Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5

Kevin C Entzminger; Jonathan K Fleming; Paul D Entzminger; Lisa Yuko Espinosa; Alex Samadi; Yuko Hiramoto; Shigeru C J Okumura; Toshiaki Maruyama

doi:10.1093/abt/tbad006

Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5

Entzminger, Kevin C; Fleming, Jonathan K; Entzminger, Paul D; Espinosa, Lisa Yuko; Samadi, Alex; Hiramoto, Yuko; Okumura, Shigeru C J; Maruyama, Toshiaki 2023-04-13 00:00:00 Antibody Therapeutics, 2023, Vol. 6, No. 2 108–118 https://doi.org/10.1093/abt/tbad006 Advance Access Publication on 13 April 2023 Research Article Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5 Kevin C. Entzminger, Jonathan K. Fleming, Paul D. Entzminger, Lisa Yuko Espinosa, Alex Samadi, Yuko Hiramoto, Shigeru C.J. Okumura and Toshiaki Maruyama Antibody Discovery, Abwiz Bio Inc., San Diego, CA 92121, USA Received: February 27, 2023; Revised: March 31, 2023; Accepted: April 5, 2023 ABSTRACT SARS-CoV-2 Omicron variant XBB.1.5 has shown extraordinary immune escape even for fully vaccinated individuals. There are currently no approved antibodies that neutralize this variant, and continued emergence of new variants puts immunocompromised and elderly patients at high risk. Rapid and cost-effective development of neutralizing antibodies is urgently needed. Starting with a single parent clone that neutralized the Wuhan-Hu-1 strain, antibody engineering was performed in iterative stages in real time as variants emerged using a proprietary technology called STage-Enhanced Maturation. An antibody panel that broadly neutralizes currently circulating Omicron variants was obtained by in vitro afﬁnity maturation using phage display. The engineered antibodies show potent neutralization of BQ.1.1, XBB.1.16, and XBB.1.5 by surrogate virus neutralization test and pM K afﬁnity for all variants. Our work not only details novel therapeutic candidates but also validates a unique general strategy to create broadly neutralizing antibodies to current and future SARS-CoV-2 variants. Statement of Signiﬁcance: We report engineering and characterization of SARS-CoV-2 therapeutic anti- bodies that neutralize all Omicron variants to date. Surrogate virus neutralization assay shows low ng/mL IC values and biolayer interferometry shows pM afﬁnity. The use of a proprietary platform technology, STage-Enhanced Maturation, is detailed. KEYWORDS: antibody engineering; neutralizing antibody; XBB.1.5; COVID-19; SARS-CoV-2 INTRODUCTION increasing the likelihood of breakthrough infections [1, 3]. At the time of writing, March 2023, XBB.1.5 and During the 3 years since the outbreak of coronavirus BQ.1.1 are dominant circulating strains, comprising 90% disease 2019 (COVID-19), the severe acute respiratory of all infections (cdc.gov). As a master of immune evasion, syndrome coronavirus 2 (SARS-CoV-2) virus has proven XBB.1.5 has been deemed the most transmissible variant exceptionally adept at mutating to evade the immune yet [4] and has the potential to dominate other variants response [1]. The emergence of the first Omicron variant worldwide. BA.1, having far more mutations in the spike receptor Antibody therapeutics are the standard of care for binding domain (RBD) compared to earlier variants, at-risk populations who are immunocompromised and dramatically increased the susceptibility of even previously thus susceptible to adverse COVID infection. Yet anti- vaccinated or infected individuals [2]. Later Omicron body therapeutics approved for treatment of COVID- sublineages BA.2 (from which XBB.1.5 is derived) and 19 have lost efficacy soon after new strains emerged. BA.4/5 (from which BQ.1.1 is derived) have continued BA.1 significantly reduced the potency of first-generation to evolve additional immune-evading mutations, further To whom correspondence should be addressed. Toshiaki Maruyama, 9823 Pacific Heights Blvd Ste J, Antibody Discovery, Abwiz Bio, San Diego, CA 92121, USA. Tel: 858-352-6911; Email: toshi@abwizbio.com © The Author(s) 2023. Published by Oxford University Press on behalf of Antibody Therapeutics. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Antibody Therapeutics, 2023 109 antibody therapeutics such as REGEN-COV (casirivimab– Spike trimers used in the study include Wuhan-Hu- imdevimab) [1]. BQ.1.1 resists neutralization by single or 1 (GenBank MN908947; Abwiz Bio Cat. #2720) as cocktail monoclonal antibody (mAb) therapies including well as variants Beta (Abwiz Bio Cat. #2652), Delta sotrovimab, bebtelovimab, bamlanivimab–etesevimab, and (Abwiz Bio Cat. #2611), BA.1 (Abwiz Bio Cat. #2672), evusheld (cilgavimab–tixagevimab) [5, 6]. XBB.1.5 has been BA.2 (Abwiz Bio Cat. #2460), BA.2.75 (Abwiz Bio Cat. shown to similarly evade neutralizing antibodies [3]. There #2664), BA.2.75.2 (Abwiz Bio Cat. #2676), BA.2.3.20 has been a call for novel, broadly active mAbs urgently (Abwiz Bio Cat. #2660), BN.1 (Abwiz Bio Cat. #2696), needed for prophylactic and/or therapeutic treatment XBB (Abwiz Bio Cat. #2700), XBB.1/XBB.1.9 (Abwiz in patients at high risk [5, 7, 8], especially given the Bio Cat. #2648), XBB.1.5/XBB.1.9.1 (Abwiz Bio Cat. danger associated with re-infection [9]. No currently U.S. #2712), XBB.1.16 (Abwiz Bio Cat. #2747), BA.4.6 (Abwiz Food & Drug Administration (FDA) approved antibody Bio Cat. #2692), BA.5 (Abwiz Bio Cat. #2688), BQ.1 therapeutics neutralize the latest Omicron variants BQ.1.1 (Abwiz Bio Cat. #2704), BQ.1.1 (Abwiz Bio Cat. #2668), and XBB.1.5. BF.7/BA.5.2.6/BF.11 (Abwiz Bio Cat. #2708), CH.1.1 Existing strategies to engineer neutralizing therapeutic (Abwiz Bio Cat. #2724), CA.3.1 (Abwiz Bio Cat. #2728), mAbs typically rely on antibody isolation from infected or and BR.2.1 (Abwiz Bio Cat. #2736). vaccinated individuals or from immunized humanized mice [10–12]. However, this process is poorly adapted for rapidly Immunization of rabbits evolving targets such as SARS-CoV-2, where the lead discovery phase must be repeated each time a new variant Three New Zealand White rabbits were immunized with emerges. Instead, we devised an approach that evolves anti- a recombinant protein including the receptor-binding body neutralization breadth in real time as the virus itself domain (RBD) of SARS-CoV-2 (Wuhan-Hu-1 strain). The evolves. Starting from a humanized lead candidate that immune sera were tested by ELISA using spike trimer, and showed strong neutralization of the Wuhan-Hu-1 strain, we the rabbit that showed the best titer was selected for library created six separate complementarity determining region construction. (CDR)-targeted libraries and selected the libraries on Wuhan-Hu-1 spike trimer to create diverse CDR pools. Library construction These CDRs were then randomly paired and iteratively selected on SARS-CoV-2 variants as they emerged, result- The bone marrow and spleen cells were collected and ing in a final panel of seven clones that strongly neutralize homogenized in TRI reagent (Molecular Research Center). all Omicron variants including XBB.1.5 and BQ.1.1. This Total ribonucleic acid (RNA) was isolated from the study not only details potential high-value therapeutic homogenate according to the manufacturer’s protocol. mAbs but also validates a general strategy to elicit broadly Messenger RNA was purified using messenger ribonucleic neutralizing mAbs for SARS-CoV-2 or other viruses. acid (mRNA) purification kit (Macherey-Nagel) according to the manufacturer’s protocol. First strand complemen- tary deoxyribonucleic acid (cDNA) was synthesized using PowerScribe MMLV Reverse Transcriptase (Monserate MATERIALS AND METHODS Biotechnology Group). cDNA was engineered using a method described in the US patent 9,890,414 and amplified Preparation of recombinant proteins with a single non-gene-specific primer. Amplified products were purified, digested, and sequentially ligated into a Spike trimers were prepared in-house for use in all proprietary rabbit Fab phagemid vector. assays. SARS-CoV-2 variants were made by overlap PCR and cloned to pCAGGS vector for production. Spike trimers include polybasic cleavage site mutation/deletion Phage panning and solubilizing mutations K986P/V987P as well as a trimerization motif and polyhistidine tag [13]. Following Library DNA was transformed into XL1-Blue cells large-scale purification, 300 μg of deoxyribonucleic acid (Agilent) for overnight growth and phage production (DNA) was transfected to HEK293T cells, with media following addition of M13KO7 Helper Phage (New Eng- harvested by centrifugation and filtration ∼7 days post land Biolabs). Phage were precipitated the following day transfection. His-tagged spike trimer was purified using from culture supernatant by addition of 4% PEG-8000/3% a Nickel Nitriloacetic acid (Ni-NTA) column: the spike NaCl on ice, followed by pelleting and resuspension in 1% trimer was washed with 50 mM imidazole then eluted with BSA/Dulbecco’s phosphate-buffered saline (DPBS) (block- 100 mM and 250 mM imidazole, followed by overnight ing buffer). High-bind microtiter wells (Immulon 4HBX) dialysis in DPBS. Quality and purity were assessed by were coated with recombinant spike trimer overnight, SDS-PAGE (sodium dodecyl sulfate–polyacrylamide gel washed with PBS, and blocked with blocking buffer for ◦ ◦ electrophoresis) gel and activity was assessed by testing for 1hat 37 C. Phage were added for incubation at 37 Cfor binding to angiotensin converting enzyme 2-human IgG 1 h, followed by washing with PBS, elution with 0.1 M Fc fusion (ACE2-Fc) by ELISA. In-house biotinylated HCl, and neutralization with 2 M Tris. Neutralized phage trimers were prepared for biolayer interferometry (BLI) were used to infect ER2738 (New England Biolabs) cells studies using EX-Link™ NHS-LC-Biotin (ThermoFisher for overnight propagation and precipitation as described A39257) followed by purification with 7 K MWCO above. Panning was performed for four or more rounds, Zeba™ Spin Desalting Columns (ThermoFisher 89882). with increasing washing stringency in later rounds to 110 Antibody Therapeutics, 2023 enrich specific binders. During