Stable nebulization and muco‐trapping properties of regdanvimab/IN‐006 support its development as a potent, dose‐saving inhaled therapy for COVID‐19

Abstract The respiratory tract represents the key target for antiviral delivery in early interventions to prevent severe COVID‐19. While neutralizing monoclonal antibodies (mAb) possess considerable efficacy, their current reliance on parenteral dosing necessitates very large doses and places a substantial burden on the healthcare system. In contrast, direct inhaled delivery of mAb therapeutics offers the convenience of self‐dosing at home, as well as much more efficient mAb delivery to the respiratory tract. Here, building on our previous discovery of Fc‐mucin interactions crosslinking viruses to mucins, we showed that regdanvimab, a potent neutralizing mAb already approved for COVID‐19 in several countries, can effectively trap SARS‐CoV‐2 virus‐like particles in fresh human airway mucus. IN‐006, a reformulation of regdanvimab, was stably nebulized across a wide range of concentrations, with no loss of activity and no formation of aggregates. Finally, nebulized delivery of IN‐006 resulted in 100‐fold greater mAb levels in the lungs of rats compared to serum, in marked contrast to intravenously dosed mAbs. These results not only support our current efforts to evaluate the safety and efficacy of IN‐006 in clinical trials, but more broadly substantiate nebulized delivery of human antiviral mAbs as a new paradigm in treating SARS‐CoV‐2 and other respiratory pathologies.


| INTRODUCTION
Most viruses that cause acute respiratory infections (ARIs), including influenza, 1-4 RSV, [5][6][7][8][9] PIV, 10 and the betacoronavirus HKU1, 11 infect almost exclusively via the apical (luminal) side of the airway epithelium, as revealed by studies using well-differentiated, polarized human airway epithelial (WD-HAE) cultures grown at the air-liquid interface. 12 In contrast, there is little-to-no productive infection when viruses are introduced into the basal (serosal) compartment in WD-HAE cultures. Importantly, infected cells appear to predominantly shed progeny viruses back into the apical compartment (i.e., into airway mucus [AM] secretions), with only rare shedding of virus into the basal compartment. This unique pathophysiology is shared by SARS-CoV-1, which only productively infects WD-HAE cultures when the virus is inoculated apically, with no appreciable infection when the same amount of virus is inoculated basally. 13 There is $1000-fold greater virus shed into the apical compartment relative to the basal compartment. The near exclusive apical infection and shedding of SARS-CoV-1 is consistent with the apical trafficking of ACE2 in airway biopsy tissues 14 and in WD-HAE cultures in vitro. 13,15,16 Not surprisingly, given that SARS-CoV-2 binds the same ACE2 receptor as SARS-CoV-1 for cellular entry, SARS-CoV-2 also undergoes preferential apical infection and shedding. 13,15,16 This pathophysiology is consistent with the substantial time window between initial appearance of upper respiratory tract symptoms and the development of pulmonary and systemic morbidities that necessitate hospitalization.
Given the concentration of SARS-CoV-2 in the respiratory tract and the resulting respiratory tract symptoms and morbidities direct delivery of potent neutralizing monoclonal antibodies (mAbs) to the site of infection should be preferred. Nevertheless, every antiviral mAb that has received full approval or emergency use authorization to date is dosed either by iv infusion or sc/im injections, despite prior studies showing only a small fraction of systemically dosed mAb reaches the respiratory tract in animal models [17][18][19][20] and in human studies. 21 The lack of efforts advancing inhaled delivery of mAb is likely due to prior work suggesting mAbs can aggregate and lose binding activity following nebulization. 22,23 This problem is particularly evident with jet and ultrasonic nebulizers, where droplet recirculation, as well as heat and shear stresses, increase the aggregation of biomolecules and leave large residual quantities of unnebulized drug. 24,25 Indeed, an earlier study delivering omalizumab using jet nebulizers for asthma was thought to possibly generate protein aggregates. 26 To date, there has been no report on stable nebulization of a fully human mAb that has been advanced through late-stage clinical trials. Nevertheless, a number of protein therapeutics have been stably nebulized using vibrating mesh nebulizers (VMN) as part of chronic treatment regimens. [27][28][29][30] This offers the potential that human mAbs, if appropriately formulated, 22 can also be stably nebulized using VMNs, with no loss in binding and no aggregation.
Given the public health urgencies of the COVID-19 pandemic, we were motivated to advance an inhaled antiviral therapy using a full length, broadly neutralizing mAb. Regdanvimab is one of few anti-SARS-CoV-2 neutralizing mAbs that have received approval from either the FDA or EMA. Regdanvimab, administered by intravenous (iv) infusion at 40 mg/kg, provided a 72% reduction in risk of hospitalization and shortened the recovery time by $5 days compared to placebo in its global Phase 3 clinical trial. 31 These results make regdanvimab a highly promising mAb for developing an inhaled  Virions that possess the greatest diffusivity (i.e., the most mobile fractions), by definition, are more likely to diffuse across the mucus layer and infect the underlying epithelium before mucus is eliminated by natural clearance mechanisms. Thus, we sought to assess the effect of IN-006 in limiting the fraction of SARS-CoV-2 VLPs that could most readily penetrate across AM. We quantitatively defined the fastmoving subpopulation of SARS-CoV-2 as virions that possess sufficient mobility to penetrate through a physiologically thick AM layer (50 μm) in 1 h, which yielded a minimum D eff ≥ 0.347 μm 2 /s. This fast-moving population was reduced from 46% in naïve AM to 4.8%

