Briefly, various CR3022 mAbs were expressed in N. benthamiana plants using “magnifection” procedure (29). Cloned expression vectors i.e., PVX-LC and TMV-IgG-HC were transformed into Agrobacterium tumefaciens strain ICF320 (Icon Genetics) and grown overnight at 28.0°CC followed by 1:000 dilution in infiltration buffer [10 mM MES (pH 5.5) and 10 mM MgSO₄]. The combinations of diluted bacterial cultures (TMV-IgG-HC + PVX-LC) were used to transfect 4 week old N. benthamiana plants (ΔXTFT glycosylation mutants) using vacuum infiltration. Using a custom-built vacuum chamber (Kentucky Bioprocessing), the aerial parts of entire plants were dipped upside down into the bacterial/buffer solution and a vacuum of 24’’ mercury was applied for 2 min. Infiltrated plants were allowed to recover and left in the growth room for transient expression of antibodies. 7 days after infiltration, plants were harvested and homogenized in extraction buffer containing 100 mM Glycine, 40 mM Ascorbic Acid, 1 mM EDTA (pH 9.5) in a 0.5:1 buffer (L) to harvested plants (kg) ratio. The resulting green juice was clarified by filtration through four layers of cheesecloth followed by centrifugation at 10,000 g for 20 min. Next, mAbs were captured from the clarified green juice using MabSelect SuRe Protein A columns (GE Healthcare). The mAbs were eluted from Protein A columns were further purified using equilibrated Capto Q columns (GE Healthcare) and flow-through fractions, which contain mAbs, were collected. The mAb-containing fractions were finally polished with CHT chromatography with type II resin (Bio-Rad). The purity of the different mAbs were verified using size-exclusion chromatography multi-angle light scattering (SEC/MALS) ( Supplementary Figure S4 ).

The size and purity of the CR3022 and its variants were evaluated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under non-reduced and reduced conditions. In brief, mAb solution was mixed with LDS sample buffer (Thermo Fisher Scientific Inc., Waltham, MA, USA), and then was incubated at 70°C for 10 min. For the preparation of reduced condition samples, TCEP solution (Thermo Fischer Scientific) was added to half of the samples and then was incubated at room temperature for 5 min. The samples were separated by SDS-PAGE (1 µg/lane) using 4-12% gradient Bis-Tris gels (Thermo Fisher Scientific). The proteins in the gel were detected by Coomassie brilliant blue staining and photographed using a gel imaging system (Chemi-Doc MP, Bio-Rad Laboratories Inc, Hercules, CA, USA).

Recombinant SARS-CoV2 spike RBD glycoprotein (Abcam plc, Cambridge, UK) or recombinant human FcRn protein (R&D Systems, Minneapolis, MN) were coated on the high-binding half-area 96-well plate at 2.5 µg/mL in bicarbonate buffer (pH 9.6) overnight at 4°C. 5% non-fat milk was used for blocking for 1 h at room temperature. CR3022 and its variants diluted in 1% non-fat milk were added to the plate, and then the plates were incubated overnight at 4°C. For the affinity assessment to RBD, PBS at pH 7.4 was utilized to prepare 1% milk, while for the affinity assessment to FcRn, the pH was adjusted to 6.0 by adding hydrochloric acid to PBS. After washing with PBS, goat anti-human IgG Fc horseradish peroxidase (Abcam, 1:5000 dilution) was added to the plate, and then the plates were incubated for 1 h at room temperature. 1-step TMB substrate solution (Thermo Fisher Scientific) was added for the enzymatic degradation of hydrogen peroxide. The reaction was stopped with 1 M hydrochloric acid, and the absorbance at 450 nm was measured using an AccuSkan FC plate reader (Fisher Scientific).

