key: cord-273613-cpiveo7j
authors: Cao, Xia; Maruyama, Junki; Zhou, Heyue; Kerwin, Lisa; Sattler, Rachel; Manning, John T.; Johnson, Sachi; Richards, Susan; Li, Yan; Shen, Weiqun; Blair, Benjamin; Du, Na; Morais, Kyndal; Lawrence, Kate; Lu, Lucy; Pai, Chin-I; Li, Donghui; Brunswick, Mark; Zhang, Yanliang; Ji, Henry; Paessler, Slobodan; Allen, Robert D.
title: Discovery and Development of Human SARS-CoV-2 Neutralizing Antibodies using an Unbiased Phage Display Library Approach
date: 2020-09-29
journal: bioRxiv
DOI: 10.1101/2020.09.27.316174
sha: 
doc_id: 273613
cord_uid: cpiveo7j

SARS-CoV-2 neutralizing antibodies represent an important component of the ongoing search for effective treatment of and protection against COVID-19. We report here on the use of a naïve phage display antibody library to identify a panel of fully human SARS-CoV-2 neutralizing antibodies. Following functional profiling in vitro against an early pandemic isolate as well as a recently emerged isolate bearing the D614G Spike mutation, the clinical candidate antibody, STI-1499, and the affinity-engineered variant, STI-2020, were evaluated for in vivo efficacy in the Syrian golden hamster model of COVID-19. Both antibodies demonstrated potent protection against the pathogenic effects of the disease and a dose-dependent reduction of virus load in the lungs, reaching undetectable levels following a single dose of 500 micrograms of STI-2020. These data support continued development of these antibodies as therapeutics against COVID-19 and future use of this approach to address novel emerging pandemic disease threats.

Global incidence of severe acute respiratory disease syndrome coronavirus 2 (SARS-CoV-2) infection has continued to rapidly increase since the virus was first detected in December 2019 1 . To date, public health agency efforts to combat the pandemic level of infection and resultant Coronavirus disease 2019 (COVID- 19) have relied mostly upon effective quarantine measures 2, 3 . To further protect at-risk populations, a worldwide effort has been undertaken to develop additional countermeasures, such as antiviral compounds, immunomodulatory agents, vaccines, and neutralizing antibodies (nAbs) 4, 5, 6 .

While vaccines have been among the most effective means in preventing and containing widespread infections caused by many pathogens, their development is often faced with drawbacks, including long development timelines, large clinical trial sizes, and uncertain efficacy due to the reliance on patientgenerated immune responses to the antigens 7, 8, 9 . Of note, it has been shown in the context of multiple virus infections that virus-specific antibodies can lead to exacerbation of disease symptoms through a process termed antibody dependent enhancement (ADE) 10, 11 . This ADE of virus infection is a phenomenon in which certain virus-specific antibodies enhance the entry of virus due to engagement between IgG Fc regions and Fc-receptors on the surface of monocytes leading to productive infection of these cells rather than virus clearance 12, 13 . In addition, the presence of immunodominant epitopes, i.e. highly immunogenic sequences, in viral antigens can skew this immune response, resulting in production of antibodies that are neither neutralizing nor protective, as seen in the vaccine development against human immunodeficiency virus (HIV) 14 . In the case of SARS-CoV-2, nAbs targeting the Spike protein have been isolated from the sera of convalescent COVID-19 patients as well as from sera donated by healthy normal patients prior to December 2019 15, 16, 17, 18, 19, 20 . While the patients' immune systems generated these nAbs, they often recognize the same region of the Spike protein or even a similar epitope due to immunodominant regions of the Spike protein 21, 22 . One way to overcome this inherent immune bias is the use of an artificial immune system in a test tube, namely phage display human antibody libraries, which allows for functional unbiased selection of Spike-binding antibodies irrespective of the underlying natural immunogenicity 23 .

Most antibody libraries are constructed in a way that allows for chain shuffling leading to random associations of heavy and light chain encoding genes and resulting in a vast repertoire of non-naturally occurring antigen specificities. In fact, few candidate anti-SARS-CoV-2 nAbs have been isolated from phage display libraries 24, 25 . Using Sorrento's G-MAB™ library 26 , a single chain variable fragment (scFv) antibody phage display library constructed from the antibody repertoire of over 600 healthy individuals, a panel of antibodies that bind the SARS-CoV-2 Spike S1 subunit was identified and characterized.

