An albumin-angiotensin converting enzyme 2-based SARS-CoV-2 decoy with FcRn-driven half-life extension

The emergence of new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mutants and breakthrough infections despite available coronavirus disease 2019 (COVID-19) vaccines calls for antiviral therapeutics. The application of soluble angiotensin converting enzyme 2 (ACE2) as a SARS-CoV-2 decoy that reduces cell bound ACE2-mediated virus entry is limited by a short plasma half-life. This work presents a recombinant human albumin ACE2 genetic fusion (rHA-ACE2) to increase the plasma half-life by an FcRn-driven cellular recycling mechanism, investigated using a wild type (WT) albumin sequence and sequence engineered with null FcRn binding (NB). Binding of rHA-ACE2 fusions to SARS-CoV-2 spike protein subdomain 1 (S1) was demonstrated (WT-ACE2 KD = 32.8 nM and NB-ACE2 KD = 31.7 nM) using Bio-Layer Interferometry and dose-dependent in vitro inhibition of host cell infection of pseudotyped viruses displaying surface SARS-CoV-2 spike (S) protein. FcRn-mediated in vitro recycling was translated to a five times greater plasma half-life of WT-ACE2 (t½ β = 13.5 h) than soluble ACE2 (t½ β = 2.8 h) in humanised FcRn/albumin double transgenic mice. The rHA-ACE2-based SARS-CoV-2 decoy system exhibiting FcRn-driven circulatory half-life extension introduced in this work offers the potential to expand and improve the anti-COVID-19 anti-viral drug armoury. Statement of significance The COVID-19 pandemic has highlighted the need for rapid development of efficient antiviral therapeutics to combat SARS-CoV-2 and new mutants to lower morbidity and mortality in severe cases, and for people that are unable to receive a vaccine. Here we report a therapeutic albumin ACE2 fusion protein (rHA-ACE2), that can bind SARS-CoV-2 S protein decorated virus-like particles to inhibit viral infection, and exhibits extended in vivo half-life compared to ACE2 alone. Employing ACE2 as a binding decoy for the virus is expected to efficiently inhibit all SARS-CoV-2 mutants as they all rely on binding with endogenous ACE2 for viral cell entry and, therefore, rHA-ACE2 constitutes a versatile addition to the therapeutic arsenal for combatting COVID-19.


Introduction
Coronavirus disease 2019 , caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) first identified in December 2019, has become a global pandemic [1] . As of February 2022, more than 424 million cases of COVID-19, includ-vector-based COVID-19 vaccines [11] . Moreover, emergence of new SARS-CoV-2 mutants may require seasonal vaccine redesign. The World Health Organization (WHO) classifies five SARS-CoV-2 variants (Alpha, Beta, Gamma, Delta and Omicron) for concern, with four shown to influence vaccine efficacy. In particular, significantly reduced neutralising antibodies are associated with the Delta and Omicron variants [12][13][14] . This highlights the need for new effective anti-viral therapeutics that can combat arising virus mutants.
SARS-CoV-2 enters cells by interaction of the receptor binding domain (RBD) in viral transmembrane spike (S) protein with cell membrane bound carboxypeptidase angiotensin converting enzyme 2 (ACE2) [15] . Recombinant human ACE2 (rhACE2) ectodomain, that resembles the endogenous viral cell entry receptor, has been used as a COVID-19 decoy and completed Phase I and II clinical trials, showing significant reduction in time on mechanical ventilation [16][17][18] . Furthermore, the therapeutic efficacy may also be enhanced through the negative regulator/modulator function of ACE2 on the peptide hormone angiotensin (Ang) II that is associated with acute lung injury linked to the severity of COVID-19 [ 18 , 19 ]. ACE2 can counterbalance Ang II functions by cleaving Ang II into Ang 1-7 and, thereby, counteract Ang II overactivity [ 18 , 20 ]. The duration of rhACE2 efficacy, however, is limited by a plasma half-life of ∼10 h in humans [16] .
Genetic fusion of rhACE2 to an immunoglobulin Fc region that interacts with the cellular recycling neonatal Fc receptor (FcRn) has been utilised as a strategy to extend the blood residence time of rhACE2 [21][22][23] . Induction of cytokine release syndrome (CRS), however, can result from high-level immune activation mediated by Fc interaction with Fc γ receptors (Fc γ Rs) on immune cells [24] .
This potentially life-threatening cytokine release can lead to cardiac dysfunction, respiratory distress syndrome and renal and/or hepatic failure [25] . Antibody-dependent enhancement (ADE) of virus infection is another mechanism by which the virus can infect host cell [26] . Virus-antibody immune complexes have shown to bind Fc γ Rs that can be internalised, leading to enhanced entry of the virus into host cells. The risk of exacerbating COVID-19 severity through Fc γ R engagement of virus-antibody complexes is a limitation for antibody-based drug designs [26] . Human serum albumin (HSA) does not engage with the classical Fc γ Rs [27] but with the non-classical Fc γ Rs, such as FcRn, to facilitate a long plasma half-life of approximately 19 days [ 27,28 ] that has been adopted in marketed albumin-based drugs [ 29 , 30 ]. Furthermore, single-point amino acid substitutions in HSA domain III have been used to tune the affinity to FcRn [31] . In this work, we designed recombinant human albumin-ACE2 (rHA-ACE2) fusions incorporating albumin sequences engineered with wild type (WT) or null-binding (NB) FcRn binding affinity in order to investigate the mechanistic role of FcRn-mediated cellular recycling in plasma half-life extension for potential exploitation in COVID-19 therapies.