Human primary tracheobronchial airway epithelial cells were obtained from individual donors without documented lung disease including, cystic fibrosis, chronic obstructive pulmonary disease, or asthma. Airway epithelial cells were isolated from surgical specimens by the University of North Carolina Cystic Fibrosis Center Tissue Culture Core Facility using Institutional Review Board-approved protocols. The isolated cells were seeded on type IV collagen-coated polyester filter inserts (pore size, 0,4 µm: area, 1.12 cm²; Transwell, Corning, NY, USA) (30) and grown in Pneumacult ALI Basal Medium (StemCell Technologies). After reaching confluence, culture conditions were changed to an air-liquid interface (ALI) and cultured in Pneumacult media containing the recommended supplements to enhance the differentiation of epithelial cells resembling the airway epithelium in vivo. After 4 weeks of culture under these conditions, abundant ciliated cells were observed using light microscopy. Transepithelial electrical resistance (TEER) was measured as ≥300 Ω·cm² using an EVOM2 instrument (World Precision Instruments, Sarasota, FL, USA).

Transport experiments were performed while maintaining ALI conditions. After the apical surface of HAE cells was washed using PBS, CR3022, CR3022-YTE, or CR3022-LS (20 or 200 µg/mL) were applied to the apical side (10 µL diluted in PBS) or basal side (2 mL diluted in medium). For recovery of cell culture apical surface washes, 200 µL of PBS was added for 15 mins and then harvested at designed time points (6, 24, 48, 72, 96 h). For basolateral sampling, 200 µL of media were collected from the basolateral side of WD-HAE at designed time points (6, 24, 48, 72, 96 h) and replaced with equal volumes of media. mAb concentrations in samples were quantified by ELISA using RBD-coated plates as described above. In the inhibition experiments, CR3022 (200 µg/mL) with intravenous immunoglobulin (IVIG, 2 mg/mL) was applied to the basal side (2 mL diluted in medium).

FCGRT expression levels were determined in LAE and SAE airway cultures by interrogating a previously published BulkRNASeq database (31). As detailed methods are contained in the original publication, only brief details are supplied here. Matched LAE and SAE cells were isolated from freshly excised normal human lungs obtained from the UNC Tissue and Cell Culture Procurement Core from transplant donors with lungs unsuitable for transplant. Briefly, LAE cells were isolated from the cartilaginous airways by protease digestion and SAE cell isolated from distal lung from the same lung as utilized for LAE cell isolation. Small airways were identified based on the absence of wall cartilage and outer diameters less than 2 mm. After expansion, LAE and SAE were cultured on human placental type IV collagen-coated, 0.4 μm pore size tissue culture membranes for 4 weeks, at which point RNA was extracted from the cells using TRI Reagent (Sigma) according to the manufacturer’s instructions. Total RNA was used for bulk RNA-seq. For Bulk RNAseq library preparation and sequencing, mRNAs were enriched using oligo (dT) beads. The library was sequenced by Novaseq 6000 platform. Raw sequence reads in FASTQ format were mapped to the reference genome GRCh38, and Gencode v32 gene and transcript annotation for exon and splice site mapping, using HISAT2 aligner. Gene expression was estimated by StringTie as Transcript Per Million (mapped reads), or TPM. Gene expression (TPM) was normalized using the voom function from Bioconductor limma package based on mean-variance modeling, and differential expression analysis was performed as linear mixed-effect models using the dream function from the Bioconductor variance Partition package. The linear mixed-effect model applied treated donor code as random effect factor on gene expression. All transcriptome data were deposited in GEO. LAE/SAE bulkRNAseq data were assigned to GSE160675.

Histological cross-sections of WD-HAE cultures and human tracheal tissues fixed in 10% buffered formalin were generated by the UNC Pathology Services Core. Histological sections were deparaffinized with xylene and ethanol gradients. Primary antibodies were added to tissue sections after heat-mediated antigen retrieval in TRIS-EDTA buffer at pH 9.0 and blocking of non-specific Ig-binding sites with 10% Normal Goat Serum (Jackson labs). A primary rabbit antibody directed against FcRn (Proteintech 16190-1-AP) or as control, an isotype rabbit IgG (Jackson labs) was applied to tissue sections at 1 ug/ml for both antibodies. Following primary antibody incubation overnight, tissue sections were then incubated with fluorescent secondary antibody (anti-rabbit IgG AlexaFluor 594) at a 1:700 dilution (Invitrogen), and stained with DAPI. Tissue sections were imaged on an Olympus VS200 virtual slide scanning system with an Allied Vision Pike 5 CCD progressive scan camera using OlyVia software (version 4.0) and images processed with Adobe Photoshop (CS6).