Herein, we report on the identification, characterization, and subsequent optimization of STI-1499, a SARS-CoV-2 neutralizing antibody isolated from the G-MAB library. STI-1499 neutralized both the WA-1/2020 isolate of SARS-CoV-2 as well as isolate 2020001, a clinical isolate of SARS-CoV-2 that encodes the G614 variant of the Spike protein. Isolates harboring the G614 Spike variant have been reported to be more infectious and occur in greater frequency than viruses such as the WA-1/2020 isolate that encode the D614 Spike variant 27 . Affinity maturation of STI-1499 resulted in identification of STI-2020, an antibody with a 35-fold increased affinity for the SARS-CoV-2 Spike receptor-binding domain (RBD) leading to a greater than 50-fold increase in virus neutralization potency against live WA-1/2020 and 2020001 viruses in vitro. Importantly, both nAbs, STI-1499 and STI-2020, provided protection against the pathogenic effects and replication of SARS-CoV-2 in the Syrian golden hamster virus challenge model 28, 29 .

The S1 subunit of the SARS-CoV-2 spike protein (amino acids 16-685) bearing a C-terminal histidine tag (ACRO Biosystems, Newark, NJ) was coated at 2 μg/ml on a Ni-NTA plate (Qiagen, Valencia, CA). After washing and blocking, 10 μg/ml antibody was added to the corresponding wells for 1-hour incubation.

The plate was washed three times and incubated with 1:5000 dilution of HRP-conjugated mouse anti-human IgG (Fc). Color development was performed with 3,3′,5,5′-Tetramethylbenzidine (TMB).

Absorbance was read at 450 nm.

Kinetic interactions between the antibodies and His-tagged receptor binding domain (RBD, amino acids 319-537) (Acro Biosystems, Newark, NJ) protein was measured at 25C using BIAcore T200 surface plasmon resonance (SPR) (GE Healthcare). STI-1499 or STI-2020 antibody was covalently immobilized on a CM5 sensor chip to approximately 500 and 100 resonance units (RU), respectively using standard N- 

STI-1499 was expressed using a two-vector transient expression protocol. Briefly, 2x10 6 /ml CHO-S cells were co-transfected with 1g of total DNA of the heavy and light chain plasmids, at 1:4 DNA:

Polyethyleneimine (PEI) ratio (g/g). After 24 hours of incubation with shaking (125 rpm) at 28°C, the culture conditions were adjusted to incubation at 37°C with shaking in a 5% CO2 chamber for 10-14 days. STI-1499 was purified from the CHO culture supernatants by affinity chromatography with MabSelectSureLX™ protein A resin using the AktaPure platform.

STI-2020 was produced using a double gene vector stable pool expression protocol. Briefly, the pXC double gene vector (Lonza) encoding both the heavy and light chain genes was used to generate polyclonal pools of stably transfected CHO-S cells, according to manufacturer's recommended transfection and selection protocols. The stable pool was cultured with shaking (140 rpm) at 37°C in a 5% CO2 incubator for 14 days.

STI-2020 was purified to homogeneity from the CHO culture supernatants first by affinity chromatography with MabSelectSureLX™ protein A resin and then by ion exchange chromatography using SPImpres resin. 

Vero E6 cells were maintained in Dulbecco's modified Eagle's medium (DMEM, Corning, NY) supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific, MA), 1% penicillinstreptomycin, and L-glutamine. The P3 stock of the SARS-CoV-2 (USA/WA-1/2020) isolate and the P1 stock of the SARS-CoV-2 (2020001) isolate were obtained from The World Reference Center for Emerging Viruses and Arboviruses (WRCEVA) at the University of Texas Medical Branch. The viruses were propagated in Vero E6 cells and cell culture supernatant of P4 and P2 stocks, respectively, were stored at -80°C under BSL3 conditions.

The day before infection, 2x10 4 Vero E6 cells were plated to 96-well plates and incubated at 37° C, 5% 

Male and female Syrian golden hamsters were obtained from Charles River Laboratories at 6 weeks of age. Hamsters were inoculated intranasally (i.n.) with 10 5 TCID50 of SARS-CoV-2 in 100 µl of sterile PBS on day 0. Antibody treatments were administered intravenously (i.v.) with monoclonal antibodies (mAbs) against SARS-CoV-2 Spike, or isotype control mAb in up to 350 µl of sterile PBS at 1 hour-post inoculation.