phage panning of STage- ImmunoResearch 300-035-240). All assays were performed Enhanced Maturation (STEM) libraries, [1] the library was in triplicate. Raw ELISA values were normalized to six first transiently heated to 65 C then cooled to improve replicate positive wells omitting IgG and to three negative thermoresistance as described previously [14], followed by wells omitting both IgG and spike trimer, with negative [2] subtraction of the library on baculovirus particle (BVP) values forced to zero. Graphpad Prism was used to prepare coated microtiter wells to remove polyreactive clones before graphs and to derive inhibition of spike trimer binding to selection on spike protein trimer. ACE2 half maximal inhibitory concentration (IC ) values by fitting to a four-parameter logistic curve. Bebtelovimab, tixagevimab, cilgavimab, and S2X324 were purchased from Fab screening Proteogenix. Following the final round of phage panning, single colonies were prepared from the eluted page. Colonies were picked STage-Enhanced Maturation antibody engineering to 96-well culture plates with 1 mL of Super Broth medium and 50 μg/mL of carbenicillin. The culture In Stage 1, Six single CDR mutant libraries were designed plates were incubated at 37 C and 1 mM Isopropyl ß- based on amino acid usage by structurally characterized D-1-thiogalactopyranoside (IPTG) was added to induce human antibodies sharing similar canonical CDR struc- overnight soluble Fab production at 30 C. For Fab ELISA, tures among a subset that possessed identical germline microtiter wells were coated with 100 μL of spike trimer genes. This design strategy further humanizes the amino at 1–2 μg/mL in PBS. Following washing and blocking, acid usage in the CDRs and helps to avoid unwanted culture supernatants containing soluble Fab were added to mutations that could negatively affect the structure and the wells for incubation at 37 C for 1 h. After washing, folding of the antibodies while constraining library size to bound Fab was detected with peroxidase conjugated goat a diversity that could be sampled using our phage libraries. anti-rabbit IgG F(ab’) (Thermo Fisher Scientific) or Mutagenic oligos were synthesized and used to construct anti-human IgG F(ab’) secondary antibodies (Jackson separate CDR libraries by fragment amplification and over- ImmunoResearch). Positive clones were Polymerase Chain lap PCR using hN2Y gene as a template. Following diges- Reaction (PCR) amplified and sequenced by Sanger tion, ligation was performed at large scale using 10 μg sequencing (Eton Biosciences). digested vector DNA. These single CDR mutant libraries were panned on Wuhan-Hu-1 spike trimer to remove non- productive clones and to enrich all possible CDR mutants IgG production and humanization of rabbit antibody that retain binding to the spike trimer. Phage panning and Light and heavy chains of rabbit clone C-A11 were PCR Fab screening was performed as described above. In Stage 2, round four output phage from Stage 1 were amplified, digested, and cloned into a bi-cistronic rabbit used to amplify targeted CDR regions and combined by IgG vector for expression. All antibodies and recombinant overlap PCR into separate light chain (Lx; paired with the ACE2-Fc (Abwiz Bio Cat. #2566) were produced by tran- wild type heavy chain) or heavy chain (Hx; paired with the sient transfection of HEK293T cells with media harvested wild type light chain) libraries. The Lx library was panned by centrifugation and filtration ∼7 days post transfection. on the Beta spike trimer, and the Hx library was panned IgGs were purified by protein A column (Cytiva) followed separately on Wuhan-Hu-1, Alpha, Beta, and Epsilon spike by overnight dialysis in DPBS. CDRs of C-A11 rabbit antibody were grafted into selected human germline genes trimers. Phage panning and Fab screening was performed (IGHV3-23∗04 and IGKV-39∗01) to create a humanized as described above. LxC1-G10 clone was identified from the clone hN2Y. An expanded structural definition of CDRs Lx library and was cloned to IgG format and characterized was used, based on observed diversity from an internal as described above. database of rabbit antibodies. Following gene synthesis, In Stage 3, round four output phage from Stage 2 hN2Y light and heavy chain genes were cloned into a were used to amplify light and heavy chain libraries for bi-cistronic human IgG vector for production as above. sequential cloning and pairing to create a single combined S728-1157 light and heavy chain genes were synthesized by library. This library was panned for six rounds on BA.5 Twist Bioscience and cloned into our IgG expression vec- followed by two rounds on BA.2.75 and one round on tor, CR3022 expression plasmids were obtained from BEI either BQ.1.1 or XBB spike trimers. Fabs were screened Resources, NIAID, NIH for expression and purification as similarly as described above and tested for binding to the described above. respective spike trimer used for selection. Phage panning and Fab screening was performed as described above. Clones possessing 6R8 or 6R9 prefixes were identified Surrogate virus neutralization test of purified IgG from this screening and were converted to IgG format for Surrogate virus neutralization tests (sVNTs) were per- characterization as described above. formed according to the protocol described previously [15] with some modifications. Microtiter wells were coated with Biolayer interferometry 100 μL of ACE2-Fc (Abwiz Cat. #2566) at 2 μg/mL in PBS and blocked with 200 μL of 1% BSA/PBS. Purified Biotinylated spike trimers were prepared in-house as IgG were mixed 1:1 with a recombinant spike trimer at RT described above except BA.1 (ACROBiosystems SPN- and added to the wells. The bound spike trimer was detected C82Ee). BLI was performed using Octet Red96e with with peroxidase conjugated anti-his tag antibody (Jackson data collection at 30 C and sample shaking at 1000 rpm Antibody Therapeutics, 2023 111 Figure 1. STEM antibody engineering strategy. (A) Rabbit immunization with Wuhan-Hu-1 SARS-CoV-2 RBD was paired with phage display to identify a potent neutralizing clone; this mAb was humanized to create the lead candidate hN2Y. Single CDR libraries were constructed and selected on Wuhan- Hu-1 spike trimer to create a pool of functional CDR sequences. hN2Y predicted structure is shown, with all CDRs colored blue and targeted CDR(s) for a given library colored red. (B) Pre-selected CDR pools were amplified from phage and combined by overlap PCR to create separate combined light and combined heavy chain libraries. The light chain library was further selected on the Beta variant, resulting in the engineered clone LxC1-G10, which showed potent neutralization of BA.1 and was thoroughly characterized by the CoVIC at La Jolla Institute of Immunology [25]. The heavy chain library was selected on a combination of Alpha, Beta, Gamma, and Epsilon variants to broaden neutralization activity. (C) Pre-selected light and heavy chain libraries were then paired to create a single combined library. This library was selected on late-stage Omicron variants including BA.2.75.2, BA.5, BQ.1.1, and XBB to create broadly neutralizing 6R8/6R9 series antibodies. during all steps. All IgGs and spike trimers were prepared in Wuhan-Hu-1 RBD containing protein, and phage libraries running buffer containing 2% BSA, 0.002% Tween-20, PBS, were selected on Wuhan-Hu-1 spike protein trimer. Fab pH 7.4. Following 4 min baseline, 3 μg/mL of biotinylated neutralization assay was used to identify the potent trimer was captured to streptavidin sensors (Sartorius 18- clone C-A11. We then grafted CDRs onto a human 5019) for 4 min followed by an additional 4 min baseline. framework to create the humanized clone hN2Y, which IgG binding was monitored for 4 min followed by 30 min strongly neutralized the Wuhan-Hu-1 strain. As evasive dissociation in fresh buffer only wells. Each IgG was tested SARS-CoV-2 variants emerged, many groups attempted in 1:2 titration series from 10 to 0.16 nM. A reference to identify novel antibodies from vaccinated or infected sensor omitting IgG was included in each assay and used patient samples [16, 17]. We instead employed our STEM for subtraction. All data were globally fit to 1:1 binding platform to broaden neutralization potency of hN2Y to model to derive kinetic and affinity values, only omitting SARS-CoV-2 variants as they emerged in real time (Fig. 1). individual traces or times due to sensor error or low signal. Single CDR mutant libraries of hN2Y were first created For the in-tandem epitope binning assay, following the (Fig. 1A). CDR libraries were designed based on observed same baseline and antigen loading steps described above, amino acid frequencies for characterized human antibodies an initial 10 min 10 nM IgG-binding step was performed to possessing a similar germline and predicted CDR canonical allow for signal saturation, followed by immediate transfer structure usage to both further humanize CDRs and to to wells containing 10 nM IgGs to monitor binding for avoid introduction of sequence liabilities. By using phage an additional 10 min. The final nm shift values after the display, each CDR library can be sampled at sizes up to second IgG incubation step were used to prepare a heatmap 2E + 10. These separate CDR libraries were first selected of results. on Wuhan-Hu-1 spike trimer to eliminate non-binders and to isolate a large pool of diverse clones. Then, CDR pools from the selected phage populations were amplified and randomly paired by overlap PCR to create secondary RESULTS combined light chain and combined heavy chain libraries Therapeutic mAbs were engineered by STEM in real time as (Fig. 1B). To increase neutralization breadth, the combined SARS-CoV-2 variants emerged light chain library was selected on the Beta variant, and the Soon after the onset of the COVID-19 pandemic, we combined heavy chain library was selected separately on used our rabbit discovery platform, which pairs rabbit Alpha, Beta, Gamma, and Epsilon variants. We identified immunization with Fab-phage display, to create a thera- an exceptional candidate from the combined light chain peutic lead candidate mAb. Rabbits were immunized with library, LxC1-G10, which showed good neutralization 112 Antibody Therapeutics, 2023 Figure 2. Broad SARS-CoV-2 neutralization for engineered mAbs by sVNT. (A) Purified mAbs were tested by sVNT [15]. Wells were coated with recombinant ACE2-Fc, and His-tagged spike trimer protein was mixed with serially diluted mAb prior to addition. Bound spike trimer was detected by anti-His tag antibody. All experiments were performed in triplicates. 