2×), 0.8 μl of each forward primer, and 0.8 μl of each reverse primer. The first-round procedure was as follows: 98°C for 2 min; 50 cycles of 98°C for 10 s, 56°C for 5 s, and 72°C for 5 s; and finally, 72°C for 5 min. The second-round procedure was as follows: 98°C for 2 min; 40 cycles of 98°C for 10 s, 60°C for 5 s, and 72°C for 5 s; and finally, 72°C for 5 min.

The first round of L chain κ PCR contained 6 μl of the RT reaction, 15 μl of PrimeSTAR Max Premix (2×), 1.0 μl of each forward primer, and 1.0 μl of each reverse primer. The second round of L chain κ PCR contained 6 μl of the first-round reaction, 20 μl of PrimeSTAR Max Premix (2×), 1 μl of each forward primer, and 1 μl of each reverse primer. The first-round PCR procedure was as follows: 98°C for 2 min; 45 cycles of 98°C for 10 s, 54°C for 5 s, and 72°C for 5 s; and finally, 72°C for 5 min. The second-round procedure was as follows: 98°C for 2 min; 45 cycles of 98°C for 10 s, 60°C for 5 s, and 72°C for 5 s; and finally, 72°C for 5 min.

The IGHV, IGKV, and IGLV subgroup distributions and the CDR3 length of the IGV domain sequences were analyzed with IMGT/V-QUEST (http://www.imgt.org). IgG expression vectors (pCMV-Rh) contain a monkey Ig gene leader peptide and multiple cloning sites upstream of the monkey Ig H chain, Ig kappa chain, and Ig lambda chain constant regions. The methods for Ab expression in mammalian cells and purification with Protein A are as follows. Briefly, a 3:2 ratio of the molar amounts of H chain plasmid and L chain plasmid were cotransfected into 293T cells for transient expression with Lipofectamine 2000 (Thermo Fisher Scientific). The supernatants were harvested 4 d after transfection. The Abs were purified with Protein A beads (Thermo Fisher Scientific) according to the manufacturer’s instructions. The purified EV-D68 virion was analyzed by SDS-PAGE. Briefly, the proteins were boiled for 10 min and then separated on 10% Tris-glycine SDS-PAGE gels that were then stained with silver or transferred to PVDF membranes for Western blot analysis. After blocking with 5% nonfat milk in 0.1% Tween 20 in PBS (PBST) overnight at 4°C, the blots were incubated for 2 h at room temperature with A6-1 (4 μg/ml). After washing in PBST, the PVDF membrane was detected by HRP-conjugated rabbit anti-human IgG (1:5000; Abcam, Cambridge, MA) for 45 min at room temperature. The signals were visualized with ECL Western blot substrate reagents.

Briefly, the 96-well plates were coated with 0.2 μg/well recombinant VP1 from EV-D68 and incubated overnight at 4°C. The plates were then blocked with 5% dry milk in PBST at room temperature for 2 h. After three washes with PBST, samples (polyclonal serum and expressed supernatant) were added to the plates, which were then incubated for 2 h. Afterwards, the plates were washed five times with PBST and incubated for 1 h with a rabbit anti-human IgG H&L HRP-conjugated secondary Ab (Abcam). The assay was developed with TMB substrates (Solarbio Science and Technology) according to the manufacturer’s instructions. The OD were measured at 450 nm. Positive binding was defined as at least 2.5-fold above the OD of the negative control. Binding capability was also tested with the same ELISA with a 4× serial Abs dilution. For peptide mapping, the plates were coated with 0.4 μg/well of each of the EV-D68 VP1 peptides at 4°C overnight and blocked with PBST containing 5% BSA for 2 h at room temperature, followed by 0.1 μg of A6-1 mAb (diluted in 100 μl of PBS) for 2 h at room temperature. After each time of incubation step, the plates were washed five times with PBS. The OD were measured at 450 nm. Positive binding was defined as at least 2.5-fold above the OD of the negative control.

