key: cord-0722336-5hhwppu7
authors: Findlay-Wilson, Stephen; Easterbrook, Linda; Smith, Sandra; Pope, Neville; Humphries, Gareth; Schuhmann, Holger; Ngabo, Didier; Rayner, Emma; Otter, Ashley David; Coleman, Tom; Hicks, Bethany; Graham, Victoria Anne; Halkerston, Rachel; Apostolakis, Kostis; Taylor, Stephen; Fotheringham, Susan; Horton, Amanda; Tree, Julia Anne; Wand, Matthew; Hewson, Roger; Dowall, Stuart David
title: Development of a cost-effective ovine antibody-based therapy against SARS-CoV-2 infection and contribution of antibodies specific to the spike subunit proteins
date: 2022-05-06
journal: Antiviral Res
DOI: 10.1016/j.antiviral.2022.105332
sha: cd3afae7f20051b4e3ed3804322be93e7a0adfec
doc_id: 722336
cord_uid: 5hhwppu7

Antibodies against SARS-CoV-2 are important to generate protective immunity, with convalescent plasma one of the first therapies approved. An alternative source of polyclonal antibodies suitable for upscaling would be more amendable to regulatory approval and widespread use. In this study, sheep were immunised with SARS-CoV-2 whole spike protein or one of the subunit proteins: S1 and S2. Once substantial antibody titres were generated, plasma was collected and samples pooled for each antigen. Non-specific antibodies were removed via affinity-purification to yield candidate products for testing in a hamster model of SARS-CoV-2 infection. Affinity-purified polyclonal antibodies to whole spike, S1 and S2 proteins were evaluated for in vitro for neutralising activity against SARS-CoV-2 Wuhan-like virus (Australia/VIC01/2020) and a recent variant of concern, B.1.1.529 BA.1 (Omicron), antibody-binding, complement fixation and phagocytosis assays were also performed. All antibody preparations demonstrated an effect against SARS-CoV-2 disease in the hamster model of challenge, with those raised against the S2 subunit providing the most promise. A rapid, cost-effective therapy for COVID-19 was developed which provides a source of highly active immunoglobulin specific to SARS-CoV-2 with multi-functional activity.

The outbreak of coronavirus disease first identified in 2019 , caused by infection with the 39 etiological agent severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), was declared a 40 pandemic on 11 th March 2020 (Cucinotta and Vanelli 2020) . It continues to blight human public 41 health, and the race to battle it with countermeasures continues apace. Whilst vaccines have made a 42 valiant effort in the control of COVID-19, none of the current vaccines offer sterilising protection and 43 there are major obstacles to overcome for global control of the virus (Kim, Marks et al. 2021) . 44 Therefore, new antiviral strategies will continue to play an important role in mitigating disease. 45

One of the first treatments explored early in the pandemic was antibody therapy sourced from 46 previously infected patients. Convalescent human plasma was authorised for emergency use by the 47 U.S. Food and Drug Administration (FDA) in August 2020 for the treatment of hospitalised patients 48 with COVID-19. However, there are at least 176 registered clinical trials assessing convalescent 49 plasma with seemingly contradictory results being reported (Piechotta, Chai et al. 2020, Ning, 50 Abagna et al. 2021). Major problems with convalescent plasma therapy include quality control and 51 standardisation, defining optimal dosing and timepoints and the risk of infection with unknown 52 blood-borne infectious agents (Ning, Abagna et al. 2021 ). These issues can be overcome by using 53 intravenous immunoglobulins (IVIG) which are sterile and purified from plasma collected from large 54 pools of donors (Cao, Liu et al. 2020). However, the same issue applies with the reliance on human 55 material as the source material. 56