cDNA design, construction and production of expression vectors
The rHA-ACE fusion proteins were designed by codon optimizing the nucleic acid sequence of the ACE2 ectodomain (amino acids 18-740 of Uniprot ID: Q9BYF1-1) for expression in a human cell line. This was subcloned by GenScript into pcDNA3.1 plasmids containing the WT and KAHQ (NB) variants [32] . Nucleic acids encoding a single GGGGS peptide linker connects the ACE2 C-terminal and the rHA N-terminal.

SDS-PAGE and Western blot analysis
Protein samples were separated on 10% SDS-PAGE and visualised by Coomassie Blue staining or blotted onto a PVDF membrane (Novex, #LC2005) for Western blot analysis. Membranes were blocked in 2% w/v skimmed milk powder in PBS (mPBS) and incubated for 2 h at room temperature (RT). For Western blot analysis of albumin, membranes were incubated with horseradish peroxidase (HRP) conjugated polyclonal goat anti-HSA antibody (Abcam, #ab19183) diluted 1:20 0 0 in 2% mPBS for 2 h at RT and subsequently washed 3 times in PBS before detection with TMB substrate (Sigma, #T0565). For Western blot analysis of ACE2, membranes were first incubated with monoclonal mouse anti-human ACE2 antibody (RnD System, #MAB933) diluted 1:250 in mPBS for 2 h at RT and washed 3 times in PBS before incubation with polyclonal HRP conjugated goat anti-mouse immunoglobulins antibody (Dako Agilent, #P0477) diluted 1:500 in mPBS for 2 h. Membranes were subsequently washed 3 times in PBS before developing with TMB substrate.

rHA-ACE2 fusion binding to SARS-CoV-2 S1 using Bio-Layer Interferometry
Binding kinetics of rHA-ACE2 to biotinylated SARS-CoV-2 S1 protein (Acro Biosystems, #S1N-C82E8) was measured by Bio-Layer Interferometry (BLI) using an Octet Red 96e system (Sartorius). Soluble ACE2 (Sigma Aldrich, #SAE0064) was used as a control. The biotinylated SARS-CoV-2 S1 protein was immobilised on Streptavidin Dip and Read TM sensor tips at a concentration of 8 nM in kinetics buffer (PBS supplemented with 0.02% Tween-20 and 0.1% bovine serum albumin (BSA)). For binding kinetics measurements rHA-ACE2 and soluble ACE2 samples were prepared in a 7-step two-fold dilution series starting at 3 μM and 125 nM, respectively, in kinetics buffer. Measurements were performed at 30 °C and 10 0 0 rpm shaking with a 200 s association and 600 s dissociation step. Tips were regenerated between samples by alternating between 10 s incubation in 10 mM Glycine at pH 3 and kinetics buffer 6 times. SARS-CoV-2 S1 protein coated streptavidin sensors in kinetics buffer was used for baseline subtraction. Data analysis was performed using the Octet data analysis software (version 10.0.1.6) using a 1:1 interaction model curve fitting.

