Development of a novel, pan-variant aerosol intervention for COVID-19

To develop a universal strategy to block SARS-CoV-2 cellular entry and infection represents a central aim for effective COVID-19 therapy. The growing impact of emerging variants of concern increases the urgency for development of effective interventions. Since ACE2 is the critical SARS-CoV-2 receptor and all tested variants bind to ACE2, some even at much increased affinity (see accompanying paper), we hypothesized that aerosol administration of clinical grade soluble human recombinant ACE2 (APN01) will neutralize SARS-CoV-2 in the airways, limit spread of infection in the lung and mitigate lung damage caused by deregulated signaling in the renin-angiotensin (RAS) and Kinin pathways. Here we show that intranasal administration of APN01 in a mouse model of SARS-CoV-2 infection dramatically reduced weight loss and prevented animal death. As a prerequisite to a clinical trial, we evaluated both virus binding activity and enzymatic activity for cleavage of Ang II following aerosolization. We report successful aerosolization for APN01, retaining viral binding as well as catalytic RAS activity. Dose range-finding and IND-enabling repeat-dose aerosol toxicology testing were conducted in dogs. Twice daily aerosol administration for two weeks at the maximum feasible concentration revealed no notable toxicities. Based on these results, a Phase I clinical trial in healthy volunteers can now be initiated, with subsequent Phase II testing in individuals with SARS-CoV-2 infection. This strategy could be used to develop a viable and rapidly actionable therapy to prevent and treat COVID-19, against all current and future SARS-CoV-2 variants.


Introduction
Early in the COVID-19 pandemic, sequencing of SARS-CoV-2 enabled recognition of the high degree of homology with SARS-CoV and the identification of ACE2 as a candidate receptor for both viruses (1,2). A series of publications in early 2020 defined the molecular details regarding structural interactions between the receptor binding domain of SARS-CoV-2 and the ACE2 receptor (3)(4)(5)(6). Blocking the Spike-ACE2 interaction provided a potential anti-SARS-CoV-2 therapeutic strategy and is the basis for virtually all successful vaccine designs (7). Proteins or peptides interacting with either of the binding partners have therapeutic potential and some very high affinity binders have been reported (8). Likewise, engineered antibodies can inhibit the Spike-ACE2 interaction, and several have received Emergency Use Authorization (EUA) as single agents or combinations from the FDA and EMA as systemic therapeutics (9)(10)(11). Recently, the EUA of a monoclonal antibody directed to the Spike protein was revoked for use as a single agent because of reduced activity against emerging viral variants (12). Use of recombinant ACE2 may prove to be a universal and robust therapeutic intervention, since all studied emerging SARS-CoV-2 variants continue to use ACE2 as the primary receptor. Importantly, although multiple entry receptors have been proposed, recent data have unequivocally shown that ACE2 is the essential SARS-CoV-2 receptor in vivo (13).
After the outbreak of the first SARS virus in 2003, soluble recombinant human ACE2 (APN01) was developed for systemic treatment of acute respiratory distress syndrome (ARDS) (14,15). In this indication, the catalytic activity of ACE2 in cleaving Ang II was exploited to reduce damage to the lung as observed in virus induced ARDS. Phase I and Phase II clinical trials demonstrated that APN01 had an acceptable safety profile and strongly reduced pathogenic Ang II levels (16).
Our group first reported in vitro SARS-CoV-2 neutralizing activity of APN01 in cells and human organoids (17). Importantly, interactions between Spike proteins of multiple variants of concern and clinical grade soluble human recombinant ACE2 (APN01) have been demonstrated to be of considerably higher affinity. Moreover, APN01 can neutralize all tested SARS-CoV-2 variants of concern and variants of interest (see accompanying manuscript). The increased affinity to both APN01 and, by implication, endogenous ACE2 receptor is thought to contribute to the enhanced infectivity and transmissibility observed for several of these variants (18). 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint Besides the emerging escape mutants, one of the difficulties for widely applicable Spike blocking therapies for people exposed to the virus or at early stages of disease, especially against variants, is their intravenous application. APN01 has recently undergone a randomized Phase 2 clinical trial for treatment of severe COVID-19 using intravenous administration (NCT04335136, manuscript in preparation). We reasoned that direct introduction of APN01 into the airways could locally neutralize the virus, limiting the spread of infection, and thereby limit damage to the lung. To probe the potential for therapeutic activity following introduction into the airways, we tested intranasal treatment in a mouse model of SARS-CoV-2 infection (13) with clinical grade APN01.
Results demonstrated strong protective activity, providing experimental proof of concept for direct APN01 administration into the airways. The critical path for such a novel intervention includes the development of an aerosol formulation of APN01 that retains both virus-binding activity and enzymatic activity. We report the successful development of inhalable APN01 and results of preclinical toxicology studies that support the safety of this intervention when administered by aerosol. These data pave the way for an inhalable universal early intervention strategy against all current and future SARS-CoV-2 variants. 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint

In vitro anti-SARS-CoV-2 activity of APN01
While anti-SARS-CoV-2 activity of soluble recombinant human ACE2 (APN01) has been reported previously (17), we first evaluated neutralizing activity of the cGMP produced, i.v. injectable APN01, in cell culture. As shown in Figure 1, one-hour exposure to concentrations of APN01 as low as 25 µg/ml completely neutralized SARS-CoV-2 as assessed in a four-day cytopathic effect (CPE) assay on Vero E6 cells.
Thus, clinical grade ACE2 exhibits antiviral activity without any apparent toxic effects. See accompanying manuscript for affinity/avidity measurements and neutralization of variants of concern. Serial dilutions of APN01 were prepared in assay medium (MEM supplemented with 2% fetal bovine serum and 50 μg/mL gentamicin) and a suspension of SARS-CoV-2 (USA-WA1/2020) was added to assess neutralization. For assessment of APN01 on viability, assay medium without virus was added. After one-hour incubation at 37°C, the dilutions were transferred to wells containing Vero E6 target cells (Multiplicity Of Infection 0.001). Incubation was continued for four days and cell numbers were assessed with a neutral red endpoint. Cytopathic Effect (CPE) was calculated as the average optical density (OD) for replicate infected and treated wells divided by average control OD X 100 (expressed as a percentage). Viability was calculated as the average 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint 6 optical density (OD) for replicate uninfected and treated wells divided by average control OD X 100 (expressed as a percentage).

APN01 protects from respiratory SARS-CoV-2 infections
We have recently developed a novel animal model that faithfully recapitulates SARS-CoV-2 infections and results in severe lung pathologies, weight loss, and, dependent on the mouse strain, in death of infected mice. This model is based on a mouse adapted SARS-CoV-2 virus (termed maVie16) and expression of ACE2 was found to be essential for infection and disease (13). We  Figure 2B). Treatment-related changes in body temperature did not reach statistical significance.
Importantly, APN01 treatment resulted in 100% survival of the SARS-CoV-2 infected mice whereas all controls succumbed to the infection (infected mice surviving to day 5 were moribund when sacrificed). Increase in lung tissue weight, a correlate of severe infection, was also significantly reduced by intranasal APN01 treatment ( Figure 2C). Thus, respiratory delivery of APN01 protects from SARS-CoV-2 infections. 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint

APN01 aerosolization
To aerosolize APN01 for preclinical studies in a way that can be scaled to future widespread and easy clinical use, we selected PARI LC PLUS nebulizers. These are widely available, effective and standardized devices already used in clinical applications. Aerosolized APN01 was collected using a custom fabricated condenser and analyzed for virus-binding activity and enzymatic activity for cleaving a fluorogenic substrate. Clinical grade APN01 is formulated for i.v. use at 5 mg/ml.
Recognizing that in vitro anti-SARS-CoV-2 activity was observed at concentrations as low as 25 µg/ml for the reference USA-WA1/2020 virus ( Figure 1) and even 10-20 times lower IC50/IC90 values for variants of concern and variants of interest (see accompanying paper), we aerosolized a range of concentrations ranging from 100 µg/ml to 2.5 mg/ml. The concentration of recovered APN01 was assessed by enzyme-linked immunosorbent assay (ELISA) measurements. Binding 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint 8 assays were conducted by coating plates with SARS-CoV-2 Spike RBD-Fc fusion protein and assessing APN01 binding ( Figure 3). Importantly, APN01 binding to SARS-CoV-2 Spike protein was almost identical after nebulization. Similar results were obtained in replicate assays for APN01 nebulized at 2.5 mg/ml and 0.1 mg/ml (Supplemental Table 1).  Table 2 presents results from replicate experiments with APN01 nebulized at 2.5 and 0.1 mg/ml. In summary, aerosolization of APN01 9 did not affect the structural integrity of APN01 in terms of its ability to bind SARS-CoV-2 Spike RBD nor its enzymatic activity.

Toxicologic assessment of assessment of aerosolized APN01
We next performed a comprehensive evaluation of the possible toxicity of twice daily inhalation administration of APN01 aerosols to beagle dogs for 14 consecutive days. Goals of this study, which was designed to support regulatory filings, included: (1) Table 1.  For treatment of experimental animals, multiple PARI LC PLUS nebulizers were multiplexed through a distribution plenum with hoses connected to oronasal masks. Analytical data (obtained by HPLC analysis of filters) demonstrated that test aerosols consistently achieved target exposure concentrations. Characteristics of the generated APN01 aerosols are summarized in Table 2.
Particle size distribution data showed that the generated particles were within the respirable size range for dogs (19) and met targets for both median mass aerodynamic diameter (MMAD) and geometric standard deviation (GSD). Table 2. Characterization of APN01 aerosols. Samples were collected at morning (AM) and afternoon (PM) exposures and characterized for APN01 concentrations to enable dose calculation and assess particle sizes to confirm respirability. GSD, geometric standard deviation; MMAD, median mass aerodynamic diameter.
The inhaled dose of APN01 following one hour exposure was calculated from the aerosol concentration, average volume inhaled (Minute Volume in L/min), exposure time and animal body weights. Calculated inhaled APN01 levels for each group after a single exposure (Study Day 1) and after repeat-dose exposure (after the first exposure on Study Day 11) are shown in Table 3.
Inhaled doses were concentration dependent and consistent over the treatment interval. Even dosed 12 at the lowest concentration, the inhaled dose was >0.3 mg/kg/exposure, far exceeding the minimum virus neutralizing activity of APN01. We also tested whether inhalation of APN01 might result in dispersion of APN01 outside the respiratory system. Systemic exposure [defined as serum levels of APN01 above the limit of quantitation (LOQ)] was very low in dogs in the low dose and mid dose groups as assessed on Days 1 and 14; in both groups, serum levels of APN01 were below the LOQ (0.5 ng/mL) in most animals at most time points (Table 4). In the high dose group, serum levels of APN01 on Days 1 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC    Table 3 for a listing of toxicology parameters studied.
No early deaths occurred during the study and no gross clinical signs of toxicity were seen in any study animal. Inhalation administration of APN01 aerosols had no effects on body weight, food consumption, clinical pathology parameters, heart rate, blood pressure, electrocardiography, blood oxygen saturation, blood pH, FOB parameters, or ophthalmology. Respiratory function evaluations (respiratory rate, tidal volume and minute volume) were inconclusive due to excitement and/or panting exhibited by study animals during measurement periods. Organ weights were comparable in all study groups. No gross or microscopic pathology was linked to APN01 administration. Moreover, no evidence of systemic or organ-specific toxicity was identified in any dog receiving twice daily 60-minute exposures to APN01 aerosols at all target concentrations of 0.019, 0.038, or 0.075 mg/L for fourteen consecutive days. On this basis, the NO(A)EL was determined to be the highest dose investigated, i.e. 0.075 mg/L. Thus, in agreement with toxicologic assessments of systemic exposure of APN01 in rodent and non-rodent species, APN01 exhibited an excellent tolerability and safety profile upon administration as an aerosol.
105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint

Discussion
Animal modeling conducted in this study with a mouse adapted SARS-CoV-2 strain has provided critical preclinical evidence for the efficacy of APN01 to prevent COVID-19 symptoms when delivered directly to the sites of respiratory infection via intranasal administration. These data are in agreement with the efficacy of ACE2 derived peptides delivered intranasally in a hamster model of COVID-19 (8), as well as novel lipopeptides, designed to inhibit viral entry, delivered as a nasal spray in a ferret model (21), all of these data pointing the way for early intervention using airway administered therapeutics. Approaches for limiting viral entry early in the course of infection via ACE2 based drugs may be complemented by approaches delivering antiviral drugs such as remdesivir and interferon β directly into the airways and lungs. Indeed combinations of remdesivir and interferon β are planned for clinical studies in the near future (NCT04647695).
As a COVID-19 intervention, APN01 offers key advantages, especially for treatment of emerging variants and APN01 exhibits markedly increased affinities/avidities to the variants RBD/Spike and enhanced neutralization activity (please see accompanying manuscript). Mutations detected in variants of concern and variants of interest cluster in the viral Spike protein and particularly in the receptor binding domain. Although these variants can in part escape neutralization by vaccines, convalescent plasma, and multiple therapeutic antibodies, all of these variants continue to use ACE2 as primary receptor. Importantly, our recent data for the first time also show that ACE2 deficient mice no longer exhibit signs of COVID-19 in our mouse adapted SARS-CoV-2 maVie16 infection model (18), demonstrating that ACE2 is essential for SARS as well as SARS-CoV-2 infections. The mouse adapted SARS-CoV-2 maVie16 strain bears two key mutations in the RBD to allow for mouse ACE2 receptor binding but remains susceptible to APN01. Thus, our novel and severe mouse COVID-19 model allowed us to assess the therapeutic activity of APN01 directly administered to the respiratory tree. Indeed, treatment with APN01 at the time of infection prevented any COVID-19 pathologies (lung), clinical symptoms (weight loss) and, importantly, completely protected from death. Similarly, intranasal therapy with murine ACE2, produced in exactly the same manner as APN01, completely prevented the SARS-CoV-2 maVie16 infection and all mice remained healthy (13), consistent with the APN01 studies. Since in contrast to therapeutic MoAbs or natural Abs (22,23) no virus escape mutants should occur in binding to ACE2, APN01 could be used as an inhalable universal therapy against all current and future SARS-CoV-2 variants. The results presented here support the high therapeutic potential and feasibility of 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint the administration of APN01 into the airways and lungs of patients for treatment of SARS-CoV-2 infection. Aerosolized APN01 retains virus binding and enzymatic activities. Furthermore, the aerosol generated using a commercial and widely used nebulizer has a particle size distribution consistent with delivery throughout the respiratory tract and could be delivered repeatedly at high doses to dogs without notable toxicities. Thus, APN01 aerosol administration should deliver effective antiviral therapy to the airways.
A limitation of our toxicology results is that only a single species of experimental animals is included. Additional confidence in the results could be generated by including non-human primates in the analysis. Of note, mice also did not show any signs of pathologies when they received APN01 or mouse soluble ACE2 into their respiratory system for 5 day efficacy studies. Escalation to twice per day and then increasing concentrations of ½ to the maximum feasible concentration is the anticipated design. A Phase I trial for safety and tolerability of aerosolized APN01 in healthy volunteers is currently underway (NCT number pending) which will be followed by Phase II trials in individuals infected with SARS-CoV-2. The latter trials will use viral clearance as the primary endpoint with severe disease and hospitalization as secondary endpoints.
In summary, our study demonstrates both the potent preclinical activity of locally administered APN01 in a mouse model of COVID-19 as well as the feasibility and excellent safety profile of APN01 delivered as an aerosol in a toxicologic assessment in dogs. These data, along with the data provided in our companion manuscript showing activity against numerous emerging variants, highlight the potential of aerosolized APN01 to serve as a potent pan-SARS-CoV-2 therapeutic. 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint

Materials and Methods
In vitro anti-SARS-CoV-2 neutralizing activity of APN01. Serial dilutions of APN01 were prepared in assay medium (MEM supplemented with 2% fetal bovine serum and 50 μg/ml gentamicin) and a suspension of SARS-CoV-2, USA-WA1/2020 was added to assess neutralization. For assessment of APN01 on viability, assay medium without virus was added.
After one-hour incubation at 37°C, the dilutions were transferred to wells containing Vero E6 Angiotenin II cleavage assay. Substrate was prepared by reconstituting MAPL-DNP (Anaspec) in DMSO to make 1 mM stock with gentle mixing by inversion to ensure dissolution. This solution was diluted in assay buffer (10 µM ZnCl2, 50 mM MES, 300 mM NaCl, 0.01% Brij L23, pH 6.5) to 1 mM and then further diluted 1:5 in assay buffer to prepare a 0.2 mM working solution of 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC MAPL-DNP. The substrate was prewarmed in an oven at 37 ºC prior to addition to the sample. APN01 samples recovered from the nebulization runs were diluted in assay buffer to 1.0 µg/ml, as quantified with hACE2 ELISA measurements. Further dilution in assay buffer was conducted to yield 100 ng/ml, 50 ng/ml, and 25 ng/ml solutions at volumes of 1 ml each. The three APN01 dilutions of each sample were loaded into the assay plate (Greiner) in quadruplicate in the center of the plate, avoiding the outer edges, while assay buffer blanks were loaded in columns 2 and 11. HPLC quantitation of APN01 on filters. The aerosol mass concentration in each oronasal inhalation exposure system was determined by collecting the aerosol on glass-fiber filters. Samples were collected at a constant flow rate equal to the port flow of the delivery tube, and the total volume of air samples was measured by a dry-gas meter. One aerosol sample per dose level was collected during each exposure. Filter samples were extracted with phosphate buffered saline (PBS) and stored refrigerated (approximately 4°C). Filter samples were analyzed for levels of APN01 using high performance liquid chromatography (HPLC) with UV wavelength detection according to a validated method. During method validation, quality control (QC) samples prepared in PBS at target APN01 concentrations of 49 and 245 μg/mL were demonstrated to be stable for at least 22 days when stored refrigerated (111% and 104% of the original value, respectively) and for at least 2 days when stored at room temperature (111% and 103% of the original value, respectively). For instrument calibration, primary standard solutions of APN01 with target 105 and is also made available for use under a CC0 license.
(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted September 20, 2021. ; https://doi.org/10.1101/2021.09.14.459961 doi: bioRxiv preprint 20 concentrations of 490 μg/mL were prepared by dilution of 0.5 mL of the test article formulation Aerosol Particle Size Distribution. Aerosol particle size distribution was determined twice per group during the study by collecting size-segregated aerosol samples using a 10-stage quartz crystal microbalance (QCM) cascade impactor (California Measurements Inc.; Sierra Madre, CA).
The aerosol output from one port of the exposure system was connected to the QCM and was sampled at least once per inhalation exposure level. The mass median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) of the test aerosol were calculated from the mass accumulated on each collection stage of the QCM by using a validated computer program (QCMSIZE) that was developed at IITRI.

Toxicology Studies.
Toxicology studies were conducted under Good Laboratory Practices. Animal studies were performed in full compliance with the Animal Welfare Act and in accordance to the NIH Guide for the Care and Use of Laboratory Animals.