6R8/6R9 clones (blue) showed highly potent and broad neutralization of all Omicron variants tested including XBB.1.5 and BQ.1.1. LxC1-G10 (purple) light chain engineered clone showed improved binding compared to lead humanized candidate hN2Y (green) or parental rabbit clone C-A11 (red), though it lost potency for later Omicron variants. Bebtelovimab (orange) showed no neutralization of later Omicron variants. (B) IC values for all variants, with non-neutralizing clones reported at 1000 ng/mL. (C) RBD residues modified in the different SARS-CoV-2 variants compared to the Wuhan-Hu-1 strain. Residues are colored by mutation. Colored side bars for each mAb note variants that are neutralized with IC <100 ng/mL. of BA.1 and BA.2. After the emergence of later BA.2 to recombinant ACE2 protein by ELISA, shows strong lineage Omicron variants, we again used our platform correlation with results from pseudovirus or live virus to further improve potency. Light and heavy chains were neutralization tests [15]. We thus used high-throughput separately amplified from the pre-selected phage pools sVNT to quantify neutralization potency of mAbs across and further combined to create a single library (Fig. 1C). SARS-CoV-2 variants. In this assay, recombinant human The library was selected on BA.5, BA.2.75.2, BQ.1.1, and Fcγ -tagged ACE2 receptor is immobilized while IgGs XBB variants, all of which possess mutation at F486. A are pre-incubated with spike trimer before addition to diverse set of seven highly potent, broadly neutralizing the blocked ACE2 coated wells. ACE2-bound spike IgGs (6R8/6R9 clones) was identified following eight or trimer is detected by affinity tag using ELISA, and IC nine rounds of panning and is characterized below. values are derived. mAbs tested include the original neutralizing rabbit clone C-A11, the parent humanized In vitro neutralization assay shows broadly neutralizing clone hN2Y, LxC1-G10 possessing mutations across light mAbs chain CDRs, 6R8/6R9 clones possessing mutations across Neutralization titers for sera or mAbs measured using both heavy and light chain CDRs, and therapeutic mAb sVNT, which assays blocking of recombinant spike trimer bebtelovimab. Antibody Therapeutics, 2023 113 Notably, humanization of the rabbit antibody did not significantly reduce efficacy, as hN2Y retained similar neu- BQ.1.1 tralization potency for Wuhan-Hu-1 and Delta variants compared to C-A11 (Fig. 2A and B and Table 1). However, both IgGs showed significantly reduced neutralization of BQ.1 the original Omicron variant BA.1. LxC1-G10 contains mutations within all three light chain CDRs compared to BA.4.6 hN2Y. Neutralization of early Omicron variants by LxC1- G10 was significantly improved to low ng/mL IC values for BA.1, BA.2, BA.2.3.20, BA2.75, and BN.1. However, all three mAbs lost efficacy against variant BA.2.75.2, which BF.7/BA.5.2.6/BF.11 is notably different compared to these others due to F486S (Fig. 2C). LxC1-G10 showed effective neutralization of BA.2.12.1 lacking F486 mutation but weak neutralization BA.5 of BA.4/5 (possessing F486V) in a pseudovirus neutraliza- tion test by CoVIC at La Jolla Institute [18] in concordance XBB.1.16 with our observation in sVNT. To recover efficacy, as described above, we engineered a panel of seven mAbs (6R8A-A2, 6R8A-E1, 6R8A-E11, XBB.1/XBB.1.9 6R8C-D10, 6R8C-G8, 6R9A-A2, and 6R9E-A8), which show strong neutralization of all Omicron variants to date including BA.1, BA.2, BA.2.3.20, BA.2.75, BN.1, XBB.1.5/XBB.1.9.1 BA.2.75.2, CH.1.1, CA.3.1, BR.2.1, XBF, XBB, XBB.1 (XBB.1.9), XBB.1.5 (XBB.1.9.1), XBB.1.16, BA.5, BF.7 XBB (BA.5.2.6/BF.11), BA.4.6, BQ.1, and BQ.1.1. To our knowledge, these are the most broadly neutralizing ACE2- blocking antibodies characterized to date against the latest XBF Omicron variants of SARS-CoV-2. Compared to both hN2Y and LxC1-G10, these IgG clones possess mutations across CDRs L1, L2, L3, H1, and H3. BR.2.1 Bebtelovimab is a therapeutic SARS-CoV-2 antibody developed by Eli Lilly and first given Emergency Use CA.3.1 Authorization by the US FDA in February 2022, though the agency revoked authorization in late November 2022 due to an inability to neutralize variants BQ.1 and BA.2.75.2 BQ.1.1. Bebtelovimab showed no efficacy against these variants by sVNT or against CH.1.1, CA.3.1, XBB, XBB.1, XBB.1.16, or XBB.1.5 (Fig. 2A and B), matching CH.1.1 expected results. S728-1157 clone was developed as a broadly neutralizing antibody against Delta and BA.1 [19]. S728-1157 additionally neutralized BA.2, BN.1, and BN.1 BA.5 by sVNT but failed to neutralize BA.2.75.2, CH.1.1, CA.3.1, BR.2.1, XBF, XBB, XBB.1.5, XBB.1.16, BQ.1, BA.2.75 or BQ.1.1 (Fig. S1). Another previously characterized, broadly neutralizing antibody S2X324, identified from human plasma [17, 20], showed strong neutralization of BA.2.3.20 Wuhan-Hu-1, Delta, BA.2, BA.2.75, BR.2.1, XBF, and BA.5 in our sVNT but failed to neutralize CH.1.1, CA.3.1, XBB.1.5, XBB.1.16, BQ.1, or BQ.1.1 (Fig. S2). In contrast, BA.2 our engineered IgGs neutralized all SARS-CoV-2 variants tested and all currently circulating globally dominant BA.1 variants to date. Binding affinity correlates with neutralization potency Delta BLI was performed by capturing biotinylated spike trimer to a streptavidin-coated sensor surface to monitor bind- Wuhan-Hu-1 ing of serially diluted mAbs, with traces globally fit to 1:1 binding models to derive kinetic and affinity values. hN2Y showed slightly reduced affinity for the Beta variant IC50 (ng/mL) compared to C-A11 (Fig. S3), but affinity was increased to Table 1. IC values derived from sVNT data. Calculated IC values from sVNT assay are provided including 95% confidence intervals. 6R8/6R9 clones show low ng/mL IC 50 50 50 values for all SARS-CoV-2 variants tested. NA, no significant neutralization activity (>1000 ng/mL); UD, undefined C-A11 10.1 11.6 NA 67.1 23.4 10.2 10.5 NA NA NA NA NA NA NA NA NA NA NA NA NA NA hN2Y 19.4 19.8 211 91.9 63.6 38.5 32.8 NA NA NA NA NA NA NA NA NA NA NA NA NA NA LxC1-G10 7.0 18.3 5.7 4.6 9.1 14.4 7.1 NA NA NA NA NA NA NA NA NA 636 NA 309.7 NA NA 6R8A-A2 11.7 12.7 6.2 5.5 11.5 9.2 6.7 31.8 16.5 15.4 18.7 31.8 17.1 19.0 20.9 37.4 10.1 13.1 12.4 15.2 10.3 6R8A-E1 11.9 8.9 5.4 5.2 11.5 8.9 6.9 26.6 14.9 13.3 12.6 18.8 20.3 22.8 20.6 33.1 10.9 14.4 12.6 13.2 9.3 6R8A-E11 11.4 11.5 5.7 6.2 12.5 9.8 7.6 33.1 20.0 17.9 15.4 27.5 29.0 21.2 28.5 39.3 10.9 17.8 13.3 14.3 15.4 6R8C-D10 12.7 11.6 5.8 5.0 12.3 8.1 6.6 25.1 13.3 19.5 19.1 16.3 11.7 14.9 16.9 20.2 10.2 14.1 12.7 11.0 12.0 6R8C-G8 12.0 9.7 6.0 5.6 11.6 8.8 7.2 34.9 14.0 15.8 15.9 26.1 19.0 15.7 21.7 39.3 11.5 16.2 11.2 12.5 11.1 6R9A-A2 9.8 10.5 6.7 6.2 8.5 10.5 7.3 33.5 8.7 24.9 20.9 33.3 10.1 14.7 17.8 39.7 9.4 10.0 8.6 9.3 14.1 6R9E-A8 13.1 12.6 5.6 5.6 13.5 8.9 5.6 18.8 13.9 17.9 11.8 12.8 19.0 15.9 19.6 28.5 9.8 16.5 12.9 12.5 7.4 bebtelovimab 8.1 9.5 55.9 7.2 407 171 48.4 NA 38.6 NA 149 79.5 NA NA NA NA 5.3 9.9 8.8 NA NA 114 Antibody Therapeutics, 2023 Figure 3. BLI shows high affinity to all SARS-CoV-2 variants for engineered mAbs. In each assay, biotinylated spike protein trimer was captured to the streptavidin-coated sensor surface and dipped into a 7-fold dilution series of IgG at a starting concentration of 10 nM. Binding was monitored for 4 min, followed by dissociation monitored for 30 min. Traces were referenced to a sensor omitting IgG. (A) For each variant tested, each experiment is shown with the same scale y-axis. C-A11, hN2Y, and LxC1-G10 lost binding activity for BQ.1.1 and XBB.1.5, but all seven 6R8/6R9 clones showed strong binding. Therapeutic mAbs tested under the same conditions showed weak or no binding. (B) BLI traces were globally fit to a 1:1 binding model to derive K 6 −1 −1 −6 −1 affinity values (pM), (C) k , on rates (× 10 M ∗s ), and (D) k , off rates (×10 s ). Some mAb affinities were below the detection limit of the assay (<1 pM) due to exceptionally slow off rates even when measured for 30 min. For some weakly binding clones, affinity, and/or kinetic values could not be determined. N.B., non-binding; n.d., not derived. Antibody Therapeutics, 2023 115 low pM binding K for LxC1-G10. LxC1-G10 possessed 24 pM K affinity for BA.1, 5-fold improved over hN2Y. However, LxC1-G10 binding affinity was itself decreased nearly 100-fold for BQ.1.1 (2.1 nM) compared to BA.1 (24 pM) (Fig. 3A and B). The seven 6R8/6R9 mAbs possessed 7 pM or lower K for BQ.1.1 and below 70 pM K for XBB.1.5. Binding affinity for these mAbs was increased nearly 10 000-fold (for BQ.1.1) or above undetectable levels (for XBB.1.5) compared to parent clone hN2Y. While on-rates were simi- lar (Fig. 3C), the difference in binding affinity for BQ.1.1 is primarily due to a nearly 1000-fold reduction in off- −3 −1 rate, with hN2Y possessing 10 s off-rate while 6R8/6R9 −6 −1 clones showed 10 s off rates (Fig. 3D). 