The human recombinant monoclonal antibody CR3022 directed against the receptor-binding-domain (RBD) of SARS-CoV-2 Spike and its variants with modified Fc regions (CR3022-YTE and CR3022-LS) were produced by transient transfection of Nicotiana plants engineered to express proteins and mAbs decorated with human N-glycosylation (29, 32, 33); the same production system has been used previously to generate mAbs for clinical trials (34, 35). The molecular size and purity of CR3022 and its variants was assessed by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and all recombinant mAbs showed the expected protein molecular size in non-reduced (150kD) and reduced conditions (50/25kD) ( Figure 1A ). The binding affinities of the mAbs to SARS-CoV2 Spike RBD glycoprotein were measured by antigen-specific enzyme-linked immunosorbent assay (ELISA). CR3022 and its variants all exhibited highly comparable binding affinities to RBD glycoprotein ( Figure 1B ), with measured EC50 representing the equilibrium dissociation constant (K _D) values of ~19, 28, and 17 ng/mL for CR3022, CR3022-YTE, and CR3022-LS, respectively. We also measured the FcRn affinities via ELISA. In good agreement with previously published studies, the YTE and LS mutations of Fc substantially improved the affinity of CR3022 for FcRn compared to the IgG1-Fc: with K _D values of CR3022, CR3022-YTE, CR3022-LS, and intravenous immunoglobulin (IVIG; included as experimental control) determined as 84.7, 38.4, 7.12, and 78.5 µg/mL, respectively ( Figure 1C ). Altogether, our panel of recombinant CR3022 and its variants all exhibited similar binding affinities to the SARS-CoV-2 Spike RBD, while exhibiting the expected variable affinities to FcRn.

We next used WD-HAE cultures to assess the basal-to-apical transepithelial transport of CR3022 and its variants over time. To achieve the typical range of systemic concentrations of antiviral mAbs achieved shortly after dosing [~20-200 μg/mL (36–40)], we introduced either 40 or 400 μg of our mAbs to the 2 mls of basal media. We then assessed the apical levels of CR3022 every 24 hrs by adding 100 microliters of PBS to the apical surfaces of HAE cultures for 15 mins, after which the PBS apical wash was harvested and used to measure the amount of CR3022 using the same highly sensitive RBD-specific ELISA to quantify CR3022. This assay eliminated background signal associated with other antibodies potentially present in serum-supplemented commercial cell culture media.

The concentrations of CR3022 and its variants in the apical washes increased over time in a dose-dependent manner. While the low mAb dose resulted in detection of low concentrations of each mAb (less than 40 ng/mL) in the apical wash at each time point, the high mAb dose resulted in appreciable quantities of mAbs in the apical washes been detected reaching hundreds of ng/mL ( Figure 2A ; Supplementary Figure S1A ). Surprisingly, we found comparable levels of CR3022, CR3022-YTE and CR3022-LS in the apical washes over time, suggesting that all our mAbs constructs exhibited similar transepithelial transport rates regardless of the differences in affinity for FcRn. The cumulative transcytosed amount over the first 4 days were consistent between all 3 mAbs, reaching ~40 ng and ~400 ng total over the first 4 days post-dosing with the low and high mAb dose, respectively ( Figure 2B ; Supplementary Figure S1B ). However, relative to the actual concentration of mAbs in the basal media, the total fraction of mAb transcytosed over the first 4 days remained exceptionally low, with roughly 0.1% of the total mAb dose transcytosed for each of the three mAbs ( Figure 2C ). Overall, the mAb concentration in apical washes over time, the cumulative amount of mAb transcytosed, and the cumulative fraction of mAbs transcytosed were not statistically different between CR3022, CR3022-YTE and CR3022-LS (see Supplementary Table 1 for statistics summary). The cumulative fraction of the mAbs transcytosed were also similar between the two different mAb doses. These results highlight the limited capacity of mAb in the basal compartment to cross the airway epithelium into the apical compartment, and that the basal-to-apical transport rate of mAbs are not dependent the affinity of mAb to FcRn.