As an alternative, animal-derived antibodies are more applicable for large scale production and 57 standardisation. Ovine immunoglobulin G (IgG)-based products have been widely used as snake 58 antivenoms (Gutierrez, Leon et al. 2011) , and have been applied as treatments to infectious diseases 59 including rabies (Redwan el, Fahmy et al. 2009 ) and tetanus (Redwan el, Khalil et al. 2005) . The 60 targeting of polyclonal antibodies to multiple epitopes enables a multitude of effector functions, 61 such as steric hindrance (preventing the virus from attaching to the host cell surface), aggregation made using a ratio of ~5mg recombinant protein per ~1g of de-hydrated, cyanogen bromide-111 activated Sepharose (Cytiva) according to manufacturer's instructions, resulting in 5mL final resin 112 per antigen. Unbound antigen was washed out with 3x 2 column volumes (CV) of 0.1M NaOAc, 113 pH4.0, 500mM NaCl, followed by 3x2 CV of 0.1M Tris-HCl, pH8.0, 500mM NaCl. This wash cycle was 114 repeated twice. For the purification of antibodies, the resins were transferred to 10mL 115 polypropylene gravity flow columns (ThermoFisher). 116

The PEG precipitated IgG-fraction was pH-adjusted by the addition of 1M HEPES, pH8.0 (1.25mL per 117 10mL of original IgG fraction). After centrifugation (10 min, 4000xg, 20° C), the clarified supernatant 118 was loaded on the respective antigen column equilibrated in 10mM HEPES p with a contact time of 119 approx. 1.5h. The columns were washed with 10 CV of equilibration buffer, followed by 10 CV 10mM 120 HEPES pH8.0, 300mM KCl. Antibodies were eluted with 100mM glycine, pH 2.5. Protein containing 121 fractions were pooled and adjusted to pH7-8 with 1M Tris-HCl, pH9.0 (1/5 th of the original volume) 122 before dialysis into DPBS. 123 124 2.5 Enzyme-linked immunosorbent assay 125 Nunc MaxiSorp microtitre plates were coated with 2 g/ml recombinant protein in bicarbonate 126 buffer overnight at 2-8 °C. Plates were washed 3 times with phosphate buffered saline (PBS) 127 containing 0.05% Tween20 (PBST)and blocked for 1 hour at 37 °C with blocking buffer (5% skimmed 128 milk powder in PBS). Plates were incubated for 1 hour at 37 °C with prediluted samples (sera 129 collected from sheep bleeds or purified antibody preparations); washed with PBST; and incubated 130 with a donkey anti-ovine IgG horseradish peroxidase (HRP) conjugate (Product 713-035-003; Jackson 131 ImmunoResearch, USA) for 1 hour at 37 °C. After further washing, TMB substrate was added and the 132 reaction was stopped by the addition of stop solution before reading the optical density at a 133 wavelength of 450 nm.

Anti-RBD antibodies were determined using the fully quantitative Roche Elecsys anti-SARS-CoV-2 S 137 assay (ACOV2 S), a species independent electrochemiluminescent immunoassay (ECLIA CoV-2 nucleocapsid. The cycling conditions were: 25 °C for 2 minutes, 50 °C for 15 minutes, 95°C for 298 2 minutes, followed by 45 cycles of 95 °C for 3 seconds, 55 °C for 30 seconds. The quantification 299 standard was in vitro transcribed RNA of the SARS-CoV-2 N ORF (accession number NC_045512.2) 300 with quantification between 1 x 10e1 and 1 x 10e6 copies/µl.

The left lung lobe and a sagittal section of the entire nasal cavity were fixed by immersion in 10% 304 neutral-buffered formalin, processed and embedded into paraffin wax. The nasal cavity was 305 decalcified using an EDTA-based solution prior to embedding. Sections of 4 µm were cut and stained 306 with hematoxylin and eosin (HE) and examined microscopically. In addition, samples were stained 307 using the RNAscope technique to visualise SARS-CoV-2 virus RNA. Briefly, tissues were pre-treated 308 with hydrogen peroxide for 10 min (room temperature), target retrieval for 15 min (98-101 °C) and Purified IgG pooled from plasmapheresis events was tested alongside antibody preparations that 339 had been affinity-purified to remove non-specific antibodies. Antibodies generated in response to 340 immunisation with full-length spike proteins and the S1 or S2 subunits all recognised recombinant 341 whole spike protein ( Figure 2a) . As expected, antibodies generated from sheep immunised with S2 342 antibodies produced low levels of S1-specific responses (Figure 2b) and similarly S1-derived 343 antibodies demonstrated low responses to S2 antigen (Figure 2c ). The affinity-purified preparations 344 consistently gave higher binding compared to the parent purified antibody preparations. 345 Wuhan-like virus with an IC50 value of 49 ng/ml. Antibodies raised against the S1 and S2 spike 351 protein also had neutralising activity with IC50 values of 1.071 and 0.428 µg/ml respectively (Table 1) . 352