| IN-006 can be stably nebulized across a range of concentrations
The most efficient method to deliver mAb to the respiratory tract is by direct inhalation 20 ; we thus tested whether we could generate IN-006-containing aerosols that are suitable for pulmonary deposition using a VMN. We first determined the aerodynamic particle size distribution (APSD) of the aerosols, as the resulting droplet sizes directly influence the site of deposition within the airways. 32 As a general guide, aerosols smaller than $2.5 μm are preferentially deposited in the deep lung, between 2.5 and 5 μm preferentially in the lower airways, and aerosols $5-10 μm in diameter preferentially in the upper airways, including nasopharyngeal and oropharyngeal regions. 32 Based on earlier unpublished work with nebulizing mAbs against RSV, we first utilized the Copley Scientific Next Generation Impactor (NGI) to measure the APSD of IN-006 nebulized at 20 and 30 mg/ml. The mass median aerodynamic diameter (MMAD) was 5.7 ± 0.08 and 5.3 ± 0.2 μm (Figure 2a), with VLPs and (f) effective diffusivities (<D eff >) of SARS-CoV-2 VLPs in each of the independent AM tested. Due to specimen volume limitations, we only assessed the muco-trapping potencies of IN-006 at 333 ng/ml for donor ID 8, 9 and 10 (ND = no data). Repeated measures one-way analysis of variance with post hoc Dunnett's test on log-transformed (diffusivity, MSD) or untransformed (% fast moving) data (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001) a fine particle fraction (FPF; particulates <5 μm) of 47 ± 1% and 51 ± 1%, respectively.
To determine whether the binding affinity of IN-006 was preserved during nebulization, we collected nebulized IN-006 using a two-stage glass twin impinger setup in accordance with European Pharmacopeia 2.9.18, and measured the binding affinity (EC 50

| Good laboratory practice nebulization characterization study of IN-006
To support a formal application to regulatory authorities to initiate human studies, we next conducted nebulization characterization studies that met good laboratory practice (GLP) guidelines. In these studies, the average MMAD of IN-006 aerosols generated across 3 independent InnoSpire Go devices, each evaluated in triplicate, was

