key: cord-300847-ycuiso0g
authors: Li, Wei; Drelich, Aleksandra; Martinez, David R.; Gralinski, Lisa; Chen, Chuan; Sun, Zehua; Schäfer, Alexandra; Leist, Sarah R.; Liu, Xianglei; Zhelev, Doncho; Zhang, Liyong; Peterson, Eric C.; Conard, Alex; Mellors, John W.; Tseng, Chien-Te; Baric, Ralph S.; Dimitrov, Dimiter S.
title: Rapid selection of a human monoclonal antibody that potently neutralizes SARS-CoV-2 in two animal models
date: 2020-06-02
journal: bioRxiv
DOI: 10.1101/2020.05.13.093088
sha: 
doc_id: 300847
cord_uid: ycuiso0g

Effective therapies are urgently needed for the SARS-CoV-2/COVID19 pandemic. We identified panels of fully human monoclonal antibodies (mAbs) from eight large phage-displayed Fab, scFv and VH libraries by panning against the receptor binding domain (RBD) of the SARS-CoV-2 spike (S) glycoprotein. One high affinity mAb, IgG1 ab1, specifically neutralized replication competent SARS-CoV-2 with exceptional potency as measured by two different assays. There was no enhancement of pseudovirus infection in cells expressing Fcγ receptors at any concentration. It competed with human angiotensin-converting enzyme 2 (hACE2) for binding to RBD suggesting a competitive mechanism of virus neutralization. IgG1 ab1 potently neutralized mouse ACE2 adapted SARS-CoV-2 in wild type BALB/c mice and native virus in hACE2 expressing transgenic mice. The ab1 sequence has relatively low number of somatic mutations indicating that ab1-like antibodies could be quickly elicited during natural SARS-CoV-2 infection or by RBD-based vaccines. IgG1 ab1 does not have developability liabilities, and thus has potential for therapy and prophylaxis of SARS-CoV-2 infections. The rapid identification (within 6 days) of potent mAbs shows the value of large antibody libraries for response to public health threats from emerging microbes.

The severe acute respiratory distress coronavirus 2 (SARS-CoV-2) (1) has spread worldwide thus requiring safe and effective prevention and therapy. Inactivated serum from convalescent patients inhibited SARS-CoV-2 replication and decreased symptom severity of newly infected patients (2, 3) suggesting that monoclonal antibodies (mAbs) could be even more effective.

Human mAbs are typically highly target-specific and relatively non-toxic. By using phage display we have previously identified a number of potent fully human mAbs (m396, m336, m102.4) against emerging viruses including severe acute respiratory syndrome coronavirus (SARS-CoV) (4) , Middle East respiratory syndrome coronavirus (MERS-CoV) (5) and henipaviruses (6, 7) , respectively, which are also highly effective in animal models of infection (8) (9) (10) (11) ; one of them was administered on a compassionate basis to humans exposed to henipaviruses and successfully evaluated in a clinical trial (12) .

Size and diversity of phage-displayed libraries are critical for rapid selection of high affinity antibodies without the need for additional affinity maturation. Our exceptionally potent antibody against the MERS-CoV, m336, was directly selected from very large (size ~10 11 clones) library from 50 individuals (5) . However, another potent antibody, m102.4, against henipavirusses was additionally affinity matured from its predecessor selected from smaller library (size ~10 10 clones) from 10 individuals (7, 13) . Thus, to generate high affinity and safe mAbs we used eight very large (size ~ 10 11 clones each) naive human antibody libraries in Fab, scFv or VH format using PBMCs from 490 individuals total obtained before the SARS-CoV-2 outbreak. Four of the libraries were based on single human VH domains where CDRs (except CDR1 which was mutagenized or grafted) from our other libraries were grafted as previously described (14) .

Another important factor to consider when selecting effective mAbs is the appropriate antigen. Similar to SARS-CoV, SARS-CoV-2 uses the spike glycoprotein (S) to enter into host cells. The S receptor binding domain (RBD) binds to its receptor, the human angiotensinconverting enzyme 2 (hACE2), thus initiating series of events leading to virus entry into cells (15, 16) . We have previously characterized the function of the SARS-CoV S glycoprotein and identified its RBD which is stable in isolation (17) . The RBD was then used as an antigen to pan phage displayed antibody libraries; we identified potent antibodies (5, 8) more rapidly and the antibodies were more potent than when we used whole S protein or S2 (unpublished). In addition, the SARS-CoV RBD based immunogens are highly immunogenic and elicit neutralizing antibodies which protect against SARS-CoV infections (18). Thus, to identify SARS-CoV-2 mAbs, we generated two variants of the SARS-CoV-2 RBD (aa 330-532) (Fig.   S1 ) and used them as antigens for panning of our eight libraries.