In this study, the neutralizing Ab titer produced by the two rhesus monkeys was detected by a neutralization assay. Before this test, all serum samples were inactivated at 56°C for half an hour in a water bath. The cell culture medium was formulated as follows: 10% new bovine serum, 1.5% penicillin/streptomycin, 2% glutamine solution (3%), 3% NaHCO₃ (6.6%), and 83.5% MEM. We first placed 50 μl of the cell culture medium (2% new bovine serum) in an empty 96-well plate and then added 50 μl of 2-fold diluted serum sample to each well of the first row. We diluted the serum as follows: we transferred 50 μl of liquid from each well in the first row to the corresponding well in the second row, then transferred 50 μl of liquid from each well in the second row to the corresponding well in the third row, and so on until the eighth row. We discarded 50 μl of liquid from each well in the eighth row. Then, the diluted serum samples were incubated for 1 h with 100 CCID₅₀ of EV-D68 virus. HeLa cells digested with 2.5% trypsin were added to each well at a density of 2 × 10⁴ cells/well and cultured in a 37°C incubator with 5% CO₂. The cytopathic effect (CPE) was tested for 7 d, and the CPE wells were counted daily.

In the standard neutralization assay, 50 μl of A6-1 mAb was incubated with 50 μl of the EV-D68 KM strain (100 CCID₅₀) for 1 h at 37°C; after which, the mixture was added to the HeLa cells. The cells incubated with the mixture were cultivated for 72 h at 37°C, and the viability was detected with a 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) reagent (Beijing Solarbio Science and Technology). The neutralization efficiency of each sample was calculated as follows: percentage of neutralization = [(OD₄₉₀ of each sample − OD₄₉₀ of the cells treated with virus only)/(OD₄₉₀ of the cells untreated − OD₄₉₀ of the cells treated with virus only)] × 100. The IC₅₀ represents the concentration of mAbs that inhibit cell death by 50% (25).

The mAbs and virus were mixed and incubated at 37°C for 1 h, and the mixture was then incubated with the cells for 2 h at room temperature. Next, the cells were washed three times with PBS and then cultivated at 37°C for 72 h. The cell viability and neutralization efficiency for each sample was determined as previously described (25).

First, 50 μl of the EV-D68 KM strain (100 CCID₅₀) in DMEM was added to the HeLa cells at room temperature for 1 h so that the virus could attach to the cell surface. The cells were washed three times to remove unbound viruses and were then incubated with 50 μl of A6-1 at 37°C for 1 h. The cells were then washed three times with PBS. After 72 h, the cell viability and the inhibitory efficiency were determined as previously described (25).

The inhibition of the mAb-mediated neutralization of EV-D68 infection by synthetic peptides (peptide 1 and peptide 25) was determined by a neutralization inhibition assay. Briefly, different amounts of peptides and A6-1 mAbs were mixed together for 1 h at 37°C. The mixtures were incubated with 100 CCID₅₀ of the EV-D68 KM strain virus for 1 h at 37°C. The A6-1-peptide/virus mixture was added in the cells. After incubation at 37°C for 72 h, the neutralization inhibition assays were evaluated using an MTS-based kit (Beijing Solarbio Science and Technology). The OD was determined at OD₄₉₀ with a microtiter plate reader. The neutralization inhibition efficiency (100%) = {1 − [(OD₄₉₀ of each sample − OD₄₉₀ of the cell treated with virus only)/(OD₄₉₀ of the cells untreated − OD₄₉₀ of the cells treated with virus only)]} × 100.

The experimental procedures were the same as those of the microneutralization assays evaluated by CPE.

Different amounts of A6-1 were mixed with 10^6.5 CCID₅₀ of EV-D68 virus in DMEM (containing 2% FBS) and then incubated at 37°C for 1 h before being added to Vero or HeLa cells (2 × 10⁵ cells/well in 12-well plates). The cells were incubated with the A6-1/virus mixture for 2 h at room temperature. After we wash the cells three times, the cell-attached viruses were used for qRT-PCR and immunofluorescence analysis. For quantification of viral loads by qRT-PCR, the cells attached to the EV-D68 virus were harvested for extracting viral RNA extraction; EV-D68 viral RNA copies were detected using the same procedure described in the section on viral load detection in rhesus monkey infection. For immunofluorescence analysis, slides of the cells were incubated with 4 μg/ml A6-1 labeled with the Lightning-Link Fluorescein Conjugation Kit (Innova Biosciences) according to the protocols provided in the kit and 5 μg/ml biotinylated Sambucus nigra lectin (α2,6-linkage) (Vector Laboratories) at room temperature for 2 h. Then, the slides were washed three times with PBS and incubated with 5 μg/ml rhodamine avidin (Vector Laboratories) for 1 h at room temperature. Finally, the slides were observed by confocal laser microscopy.