| DISCUSSION
Many in the scientific community believe that COVID-19 will become endemic, despite the availability of effective vaccines. 33 This underscores the need to ensure broad availability and easy access to effective treatments that can prevent progression to severe COVID and hospitalization, particularly given the sizable population of individuals with vaccine hesitancy around the world. While highly effective mAb therapies were initially quickly advanced (e.g., REGN-COV2, 34 bamlanivimab + etesevimab, 35 sotrovimab 36 ), prior mAb treatment options enjoyed limited adoption because of critical access issues, as well as the large doses required. First, infusions at dedicated facilities not only create a substantial burden on healthcare system and subject healthcare workers to infection risk, but also require substantial time to administer (including time for wait/registration, health check, infusion times that can last $30-60 min, and then an additional 60 min of monitoring time postinfusion 34 ). Until the creation of temporary infusion centers, many unused mAbs accumulated in healthcare facilities as physicians were simply unable to administer them to enough patients. 37 While select mAbs were allowed to be given by multiple subcutaneous (sc) injections instead of iv (e.g., REGEN-COV2), that change in delivery does not appreciably ease the burden on the healthcare system, in part because patients must still spend time to register, be screened, be treated, and then still must undergo at least 1 hour of post-injection monitoring. To improve access, we believe it is essential to have an effective therapeutic intervention that can be used by patients in an outpatient setting immediately following a positive diagnosis using point-of-care or otherwise rapid diagnostics. Second, the systemic delivery of mAbs necessitates large mAb doses. This creates a major supply chain issue, as the sheer number of COVID-19 patients, coupled with the large mAb dose required per patient, translates to multimetric tons of mAbs needed, which in turn creates marked constrains on global mAb manufacturing capacity.
These realities suggest an ideal treatment should not require iv, intramuscular (im), or sc administrations that must be carried out by healthcare workers, or a period of monitoring following injection.
Instead, the ideal treatment should be easily self-administered soon after diagnosis, allowing patients to be treated at home while still during early stages of the disease, without significant pulmonary morbid- Finally, both regimens involve swallowing a substantial number of pills (40 for molnupiravir, 30 for Paxlovid) over the course of the treatment regimen, which may prove to be difficult for pediatric, geriatric, and select populations with underlying conditions. 41 While the overall prevalence of dysphagia in the Midwestern US population has been reported to be 6%-9%, 42 its prevalence in people over age 50 years is estimated to be closer to 15%-22%, [43][44][45] and possibly as high as 40%-60% of residents in assisted living facilities and nursing homes. 41,45,46 For these reasons, even with the approval of Paxlovid and molnupiravir, we continue to believe inhaled treatment could address an important unmet need among COVID patients.
One of the key benefits of nebulized mAb therapy is that a much greater fraction of the overall drug dose is delivered directly to the primary site of infection and morbidity. Prior studies of antiviral mAbs for treatment of ARIs have encountered substantial roadblocks in clinical translation, likely due to inefficient and inadequate transudation of systemically administered mAb into the respiratory tract. Indeed, in a clinical study of CR6261 (an anti-influenza mAb), 21 the C max in the nasal swab samples was not achieved until Days 2-3 following iv dosing, in stark contrast to peak serum concentrations of 1 Â 10 6 ng/ml reached within 15 min after infusion. More importantly, the mean peak concentration of CR6261 from nasal swabs was only $600 ng/ ml, or $1700-fold lower concentrations in the nasal mucosa than in plasma. 21 CR6261 is not alone in its limited distribution into the lung airways; mAbs are large molecules with generally small volumes of distribution that tend to remain in the serum and peripheral fluids in the absence of mechanisms of active transport. Previous nonhuman primate studies comparing the pharmacokinetics and biodistribution of systemically administered mAbs have consistently shown both slow and limited pulmonary distribution, despite achieving high concentrations in the plasma. Indeed, the concentration of mepolizumab in BALF, following iv injection, was $500-fold lower than the concentration in plasma. 17 Even greater differences in BALF versus plasma concentration were noted in biodistribution studies of motavizumab (anti-RSV mAb) in cynomolgus monkeys, where a 2000-fold difference between BALF ($100 ng/ml) and plasma concentrations ($200,000 ng/ ml) was measured 4 days following an iv dose at 30 mg/kg. 18 The poor distribution of motavizumab to the respiratory tract is likely responsible for its lack of efficacy as a treatment. In an article describing the inability of bamlanivimab to provide significant therapeutic benefit in COVID patients with severe disease, the authors hypothesized that the failure might be attributed to limited penetration into infected tissues. 47 In the case of rapidly multiplying viral infections, intervening early in the course of infection is key to effective therapy, and a few days of delay before achieving efficacious therapeutic concentrations at the site of infection may well represent the difference between clinical success and failure.