6R8C-G8 and 6R9E-A8 were also tested on XBF, CH.1.1, and CA.3.1 variants; these mAbs showed strong binding to all three, though weaker relative binding to CA.3.1. Bebtelovimab showed weak binding to BQ.1.1 by BLI and no binding at all to XBB.1.5. Evusheld is an antibody cocktail of Tixagevimab and Cilgavimab first given Emer- Figure 4. BLI epitope binning shows no change in binding site for gency Use Authorization by the US FDA in December engineered mAbs. In an in-tandem assay, 10 nM of primary mAb was first saturated on the spike protein trimer for 10 min, followed by dipping into 2021, though the agency revoked authorization January fresh 10 nM solution of secondary mAb. The nm shift values obtained 2023 due to the inability to neutralize the latest Omi- at the end of the secondary mAb incubation are reported and colored cron variants. Both tixagevimab and cilgavimab showed by heatmap. hN2Y, LxC1-G10, 6R8C-G8, and 6R9E-A8 all binned into no binding to BQ.1.1 or to XBB.1.5 (Fig. 3A and B). In the same cluster. CR3022 did not cross-compete with any other mAb as contrast, our engineered IgGs show low pM binding K for expected, while Bebtelovimab and Cilgavimab clustered together. these dominant SARS-CoV-2 Omicron variants currently circulating in the US. engineered a single IgG, originally developed for neutral- ization of the Wuhan-Hu-1 strain. Engineered mAbs recognize the same epitope as the parent Our humanized clone hN2Y was derived from a rabbit antibody hN2Y lead candidate. Rabbits are an excellent source of potent antibodies due to the exceptionally high diversity of their We then tested whether broader neutralization potency immune repertoire by use of somatic gene conversion and was achieved by significantly altering the epitope. BLI was somatic hypermutation [22]. C-A11 rabbit mAb possesses used for in-tandem epitope binning. In total, 10 nM of a 19 amino acid CDR H3 and 9 amino acid CDR L3. IgG was first saturated on the sensor for 600 s, followed Long CDR H3 loops, typically defined as ≥24 amino acids, by dipping the sensor into new solutions of 10 nM IgGs. are a common feature of broadly neutralizing antiviral The results are presented as a heatmap using the final nm mAbs [23]. Though not technically classified as ‘long,’ the shift values obtained after incubation with the second IgG longer CDR H3 may have contributed to C-A11’s potent (Fig. 4). hN2Y, LxC1-G10, 6R8C-G8, and 6R9E-A8 all neutralization of early variants through Delta. Despite high competed with each other, suggesting that they bind to CDR diversity, humanization of rabbit clones is relatively identical or overlapping epitopes. CR3022, a neutralizing straightforward due to their restricted use of a limited antibody that does not block ACE2 receptor binding [21], number of germline genes [24]. After humanization, hN2Y does not cross-compete with any antibody as expected. showed almost no change in affinity or neutralization Interestingly, both Bebtelovimab and Cilgavimab, which activity compared to C-A11 and showed even stronger block ACE2 receptor binding but do not bind BQ.1.1 or neutralization of some variants including BA.1 and XBB.1.5, were binned into a distinct cluster that does not BA.2. compete with our engineered mAbs. This demonstrates that Omicron variant BA.1 reduced the neutralization broader neutralization activity of the engineered IgGs was potency of hN2Y, as observed for many other therapeutic likely achieved by optimizing binding to conserved SARS- antibodies developed prior to December 2021. As detailed CoV-2 residues within the same epitope as the parent mAb, above, LxC1-G10 was selected by panning a combined light rather than by significantly changing the binding site. chain library on the Beta variant, which possesses RBD mutations K417N/E484K/N501Y (Fig. 2C). LxC1-G10 showed sub-pM affinity for the Beta variant, >100-fold DISCUSSION increased compared to hN2Y (Fig. 3B). This increased We have detailed the engineering and activity of broadly binding was sufficient to improve neutralization of early neutralizing mAbs that show potent efficacy against all Omicron subvariants BA.1, BA.2, BA.2.3.20, BA.2.75, and dominant Omicron circulating variants including XBB.1.5 BN.1, despite those variants possessing many additional and BQ.1.1. Uniquely, instead of identifying a novel clone mutations within the spike protein. LxC1-G10 shares an from infected or immunized samples, we rather iteratively identical heavy chain as hN2Y but possesses over 14 amino 116 Antibody Therapeutics, 2023 acid differences distributed across all light chain CDRs. is designed to sample the broadest sequence diversity Light chain CDR mutations alone were thus sufficient including inter-CDR cooperative mutations and directly to confer significant improvement in binding affinity and incorporates developability filters early during phage neutralization potency. panning. However, LxC1-G10 lost activity against all Omicron Comparing hN2Y to 6R8/6R9 mAbs, there are four variants possessing a mutation at F486. F486 mutations are to five amino acid differences in CDR L1, four in CDR found in all later Omicron variants and have been linked L2, four in CDR L3, six to nine in CDR H1, and two to immune evasion [1]. The further engineered 6R8/6R9 in CDR H3, totaling 21–24 amino acid mutations accu- series mAbs, selected from panning a combined light mulated during affinity maturation. This is a very large and heavy chain library on Omicron variants possessing number of changes which would be difficult to impossi- F486S (BA.2.75.2 and XBB) and F486V (BA.5 and ble to obtain by any other established affinity maturation BQ.1.1) mutations, showed exceptional neutralization method. Notably, none of the CDR sequences from the potency and binding for all SARS-CoV-2 variants tested, 6R8/6R9 clones were identified during screening of single including those with alternate amino acids at this site CDR libraries (panned on Wuhan-Hu-1) or separate com- such as F486I (BR.2.1) and F486P (XBF and XBB.1.5). bined CDR libraries (separately panned on Alpha, Beta, Compared to both hN2Y and LxC1-G10, 6R8/6R9 mAbs Gamma, Delta, or Epsilon), making it unlikely that a CDR possess two mutations in CDR H3 (which is identical walking [33] or parallel CDR optimization [26] approach among all 6R8/6R9 clones) and 6–9 mutations in CDR would be successful. Alternative methods combine satu- H1. Compared to LxC1-G10, 6R8/6R9 mAbs possess rated mutagenesis libraries by DNA shuffling [34], but the 11–13 mutations across all light chain CDRs. Extensive likelihood of these methods resulting in two or more tan- mutagenesis across all CDRs except H2 was thus required dem mutations within an individual CDR, as observed in to confer the increased neutralization activity and binding all 6R8/6R9 clones, is very low due to the improbability of affinity observed. recombination within such close proximity. Computational Both our early lead candidate hN2Y (CoVIC-359), and methods used for affinity maturation have shown promise, LxC1-G10 (CoVIC-362) were submitted to the CoVIC but the quality of prediction begins to fall and the complex- at La Jolla Institute of Immunology (https://covicdb. ity in modeling begins to increase dramatically beyond five lji.org/) for analysis and testing against over 350 other amino acid mutations, making it impossible to predict the final 6R8/6R9 sequences in silico [35]. Taken together, only therapeutic antibody candidates [18, 25]. CoVIC is an by applying our STEM platform, which samples excep- international effort to conduct side-by-side analyses of tionally large diversity across all CDRs, could these potent candidate antibody therapeutics targeting the SARS-CoV- neutralizing antibodies be created. The absence of competi- 2 spike protein in standardized assays. Both antibodies tor mAbs showing similar potency further underscores our showed nearly 100% protection against cell infection in approach. pseudovirus neutralization assay for Beta, Delta, BA.1, We have characterized an antibody panel that outper- BA.1.1, and BA.2 variants, and showed efficacy with live forms all existing FDA approved antibody therapeutics Wuhan-Hu-1 virus in cell-based neutralization assay and against all globally dominant circulating SARS-CoV-2 during in vivo mouse challenge [25]. Both were binned into strains to date. This study validates our unique STEM the same epitope class, RBD-2a, characterized by bivalent platform. Though our lead humanized clone showed no intra-spike binding, where both arms of the IgG bind to binding to or neutralization of XBB.1.5, we successfully two spike units within a single trimer. This binding motif recovered potency by engineering through repeated sam- was further validated by cryo-EM in the CoVIC study, pling and combination of CDRs progressively against in which LxC1-G10 was observed to bind to both spike SARS-CoV-2 variants. The ability to recover neutral- units in the ‘up’ conformation, a conformational state that ization activity from a single lead mAb demonstrates exposes the ACE2 receptor-binding site [25]. Our epitope the exceptional CDR sequence diversity that can be binning assay (Fig. 4) demonstrated that the engineered obtained by applying an iterative selection and library 6R8/6R9 mAbs likely bind in the same RBD-2a epitope creation strategy. By continuing to engineer a single class. The increased avidity conferred by bivalent binding mAb, lead time was significantly reduced and significant likely contributes to the strong neutralization activity and resources were saved by avoiding the need to repeat the high affinity for 6R8/6R9 mAbs. discovery process by immunization or screening. In vitro In traditional humanization, animal-derived CDRs are affinity maturation of a single neutralizing antibody has grafted into human framework germline genes. Then, the potential to be far more efficient and cost-effective further engineering is performed using various methods to compared to discovery campaigns reliant on in vivo affinity obtain the best lead candidate for therapeutic mAb devel- maturation by successive vaccination and/or infection. In opment [26–32]. These methods suffer from notable set- patient samples, broadly neutralizing antibodies must be backs including [1] the inability to easily sample mutations identified from a large background of non-neutralizing that work cooperatively within or across CDRs such as in anti-spike protein antibodies, while affinity maturation of saturation mutagenesis and CDR walking, [2] introduction a neutralizing antibody allows for fine tuning around a of unwanted framework mutations such as in error prone known neutralizing epitope. As