Panels of high-affinity binders to RBD in Fab, scFv and VH domain formats were identified. There was no preferential use of any antibody VH gene (an example for a panel of binders selected from the scFv library is shown in Fig. S2A ) and the number of somatic mutations was relatively low (Fig. S2B , for the same panel of binders as in Fig. S2A ). For nine of the highest affinity mAbs a provisional patent application was filed on March 12, 2020 by the University of Pittsburgh. Those high affinity mAbs can be divided into two groups in terms of their competition with hACE2. Two representatives of each group are Fab ab1 and VH ab5. To further increase their binding through avidity effects and extend their half-live in vivo they were converted to IgG1 and VH-Fc fusion formats, respectively. Ab1 was characterized in more details because of its potential for prophylaxis and therapy of SARS-CoV-2 infection.

The Fab and IgG1 ab1 bound strongly to the RBD (Fig. 1A ) and the whole SARS-CoV-2 S1 protein (Fig. 1B) as measured by ELISA. The Fab ab1 equilibrium dissociation constant, Kd, as measured by the biolayer interferometry technology (BLItz), was 1.5 nM (Fig. 1C) . The IgG1 ab1 bound with high (Kd =160 pM) avidity to recombinant RBD (Fig. 1D) . IgG1 ab1 bound cell surface associated native S glycoprotein suggesting that the conformation of its epitope on the RBD in isolation is close to that in the native S protein (Fig. 2, S3 ). The binding of IgG1 ab1 was of higher avidity than that of hACE2-Fc (Fig. 2B) . Binding of ab1 was specific for the SARS-CoV-2 RBD; it did not bind to the SARS-CoV S1 (Fig. 3A ) nor to cells that do not express SARS-CoV-2 S glycoprotein ( Fig. 2A ). Ab1 competed with hACE2 for binding to the RBD ( Fig.   3B and C) indicating possible neutralization of the virus by preventing binding to its receptor. It did not compete with the CR3022 (Fig. 3D and E) , which also binds to SARS-CoV (19) and with ab5 ( Fig. 3F ).

IgG1 ab1 potently neutralized SARS-CoV-2 pseudovirus with an IC50 of 10 ng/ml ( Fig   4A) . It did not enhance pseudovirus infection of FcγRIA overexpressing 293T-hACE2 cells at any concentration ( Fig 4B) . It also did not mediate pseudovirus infection of FcγRII expressing K562 cells ( Fig S4B) . Importantly, IgG1 ab1 exhibited potent neutralizing activity against authentic SARS-CoV-2 in two independent assays -a microneutralization-based assay (100% neutralization at < 400 ng/ml) (Fig. 4C ) and a luciferase reporter gene assay (IC50 = 200 ng/ml) ( Fig. 4D ). In agreement with the specificity of binding to the SARS-CoV-2 S1 and not to the SARS-CoV S1 the IgG1 ab1 did not neutralize live SARS-CoV (Fig. 4C ). The IgG1 m336 (5) control which is a potent neutralizer of MERS-CoV, did not exhibit any neutralizing activity against SARS-CoV-2 (Fig. 4C ). The VH ab5 and VH-Fc ab5 bound the RBD with high affinity and avidity (Fig. S5A .B) but did not compete with hACE2 ( Fig. S5C ) or neutralize SARS-CoV-2 ( Fig. 4D) , indicating that not all antibodies targeting epitopes on the RBD affect virus replication.