For instance, treatment with oseltamivir (i.e., Tamiflu) and baloxavir marboxil (i.e., Xofluza), two oral anti-influenza antivirals, must be initiated within 48 h (and preferably within 24 h) of the emergence of symptoms in order to be efficacious. 48 Since IN-006 can achieve pulmonary C max virtually instantaneously, we believe it is exceptionally suited as an early intervention to prevent progression to severe COVID.
The frequent reports of mAb aggregation following nebulization 22 have led many in the field to believe fully human mAbs are too large or too unstable to be nebulized, and that smaller protein binders such as nanobodies (camelid antibodies that consist of only heavy chains, without light chains and without effector functions) are more suitable for nebulized delivery. In sharp contrast to this prevailing dogma, we showed here that nebulization of IN-006, a fully human IgG1 mAb, did not result in any appreciable increase in aggregation or fragmentation, a loss of binding/neutralization activity, or other impacts on physical integrity (e.g., monomer content). Although mAbs and nanobodies would both be expected to provide potent neutralization, the presence of an Fc domain on the full mAb confers additional effector functions, including opsonization, ADCC and ADC, and muco-trapping. Furthermore, since nanobodies are camelid in origin, there may be substantial immunogenicity, as previously reported for inhaled nanobodies for RSV treatment. For these reasons, we believe nebulized treatments using full length human IgG 1 mAbs may confer additional benefits over nebulized therapies based on nanobodies or Fabs.
The potential muco-trapping effector function of IgG-to physically trap pathogens in mucus-has only been appreciated recently.
An inherent assumption by the field has been that, to trap a pathogen, Ab must bind tightly to mucins. However, many investigators reported seemingly negligible affinity of IgG to mucins [49][50][51][52][53] ; for instance, the diffusion rates of IgG in human cervical mucus are slowed only $10% versus their rates in buffer, implying that 90% of the time an IgG is simply not bound to mucins. 54 This led most researchers to conclude that Ab have no meaningful function in mucus besides neutralization.
Instead, our discovery of the muco-trapping potential of IgGs is based on multiple weak and transient bonds between IgGs and mucins, and highlights two pivotal concepts: First, many IgGs can bind to the surface of a pathogen, and the resulting array of IgGs can generate high binding avidity to mucins (analogy: a Velcro patch can tightly bind two surfaces despite individually weak hooks). Second, IgG must possess a narrow range of weak and transient affinity to quickly accumulate on the invading pathogen surface while minimizing the number of pathogen-bound IgG needed to trap the pathogen [55][56][57][58] ; mAbs that bind too tightly to mucins would not be able to travel through mucus to bind pathogens. We have previously shown that IgG possessing suitable N-glycans on IgG-Fc is capable of immobilizing viruses in various mucus secretions, resulting in rapid clearance from the respiratory tract 59 and can provide effective protection against vaginal Herpes transmission. 54,57 In good agreement with our previous findings, we showed here that IN-006 was able to effectively immobilize SARS-CoV-2 in AM. Trapped virions are unable to diffuse through AM to infect cells and are expected to be quickly eliminated from the respiratory tract by natural muco-ciliary or cough-driven mucus clearance mechanisms for sterilization in the low pH gastric environments. Indeed, trapped virions are cleared from the lung airways minutes to hours for sterilization in the low pH gastic environments, 59 and trapping viruses in mucus affords sterilizing immunity against mucosal transmission. 57 Thus, mAbs capable of this muco-trapping effector function provide a mechanism to physically remove viruses from infected airways. infection, we believe there is a need to deliver the drug to all parts of the respiratory tract, rather than target only the lungs (e.g., with dry powder inhalers) or target only the upper airways (e.g., with nasal drops). We believe nebulization, by generating diverse droplet sizes that enable simultaneous delivery to all parts of the respiratory tract, is uniquely suited for inhaled delivery of antivirals.
Finally, our study had some limitations. First, since collection of BALF is a terminal procedure in rats, we could collect BALF at the last two timepoints in the toxicokinetic portions of the GLP rat study (8 and 12 h after last dose). We are thus estimating the systemic halflife by extrapolating pulmonary elimination rates across a duration shorter than the actual half-life estimated. We do wish to point out that, while the rough estimates of lung PK are interesting, the key conclusion from these pulmonary measurements is not the elimination rate, but rather the relative concentrations achieved in the lung versus in serum following nebulized delivery. More rigorous investigation of the pulmonary and systemic PK of IN-006 will be conducted in clinical studies. Second, the ex vivo mucus trapping studies (Figure 1) were conducted on AM samples from otherwise healthy individuals who did not have SARS-CoV-2 infection. We believe that the AM in patients with mild and moderate COVID-19 is likely similar to that of healthy individuals in this study, given that occasional reports of increased mucus accumulation is only observed with extensive pneumonia and bronchiolitis. 61 Fortunately, even in cases of extreme hyperviscoelastic mucus (e.g., sputum from patients with cystic fibrosis), the microrheology still approaches that of water, 62 and virus-sized nanoparticles that do not stick to mucins can still undergo rapid diffusion in CF sputum. 63