in this study, filters can be PCR, [3] lack of secondary selection pressure when combin- included during phage panning, including heat treatment ing beneficial mutations such as in parallel CDR optimiza- to remove unstable clones [14] and negative subtraction on tion [26], and [4] almost all omit selection pressure for well- BVP to remove polyreactive clones, to ensure that the final behaved clones. In contrast, our in vitro STEM platform Antibody Therapeutics, 2023 117 resulting candidates retain good developability profiles. project, and K.C.E. and T.M. wrote the manuscript, with This approach can be adapted for future therapeutic mAb all authors providing comments and edits. development against SARS-CoV-2 or other viruses, even Kevin C. Entzminger (Conceptualization, Data curation, to allow for rapid development of broadly neutralizing Formal analysis, Investigation, Methodology, Project therapeutic mAbs in real time. administration, Supervision, Validation, Writing—original draft, Writing—review & editing [Equal]), Jonathan K. Fleming (Data curation, Formal analysis, Investigation, ACKNOWLEDGMENTS Methodology, Writing—review & editing [Equal]), Paul D. Entzminger (Data curation, Formal analysis, Investigation, We thank Drs Erica Ollmann Saphire and Sharon Schendel Methodology, Writing—review & editing [Equal]), Lisa at CoVIC of La Jolla Institute for testing and providing Yuko Espinosa (Data curation, Investigation [Equal]), us the data for our antibodies. CoVIC is an international Alex Samadi (Data curation, Investigation, Methodology, effort to conduct side-by-side analyses of candidate anti- Writing—review & editing [Equal]), Yuko Hiramoto body therapeutics targeting the SARS-CoV-2 spike protein (Data curation, Investigation, Writing—review & editing in standardized assays. [Equal]), C.J. Okumura (Data curation, Funding acquisi- tion, Project administration, Writing—review & editing [Equal]), and Toshiaki Maruyama (Conceptualization, SUPPLEMENTARY DATA Data curation, Formal analysis, Investigation, Method- Supplementary data are available at ABT online. ology, Project administration, Supervision, Writing— original draft, Writing—review & editing [Equal]) FUNDING Funding for this study was provided by Abwiz Bio Inc. ETHICS AND CONSENT STATEMENT The CoVIC was supported by the COVID-19 Therapeutics Accelerator (INV-006133), the Bill and Melinda Gates Not applicable. Foundation (OPP1210938), a supplement to National Institutes of Health/National Institute of Allergy and Infectious Diseases (grant U19 AI142790-S1), and the ANIMAL RESEARCH STATEMENT GHR Foundation. Rabbit immunization was performed by ProSci (San Diego, CA), following all Institutional Animal Care and Use Com- mittee guidelines. Care was taken throughout the study to CONFLICTS OF INTEREST STATEMENT prevent any undue animal suffering. All authors are employees of Abwiz Bio Inc. Abwiz Bio has filed patent applications relating to anti-SARS-CoV- 2 antibodies and STEM technology with K.C.E., C.O., and T.M. as inventors. Some mAbs in this study are being REFERENCES considered for development of therapeutic antibodies. 1. Cao, Y, Yisimayi, A, Jian, F et al. BA.2.12.1, BA.4 and BA.5 escape antibodies elicited by Omicron infection. Nature 2022; 608: 593–602. 2. Willett, BJ, Grove, J, MacLean, OA et al. SARS-CoV-2 Omicron is an immune escape variant with an altered cell entry pathway. Nat DATA AVAILABILITY Microbiol 2022; 7: 1161–79. The datasets used and/or analyzed during the current study 3. Qu, P, Faraone, JN, Evans, JP et al. Extraordinary evasion of are available from the corresponding author on reasonable neutralizing antibody response by Omicron XBB.1.5, CH.1.1 and request. CA.3.1 variants. 2023. https://doi.org/10.1101/2023.01.16.524244. 4. Callaway, E. Is coronavirus variant XBB.1.5 a global threat? Nature 2023; 613: 222–3. 5. Arora, P, Kempf, A, Nehlmeier, I et al. Omicron sublineage BQ.1.1 AUTHORS’ CONTRIBUTIONS resistance to monoclonal antibodies. Lancet Infect Dis 2022; 23: K.C.E. isolated the neutralizing clone C-A11 from the 22–3. 6. Imai,M,Ito,M,Kiso, M et al. Efficacy of antiviral agents against rabbit library, designed CDR mutant libraries, performed omicron subvariants BQ.1.1 and XBB. New Engl J Med 2023; 388: selection of clones that bind multiple variants of COVID- 89–91. 19, and performed BLI studies. T.M. designed the project, 7. Planas, D, Bruel, T, Staropoli, I et al. Resistance of omicron cloned RBD containing spike protein for immunization, subvariants BA.2.75.2, BA.4.6 and BQ.1.1 to neutralizing built the rabbit phage libraries, performed panning and antibodies. 2022. https://doi.org/10.1101/2022.11.17.516888. 8. Cao, Y, Jian, F, Wang, J et al. Imprinted SARS-CoV-2 humoral screening of the libraries, humanized the neutralizing immunity induces convergent Omicron RBD evolution. Nature 2022; clone C-A11, and performed sVNTs. J.K.F. produced 614: 521–9. https://doi.org/10.1038/s41586-022-05644-7. all spike trimers, ACE2 protein, and antibodies. P.D.E. 9. Bowe, B, Xie, Y, Al-Aly, Z. Acute and postacute sequelae associated performed sVNTs and analyzed the data. L.Y.E. performed with SARS-CoV-2 reinfection. Nat Med 2022; 28: 2398–24055. phage panning on Omicron variants. A.S. preformed 10. Brouwer, PJM, Caniels, TG, van der Straten, K et al. Potent neutralizing antibodies from COVID-19 patients define multiple manufacturability and biophysical risk assessment with targets of vulnerability. Science (1979) 2020, 1979; 369: 643–50. J.K.F., Y.H., and T.M. created spike trimer variant vectors, 11. Taylor, PC, Adams, AC, Hufford, MM et al. Neutralizing cloned antibodies into IgG expression vector, and prepared monoclonal antibodies for treatment of COVID-19. Nat Rev plasmid DNA for transfection. T.M. and C.O. oversaw the Immunol 2021; 21: 382–93. 118 Antibody Therapeutics, 2023 12. Morales-Núñez, JJ, Muñoz-Valle, JF, Torres-Hernández, PC et al. 24. Knight, KL. Restricted VH gene usage and generation of antibody Overview of neutralizing antibodies and their potential in diversity in rabbit. Annu Rev Immunol 1992; 10: 593–616. COVID-19. Vaccine 2021; 9: 1376. 25. Callaway, HM, Hastie, KM, Schendel, SL et al. Bivalent intra-spike 13. Keating, SM, Mizrahi, RA, Adams, MS et al. Generation of binding provides durability against emergent omicron lineages: recombinant hyperimmune globulins from diverse B-cell repertoires. results from a global consortium. Cell Rep 2023; 42: 112014. Nat Biotechnol 2021; 39: 989–99. 26. Barbas, CF III, Burton, DR. Selection and evolution of high-affinity 14. Entzminger, KC, Johnson, JL, Hyun, J et al. Increased fab human anti-viral antibodies. Trends Biotechnol 1991; 14:230–4. thermoresistance via VH-targeted directed evolution. Protein Eng 27. Simons, JF, Lim, YW, Carter, KP et al. Affinity maturation of Des Selection 2015; 28: 365–77. antibodies by combinatorial codon mutagenesis versus error-prone 15. Tan, CW, Chia, WN, Qin, X et al. A SARS-CoV-2 surrogate virus PCR. MAbs 2020; 12: e1803646. neutralization test based on antibody-mediated blockage of 28. Kennedy, PJ, Oliveira, C, Granja, PL et al. Monoclonal antibodies: ACE2–spike protein–protein interaction. Nat Biotechnol 2020; 38: technologies for early discovery and engineering. Crit Rev Biotechnol 1073–8. 2018; 38: 394–408. 16. Westendorf, K, Zentelis, S, Wang, L et al. LY-CoV1404 29. Rathore, AS, Sarker, A, Gupta, RD. Recent developments toward (bebtelovimab) potently neutralizes SARS-CoV-2 variants. Cell Rep antibody engineering and affinity maturation. Protein Pept Lett 2022; 39: 110812. 2018; 25: 886–96. 17. Cameroni, E, Bowen, JE, Rosen, LE et al. Broadly neutralizing 30. Tabasinezhad, M, Talebkhan, Y, Wenzel, W et al. Trends in antibodies overcome SARS-CoV-2 omicron antigenic shift. Nature therapeutic antibody affinity maturation: from in-vitro towards 2022; 602: 664–70. next-generation sequencing approaches. Immunol Lett 2019; 212: 18. Hastie, KM, Li, H, Bedinger, D et al. Defining variant-resistant 106–13. epitopes targeted by SARS-CoV-2 antibodies: a global consortium 31. Lim, CC, Choong, YS, Lim, TS. Cognizance of molecular methods study. Science 2021; 374: 472–8. for the generation of mutagenic phage display antibody libraries for 19. Changrob, S, Halfmann, PJ, Liu, H et al. Site of vulnerability on affinity maturation. Int J Mol Sci 2019; 20: 1861. SARS-CoV-2 spike induces broadly protective antibody to 32. Prassler, J, Steidl, S, Urlinger, S. In vitro affinity maturation of antigenically distinct omicron SARS-CoV-2 subvariants. 2022. HuCAL antibodies: complementarity determining region exchange https://doi.org/10.1101/2022.10.31.514592. and RapMAT technology. Immunotherapy 2009; 1: 571–83. 20. Park,Y,Pinto,D,Walls,AC et al. Imprinted antibody responses 33. Yang, W, Green, K, Pinz-Sweeney, S et al. CDR walking against SARS-CoV-2 Omicron sublineages. Science 2022; 378: mutagenesis for the affinity maturation of a potent human 619–27. anti-HIV-1 antibody into the picomolar range. J Mol Biol 1995; 254: 21. Tian, X, Li, C, Huang, A et al. Potent binding of 2019 novel 392–403. coronavirus spike protein by a SARS coronavirus-specific human 34. Banach, BB, Tripathi, P, Pereira, LDS et al. Highly protective monoclonal antibody. Emerg Microbes Infect 2020; 9: 382–5. antimalarial antibodies via precision library generation and yeast 22. Weber, J, Peng, H, Rader, C. From rabbit antibody repertoires to display screening. JExp Med 2022; 219: e20220323. rabbit monoclonal antibodies. Exp Mol Med 2017; 49: e305. 35. Bachas, S, Rakocevic, G, Spencer, D et al. Antibody optimization 23. Sok, D, Burton, DR. Recent progress in broadly neutralizing enabled by artificial intelligence predictions of binding affinity and antibodies to HIV. Nat Immunol 2018; 19: 1179–88. naturalness. 2022. https://doi.org/10.1101/2022.08.16.504181. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Antibody Therapeutics Oxford University Press http://www.deepdyve.com/lp/oxford-university-press/rapid-engineering-of-sars-cov-2-therapeutic-antibodies-to-increase-0o1o5LZVJI

Loading next page...