Virus-like particles inhibition assay
The inhibitory effect of the rHA-ACE2 fusion on viral cell entry was investigated using virus-like particles (VLPs) decorated with SARS-CoV-2 S protein, as previously described [33] . In brief, HEK293T cells were transfected with lentiviral packaging vectors encoding the gag-pol, REV and eGFP reporter gene (a gift from Prof. Jacob Giehm Mikkelsen, Aarhus University, Denmark) and a SARS-CoV-2 S protein plasmid (a gift from Zhaohui Qian, Addgene plasmid, #145780) at a ratio of 3:6:4:4 by calcium phosphate precipitation. Media was exchanged after 24 h and VLPs were harvested from the cells 48 h post transfection, filtered (0.45 μm), and incubated with rHA-ACE2 fusion, soluble ACE2 or PBS as a control for 1.5 h at 37 °C and 5% CO 2 . The pre-incubated VLPs were subsequently transferred to HEK293T cells, previously transfected with TMPRSS2 (a gift from Roger Reeves, Addgene plasmid #53887) and human ACE2 (a gift from Hyeryuen Cho, Addgene plasmid #1736) by calcium phosphate precipitation, seeded at a density of 10,0 0 0 cells/well at a 1:1 ratio. The media was exchanged after 24 h, and cells were further cultured for 72 h. Transduced cells were analysed for eGFP fluorescence by fluorescence microscopy (IX73 inverted microscope, Olympus equipped with DP73 camera, Olympus) and flow cytometry (Cytoflex, Beckman Coulter) and analysed using the CytExpert software. The infectivity of VLPs pre-incubated with rHA-ACE2 fusions or soluble ACE2 was normalized to the infectivity of VLPs pre-incubated with PBS. IC 50 values were determined using the Origin 2018 software.
The cellular recycling assay was performed as previously described [28] . In brief, HMEC-1-FcRn cells, stably overexpressing human FcRn, were seeded at a density of 1 × 10 5 cells per well in 48 well plates coated with 1:50 dilution of Geltrex (Gibco, #2248603) in PBS. When near confluent, cells were washed twice with preheated PBS and incubated for 1 h at 37 °C and 5% CO 2 with 0.15 μM of rHA-ACE2 fusions or rHA controls in Hank's buffered salt solution (HBSS) adjusted to pH 6 with 1 M 2-(Nmorpholino)ethanesulfonic acid (MES) buffer. Cells were subsequently washed 5 times with ice cold PBS and left for 1 h at 37 °C and 5% CO 2 in 160 μL/well release media (complete media without FBS). Supernatant was harvested and analysed by sandwich ELISA.
For the sandwich ELISA analysis of the supernatants, Maxisorp plates (Thermo Fisher, #442404) were coated for 2 h at RT with polyclonal goat anti-HSA antibody (Sigma Aldrich, #A-7544), diluted 1:10 0 0 in PBS. Coated plates were then blocked with 2% mPBS for 2 h at RT followed by washing three times with PBS supplemented with 0.05% Tween (PBST). Supernatant samples and albumin controls were added to the plates and incubated overnight at 4 °C. For detection, plates were incubated for 1 h at RT with polyclonal HRP conjugated goat anti-HSA antibody (Abcam, # ab19183) followed by washing three times with PBST and developing with TMB PLUS2 substrate (Kem-En-Tec Diagnostics, #4395). All samples were performed in quadruplicates and the experiment was repeated five times.

Animals
Female and male double transgenic human FcRn (hFcRn + / + )/human serum albumin (hAlb + / + ) with a C57BL/6 background [34] were supplied by GenOway (Lyon, France). Pharmacokinetic studies were performed at the Animal Facility, Department of Biomedicine, Aarhus University, in accordance with the ethical approval in national guidelines for care and use of laboratory animals under the Danish Animal Experiment Inspector license #2018-15-0201-01399. Animals were housed in polycarbonate cages with wire tops, wood chip bedding, and access to ad libitum food and tap water.