PCR-amplified EV-D68 VP1 gene fragments (bp 1–453, bp 301–750, and bp 628–927) were cloned into the pSNAP_f Vector (New England Biolabs). The A6-1 mAb plasmid was cotransfected into 293T cells with the above VP1 fragment plasmids. The fluorescence lifetime imaging (FLIM)/fluorescence resonance energy transfer (FRET) and stimulated emission depletion (STED) microscopy imaging experiments were performed 36 h after transfection. After 36 h of transfection, a labeling medium consisting of 3 μM SNAP-Cell TMR-Star dye (New England Biolabs) was used to replace the medium on cells expressing a SNAP fusion protein, which were then incubated at 37°C with 5% CO₂ for 10 min. The cells were then washed three times with complete culture medium containing 10% FBS and were then incubated in fresh medium for 30 min. Next, the medium was replaced once to remove unreacted SNAP-tag substrate that had diffused out of the cells. Then, the cells were fixed with 4% paraformaldehyde (Beijing Solarbio Science and Technology) for 15 min at room temperature and rinsed three times with PBS. Next, the cells were permeabilized with 0.1% Triton in PBS for 10 min, washed three times with PBS, and then blocked with 10% FBS in PBS overnight at 4°C. The next day, the cells were incubated with goat anti-human IgG Fc (DyLight 488) (Abcam) for 1 h at room temperature. Finally, the cell nuclei were counterstained with DAPI (Abcam), and the slides were kept at 4°C until use. For FLIM/FRET experiments (26), the “donor only” sample consisted of the same cells transfected with only the A6-1 mAb plasmids. Then, the average fluorescence lifetimes of the donor-only and FRET samples were measured in fast FLIM mode. The measurement was completely controlled via Leica Application Suite Advanced Fluorescence. The FRET efficiency was calculated from the ratio of the FRET donor lifetime τ_quench and the non-FRET lifetime τ_quench as follows:

For the STED microscopy imaging, the cell slides were visualized and captured with an appropriate filter set under a Leica TCS SP8 microscope equipped with a STED laser (Leica Microsystems). Finally, we analyzed the colocation between the A6-1 mAb and different VP-truncated proteins.

The mice protection experiments were approved by the IACUC of the IMB, CAMS. VRC01 used as the irrelevant Ab in this study is a neutralizing human mAb against HIV-1 (27). Three groups of 2-d-old suckling mice were i.p. injected, respectively, with purified A6-1 Ab (30 μg/mouse), VRC01 (30 μg/mouse) as irrelevant Ab, or PBS 12 h before the nasal administration of the virus. For viral infection, 10 μl of the EV-D68 KM strain (5 × 10⁵ CCID₅₀/ml) was micropipetted onto the noses of each suckling mouse two times with a 30-min interval between exposures. Then, three mice in each group were randomly sacrificed at four time points (0, 3, 7, and 10 dpi) to measure the viral load. The lung and brain tissues of the mice were sampled for histopathological examination and immunofluorescence analysis at 0, 3, and 7 dpi.

EV-D68 acquired from lung or brain tissues of virus-infected suckling mice were analyzed by a microtitration assay (28). In brief, the virus supernatant was extracted from lung and brain homogenates. Then, the supernatants were passed through a 0.22-μm sterile filter, and HeLa cells in 96-well plates were inoculated and cultivated with the supernatants at 37°C for 7 d. The CPE was observed each day.

The brain and lung tissue samples were embedded in an optimal cutting temperature compound and frozen in liquid nitrogen. The frozen tissues were cut into 6-μm sections and were placed on the poly-ι-lysine–coated glass slides and fixed in 4% paraformaldehyde. The glass slides containing tissue samples were stained with H&E according to the manufacturer’s instructions (Beijing Solarbio Science and Technology). For the EV-D68 viral immunofluorescence analysis, the prepared frozen sections were stained with anti–EV-D68 VP1 Ab (GeneTex), followed by a goat anti-rabbit secondary Ab labeled with Alexa Fluor 647 dye (Thermo Fisher Scientific) and observed by confocal microscopy (Leica Microsystems) (29).