References (39)

M. Imai, Mutsumi Ito, Maki Kiso, S. Yamayoshi, R. Uraki, S. Fukushi, Shinji Watanabe, Tadaki Suzuki, Ken Maeda, Yuko Sakai-Tagawa, K. Iwatsuki-Horimoto, P. Halfmann, Y. Kawaoka (2022)
Efficacy of Antiviral Agents against Omicron Subvariants BQ.1.1 and XBB
The New England Journal of Medicine
Benjamin Bowe, Yan Xie, Z. Al‐Aly (2022)
Acute and postacute sequelae associated with SARS-CoV-2 reinfection
Nature Medicine, 28
H. Callaway, K. Hastie, S. Schendel, Haoyang Li, Xiaoying Yu, Jeremy Shek, Tierra Buck, S. Hui, Daniel Bedinger, C. Troup, S. Dennison, Kan Li, M. Alpert, Charles Bailey, S. Benzeno, J. Bonnevier, Jinqiu Chen, Charm Chen, Hyeseon Cho, Peter Crompton, V. Dussupt, Kevin Entzminger, Yassine Ezzyat, J. Fleming, N. Geukens, A. Gilbert, Y. Guan, Xiaojian Han, Christopher Harvey, Julia Hatler, Bryan Howie, Chao Hu, Ailong Huang, M. Imbrechts, A. Jin, Nik Kamachi, Gladys Keitany, M. Klinger, J. Kolls, S. Krebs, Tingting Li, Fei-yang Luo, T. Maruyama, M. Meehl, Letzibeth Mendez-Rivera, Andrea Musa, C. Okumura, B. Rubin, A. Sato, Meiying Shen, Anirudh Singh, Shuyi Song, Joshua Tan, J. Trimarchi, Dhruvkumar Upadhyay, Yingming Wang, Lei Yu, Tom Yuan, E. Yusko, Bjoern Peters, G. Tomaras, E. Saphire (2023)
Bivalent intra-spike binding provides durability against emergent Omicron lineages: Results from a global consortium
Cell Reports, 42
D. Sok, D. Burton (2018)
Recent progress in broadly neutralizing antibodies to HIV
Nature Immunology, 19
Yunlong Cao, A. Yisimayi, F. Jian, Weiliang Song, T. Xiao, Lei Wang, Shuo Du, Jing Wang, Qian-qian Li, Xiaosu Chen, Yuanling Yu, Peng Wang, Zhiying Zhang, Pulan Liu, R. An, Xiaohua Hao, Yao Wang, Jing Wang, Rui Feng, Haiyan Sun, Li-Jin Zhao, Wen Zhang, D. Zhao, Jiang Zheng, Lingling Yu, Can Li, Na Zhang, Rui Wang, Xiaowei Niu, Sijie Yang, Xuetao Song, Yangyang Chai, Ye Hu, Yan-mei Shi, Linlin Zheng, Zhiqiang Li, Qingqing Gu, Fei Shao, Wei-jin Huang, R. Jin, Zhongyang Shen, Youchun Wang, Xiangxi Wang, Junyu Xiao, X. Xie (2022)
BA.2.12.1, BA.4 and BA.5 escape antibodies elicited by Omicron infection
Nature, 608
B. Willett, J. Grove, O. MacLean, Craig Wilkie, G. Lorenzo, Wilhelm Furnon, D. Cantoni, S. Scott, N. Logan, Shirin Ashraf, Maria Manali, Agnieszka Szemiel, V. Cowton, Elen Vink, William Harvey, Chris Davis, Patawee Asamaphan, Katherine Smollett, Lily Tong, R. Orton, Joseph Hughes, P. Holland, Vanessa Silva, D. Pascall, K. Puxty, Ana Filipe, G. Yebra, Sharif Shaaban, Matthew Holden, R. Pinto, R. Gunson, K. Templeton, P. Murcia, Arvind Patel, P. Klenerman, S. Dunachie, Susanna Kronsteiner, S. Dunachie, P. Klenerman, E. Barnes, A. Brown, S. Adele, B. Kronsteiner, Sam Murray, P. Abraham, A. Deeks, M. Ansari, Thushan Silva, L. Turtle, S. Moore, J. Austin, Alex Richter, C. Duncan, R. Payne, Amy Curra, Amy Ash, C. Koshy, B. Kele, T. Cutino-Moguel, D. Fairley, J. McKenna, T. Curran, Helen Adams, C. Fraser, D. Bonsall, H. Fryer, K. Lythgoe, Laura Thomson, T. Golubchik, Abigail Murray, D. Singleton, Shaun Beckwith, Anna Mantzouratou, Magdalena Barrow, S. Buchan, N. Reynolds, B. Warne, Joshua Maksimovic, Karla Spellman, Kathryn Mccluggage, M. John, Robert Beer, Safiah Afifi, Sian Morgan, A. Mack, A. Marchbank, A. Price, A. Morriss, Catherine Bresner, C. Kitchen, I. Merrick, Joel Southgate, Martyn Guest, Owen Jones, R. Munn, T. Connor, Thomas Whalley, T. Workman, William Fuller, A. Patel, Bindi Patel, G. Nebbia, J. Edgeworth, L. Snell, R. Batra, T. Charalampous, A. Beckett, E. Shelest, S. Robson, A. Underwood, Ben Taylor, C. Yeats, D. Aanensen, Khalil AbuDahab, Mirko Menegazzo, A. Joseph, G. Clark, H. Howson-Wells, L. Berry, Manjinder Khakh, M. Lister, T. Boswell, Vicki Fleming, C. Holmes, C. McMurray, Jessica Shaw, Julian Tang, K. Fallon, Minaxshi Odedra, Nicholas Willford, P. Bird, Thomas Helmer, L. Williams, Nicola Sheriff, Sharon Campbell, Veena Raviprakash, Victoria Blakey, Christopher Moore, Fei Sang, Johnny Debebe, M. Carlile, M. Loose, N. Holmes, Victoria Wright, M. Torok, W. Hamilton, A. Carabelli, A. Jermy, B. Blane, C. Churcher, C. Ludden, D. Aggarwal, Elaine Westwick, E. Brooks, Georgina McManus, Katerina Galai, Kenneth Smith, Kim Smith, MacGregor Cox, Mireille Fragakis, P. Maxwell, Sarah Judges, S. Peacock, T. Feltwell, A. Kenyon, Sahar Eldirdiri, Thomas Davis, Joshua Taylor, Ngee Tan, A. Zarebski, B. Gutiérrez, J. Raghwani, L. Plessis, M. Kraemer, O. Pybus, Sarah François, S. Attwood, T. Vasylyeva, A. Jahun, I. Goodfellow, I. Georgana, M. Pinckert, M. Hosmillo, Rhys Izuagbe, Y. Chaudhry, Felicity Ryan, Hannah Lowe, S. Moses, L. Bedford, J. Cargill, W. Hughes, J. Moore, Susanne Stonehouse, D. Shah, Jack Lee, Julianne Brown, K. Harris, Laura Atkinson, N. Storey, M. Spyer, F. Flaviani, A. Alcolea-Medina, J. Sehmi, J. Ramble, Natasha Ohemeng-Kumi, P. Smith, B. Bertolusso, Claire Thomas, G. Vernet, J. Lynch, N. Moore, N. Cortes, Rebecca Williams, S. Kidd, L. Levett, Monika Pusok, P. Grant, S. Kirk, W. Chatterton, Li Xu-McCrae, Darren Smith, G. Young, M. Bashton, Katie Kitchman, K. Gajee, K. Eastick, P. Lillie, P. Burns, William Everson, A. Cox, A. Holmes, F. Bolt, J. Price, M. Pond, P. Randell, P. Madona, S. Mookerjee, E. Volz, L. Geidelberg, M. Ragonnet-Cronin, O. Boyd, Robert Johnson, C. Pope, A. Witney, I. Monahan, K. Laing, Katherine Smollett, A. McNally, C. McMurray, Joanne Stockton, J. Quick, N. Loman, Radoslaw Poplawski, S. Nicholls, W. Rowe, Anibolina Castigador, E. Macnaughton, K. Bouzidi, M. Sudhanva, T. Lampejo, R. Nunez, Cathy Breen, Graciela Sluga, Karen Withell, N. Machin, R. George, Shazaad Ahmad, David Pritchard, D. Binns, Nick Wong, Victoria James, Cheryl Williams, C. Illingworth, Christopher Jackson, D. Angelis, D. Pascall, Afrida Mukaddas, A. Broos, Ana Filipe, Daniel Mair, David Robertson, Derek Wright, E. Thomson, Igor Starinskij, Ioulia Tsatsani, J. Shepherd, J. Nichols, Joseph Hughes, K. Nomikou, Lily Tong, R. Orton, S. Vattipally, William Harvey, R. Sanderson, S. O’Brien, S. Rushton, Jon Perkins, Rachel Blacow, R. Gunson, Abbie Gallagher, E. Wastnedge, K. Templeton, M. McHugh, R. Dewar, S. Cotton, L. Coupland, Rachael Stanley, S. Dervisevic, Lewis Spurgin, Louise Smith, Clive Graham, D. Padgett, E. Barton, Garren Scott, A. Cross, M. Mirfenderesky, E. Swindells, J. Greenaway, Rebecca Denton-Smith, Robyn Turnbull, Giles Idle, K. Cole, Amy Hollis, Andrew Nelson, C. McCann, J. Henderson, M. Crown, W. Yew, W. Stanley, Nichola Duckworth, P. Clarke, Sarah Walsh, T. Sloan, K. Bicknell, R. Impey, S. Wyllie, Scott Elliott, S. Glaysher, D. Bradley, N. Killough, Tim Wyatt, A. Bosworth, B. Vipond, C. Pearson, E. Allara, E. Robinson, Hannah