All preparations had strong neutralisation activity to both SARS-CoV-2 strains, although ~ 10-fold 353 higher antibody concentrations were required to neutralise the Omicron BA.1 variant. 354

To determine the breadth of in vitro activity of antibodies, a range of tests were undertaken. Firstly, 355 binding specific to the receptor-binding site was assessed, with both the full-length and S1-specific 356 antibodies demonstrating strong recognition, whereas the S2-specific antibody did not bind ( Figure  357 3a). Antibody-dependent complement deposition was observed with all three antibody 358 preparations, with lower but similar levels observed for S1-and S2-specific preparations compared 359 to those derived from sheep immunised with full-length spike protein (Figure 3b ). Using an antibody-360 dependent neutrophil phagocytosis assay, the full-length produced antibodies had higher activity 361 than S1-derived antibodies, whereas the S2-derived antibodies were less opsonophagocytotic 362 ( Figure 3c ). Antibodies raised against the full-length spike protein reacted strongly to the RBD and 363 spike antigens from different SARS-CoV-2 variants, including alpha (B.1.1.7), beta (B.1.351) and 364 gamma (P.1), whereas the S1-raised preparation cross-reacted only with the RBD of the variants, and 365 S2 showed limited cross-reactivity ( To assess the ability of antibodies to protect against SARS-CoV-2 disease, hamster received 2mg 381 antibody 24 hours before challenge. In addition, a further group received antibodies specific for full-382 length spike protein 3 days post-challenge for therapeutic effect. Results demonstrated that in 383 comparison to a PBS control group, hamsters that received the antibodies lost less weight, with the 384 groups that received full-length or S2 specific antibodies showing no evidence of any weight loss 385 ( Figure 4a ). When the maximum weight loss of individual animals was analysed, all control animals 386 lost weight, whereas weight loss in other groups was less marked (Figure 4b ). The differences 387 between the PBS group and the full-length and S2 groups were statistically significant (P=0.0202 and 388 0.0051, respectively) whereas significance was not reached for the S1 and delayed full-length 389 treated groups (P=0.0656 and 0.02623, respectively). The clinical score of animals which received 390 either the full-length spike, S1 subunit and S2 subunit antibodies were markedly lower than the PBS 391 control group (Figure 4c) . Observations in the S2 subunit group were only the minimal sign of ruffled 392 fur which was seen in two animals. In the PBS control group, three animals met humane endpoints 393 during the course of the study; but this difference was just outside of statistical significance 394 (P=0.056, Log-Rank survival test). 395 396

Prior to challenge, a blood sample was collected to enable the concentration of circulating antibody 398 to be evaluated. Results showed that that in the full-length spike and S1 subunit groups, one animal 399 had low amounts of antibodies (Figure 5a ). To determine whether this was the cause of a failure to 400 protect against disease, a correlation was plotted with the maximum weight loss which 401 demonstrated that the animals with low ovine-specific antibody concentrations were those which 402 showed a more severe clinical disease (Figure 5b) . 403