Animals were monitored for illness and mortality for 10 days post-inoculation and clinical observations were recorded daily. Body weights and temperatures were recorded at least once every 48 hours throughout the experiment. On day 5 post-infection, 5 animals from each treatment group were sacrificed and virus titers in the lungs of these animals were measured using a SARS-CoV-2 virus TCID50 assay.

Average % weight change on each experimental day were compared with the isotype control mAb-treated group using 2-way ANOVA following Dunnett's multiple comparisons test. All animals were housed in animal biosafety level-2 (ABSL-2) and ABSL-3 facilities in Galveston National Laboratory at the University of Texas Medical Branch. All animal studies were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Texas Medical Branch and were conducted according to the National Institutes of Health guidelines.

Prior to initiation of the experiment, five animals from each treatment group (n=10) were designated for virus titration in lung tissue at 5 days post inoculation. On the assigned day, animals were euthanized, lung tissue samples were collected from each animal, and a portion of the tissue (0.08-0.3g) was placed into pre-labeled microcentrifuge tubes containing 5 mm stainless steel beads (Qiagen Inc., CA). Lung samples were homogenized with DMEM + 2% FBS in a TissueLyser (Qiagen Inc., CA) operated at 25-30 Hz for four minutes. Tubes were centrifuged and clarified homogenate was serially diluted 10-fold with DMEM+2% FBS. From material representing each serial dilution step, 100 l was transferred to each of four wells of a 96-well plated seeded with Vero E6 cells. Plates were then incubated for 72-96 hours at 37° C, 5% CO2. Cells were subsequently fixed with 10% formalin and stained with 0.25% crystal violet solution. TCID50 values were calculated by the method of Reed and Muench 30 .

The G-MAB library was panned using a recombinant His-tagged SARS-CoV-2 Spike S1 subunit to allow for selection of candidates with high affinity to the target antigen, as outlined in Figure 1 . In summary, following confirmation of S1 binding by ELISA, clonal scFv preparations were tested in a competition ELISA format for disruption of Spike S1:ACE2 binding. Candidate scFvs with high S1 binding affinity and/or the capacity to block Spike S1:ACE2 binding were converted into and expressed as full length human IgG1 antibodies.

In an effort to mitigate the risk of ADE resulting from administration of G-MAB-derived nAbs, the IgG1 

The nAb candidate termed STI-1499 displayed potent SARS-CoV-2 neutralizing activity superior to that of other G-MAB nAb hits, with an IC50 in the virus neutralization assay of 3.2 g/ml (Figure 2A) . In parallel to STI-1499 preclinical profiling and cell line development efforts, affinity maturation of STI-1499 was undertaken, with systematic amino acid changes engineered within complementarity determination regions (CDRs) of the STI-1499 heavy and light chains. For individual CDR regions from STI-1499, a library of variants was screened for binding to the Spike S1 subunit by ELISA ( Figure 2B ). Variants bearing combinations of affinity-enhancing CDR amino acid changes were then systematically engineered and expressed as IgG1 LALA antibodies. Based on Spike S1 ELISA binding results, the most potent clone of the single CDR and combination CDR variants, STI-2020, was further profiled for binding affinity using surface plasmon resonance (SPR). STI-2020 bound to the RBD region of the Spike protein with an affinity of 1.3 nM, a 35-fold affinity improvement over that of the parental STI-1499 ( Figure 2C ).

STI-1499 and STI-2020 were tested for binding to full length Spike proteins derived from naturally emerging viruses including the USA/WA-1/2020 (WA-1) isolate as well as an emerging SARS-CoV-2 isolate that bears a D →G mutation at amino acid residue 614 in the Spike protein (D614G). Binding studies demonstrated equivalent binding affinity for both nAbs against cell surface-expressed Spike protein with the D614G mutation as well as that of the WA-1 isolate ( Figure 3A ).