Pymont, H. Osman, Peter Muir, R. Hopes, S. Hutchings, M. Curran, S. Parmar, A. Thornton, A. Lackenby, Chloe Bishop, D. Bibby, David Lee, E. Gallagher, G. Dabrera, Ian Harrison, Jonathan Hubb, K. Twohig, M. Chand, N. Ellaby, Nikos Manesis, R. Myers, S. Platt, T. Mbisa, Vicki Chalker, G. Yebra, Matthew Holden, Sharif Shaaban, Stefan Rooke, Alec Birchley, Alexander Adams, A. Davies, Amy Gaskin, Bree Gatica-Wilcox, C. McKerr, Catherine Moore, Catryn Williams, Chris Williams, David Heyburn, E. Lacy, Ember Hilvers, Fatima Downing, G. Pugh, H. Jones, H. Asad, J. Coombes, Jessica Hey, Jessica Powell, J. Watkins, John Evans, L. Fina, Laura Gifford, L. Gilbert, Lee Graham, Malorie Perry, M. Morgan, Matthew Bull, Nicole Pacchiarini, N. Craine, S. Corden, Sara Kumžiene-Summerhayes, S. Rey, Sarah Taylor, S. Cottrell, Sophie Jones, Sue Edwards, Tara Annett, A. Trotter, A. Mather, Alp Aydin, A. Page, David Baker, Ebenezer Foster-Nyarko, Gemma Kay, J. O’Grady, L. Martins, Lizzie Meadows, N. Alikhan, Sophie Prosolek, Steven Rudder, Thanh Le-Viet, A. Casey, Liz Ratcliffe, Aditi Singh, Arun Mariappan, Chris Baxter, Clara Radulescu, David Simpson, Deborah Lavin, Fiona Rogan, Julia Miskelly, Marc Fuchs, Miao Tang, Sílvia Carvalho, S. Bridgett, T. Skvortsov, Z. Molnár, Newara Ramadan, B. Knight, Christopher Jones, C. Auckland, Helen Morcrette, Jennifer Poyner, Dianne Irish-Tavares, Eric Witele, J. Hart, T. Mahungu, T. Haque, Y. Bourgeois, C. Fearn, Kate Cook, Katie Loveson, S. Goudarzi, C. Evans, D. Partridge, Katie Johnson, Mehmet Yavus, M. Raza, Craig Mower, P. Baker, S. Essex, S. Bonner, Leanne Murray, Louisa Watson, Steven Liggett, Andrew Lawton, R. Lyons, Brendan Payne, G. Eltringham, J. Collins, S. Waugh, S. Burton-Fanning, Y. Taha, Christopher Jeanes, Andrea Gomes, D. Murray, Maimuna Kimuli, D. Dobie, Paula Ashfield, A. Best, B. Percival, E. Moles-Garcia, F. Ashford, J. Mirza, L. Crawford, M. Mayhew, Nicola Cumley, O. Megram, D. Frampton, J. Heaney, Matthew Byott, C. Houlihan, Charlotte Williams, E. Nastouli, Helen Lowe (2022)
SARS-CoV-2 Omicron is an immune escape variant with an altered cell entry pathway
Nature Microbiology, 7
K. Hastie, Haoyang Li, Daniel Bedinger, S. Schendel, S. Dennison, Kan Li, V. Rayaprolu, Xiaoying Yu, Colin Mann, M. Zandonatti, Ruben Avalos, D. Żyła, Tierra Buck, S. Hui, K. Shaffer, C. Hariharan, J. Yin, Eduardo Olmedillas, Adrian Enriquez, D. Parekh, Milite Abraha, E. Feeney, Gillian Horn, Y. Aldon, Hanif Ali, Sanja Aracic, R. Cobb, Ross Federman, Joseph Fernandez, J. Glanville, Robin Green, G. Grigoryan, Ana Hernandez, D. Ho, Kuan-Ying Huang, J. Ingraham, Weidong Jiang, P. Kellam, Cheolmin Kim, Minsoo Kim, Hyeong Kim, Chao Kong, S. Krebs, F. Lan, Guojun Lang, Sooyoung Lee, Cheuk Leung, Junli Liu, Yanan Lu, Anna MacCamy, A. McGuire, A. Palser, T. Rabbitts, Z. Tehrani, Mohammad Sajadi, R. Sanders, A. Sato, L. Schweizer, Jimin Seo, Bingqing Shen, Jonne Snitselaar, L. Stamatatos, Yongcong Tan, M. Tomic, M. Gils, S. Youssef, Jian Yu, Tom Yuan, Qian Zhang, B. Peters, G. Tomaras, T. Germann, E. Saphire (2021)
Defining variant-resistant epitopes targeted by SARS-CoV-2 antibodies: A global consortium study
Science, 374
P. Taylor, A. Adams, M. Hufford, I. Torre, K. Winthrop, R. Gottlieb (2021)
Neutralizing monoclonal antibodies for treatment of COVID-19
Nature Reviews. Immunology, 21
Data curation, Formal analysis, Investigation, Methodology, Writing-review & editing
C. Tan, W. Chia, X. Qin, Pei Liu, M. Chen, Charles Tiu, Zhiliang Hu, V. Chen, B. Young, W. Sia, Y. Tan, Randy Foo, Yongxiang Yi, D. Lye, D. Anderson, Lin‐Fa Wang (2020)
A SARS-CoV-2 surrogate virus neutralization test based on antibody-mediated blockage of ACE2–spike protein–protein interaction
Nature Biotechnology, 38
Wei-ping Yang, Kimberly Green, S. Pinz-Sweeney, Amelia Briones, D. Burton, C. Barbas (1995)
CDR walking mutagenesis for the affinity maturation of a potent human anti-HIV-1 antibody into the picomolar range.
Journal of molecular biology, 254 3
Siriruk Changrob, P. Halfmann, Hejun Liu, J. Torres, J. McGrath, G. Ozorowski, Lei Li, M. Kuroda, Tadashi Maemura, Min Huang, G. Wilbanks, Nai-Ying Zheng, H. Turner, S. Erickson, Yanbin Fu, Gagandeep Singh, F. Krammer, D. Sather, A. Ward, I. Wilson, Y. Kawaoka, P. Wilson (2022)
Site of vulnerability on SARS-CoV-2 spike induces broadly protective antibody to antigenically distinct omicron SARS-CoV-2 subvariants
bioRxiv
Justus Weber, Haiyong Peng, C. Rader (2017)
From rabbit antibody repertoires to rabbit monoclonal antibodies
Experimental & Molecular Medicine, 49
wrote the manuscript
D. Planas, T. Bruel, I. Staropoli, F. Guivel-Benhassine, F. Porrot, P. Maes, L. Grzelak, M. Prot, Said Mougari, C. Planchais, J. Puech, M. Saliba, Riwan Sahraoui, F. Fémy, N. Morel, Jérémy Dufloo, R. Sanjuán, H. Mouquet, E. André, L. Hocqueloux, E. Simon-Lorière, D. Veyer, T. Prazuck, H. Péré, O. Schwartz (2022)
Resistance of Omicron subvariants BA.2.75.2, BA.4.6 and BQ.1.1 to neutralizing antibodies
bioRxiv
K. Knight (1992)
Restricted VH gene usage and generation of antibody diversity in rabbit.
Annual review of immunology, 10
Prerna Arora, Amy Kempf, Inga Nehlmeier, Sebastian Schulz, H. Jäck, S. Pöhlmann, M. Hoffmann (2022)
Omicron sublineage BQ.1.1 resistance to monoclonal antibodies
The Lancet. Infectious Diseases, 23
P. Kennedy, Carla Oliveira, P. Granja, B. Sarmento (2018)
Monoclonal antibodies: technologies for early discovery and engineering
Critical Reviews in Biotechnology, 38
P. Brouwer, Tom Caniels, K. Straten, Jonne Snitselaar, Y. Aldon, S. Bangaru, J. Torres, Nisreen Okba, M. Claireaux, Gius Kerster, A. Bentlage, M. Haaren, D. Guerra, J. Burger, Edith Schermer, K. Verheul, Niels Velde, A. Kooi, J. Schooten, M. Breemen, Tom Bijl, K. Sliepen, A. Aartse, R. Derking, I. Bontjer, N. Kootstra, W. Wiersinga, G. Vidarsson, B. Haagmans, A. Ward, G. Bree, R. Sanders, M. Gils (2020)
Potent neutralizing antibodies from COVID-19 patients define multiple targets of vulnerability
Science (New York, N.y.), 369
José Morales-Núñez, J. Muñóz-Valle, Paola Torres-Hernández, J. Hernández-Bello (2021)
Overview of Neutralizing Antibodies and Their Potential in COVID-19
Vaccines, 9
P. Qu, Julia Faraone, John Evans, Yi-Min Zheng, Claire Carlin, M. Anghelina, P. Stevens, Soledad Fernandez, Daniel Jones, A. Panchal, L. Saif, E. Oltz, Kai Xu, R. Gumina, Shan-Lu Liu (2023)
Extraordinary Evasion of Neutralizing Antibody Response by Omicron XBB.1.5, CH.1.1 and CA.3.1 Variants
bioRxiv
C. Barbas, D. Burton (1996)
Selection and evolution of high-affinity human anti-viral antibodies.