Nasal wash and a pharyngeal swab were collected 2 days post-challenge to assess the levels of live 406 SARS-CoV-2. In animals which received antibody treatment prior to challenge, there were lower 407 levels of live virus in the nasal wash and pharyngeal swabs samples ( Figure 6a) ; however, these did 408 not reach statistical significance with the full-length group being the closest in the nasal wash sample 409 Antibodies were raised against three antigens: whole spike protein and its two constitutive parts, 454 the S1 and S2 subunits. This contrasts with others who have focused on generating responses to the 455 RBD region. An equine polyclonal preparation using the RBD of SARS-CoV-2 as the immunogen 456 demonstrated neutralisation activity but without efficacy testing in a disease model (Zylberman, 457 Sanguineti into cells, such as the fusion protein, and thus has conserved sequences which may explain the 494 reactivity observed in some samples collected prior to the COVID-19 pandemic (Shrock, Fujimura et 495 al. 2020 ). The S2-specific neutralisation may be due to several mechanisms, including the S2 subunit 496 altering the structural change of the spike affecting the binding between the virus and its receptor 497 which may prevent formation of a complementary surface between the two required for efficient 498 neutralisation (Li, Li et al. 2005) . Other mechanisms include the S1 subunit and receptor binding 499 being reversible so the virus may fail to stably dock without subsequent insertion of the fusion 500 antibodies binding the S2 region hindering binding to the receptor (Zeng, Hon et al. 2006) . 502

We report two animals which underwent antibody administration via the intraperitoneal route 503 which did not translate into increased levels of circulating antibody in the blood. These animals were 504 maintained in the dataset and analysis due to undergoing all procedures as planned and no 505 substantial reason for exclusion outside of the subsequent measurement of antibody levels. It is not 506 known why the antibody failed to enter the circulation. Using the intravenous route, which is the 507 likely delivery route for use in human administration, is possible in small animal models but requires 508 either very small samples being delivered or catheterisation of animals (Dowall, Taylor et al. 2013) . 509

In providing interim recommendations on assessing the effectiveness of convalescent plasma In addition to testing as a prophylactic, we also assessed giving the affinity-purified full-length spike 523

Ab 3 days post-challenge to look at therapeutic effect. This timepoint was chosen due to clinical 524 signs first being present in infected hamsters at this timepoint (Dowall, Salguerio et al. 2021 ). In 

COVID-19 convalescent plasma: Interim recommendations from the AABB

WHO Declares COVID-19 a Pandemic

Development of a hamster natural transmission model of SARS-CoV-2 infection

Development of a Hamster Natural Transmission Model of SARS-CoV-2 Infection

Catheterized guinea pigs 606 infected with Ebola Zaire virus allows safer sequential sampling to determine the pharmacokinetic 607 profile of a phosphatidylserine-targeting monoclonal antibody

Ebola virus disease in guinea pigs using EBOTAb, an ovine antibody-based therapeutic

Development of a Cost-effective Ovine Polyclonal Antibody-Based Product, EBOTAb, 615 to Treat Ebola Virus Infection

Post-exposure treatment of 618 non-human primates lethally infected with Ebola virus with EBOTAb, a purified ovine IgG product

Nonneutralizing functional 621 antibodies: a new "old" paradigm for HIV vaccines

Potential use of convalescent plasma for SARS-CoV-2 623 prophylaxis and treatment in immunocompromised and vulnerable populations

Evaluation of the 628 safety and efficacy of XAV-19 in patients with COVID-19-induced moderate pneumonia: study 629 protocol for a randomized, double-blinded, placebo-controlled phase 2 (2a and 2b) trial

Antivenoms for snakebite 632 envenomings

variants, spike mutations and immune escape

Recombinant polyclonal antibodies: the next generation of antibody 637 therapeutics?

Convalescent Plasma Antibody Levels 646 and the Risk of Death from Covid-19

Looking beyond COVID-19 vaccine phase 3 trials

Diphtheria to Cancer, COVID-19, and Beyond

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor

Metagenomic Nanopore Sequencing 658 of Influenza Virus Direct from Clinical Respiratory Samples

Structure of SARS coronavirus spike receptor-660 binding domain complexed with receptor

The Impact of Mutations in SARS

Early High-Titer Plasma Therapy to Prevent Severe Covid-19 in

Older Adults

Non-Neutralizing Antibodies Directed against HIV and Their 674 Functions

Development and application of 676 therapeutic antibodies against COVID-19

Convalescent plasma or hyperimmune 679 immunoglobulin for people with COVID-19: a living systematic review