| Non-GLP nebulization study
To test the feasibility of nebulizing IN-006 across a range of concentrations, IN-006 was formulated at various concentrations from 5 to 60 mg/ml and nebulized using an InnoSpire Go VMN. Generally, USP <1601> was adhered to for generation of data and calculation of MMAD and geometric standard deviation (GSD). As discussed in this report, FPF refers to particles collected on Stages 4 and smaller of the NGI (<5.39 μm at 15 L/min). Briefly, the NGI (MSP Corp) and collection stages were precooled to 4 C for at least 90 min before experiments. The nebulizer was loaded with enough mAb solution to ensure replicates could be performed sequentially, while avoiding sputtering (i.e., remaining above the manufacturer's minimum recommended volume). A custom silicone mouthpiece molded to the nebulizer/NGI inlet interface was used to affix the nebulizer to the inlet with a tight seal. A solenoid in line with the NGI and vacuum (set to 15 L/min) was used to collect sufficient nebulized mAb at a given concentration. The nebulizer was actuated, and the solenoid was switched on to begin collection. Following nebulization, the vacuum and nebulizer were switched off and the NGI stages and inlet were removed. Quickly, the next set of stages and inlet were swapped in to perform a second replicate nebulization. To collect deposited mAb, stages were rinsed with 5 ml of the formulation buffer, matching the buffer of nebulized material, and assayed for mAb mass deposition at A280. APSDs were plotted as cumulative percentage of drug mass undersize against aerodynamic stage cut-off diameter for IN-006 on a logprobability scale. The MMAD, GSD, and FPF were determined from this data.
In assessing APSD, the MMAD was defined as the aerosol diameter cut-off at which 50% of the mass of drug was in larger aerosols and 50% was in smaller particles. The FPF is calculated as the mass of drug contained in particles smaller than $5 μm divided by the total emitted dose to roughly estimate the fraction of nebulized therapeutic that would be delivered to the lower airways.