To determine the neutralizing activity of STI-1499 and STI-2020 against the aforementioned SARS-CoV-2 clinical isolates (vide supra), nAbs were evaluated in the virus neutralization assay. Both nAbs exhibited equipotent neutralization of the virus isolates. In keeping with observed improved binding properties, STI-2020 neutralized both isolates with an IC50 of 0.055 g/ml and an IC99 of 0.078 g/ml, a greater than 50fold enhancement of neutralizing potency over the parental STI-1499 nAb ( Figure 3B ). Animals in the STI-2020 treatment groups were administered a single dose of either 30, 100, 300, or 500 g IV. Administration of a 500 g dose of STI-2020 resulted in maximum average percentage body weight loss of -1.9%, which occurred on 2 d.p.i.. After that, animals treated with 500 ug STI-2020 maintained an average body weight that as a percentage of day 0 weight was significantly different on days 4, 5, 6, 7, 8, and 10 than the average weight measured among IsoCtl-treated animals. As with STI-1499-treated hamsters, virus titers in the lung were measured in 5 of 10 animals from each treatment group. In IsoCtltreated animals, an average of 1.6x10 3 TCID50/g of lung tissue was detected. Treatment with 500 g STI-2020 resulted in reduction of lung titers below the level of detection in all animals tested, a STI-2020treatment-related lung titer reduction of 80-fold at minimum. In the 300 g STI-2020 treatment group, the average lung titer was reduced below the level of detection in 3 of 5 animals tested, while 2 of 5 animals had lung titers of similar magnitude to those measured in IsoCtl-treated animals. Animals with undetectable lung virus titers in the 300 g treatment group also experienced only moderate weight loss compared to animals with detectable lung virus titers in this group (d5 values of -11.1% and -16.7%) No changes in average lung titer compared to IsoCtl-treatment were detected in animals from the 100 g STI-2020 treatment group.

In summary, STI-1499 demonstrated promising protective efficacy in the hamster model at a dose of 2000 g. The observed overall enhancement of in vitro potency for the affinity matured nAb, STI-2020, translated into increased protection against disease in the hamster challenge model, with a greater than 4-fold increase in protective efficacy over that of STI-1499.

In this study, we detail the initial discovery and profiling of a SARS-CoV-2 nAb isolated from a phage display antibody library derived from the B-cell repertoire of over 600 healthy normal individuals. This approach complements the isolation of nAbs collected from natural sources, such as convalescent patients or vaccinated healthy individuals, and allows for isolation of nAbs recognizing neutralizing epitopes developed outside the temporal and biological context of the pathogen/antigen-specific immune response. The parental nAb STI-1499 and the affinity-matured derivative STI-2020 were characterized for their biochemical and function properties in vitro and in vivo against prevalent clinical isolates of SARS CoV-2. Both nAbs demonstrated potent functionality in these assays, which is strongly supportive of continued preclinical and clinical development. Notably, demonstration of improvement in neutralizing potency in vitro for STI-2020 over that of STI-1499 was predictive of increased protective efficacy in the hamster challenge model following post-exposure therapeutic treatment. In addition, in vivo efficacy was achieved in the presence of the ADE-mitigating LALA Fc modification, providing further confidence that this modification might add to the utility of nAbs under clinical evaluation for use in the treatment of COVID-19. Studies aimed at establishing the therapeutic and prophylactic treatment window in the hamster model as well as demonstrating efficacy following administration of nAb proteins by nonparenteral routes are currently underway.

In addition, introduction of nAb-encoding DNA plasmids provides a low cost, long-acting means of establishing antibody-mediated immunity, and experiments in rodents are ongoing and will determine the suitability of STI-2020 for use in this context. Treatments using STI-2020 nAb protein or DNA-encoding of nAbs in combination with other identified nAbs targeting non-overlapping epitopes on the Spike S1 subunit to constitute a nAb cocktail approach will provide a means of increasing the barrier to treatment resistance as the prevalence of naturally emerging and treatment-associated virus variants increases. This approach is under investigation and in preclinical development.

As emergence of novel pathogens will continue to pose a significant threat to humanity, the rapid discovery of potent nAbs against these yet-unknown microbes will remain paramount to global health efforts. The discovery and development of anti-SARS-CoV-2 nAbs STI-1499 and STI-2020 in less than 6 months without the need of samples from infected individuals could serve as blueprint for future rapid response efforts.