The receptor binding domain of the viral spike protein is an immunodominant and 685 highly specific target of antibodies in SARS-CoV-2 patients

Ovine anti-rabies 687 antibody production and evaluation

Production and purification of ovine anti-689 tetanus antibody

A double edged-sword -The Complement System during SARS-CoV-2 infection

Cell entry mechanisms of

SARS-CoV-2

Viral epitope profiling 700 of COVID-19 patients reveals cross-reactivity and correlates of severity

Pathogenesis and transmission of SARS-CoV-703 2 in golden hamsters

Structural and Functional Basis of SARS-CoV-2 Entry by Using Human 706 ACE2

Escape from neutralizing antibodies by SARS-CoV-2 spike protein 711 variants

Antibody-713 dependent enhancement of coronavirus

Molecular Architecture of the SARS-CoV-2 Virus

Quantitative comparison of the efficiency of antibodies against S1 and S2 subunit of SARS 719 coronavirus spike protein in virus neutralization and blocking of receptor binding: implications for 720 the functional roles of S2 subunit

Development of a hyperimmune equine serum 728 therapy for COVID-19 in Argentina

Reactivity to whole spike protein. (b) Reactivity to S1 subunit protein. 739 (c) Reactivity to S2 subunit protein. Lines show mean values. Solid line, purified antibodies

Figure 3. Functional activity of affinity-purified antibodies produced against recombinant SARS

CoV-2 glycoproteins. (a) Binding to the SARS-CoV-2 RBD. Bars show mean values with error bars 744 denoting standard error from triplicate samples. (b) Antibody-dependent complement deposition

Results from a single assay are shown. (b-c) Data is 746 calibrated to the NIBSC 20/162 SARS-CoV-2 antibody diagnostic calibrant

Clinical outcomes of hamsters receiving antibody preparations after challenge with SARS

CoV-2. (a) Weight of animals. Lines show mean value with error bars denoting standard error

Three animals in the PBS group met humane clinical endpoint as 752 indicated by skull and crossbones symbol. (b) Maximum weight loss of animals, with line and whisker 753 plots showing mean value and standard error. (c) Clinical score. Lines show mean values with error 754 bars denoting standard error

Results show mean absorbance level 758 from duplicate wells from each animal tested at a 1:100 dilution. Bar and whisker plots denote mean 759 and standard error. (b) Comparison of animal level at time of challenge with maximal weight loss 760 observed after challenge with SARS-CoV-2

Figure 6. Virology readouts of hamsters receiving ovine antibody preparations after challenge with

Quantification of live virus detected by focus-forming assay in nasal wash and 764 pharyngeal swab samples collected 2 days post-challenge with SARS-CoV-2. Bars show mean values 765 with error bars denoting standard error. No statistical significance between groups receiving 766 antibodies compared to PBS control (P>0.05). (b) Viral RNA levels in pharyngeal swabs. Bars show 767 mean values with error bars denoting standard error. *, P<0.05. (c) Viral RNA levels in lung tissue 768 collected at necropsy. Open circles indicate animals which met humane clinical endpoints

Representative microscopic images of lungs and nasal cavities of hamsters receiving 772 ovine antibodies preparations after challenge with SARS-CoV-2. Top row, lung-multifocal to 773 patchy areas of pneumonic consolidation (arrows) (H&E); middle row, nasal cavity-inflammation 774 and degeneration of the mucosa with variable luminal exudate (asterisks). Inset, higher power 775 images of nasal epithelium (x 800 magnification) (H&E)

CoV-2 viral RNA in the mucosa and luminal exudate (in situ hybridisation). challenge with SARS-CoV-2. (a) Area of consolidation in the lung as a percentage. 780 (b) Total pathology score in the lung. (c) Total pathology score in the nasal cavity. Line and whisker 781 plots show mean value and standard error

Supplementary figure 2: Gating strategy for the ADNP assay