Circulatory half-life of rHA-ACE2 fusions
Male and female mice were randomly divided into the following experimental groups; WT-ACE2 ( N = 4), NB-ACE2 ( N = 5) and soluble ACE2 ( N = 6). Three male mice were used for a PBS control. Mice were intravenously injected in the tail vein with either 4 mg/kg rHA-ACE2 fusion or an equimolar concentration of For serum detection by sandwich ELISA, Maxisorp plates (Thermo Fisher, #442404) were coated with capture antibodies (rHA-ACE2: monoclonal mouse anti-human ACE2 (RnD Systems, #MAB933) diluted 1:500 and for ACE2: goat anti-ACE2 (Abcam, #AF933) diluted 1:500) in PBS for 2 h at RT. Plates were first washed once with PBST and three times with PBS. Plates were then blocked with 2% mPBS for 2 h at RT and subsequently washed once in PBST and three times in PBS. Mouse serum samples or rHA-ACE2 standard series were then added and incubated overnight at 4 °C. Plates were washed once in PBST and three times in PBS before incubation with detection antibodies conjugated with HRP (rHA-ACE2: polyclonal goat anti-HSA (Abcam, #ab19183) diluted 1:10,0 0 0 and ACE2: monoclonal anti-poly Histidine-Peroxidase (Sigma-Aldrich, #A7058-1VL) in 2% mPBS for 2 h at RT. After washing once with PBST and three times with PBS, TMB PLUS2 (Kem-En-Tec Diagnostics, #4395) was used for plate development. Data was analysed in GraphPad Prism software version 9.3.1. by a two-phase decay model and interpolation to the associated standard series.

Statistical analysis
Experimental data was analysed using the Origin 2018 software and GraphPad Prism software v.9. The respective conducted statistical tests are indicated in the corresponding figure and table legends. One-way ANOVA and Turkey's test were used for statistical analysis of BLI experiments and viral inhibition assay. Unpaired ttest was used for statistical analysis of the recycling assay. A minimum value of p < 0.05 was considered statistically significant. To ensure reproducibility, all in vitro experiments were repeated at least 3 times.

rHA-ACE2 fusion expression and purification
Recombinant ACE2 ectodomain (residues 18-740) was genetically fused to the N-terminus of rHA WT, or an rHA NB variant engineered for null FcRn engagement by point mutations in the FcRn binding interface ( Fig. 1 ). The fusion proteins expressed in HEK293E cells were purified by HSA affinity chromatography (Supplementary Fig. 1). Identification of rHA-ACE2 fusion protein using SDS-PAGE showed the expected molecular weight of ∼150 kDa ( Fig. 2 A) and albumin ( Fig. 2 B) and ACE2 components ( Fig. 2 C) were verified respectively by Western blot analysis.

Binding of SARS-CoV-2 S1 protein
Binding of rHA-ACE2 to immobilised SARS-CoV-2 S1 protein was investigated by BLI and compared to binding of soluble ACE2 ( Fig. 3 ). Measurements showed high affinity binding of the two rHA-ACE2 fusions (WT-ACE2 and NB-ACE2) ( Table 1 ). The calculated binding kinetics based on a 1:1 fitting model revealed a K D of 32.8 nM and 31.7 nM for WT-ACE2 and NB-ACE2, respectively, with no significant differences in binding affinities between the fusions. The K D of 0.9 nM for soluble ACE2 showed a higher binding affinity compared to rHA-ACE2 fusion proteins, that can be attributed to decreased k on rates of WT-ACE2 (7.86E + 03 1/Ms) and NB-ACE2 (9.64E + 03 1/Ms) compared to soluble ACE2 (1.84E + 05 1/Ms) ( Table 1 ).

In vitro VLP inhibition
Lentiviral VLPs decorated with the SARS-CoV-2 S protein were used as a model to test in vitro viral cell entry inhibition of rHA-ACE2 fusions. HEK293T cells, overexpressing ACE2 and TMPRSS2 required for SARS-CoV-2 cell entry, were used as permissive cell line in combination with an integrated eGFP reporter gene to allow detection of transduced cells ( Fig. 4 B). Flow cytometric analysis of transduced cells showed efficient dose-dependent inhibition of VLP entry by the rHA-ACE2 fusions and soluble ACE2 ( Fig. 4 A,  Supplementary Fig. 2). The WT-ACE2 and NB-ACE2 fusions exhibited an IC 50 of 4 4 4.0 ± 217.7 nM and 748.0 ± 158.9 nM, respectively, whilst the IC 50 of soluble ACE2 was 38.5 ± 24.6 nM, that correlated with the binding kinetics observed by BLI. There was no statistically significant difference observed between the IC 50 values of the two rHA-ACE2 fusions (WT-ACE2 and NB-ACE2). Control WT rHA did not show any inhibitory effect on the VLPs at any of the investigated concentrations.