We established an animal model of Chinese rhesus macaques infected with the KM strain of EV-D68. Two 6-mo-old rhesus macaques were successfully infected via the nasal tract (10^4.5 CCID₅₀/monkey). To assess the Ab responses against EV-D68 in these two animals, plasma samples harvested during a viral exposure challenge were assayed for their neutralizing activity. The viral replication was typical of EV-D68 replication during a primary infection (23, 30), and the virus replication kinetics did not differ significantly between these two animals.

We observed that the viral loads in both rhesus monkeys began to increase on the third day after EV-D68 infection. The average viral load in the blood peaked at 4740 copies/ml 9 dpi (Fig. 1A). However, the viral loads in the nasal washes and feces peaked on the fifth day, at 14,866 copies/100 mg of feces and 2697 copies/ml in the nasal washes, and then decreased rapidly after the seventh day (Fig. 1B). Either the viremia in the blood or the high viral load in the nasal swabs and feces was sufficient to indicate that both monkeys were infected by EV-D68, although the clinical features were slight (data not shown).

For the immune reaction induced by the viral infection, the neutralizing Ab against EV-D68 was detected positively 14 dpi in both rhesus monkeys, with a titer of 2³ and 2⁴ (Fig. 1C), whereas the total Ab binding to EV-D68 as analyzed by ELISA also showed increasing OD₄₅₀ values of 0.375 and 0.225 (Fig. 1D). These results are similar to our previous findings in EV-D68–infected ferrets (23).

Based on the neutralizing Ab responses, we performed FACS on PBMCs from donor monkeys 15247 and 15249. Compared with the very small percentage of memory B cells (CD20⁺/CD27⁺/EV-D68 VP1⁺) detected in monkeys before EV-D68 infection, ∼3.56 and 2.02% of the memory B cells from donor monkeys 15247 (Fig. 1E) and 15249 (Fig. 1F), respectively, were VP1 specific, indicating an obvious increase in EV-D68 Ag presentation by BCR after EV-D68 infection.

To determine the Ab properties of the isolated memory B cells, we amplified the Ab gene by cDNA synthesis from single memory B cells followed by nested PCR (24, 31, 32). Two rounds of specific PCR were performed with two sets of optimal primers for monkey IgG genes (Table I). Of the 301 single cells tested, 31 cells produced both H and L chain products, representing a 10.3% cloning efficiency among the positive wells.

To assess the proportions of cloned IgGs that could recognize EV-D68 immunogens, we expressed these 31 IgGs in a 293T cell system. All mAbs secreted into the cell culture supernatant were used for the VP1-binding ELISA screening (Fig. 2A). Of the total Abs tested, A6-1, E2-2, F4-3, and D7-4 showed obvious positive Ag binding against purified EV-D68 VP1. We then assessed whether A6-1 could bind to KM EV-D68 viral proteins by Western blot analysis because A6-1 bound to VP1 with the highest binding capability. The results demonstrated that mAb could bind to EV-D68’s VP1 with specificity, indicating that the A6-1 candidates could also recognize viral particles (Fig. 2B).

We next explored the EV-D68–neutralizing capacity of these four Ab candidates. A6-1 fully neutralized two genotypes of the EV-D68 strain (Fermon as genotype A and KM as genotype B), achieving 100% inhibition, with IC₅₀ values of 0.6 μg/ml for the KM strain (Fig. 3A) and 1.57 μg/ml for the Fermon strain (Fig. 3B). A6-1 exhibited greater potency than the F4-3, E2-2 and D7-4 mAbs, which all had similar IC₅₀ values against both strains (in the range of 2–6 μg/ml); none of these three Abs achieved a maximum percent inhibition (MPI) of 100%.