As we did not detect enhanced transcytosis of both CR3022-YTE and CR3022-LS, we sought to confirm FcRn was present in the WD-HAE cultures. We did so by first assessing the expression of Fc Gamma Receptor and Transporter (FCGRT), the gene encoding FcRn, in WD-HAE cultures. We note that WD-HAE cultures in our studies utilized cells harvested from the large cartilaginous proximal airways. Thus, we first interrogated published Bulk RNA sequencing data (BulkRNAseq) previously generated at UNC on WD-HAE cultures derived from large airway epithelial (LAE) cells in 10 individual tissue donors (31), and found FCGRT to be abundantly expressed in WD-HAE from all 10 tissue donors ( Figure 3A ).

The previously published BulkRNASeq study also had data from small non-cartilaginous airway epithelium (SAE) from the same 10 tissue donors, which provided a unique opportunity to compare FCGRT expression levels in LAE and SAE derived from the same tissue donor. While FCGRT expression was similar in magnitude in SAE as in LAE, the expression levels in SAE were more variable, with 7 of 10 tissue donors expressing FCGRT in SAE at levels comparable to that in LAE ( Figure 3B ).

To further ascertain expression of the FcRn protein in WD-HAE, we performed immunohistochemistry on histologic sections of WD-HAE and human tracheal tissue ( Figure 3B ). Using immunofluorescence as our detection method, we found the airway epithelial cells in WD-HAE cultures were strongly positive for FcRn immunoreactivity, with more robust signal at the apical aspects of the luminal epithelial cells. Similarly, in sections of human tracheal tissue, FcRn immunoreactivity was largely found in the epithelial cells with increased immunoreactivity at the apical aspects of the epithelium. These results clearly indicate that FcRn is highly expressed in WD-HAE cultures, and thus the failure to observe increased transcytosis with Fc variants is not due to the absence of FcRn in our WD-HAE culture model.

To functionally assess the involvement of FcRn in basal-to-apical transcytosis in WD-HAE cultures, we next performed a competitive inhibition experiment using 10-fold excess of a clinical preparation of human IgG (4 mg of intravenous immunoglobulin (IVIG)) to the high dose of CR3022 (400 μg). While we were able to detect statistically significant differences in CR3022 concentrations in apical washes by 24 hrs after dosing ( Figures 3C–E ), the reduction in the amount of CR3022 transcytosed was markedly less than expected from the stoichiometric ratio of input IVIG vs. CR3022. Specifically, CR3022 concentrations and cumulative amount transcytosed were only decreased by ~50% in the presence of 10-fold excess IVIG, which should have reduced apical CR3022 by ~90% if FcRn-mediated transcytosis is the sole mechanism of basal-to-apical transport. These results suggest FcRn likely has a significant but not exclusive role in the basal-to-apical transcytosis of mAbs across WD-HAE cultures.

To ascertain the impact of having excess IVIG in the environment, we performed additional studies with the YTE and LS mAb variants in the presence of excess IVIG competition ( Supplementary Figure S2 ). Unlike with wildtype CR3022, excess IVIG reduced the two mAbs to even less extent, with virtually no detectible difference in the amount of CR3022-YTE that underwent basal-to-apical transcytosis, and only a very modest (~20%) reduction in the amount of CR3022-LS that underwent basal-to-apical transcytosis. These results are consistent with a model where greater FcRn affinity does not increase kinetics or rate of FcRn transcytosis. It suggests that, in the presence of high IVIG, increased FcRn affinity may allow mAbs to better compete for available FcRn and very modestly increase basal-to-apical transport.