To evaluate the efficacy of IgG1 ab1 in vivo we used two animal models. The first one is based on the recently developed mouse ACE2 adapted SARS-CoV-2 which has two mutations Q498T/P499Y at the ACE2 binding interface on RBD (20). IgG1 ab1 protected mice from high titer intranasal SARS-CoV-2 challenge (10 5 pfu) of BALB/c mice in a dose dependent manner ( Fig 5A) . There was complete neutralization of infectious virus at the highest dose of 0.9 mg, and statistically significant reduction by 100-fold at 0.2 mg; there was a trend for reduction at 0.05 mg dose but did not reach statistical significance. The IgG1 m336 which potently neutralizes the MERS-CoV in vivo was used as an isotype control because it did not have any activity in vitro. These results also suggest that the RBD double mutations Q498T/P499Y do not affect IgG1 ab1 binding. The second model we used is the transgenic mice expressing human ACE2 (hACE2) (21). Mice were administered 300 ug of IgG1 ab1 prior to wild type SARS-CoV-2 challenge followed by detection of infectious virus in lung tissue 2 days later. Replication competent virus was not detected in four of the five mice which were treated with IgG1 ab1 ( Fig   5B) . All six control mice and one of the treated mice had more than 10 3 PFU per lung. These results show clear evidence of a potent preventive effect of IgG1 ab1 in vivo. The reason for absence of virus neutralization in one of the mice is unclear but may be due to individual variation in antibody transfer from the peritoneal cavity where it was administered to the upper and lower respiratory tract. Our previous experiments with transgenic mice expressing human DPP4 and treated with two different doses of m336 (0.1 and 1 mg per mouse) showed similar lack of protection of one (out of four) mice at the lower dose but at the higher dose all four mice were protected (9) similarly to the results obtained with the mouse adapted SARS-CoV-2. The in vivo protection also indicates that IgG1 ab1 can achieve neutralizing concentrations in the respiratory tract. This is the first report of in vivo activity of a human monoclonal antibody against SARS-CoV-2 by using two different mouse models. The results also show some similarity between the two models in terms of evaluation of antibody efficacy. In both models about the same dose of antibody (0.2-0.3 mg) reduced about 100-fold the infectious virus in the lungs. This result now suggests that testing of antibody efficacy could be performed at a larger scale than testing with the hACE2 transgenic mice due to the availability of wild type mice. It also shows robust neutralizing activity of IgG1 ab1 in two different models of infection.

Interestingly, Fab ab1 had only several somatic mutations compared to the closest germline predecessor genes. This implies that ab1-like antibodies could be elicited relatively quickly by using RBD-based immunogens especially in some individuals with naïve mature B cells expressing the germline predecessors of ab1. This is in contrast to the highly mutated broadly neutralizing HIV-1 antibodies that require long maturation times, are difficult to elicit and their germline predecessors cannot bind native HIV-1 envelope glycoproteins (22, 23). The RBD of the MERS-CoV S protein was previously shown to elicit neutralizing antibodies (24, 25). For SARS-CoV-2 only a few somatic mutations would be sufficient to generate potent neutralizing antibodies against the SARS-CoV-2 RBD which is a major difference from the elicitation of broadly neutralizing antibodies against HIV-1 which requires complex maturation pathways (22, 26-29). The germline-like nature of the newly identified mAb ab1 also suggests that it has excellent developability properties that could accelerate its development for prophylaxis and therapy of SARS-CoV-2 infection (30).

To further assess the developability (drugability) of ab1 its sequence was analyzed online (opig.stats.ox.ac.uk/webapps/sabdab-sabpred/TAP.php); no obvious liabilities were found. In addition, we used dynamic light scattering (DLS) and size exclusion chromatography to evaluate its propensity for aggregation. IgG1 ab1 at a concentration of 2 mg/ml did not aggregate for six days incubation at 37°C as measured by DLS (Fig. 6A) ; there were no high molecular weight species in freshly prepared IgG1 ab1 also as measured by size exclusion chromatography (SEC) (Fig. 6B ). IgG1 ab1 also did not bind to the human cell line 293T ( Fig. 2A ) even at very high concentration (1 μM) which is about 660-fold higher than its Kd indicating absence of nonspecific binding to many membrane-associated human proteins. The IgG1 ab1 also did not bind to 5,300 human membrane-associated proteins as measured by a membrane proteome array (Fig.   6C ).