FcRn-mediated cellular recycling
A cellular recycling assay established by our group [ 28 , 35 ] was utilised to investigate FcRn-mediated cellular recycling of rHA-ACE2 in HMEC-1-FcRn overexpressing cells. Different rHA standard variants with engineered FcRn affinities, referred to as NB rHA and WT rHA, were used to validate the assay ( Fig. 5 ). The level of recycled protein in the media was determined by sandwich ELISA detecting the rHA component of the rHA-ACE2 fusions. The cellular recycling showed the expected difference of WT-ACE2 > NB-ACE2, that was greater than the cellular recycling detected for WT rHA ( Fig. 5 ).

In vivo circulatory half-life of rHA-ACE2 fusions
The circulatory half-life of rHA-ACE2 fusions were benchmarked against soluble ACE2 in the double transgenic mouse model with both human FcRn and HSA genes (hFcRn + / + , hAlb + / + ) under the control of an endogenous promoter [34] . Proteins were intravenously administered by tail vein injection, and blood subsequently collected at selected time points to determine intact fullsize protein levels in serum by sandwich ELISA. The circulatory half-life calculated from the elimination phase (t ½ β) was shown     to be 13.5 h and ∼3.1 h for WT-ACE2 and NB-ACE2 fusions, respectively, compared to the t ½ β of approximately 2.8 h for soluble ACE2 ( Fig. 6 ).