Direct inhaled delivery of mAbs is quickly becoming a reality for treatment of acute respiratory infections, as illustrated by our recent Phase 1 study of an inhaled antiviral mAb therapy for COVID-19 (41) and recent large animal studies for RSV and SARS-CoV-2 (42, 43). This motivated us to determine how quickly our mAbs delivered to the apical surface of the airway would undergo apical-to-basal transport across HAE. To test this, we added either 0.2 or 2 μg of CR3022 or each of the CR3022 variants directly to the apical compartment of WD-HAE, then quantified the concentrations of mAbs in the basal media over time by sampling a small fraction of the total volume of media.

Similar to the earlier basal-to-apical transport study above, we found limited distribution of mAbs into the basal media over the first 4 days, with concentrations reaching only ~1-2 ng/mL 96 hrs following apical dosing of 0.2 μg of CR3022 ( Figure 4A ; Supplementary Figure S3A ). At the higher apical mAb dose of 20 μg CR3022, we saw concentrations in basal media reaching ~10-15 ng/mL ( Figure 4A ), roughly proportional to the 10-fold greater input mAb dose. At both input doses, CR3022 and its YTE and LS variants exhibited highly comparable cumulative fraction of mAb transcytosed to the basal media ( Figure 4B ; Supplementary Figure S3B ). When we adjust for the input dose, the cumulative fraction of mAb transcytosed over the first 4 days reached ~1-3% for all three mAbs at the high input dose ( Figure 4C ); for CR3022 and CR3022-YTE but not CR3022-LS, the cumulative fraction transcytosed at the lower dose was even higher, although the difference did not reach statistical significance (see summary statistics in Supplementary Table S2 ). Indeed, similar to the studies on basal-to-apical mAb distribution, there was overall limited distribution of mAbs from the apical surface to the basal compartment, and no statistically significant differences were found between the three mAbs tested. The results here are consistent with the limited distribution of mAbs into the submucosa and the systemic circulation following nebulized inhaled delivery in our Phase 1 study. These results here also suggest minimal dependence of FcRn-affinity on apical-to-basal transport of mAbs across human airway epithelium.

Finally, we sought to directly compare the concentrations in the apical and basal compartments of the three CR3022 mAb variants following apical vs. basal delivery. Not surprisingly, the mAb concentrations in the apical compartment remained high following apical delivery, with mAb concentrations reaching ~1 and ~10 μg/mL in the apical wash for the 0.2 and 2 μg doses, respectively. For mAbs dosed into the basal compartment to reach the same concentration in apical washes as the low apical mAb dose, a ~2000-fold increase in CR3022 (i.e. 400 μg) was necessary ( Figure 5A ). The superiority of direct pulmonary delivery is also illustrated by the fact that a 200-fold lower mAb dose delivered apically still achieved ~10-fold greater mAb concentrations in the apical washes (p<0.01); the much greater mAb levels following apical dosing were consistent regardless of the FcRn-affinity of the various CR3022 mAbs.

We next compared the concentration of CR3022 mAbs in the basal compartment. In good agreement with our expectations, the basal concentration of CR3022 administered apically was over three orders of magnitude lower than when the same mAbs were applied to the basal side, albeit at a much higher dose ( Figure 5B ). Given the vast difference in the amount of mAbs dosed apically vs. basally, we estimated the delivery efficiency based on the ratios for the apical vs. basal mAb concentrations ( Figure 5C ). Apical delivery of mAb, regardless of the mAb FcRn-affinity and the actual dose, achieved an apical:basal ratio of ~500:1 (i.e. the concentration in apical washes is ~500-fold higher than in basal media). In contrast, basal delivery resulted in a ratio close to 1:500 regardless of the FcRn-affinity and the actual dose of the mAbs. The difference in dosing route yields a nearly 5-log difference in distribution efficiency. Since viruses that cause acute respiratory infections are predominantly present in the airway lumen, achieving effective inhibitory concentrations of mAb on the airway luminal surface is essential for maximizing therapeutic efficacy. Our results underscore the route of mAb dosing, rather than Fc engineering, as the key to achieving high apical concentrations as part of the intervention against respiratory viruses early in infection.