Panning of the G-MAB library consisting of fully human antibody sequences allowed for the expeditious progression of lead nAb candidates from initial screening to the production cell line generation stage, avoiding additional development steps such as subsequent humanization of the identified antibodies and associated immunogenicity concerns of non-human antibody-based therapeutics. This has enabled an efficient and expedited discovery and development of SARS-CoV-2 nAbs starting in March 2020, thus taking less 6 months from initiation of the project to completed IND filing. Notably, STI-1499 has recently received FDA clearance to commence first-in-human studies making it the first clinical-stage anti-SARS-CoV-2 nAb derived not from convalescent patients but from a fully human phage display antibody library, and it is anticipated that dosing of the first patients will begin in October 2020. panned for SARS-CoV-2 Spike S1 subunit-binding scFv fragments. Following confirmation of binding activity and blocking of S1:ACE2 interactions by candidate scFvs, the most potent of these candidates were converted to IgG1 antibodies bearing the LALA Fc modification. Candidate nAbs were characterized for binding of Spike S1 subunit and neutralization of related clinical SARS-CoV-2 isolates. Affinity maturation of potent nAbs was carried out in parallel to biophysical profiling, cell line development, and evaluation of protective efficacy for the parental nAb, STI-1499. and STI-2020. The antibody affinities were measured using SPR on a BIAcore T200 instrument. Competing interests: Sorrento authors own options and/or stock of the company. This work has been described in one or more provisional patent applications. HJ is an officer at Sorrento Therapeutics, Inc.. 

A Novel Coronavirus from Patients with Pneumonia in China

COVID-19: towards controlling of a pandemic

Public Health Response to the Initiation and Spread of Pandemic COVID-19 in the United States

Pandemic Preparedness: Developing Vaccines and Therapeutic Antibodies For COVID-19

The Current and Future State of Vaccines, Antivirals and Gene Therapies Against Emerging Coronaviruses

Countermeasures to Coronavirus Disease 2019: Are Immunomodulators Rational Treatment Options-A Critical Review of the Evidence

Immunological considerations for COVID-19 vaccine strategies

The Long Road Toward COVID-19 Herd Immunity: Vaccine Platform Technologies and Mass Immunization Strategies

SARS-CoV-2 Vaccines: Status Report

Antibody-dependent enhancement and SARS-CoV-2 vaccines and therapies

Antibody-dependent enhancement of Coronavirus

Antibody-dependent SARS coronavirus infection is mediated by antibodies against spike proteins

Anti-severe acute respiratory syndrome coronavirus spike antibodies trigger infection of human immune cells via a pH-and cysteine protease-independent FcgammaR pathway

The Hard Way towards an Antibody-Based HIV-1 Env Vaccine: Lessons from Other Viruses

Potent neutralizing antibodies directed to multiple epitopes on SARS-CoV-2 spike

A human monoclonal antibody blocking SARS-CoV-2 infection

Convergent antibody responses to SARS-CoV-2 in convalescent individuals

Longitudinal Isolation of Potent Near-Germline SARS-CoV-2-Neutralizing Antibodies from COVID-19 Patients

Cross-neutralization of SARS-CoV-2 by a human monoclonal SARS-CoV antibody

A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2

Identification of immunodominant linear epitopes from SARS-CoV-2 patient plasma

Two linear epitopes on the SARS-CoV-2 spike protein that elicit neutralising antibodies in COVID-19 patients

Manufacturing immunity to disease in a test tube: the magic bullet realized

Isolation of and Characterization of Neutralizing Antibodies to Covid-19 from a Large Human Naïve scFv Phage Display Library. bioRxiv

Blocking antibodies against SARS-CoV-2 RBD isolated from a phage display antibody library using a competitive biopanning strategy. bioRxiv

Recombinant antibody libraries and selection technologies

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the COVID-19 Virus

Simulation of the clinical and pathological manifestations of Coronavirus Disease 2019 (COVID-19) in golden Syrian hamster model: implications for disease pathogenesis and transmissibility

Animal models for COVID-19

A simple method of estimating fifty percent endpoints

Novel human IgG1 and IgG4 Fc-engineered antibodies with completely abolished immune effector functions

A novel angiotensin-converting enzyme-related carboxypeptidase (ACE2) converts angiotensin I to angiotensin 1-9

Tissue distribution of ACE2 protein, the functional receptor for SARS coronavirus. A first step in understanding SARS pathogenesis

Quantitative mRNA expression profiling of ACE 2, a novel homologue of angiotensin converting enzyme