Discussion
One of the first compounds investigated for treatment of COVID-19 was soluble rhACE2, used to neutralise SARS-CoV-2 and minimise organ injury related to increased Ang II concentrations [18] . The potential of rhACE2 as a decoy for SARS-CoV-2 by mimicking its cognate cell entry receptor is exemplified in a recent phase II clinical trial and case study [ 17 , 18 ]. Fusion of the rhACE2 protein to an immunoglobulin Fc region has led to improved circulatory half-life of the decoy due to FcRn-mediated cellular recycling of the protein in vivo [21][22][23] . In recent work, the humanised decoy antibody ACE2-Fc fusion protein abrogated virus replication in ACE2-expressing lung cells and lung organoids [36] . In another study, recombinant ACE2-IgG1 Fc or ACE2-IgA Fc fusion proteins incorporated an ACE2 sequence optimised for enhanced binding affinity to S protein of different virus variants [37] . A rhACE2-Fc fusion protein has also been investigated in a transgenic hACE2 mouse model and showed dose-dependent prophylactic and therapeutic efficacy compared to soluble ACE2 [38] . Interaction of the Fc fragment with Fc γ Rs on immune cells, however, can potentially elicit adverse effects by exaggerating an adverse immune response in COVID-19 [39] . In this work, we adopted albumin as an alternative half-life extension technology to overcome the risks associated with Fc fusions. An rHA-ACE2 fusion containing a wild type albumin sequence, and an FcRn null-binding albumin sequence [32] , previously shown by our laboratory to provide a tool to investigate mechanistic FcRn-driven cellular recycling of an alternative albumin fusion [40] , were produced. Inclusion of an NB-ACE2 variant allowed investigations into the influence of the fusion on the engagement of rHA with FcRn and confirmation of a predominant FcRn-driven half-life extension rather than a molecular weight effect.
The mechanism by which SARS-CoV-2 infects host cells is dependent on the virus transmembrane S protein targeting membrane bound ACE2 [41] , therefore, complex formation between the viral S protein and designed ACE2 decoy is crucial. BLI analysis of rHA-ACE2 binding affinity to SARS-CoV-2 S1 protein showed the equilibrium dissociation constant to be in the nanomolar range, but with lower affinity compared to soluble ACE2. The reduced binding affinities of rHA-ACE2 fusions most likely arise from fusing the ACE2 ectodomain to albumin, which may cause some steric hindrance on binding. The binding affinities, however, are still similar to ACE2-Fc fragment fusions previously reported in preclinical studies [ 21 , 23 ]. The rHA-ACE2 fusions were able to prevent VLP infection of host cells, even though the IC 50 values were increased compared to soluble ACE2, possibly due to steric hindrance effects on VLP engagement from incorporation of albumin. Optimisation of the linker flexibility and length between the rHA and ACE2 domain in the fusion could potentially reduce steric hindrance and increase the inhibitory effect of the fusion protein. Furthermore, the use of SARS-CoV-2 infection of primary target cells from human airway epithelium or organoids could also be used to further evaluate the natural virus inhibition potential [42] . However, the FcRnmediated increase in half-life of the fusion protein is expected to more than compensate for the decreased inhibitory efficiency observed in vitro.
IC 50 values, determined by an in vitro VLP inhibition assay, between the two fusions were not significantly different, indicating no enhancement in cellular entry due to FcRn engagement. Moreover, this supports ACE2 decoy function not being influenced by amino acid mutations in the predominant FcRn-binding region lying in albumin domain III. Greater cellular recycling of WT-ACE2 was observed in FcRn-overexpressing endothelial cells compared to NB-ACE2 that was included as a negative control to investigate FcRn-driven cellular recycling. This is in accordance with our previous albumin fusion recycling work using the same WT and NB sequence [40] . The concentration of detected protein in medium was increased for the rHA-ACE2 fusion compared to control rHA, possibly due to engagement of the ACE2 domain with other receptors involved in the renin-angiotensin system [43] .
The expected FcRn-mediated difference showing WT-ACE2 > NB-ACE2 was maintained in a physiologically relevant model, supporting the thesis that the half-life of the fusion protein is highly dependent on FcRn engagement. The WT-ACE2 fusion had a 4.8-(t ½ β = 13.5 h) and 4.4-fold higher circulatory half-life compared to soluble ACE2 (t ½ β = 2.8 h) and NB-ACE2 (t ½ β = 3.1 h), respectively. The circulatory half-life of the rHA-ACE2 fusions were lower than previous studies performed with a bispecific T-cell engageralbumin WT fusion in the same double transgenic (hFcRn + /+ , hAlb + /+ ) mouse model, where a circulatory half-life of 26 h was obtained [40] .
ACE2 is a dipeptidyl carboxypeptidase known to cleave Cterminal dipeptides from angiotensin I to produce angiotensin II [44] and functions similar to carboxypeptidase A (CPA) shown to cleave C-terminal peptides on HSA resulting in reduced FcRn affinity and circulatory half-life [45] . Genetic fusion of ACE2 to the Cterminal of rHA may, therefore, be favourable in future designs. Alternatively, a fusion protein could be designed containing mutations in the ACE2 catalytic active sites that inactivates the enzyme. This approach was applied in previous work by Tada et al. in which the inactivated ACE2 ectodomain fused to Fc domain 3 of the Ig heavy chain could still inhibit viral host cell entry in vivo [22] . In addition, serum stability analysis ( Supplementary Fig. 3) of the rHA-ACE2 fusion indicated a decline in protein integrity over days in 90% fetal bovine serum. Future work could focus on optimisation of stability parameters such as fusion linker engineering, shown previously to influence the stability of albumin fusion proteins [46] . The substantial increase in circulatory half-life of the WT-ACE2 fusion over soluble ACE2, obtained in our work, is still expected to translate into less frequent dosage and longer lasting effects. However, the therapeutic effect of the fusion protein compared to soluble ACE2 needs to be evaluated in future studies in established COVID-19 models. To better investigate the role of human FcRn-mediated half-life extension on the therapeutic effect, an in vivo COVID-19 model with a humanised FcRn and albumin background could be developed. In addition to the therapeutic efficacy, the safety of the fusion protein should be investigated in future work, to evaluate potential adverse effects that may arise from prolonged circulation. However, based on promising clinical trial results on the administration of ACE2 [ 16 , 17 ] and the approval of albumin protein fusions [ 29 , 30 ], minimal side-effects and immunogenicity are anticipated for the rHA-ACE2 fusion protein.
Emergence of new SARS-CoV-2 variants, can potentially decrease the efficacy of current vaccines and antibody treatments [11] , and place patients with compromised immune system or other pre-existing medical conditions at higher risk for more severe COVID-19 effects. This highlights the need for a therapeutic, effective against both, existing and future variants of the virus. Potent binding and inhibition of different virus variants has been shown for several ACE2 based decoys [ 22 , 37 , 38 ], due to the resemblance to the endogenous host cell receptor. The rHA-ACE2 fusion protein, therefore, is also expected to neutralise different mutants with similar potency. The albumin fusion protein may, therefore, provide a valuable tool to effectively treat new emerging SARS-CoV-2 variants. In summary, this work introduces an rHA-ACE2based SARS-CoV-2 decoy with FcRn-driven half-life extension that can potentially be used to maximise therapeutic effects and expand the required armoury to combat COVID-19.

Data availability
The data supporting the presented findings are provided in the supplementary information. Detailed raw and progressed data can be made available upon reasonable request.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.