The high affinity/avidity and specificity of IgG1 ab1 along with potent neutralization of virus and good developability properties suggests its potential use for prophylaxis and therapy of SARS-CoV-2 infection. Because it strongly competes with hACE2 indicating a certain degree of mimicry, one can speculate that mutations in the RBD may also lead to inefficient entry into cells and infection. In the unlikely case of mutations that decrease the ab1 binding to RBD but do not affect binding to ACE2 it can be used in combination with other mAbs including those we identified or in bi(multi)specific formats to prevent infection of such SARS-CoV-2 isolates. Ab1 could also be used to select appropriate epitopes for vaccine immunogens and for diagnosis of CoV-specific IgG1 m336 antibody was expressed in human mammalian cell as described

previously (5). The ACE2 gene was ordered from OriGene (Rockville, MD). The RBD domain (residues 330-532) and S1 domain (residues 14-675) and ACE2 (residues 18-740) genes were cloned into plasmid which carries a CMV promotor with an intron, human IgG1 Fc region and Woodchuck posttranscriptional regulatory element (WPRE) to generate the RBD-Fc, S1-Fc and ACE2-Fc expression plasmids. The RBD-avi-his protein with an avi tag followed by a 6×His tag at C-terminal was subcloned similarly. These proteins were expressed with Expi293 expression system (Thermo Fisher Scientific) and purified with protein A resin (GenScript) and by Ni-NTA resin (Thermo Fisher Scientific). The Fab CR3022 antibody gene with a His tag was cloned into pCAT2 plasmid (developed in house) for expression in HB2151 bacteria and purified with Ni-NTA resin. Protein purity was estimated as >95% by SDS-PAGE and protein concentration was measured spectrophotometrically (NanoVue, GE Healthcare). BLItz. Antibody affinities and avidities were analyzed by the biolayer interferometry BLItz Binding EC50 was obtained by using the non-linear mode in Graphpad Prism 7. IgG1 ab1 showed higher binding avidity to 293T-S cells than hACE2-Fc (0.25 nM v.s. 0.52 nM for IgG1 ab1 and hACE2-Fc to achieve 50% binding, respectively). followed by PBST washing. For detection, an HRP conjugated anti mouse Fc antibody was used.

Competition of ab1 with hACE2 tested by Blitz. 100 nM hACE2-Fc was monitored to bind ab1 saturated sensors (red line), which is compared to its independent binding signal to RBD sensor in the absence of ab1(green line were defined as the sample concentration at which a 50% reduction in RLU was observed relative to the average of the virus control wells. 

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Convalescent plasma as a potential therapy for COVID-19. The Lancet Infectious Diseases

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Structure of severe acute respiratory syndrome coronavirus receptor-binding domain complexed with neutralizing antibody

Exceptionally potent neutralization of Middle East respiratory syndrome coronavirus by human monoclonal antibodies

Potent neutralization of Hendra and Nipah viruses by human monoclonal antibodies

Exceptionally potent cross-reactive neutralization of Nipah and Hendra viruses by a human monoclonal antibody

Potent cross-reactive neutralization of SARS coronavirus isolates by human monoclonal antibodies

Passive Transfer of A Germline-like Neutralizing Human Monoclonal Antibody Protects Transgenic Mice Against Lethal Middle East Respiratory Syndrome Coronavirus Infection

A neutralizing human monoclonal antibody protects against lethal disease in a new ferret model of acute nipah virus infection

A neutralizing human monoclonal antibody protects african green monkeys from hendra virus challenge

Safety, tolerability, pharmacokinetics, and immunogenicity of a human monoclonal antibody targeting the G glycoprotein of henipaviruses in healthy adults: a first-in-human, randomised, controlled, phase 1 study

Construction of a large naive human phage-displayed Fab library through one-step cloning

Construction of a large phagedisplayed human antibody domain library with a scaffold based on a newly identified highly soluble, stable heavy chain variable domain

Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2

An emerging coronavirus causing pneumonia outbreak in Wuhan, China: calling for developing therapeutic and prophylactic strategies

The SARS-CoV S glycoprotein: expression and functional characterization

Biotinylated hACE2-Fc (10 nM) was incubated with RBD-Fc in the presence of different concentrations of VH ab5. After washing, bound hACE2-Fc was detected by using HRP conjugated streptavidin

After washing, bound CR3022 was detected by using HRP conjugated anti human Fc antibody. Ab5 showed weak competition with CR3022 for binding to SARS-CoV-2 RBD. All the ELISA experiments were performed in duplicate and the error bars denote ± SD