Notably, the H chain complementarity-determining regions (CDRHs) of mAbs F4-3, A6-1, and E2-2 pertain to the progenitor IGHV3.9. An analysis of the variable Ab genes revealed that these three pairs of Abs shared almost identical somatic mutations, with 36–37% of the mutations in the V_H. F4-3 and A6-1 possess 18–19% of the Λ V_L, which is different from the speculative germline gene sequences; the L chain of E2-2 was derived from a different germline, VK4.1. All three Abs contained a relatively long CDRH3 loop of 21–22 aa (Fig. 3C) and mutations at residue F97 and S109 in A6-1 (Fig. 3C), which likely introduced potential differences into their binding affinities. The results of ELISA indicated that A6-1 still displayed efficient affinity to VP1 with a low concentration of 0.05 μg/ml compared with the other three mAbs (Fig. 3D). We also noticed that A6-1 and F4-3 shared the same germline with only 2 aa differences in the H chain. Moreover, compared with F4-3, the homology-modeling three-dimensional structure of A6-1 indicated an S109 deletion in CDRH3, forming a longer loop that affected its neutralizing activity (Fig. 3E).

The H chain of D7-4 was derived from a different progenitor, V_H1.36, but its L chain was derived from the same V_L3.10 alleles as A6-1. D7-4 had a lower somatic mutation rate, with 15% of V_H and 19% of V_L nucleotides different from the putative germline gene sequences. The CDRH3 of D7-4 contained 14 aa. Comparatively, A6-1 from the progenitor IGHV3.9 showed the best neutralizing activity against EV-D68, suggesting its highlighted potency for our further study (Fig. 3C).

Long CDRH3 loops are relevant to Ab specificity and cross-reactivity (33); therefore, we tested all four Abs for reactivity against a panel of two genotypes of EV-D68 and two other enteroviruses, EV71 and CA16. None of these four Abs showed any substantial binding capability or neutralization activity toward the EV71 or CA16 strains tested, confirming the specificity of these Abs to target EV-D68 (Fig. 3F, Table II).

We speculated that A6-1, E2-2, and F4-3 might recognize the same or similar epitopes because these three Abs are somatic variants with many similarities. As a representative, we used A6-1 to examine the binding affinity with VP1 subregions (VP1-1, aa 1–151; VP1-2, aa 101–250; and VP1-3, aa 210–309) by STED. Confocal imaging showed that A6-1 bound strongly to EV-D68 VP1-1 and VP1-2, which both overlap with the DE loop region at 101–151 aa, but did not bind strongly to VP1-3 (Fig. 4A, 4B).

We next investigated if the great neutralization potency of A6-1 resulted in a high binding affinity to specific Ags VP1, VP1-1, VP1-2, and VP1-3 of the EV-D68 Fermon and KM strains using a FLIM/FRET assay (27, 34–36). FRET can be observed in the FLIM/FRET process; FRET activity shortens the lifetime of the donor fluorophore when an acceptor fluorophore exists because of energy transfer. Therefore, by immunostaining a VP1-truncated protein (acceptor) and an A6-1 mAb (donor) with fluorophores at emission wavelengths of 580 and 524 nm, respectively, the distance between the VP1-truncated protein and the A6-1 protein can be detected by FLIM through calculating the fluorescence lifetimes of the donor fluorophores. We can measure the proximity between two proteins by calculating the FRET efficiency, revealing the characteristics of protein/protein interaction. The FRET efficiency values of A6-1 for KM VP1, VP1-1, VP1-2, and VP1-3 were 30.3, 33.7, 14.8, and 5%, respectively (Fig. 4C), and the FRET efficiency values of A6-1 for Fermon strains VP1, VP1-1, VP1-2, and VP1-3 were 21.5, 21.2, 14, and 2.4%, respectively (Fig. 4D). At the same time, the change in the pseudo-color images of A6-1 (donor) from green to blue demonstrated that the donor’s fluorescence lifetime decreased with increasing FRET efficiency (Fig. 4C bottom, Fig. 4D bottom). The FRET efficiency values showed that the distances between the A6-1 mAb and VP1 or VP1-1 from different strains were less than those of the other truncated VP1 proteins, demonstrating that A6-1 had a better affinity for VP1-1 and possibly explaining why it had the highest neutralizing activity against EV-D68.

To further map the A6-1 epitope within VP1, we screened a panel of 60 overlapping peptides spanning the sequences of the entire VP1 gene of the EV-D68 KM strain for reactivity. As shown in Fig. 4E, A6-1 reacted strongly with peptide 25 (aa 121–135) but less so with other peptides. Considering certain differences in the sequences of the BC, DE, and GH loops in the EV-D68 Fermon strain VP1 gene, we also screened an extra panel of 15 peptides covering these regions of the Fermon strain (Fig. 4F). The A6-1 also bound to the same region with the sequence RFDAEITILT (Fig. 4G), which harbors the neutralizing immunogen sites NIm-IA and NIm-IB (22).