| DLS for determining aggregates in non-GLP study
To detect the presence of aggregates following nebulization, postnebulization samples were collected from the NGI and measured via DLS (Zetasizer; Malvern Instruments; particle quantitation limit 0.3 nm-10 μm). Postnebulized IN-006 samples were added to a cuvette at the collected concentration ($1 mg/ml after dilution during sample recovery), and particle size polydispersity and average diameters were determined using volume-weighted analyses.

| GLP nebulization study and characterization
Three InnoSpire Go devices were tested in triplicate into a cascade impactor (Copley NGI), operated with an extraction flow of 15 ± 0.5 L/min and temperature of 5 ± 2 C. Each device was loaded with a total of 4.2 ml of IN-006. Gravimetric weights were recorded to enable full mass balance calculations. Devices were operated for 90 s into the NGI. At the end of the run, samples were collected from all stages of the NGI and analyzed by UV microplate reader (A280). This nebulization method was used to determine delivered dose, APSD, and the collected samples were used to assess drug integrity following nebulization (e.g., ELISA, neutralization assay, and HPLC).

| SARS-CoV-2 RBD binding ELISA (for GLP nebulization characterization study)
In switching to these GLP assays, we also changed the antigen coat in ELISA assay from full Spike protein to Spike RBD that was produced to meet the more rigorous characterization requirements of regulatory authorities (Figure 3b). Post-nebulized samples were serially diluted (1200-0.00239 ng/ml, 10 points) and added to a 96-well microplate previously coated with 0.05 μg/ml of SARS-CoV-2 RBD manufactured by Celltrion. The bound primary sample was detected using antihuman IgG Fc-HRP conjugate. The signal was measured after TMB (3,3 0 ,5,5 0 -tetramethylbenzidine) treatment and acid stopping. The optical density values were fitted using 4 parameter logistic regression analysis and the relative binding activity of samples to SARS-CoV-2 RBD was determined from comparison of the EC 50 value of samples and that of CT-P59 in-house reference standard by PLA software. 5.9 | Size exclusion chromatography, high pressure liquid chromatography SEC was performed to evaluate the relative abundance of aggregates, monomers, and fragments under non-denaturing conditions for preand post-nebulization samples. This assay was performed using a Waters HPLC Alliance system on a TOSOH TSKgel G3000SWXL column (7.8 mm Â 300 mm) using aqueous buffered mobile phase at ambient temperature. The isocratic elution profile at a constant flow rate 1.0 ml/min was monitored using UV detection at 214 nm.

| Pseudovirus neutralization
5.10 | Ion exchange chromatography, high pressure liquid chromatography IEC was performed to evaluate the distribution of charge variants preand post-nebulization samples using cation exchange chromatography. The Waters HPLC Alliance system was equipped with a BioPro IEX SF analytical column (4.6 Â 100 mm) at ambient temperature. Gradient NaCl elution was performed at a constant flow rate of 0.8 ml/min and UV signals were obtained at 280 nm.

| SVPs using light obscuration
The number of SVPs in pre-and post-nebulization samples were measured by light obscuration method using the high accuracy (HIAC) liquid particle counting system. Analysis was performed by HACH ULTRA analytics, HIAC 9703 liquid particle counter equipped with 1.0 ml syringe and small bore probe at ambient temperature. Processing was performed using PharmSpec software. A minimum 2-week acclimation period was allowed between receipt of the animals and the start of treatment to accustom the rats to the laboratory environment. For respiratory parameters, five main male animals were conditioned to restraint tubes used for the respiratory function measurements in order to accustom the animals to the experimental procedures. This conditioning occurred over a minimum 10-day period before test/control item administration and was performed for a period of 30 min per day.

| GLP rat nebulization study
Any reaction noted during acclimation to the restraint tubes was noted.
The exposure system used was a flow-past rodent inhalation exposure system that allowed for inhalation by nose-only exposure.
The aerosol produced by the Aeroneb Solo nebulizers was discharged through a 40-mm diameter tube into a flow-past inhalation exposure system. The airflow rate through the exposure system was monitored and recorded manually during the aerosol generation. Airflow to the exposure system was controlled by the absolute volume of air supplying the aerosol generators using variable area flow meters. Control of the aerosol exhaust flow from the animal exposure system was achieved using a diaphragm valve, and the overall balance of airflows in the exposure system was monitored using pressure gauges. The including respiratory tract tissues (including carina, nasal cavity, nasopharynx, larynx, trachea, lungs and bronchi, and tracheobronchial lymph node) and all gross lesions from all main and recovery animals.

| Urea measurement for determination of BALF dilution factor
We sought to determine the extent to which the BALF samples were diluted during collection (due to rinsing of the lungs with saline). During this collection process, the airway fluid is inherently diluted, artificially decreasing the apparent concentration of therapeutics in the recovered BALF. We therefore measured the concentration of urea in the BALF samples and in the serum samples to determine the average extent of dilution (before dilution, urea concentrations are equal in the lung fluid and in the serum, allowing for the calculation of a dilution factor 64

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.