To assess the importance of individual binding site residues for the epitope recognition of the A6-1 Ab, alanine scanning analysis was performed against peptide 25. As shown in Fig. 4H, alanine substitution at residues 127, 128, 130, 132, 134, and 135 did not strongly affect the binding of the corresponding peptides to A6-1, whereas point mutations at positions 126, 131, and 133 weakened A6-1 binding considerably, indicating that the three amino acids R126, I131, and I133 were crucial for A6-1 binding (Fig. 4H). Considering that A6-1 recognized the VP1-1 in cells and bound the DE loop in the epitope screening, we speculated that the specific peptide corresponding to the DE loop affected the A6-1–mediated neutralization of the KM and Fermon strains of EV-D68.

To determine whether these Ab-binding peptides were indeed involved in Ab recognition and function, two representative peptides, 1 and 25, were used to block A6-1 neutralization. As shown in Fig. 5A, a low concentration (2.5 μg/ml) of peptide 25 significantly reduced the neutralization activity of A6-1, whereas as a control sample, peptide 1 had no significant inhibitory effect, even at the highest concentration used (20 μg/ml) (Fig. 5A).

To further understand the potential mechanism of the inhibitory functions of A6-1, we investigated whether A6-1 could block the binding of EV-D68 to its receptors. An α2,6-linked sialic acid is an attachment receptor for EV-D68 (22, 37), and we examined whether A6-1 treatment could interfere with the interaction between the virus and the receptor. A6-1 treatment potently reduced the amount of the virus pulled down in an Ab dose-dependent pattern, resulting in an up to 80% decrease when ∼2.5 μg of A6-1 was incubated with HeLa cells (Fig. 5B). A similar trend of Ab dose-dependent inhibition was also observed in Vero cells (Fig. 5C), indicating that A6-1 can interfere with EV-D68 binding to its α2,6-linked sialic acid receptor and thereby inhibit virus entry. In the different cell types (HeLa or Vero cells) treated with the virus only (Fig. 4D, 4E middle), the cell surface was surrounded by the EV-D68 virus (green), whereas the amount of virus attached to the cells significantly decreased when the cells were treated with the A6-1 and virus mixture (Fig. 4D, 4E bottom). This indicated that A6-1 prevented the adhesion of the EV-D68 virus to Vero and Hela cells. Moreover, colocalization of the virus to the α2,6-linked sialic acid (red) showed that A6-1 reduced the attachment of the virus to the a2,6-linked sialic acid. In conclusion, A6-1 interferes with the attack of EV-D68 on α2,6-linked sialic acid (Fig. 4D, 4E).

To more comprehensively understand the possible multiple molecular mechanisms by which A6-1 inhibits EV-D68, we further examined the preattachment and postattachment effects by preincubating the Ab and virus together before applying them to the cells and by applying the Ab to the cells pi with the virus. The Ab pretreatment clearly rendered the virus highly sensitive to A6-1 in the KM strain, with an IC₅₀ of 1.39 μg/ml (Fig. 5F), whereas posttreatment did not exhibit any neutralization (Fig. 5G) even though A6-1 bound to the virion attached to the target cells (Figs. 5H, 6). This indicated that the Ab did not inhibit the virus with postattachment neutralization manner. Together, these data show that the primary anti–EV-D68 activity of A6-1 arises from blocking the binding of the virus to its receptor.

EV-D68 causes infection and death in some reports of mouse infection models (38–41). Thus, we evaluated the protective efficacy of A6-1 in 2-d-old suckling C57BL/6 mice. After nasal challenging with EV-D68 KM strain, the virus was undetectable from 0 through 10 dpi in the mice treated with A6-1. However, in the PBS or irrelevant Ab–treated group, the virus can be detected in the lung and brain tissues at 3 dpi and showed a further increasing at 7 dpi and then declined, becoming undetectable at 10 dpi (Fig. 7A, 7B). Meanwhile, compared with the A6-1–treated mice, in histopathological changes such as hemorrhaging, inflammatory cells aggregation were observed in the lung tissue at 3 or 7 dpi in PBS or irrelevant Ab–treated group. Similar results were found in brain tissue, with an observation of infiltration of inflammatory cells and accumulation of the glial cells at 3 or 7 dpi (Fig. 7C). No obvious pathological changes were observed in the brain tissue of the mice treated with A6-1 (Fig. 7C). These results indicated that A6-1 inhibited the EV-D68 intranasal infection and alleviated the pathological outcomes of the infected mice.

We isolated a class of EV-D68 entry–blocking neutralizing Abs from infected monkeys. Low concentrations of these mAbs neutralized KM and Fermon strains in vitro. One mAb, A6-1, was also effective in an EV-D68 intranasally infected mouse model. To the best of our knowledge, no previous study has reported any mAbs that could neutralize EV-D68 isolates.

In this study, we infected Chinese rhesus macaques with the EV-D68 virus and isolated their mAbs via single memory B cell sorting using FACS, RT-PCR, and nested PCR of their PBMCs after 14 d of infection. Of the total isolated Abs induced by EV-D68 infection, 10.3% percent were specific to VP1.

The mAbs isolated in our study exhibited unique features and different activities. The H chains of F4-3, A6-1, and E2-2 possess extensive somatic mutation compared with the D7-4 Ab, resulting in a higher affinity for VP1 of EV-D68. Moreover, A6-1 exhibited a highly potent neutralizing activity, reaching 100% MPI. Previous studies showed that potent neutralizing activity is usually associated with a high affinity (42–44). The high affinities of the A6-1 identified in this study also render it effective against EV-D68.

We attribute the A6-1’s antiviral activity to its attack on the EV-D68 receptor. The Ab binds with a high capacity to EV-D68 VP1-1, thereby interfering with its interaction with the target cell surface for viral entry. Our results (Fig. 5D, 5E) and recent studies (21, 23, 37, 45) show that EV-D68 prefers to recognize α2,6-linked sialic acid receptors. According to recent research (21), the structure of EV-D68 shows that its surface depression (canyon) consists of VP1 protein, which is the site of receptor binding (22, 46, 47). Although the floor of the canyon formed by the GH loop of VP1 and partial segments of VP3 is an important position for viral attachment to the cell receptors (21, 48, 49), the BC loop and DE loop of VP1 are also important to its binding, because both are near the canyon and close to 5-fold axes (22, 50). In this study, the A6-1–binding epitope AEITILT (aa 129–136), located in the DE loop region (22), is conserved in A, B, and C clades of EV-D68 (22, 51). Whereas this region is near the binding site of the sialic acid receptor that exhibited highly potent neutralizing activity reaching 100% MPI, we conclude that A6-1 may interfere with EV-D68, attaching to the α2,6-linked sialic acid host entry receptor by binding to the DE loop, causing a steric hindrance and thereby achieving viral neutralization (Fig. 6).

The identification of the DE loop of EV-D68 as a primary epitope suggests the possibility that the mechanism of action of A6-1 involves targeting a specific Ag location and causing a steric clash. Notably, A6-1 has a relatively long CDRH3, containing 21 aa, which might affect the steric constraints of the DE loop and thereby sterically disrupt the key step in EV-D68 entry. Our observations indicate that Abs against EV-D68 that bind to the important residues R126, I131, and I133, which are the ones spatially closest to the DE loop, exhibit excellent antiviral effects, but they do not prove that A6-1 blocks EV-D68 infection by means of postattachment neutralization. This activity of blocking viral entry was confirmed in vivo. Mice pretreated with A6-1 were protected from EV-D68 infection, indicating that A6-1 could be a prophylactic for fighting EV-D68 infection, which is an urgent clinical need. Meanwhile, our results also show that A6-1 can inhibit EV-D68 from infecting the brain tissue (Fig. 7B), which shows that the A6-1 plays some role in preventing the CNS from being infected with EV-D68. Therefore, we think that A6-1 can be further evaluated for its activities in other animal models such as acute flaccid paralysis (37) or asthma induced by EV-D68 infection (52).

The use of this technology to isolate and characterize mAbs induced by an infection or immunization is feasible, especially in response to new pathogens. The Ab/virus interaction at structural regions such as the DE loop might be regarded as a key to designing an efficient EV-D68 vaccine through genetic engineering via reverse vaccinology (53, 54).

The authors have no financial conflicts of interest.