Trends in biotechnology, 14 7
Maryam Tabasinezhad, Y. Talebkhan, W. Wenzel, H. Rahimi, E. Omidinia, F. Mahboudi (2019)
Trends in therapeutic antibody affinity maturation: From in-vitro towards next-generation sequencing approaches.
Immunology letters
J. Simons, Yoong Lim, Kyle Carter, Ellen Wagner, Nicholas Wayham, A. Adler, David Johnson (2020)
Affinity maturation of antibodies by combinatorial codon mutagenesis versus error-prone PCR
mAbs, 12
Kevin Entzminger, J. Johnson, Jeong-min Hyun, R. Lieberman, J. Maynard (2015)
Increased Fab thermoresistance via VH-targeted directed evolution.
Protein engineering, design & selection : PEDS, 28 10
(2023)
Is coronavirus variant XBB.1.5 a global threat
Entzminger (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Supervision, Validation, Writing-original draft, Writing-review & editing
S. Keating, Rena Mizrahi, M. Adams, Michael Asensio, Emily Benzie, Kyle Carter, Yao Chiang, Robert Edgar, Bishal Gautam, Ashley Gras, Jackson Leong, Renee Leong, Yoong Lim, Vishal Manickam, Angélica Medina-Cucurella, Ariel Niedecken, J. Saini, J. Simons, Matthew Spindler, Kacy Stadtmiller, Brendan Tinsley, Ellen Wagner, Nicholas Wayham, L. Tracy, C. Lundberg, D. Büscher, J. Terencio, L. Roalfe, Emma Pearce, Hayley Richardson, D. Goldblatt, Anushka Ramjag, C. Carrington, G. Simmons, M. Muench, S. Chamow, Bryan Monroe, C. Olson, Thomas Oguin, Heather Lynch, R. Jeanfreau, Rachel Mosher, Matthew Walch, C. Bartley, C. Ross, E. Meyer, A. Adler, David Johnson (2021)
Generation of recombinant hyperimmune globulins from diverse B-cell repertoires
Nature biotechnology, 39
Kathryn Westendorf, Stefanie Žentelis, Lingshu Wang, Denisa Foster, P. Vaillancourt, M. Wiggin, E. Lovett, R. Lee, J. Hendle, A. Pustilnik, J. Sauder, Lucas Kraft, Yuri Hwang, R. Siegel, Jinbiao Chen, B. Heinz, R. Higgs, N. Kallewaard, Kevin Jepson, R. Goya, Maia Smith, D. Collins, Davide Pellacani, Ping Xiang, Valentine Puyraimond, Markéta Řičicová, L. Devorkin, Caitlin Pritchard, Aoise O’Neill, Kush Dalal, P. Panwar, H. Dhupar, Fabian Garces, Courtney Cohen, J. Dye, K. Huie, C. Badger, D. Kobasa, J. Audet, J. Freitas, Saleema Hassanali, I. Hughes, Luis Munoz, Holly Palma, Bharathi Ramamurthy, R. Cross, T. Geisbert, Vineet Menacherry, K. Lokugamage, V. Borisevich, I. Lanz, Lisa Anderson, Payal Sipahimalani, Kizzmekia Corbett, E. Yang, Yi Zhang, W. Shi, T. Zhou, M. Choe, J. Misasi, P. Kwong, N. Sullivan, B. Graham, Tara Fernandez, C. Hansen, E. Falconer, J. Mascola, Bryan Jones, B. Barnhart (2022)
LY-CoV1404 (bebtelovimab) potently neutralizes SARS-CoV-2 variants
Cell Reports, 39
Y Cao, F Jian, J Wang (2022)
Imprinted SARS-CoV-2 humoral immunity induces convergent Omicron RBD evolution
, 614
Chia Lim, Y. Choong, T. Lim (2019)
Cognizance of Molecular Methods for the Generation of Mutagenic Phage Display Antibody Libraries for Affinity Maturation
International Journal of Molecular Sciences, 20
Bailey Banach, P. Tripathi, Lais Pereira, J. Gorman, T. Nguyen, Marlon Dillon, Ahmed Fahad, P. Kiyuka, Bharat Madan, Jacy Wolfe, Brian Bonilla, B. Flynn, J. Francica, N. Hurlburt, Neville Kisalu, Tracy Liu, Li Ou, Reda Rawi, A. Schön, Chen-Hsiang Shen, I-Ting Teng, Baoshan Zhang, M. Pancera, A. Idris, R. Seder, P. Kwong, B. Dekosky (2022)
Highly protective antimalarial antibodies via precision library generation and yeast display screening
The Journal of Experimental Medicine, 219
E. Cameroni, John Bowen, Laura Rosen, C. Saliba, Samantha Zepeda, K. Culap, D. Pinto, L. VanBlargan, A. Marco, J. Iulio, F. Zatta, Hannah Kaiser, Julia Noack, Nisar Farhat, N. Czudnochowski, C. Havenar-Daughton, Kaitlin Sprouse, J. Dillen, Abigail Powell, Alex Chen, Cyrus Maher, Li Yin, David Sun, Leah Soriaga, J. Bassi, C. Silacci-Fregni, C. Gustafsson, Nicholas Franko, J. Logue, N. Iqbal, I. Mazzitelli, J. Geffner, R. Grifantini, H. Chu, A. Gori, A. Riva, O. Giannini, A. Ceschi, P. Ferrari, P. Cippà, A. Franzetti-Pellanda, C. Garzoni, P. Halfmann, Y. Kawaoka, C. Hebner, L. Purcell, L. Piccoli, M. Pizzuto, A. Walls, M. Diamond, A. Telenti, H. Virgin, A. Lanzavecchia, G. Snell, D. Veesler, D. Corti (2021)
Broadly neutralizing antibodies overcome SARS-CoV-2 Omicron antigenic shift
Nature, 602
Young-Jun Park, D. Pinto, A. Walls, Zhuoming Liu, A. Marco, F. Benigni, F. Zatta, C. Silacci-Fregni, J. Bassi, Kaitlin Sprouse, A. Addetia, John Bowen, Cameron Stewart, M. Giurdanella, C. Saliba, B. Guarino, M. Schmid, Nicholas Franko, J. Logue, H. Dang, Kevin Hauser, J. Iulio, William Rivera, G. Schnell, Anushka Rajesh, Jiayi Zhou, Nisar Farhat, Hannah Kaiser, Martin Montiel-Ruiz, Julia Noack, Florian Lempp, Javier Janer, R. Abdelnabi, P. Maes, P. Ferrari, A. Ceschi, O. Giannini, G. Melo, Lauriane Kergoat, H. Bourhy, J. Neyts, Leah Soriaga, L. Purcell, G. Snell, S. Whelan, A. Lanzavecchia, H. Virgin, L. Piccoli, H. Chu, M. Pizzuto, D. Corti, D. Veesler (2022)
Imprinted antibody responses against SARS-CoV-2 Omicron sublineages
bioRxiv
S. Bachas, G. Rakocevic, David Spencer, Anand Sastry, Robel Haile, John Sutton, George Kasun, A. Stachyra, Jahir Gutierrez, Edriss Yassine, Borka Medjo, Vincent Blay, Christa Kohnert, Jennifer Stanton, Alexander Brown, Nebojsa Tijanic, Cailen McCloskey, Rebecca Viazzo, Rebecca Consbruck, Hayley Carter, Simon Levine, S. Abdulhaqq, J. Shaul, A. Ventura, Randal Olson, Engin Yapici, J. Meier, S. McClain, Matthew Weinstock, G. Hannum, Ariel Schwartz, Miles Gander, R. Spreafico (2022)
Antibody optimization enabled by artificial intelligence predictions of binding affinity and naturalness
bioRxiv
A. Rathore, A. Sarker, R. Gupta (2018)
Recent Developments Toward Antibody Engineering and Affinity Maturation.
Protein and peptide letters, 25 10
J. Prassler, S. Steidl, S. Urlinger (2009)
In vitro affinity maturation of HuCAL antibodies: complementarity determining region exchange and RapMAT technology.
Immunotherapy, 1 4
Xiao-long Tian, Cheng Li, Ailing Huang, Shuai Xia, Sicong Lu, Z. Shí, Lu Lu, Shibo Jiang, Zhenlin Yang, Yanling Wu, T. Ying (2020)
Potent binding of 2019 novel coronavirus spike protein by a SARS coronavirus-specific human monoclonal antibody
Emerging Microbes & Infections, 9
(2022)
Peer Review File Manuscript Title: Imprinted SARS-CoV-2 humoral immunity induces convergent Omicron RBD evolution Reviewer Comments & Author Rebuttals

Publisher: Oxford University Press
Copyright: © The Author(s) 2023. Published by Oxford University Press on behalf of Antibody Therapeutics.
eISSN: 2516-4236
DOI: 10.1093/abt/tbad006
Publisher site: See Article on Publisher Site

Abstract

Antibody Therapeutics, 2023, Vol. 6, No. 2 108–118 https://doi.org/10.1093/abt/tbad006 Advance Access Publication on 13 April 2023 Research Article Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5 Kevin C. Entzminger, Jonathan K. Fleming, Paul D. Entzminger, Lisa Yuko Espinosa, Alex Samadi, Yuko Hiramoto, Shigeru C.J. Okumura and Toshiaki Maruyama Antibody Discovery, Abwiz Bio Inc., San Diego, CA 92121, USA Received: February 27, 2023; Revised: March 31, 2023; Accepted: April 5, 2023 ABSTRACT SARS-CoV-2 Omicron variant XBB.1.5 has shown extraordinary immune escape even for fully vaccinated individuals. There are currently no approved antibodies that neutralize this variant, and continued emergence of new variants puts immunocompromised and elderly patients at high risk. Rapid and cost-effective development of neutralizing antibodies is urgently needed. Starting with a single parent clone that neutralized the Wuhan-Hu-1 strain, antibody engineering was performed in iterative stages in real time as variants emerged using a proprietary technology called STage-Enhanced Maturation. An antibody panel that broadly neutralizes currently circulating Omicron variants was obtained by in vitro afﬁnity maturation using phage display. The engineered antibodies show potent neutralization of BQ.1.1, XBB.1.16, and XBB.1.5 by surrogate virus neutralization test and pM K afﬁnity for all variants. Our work not only details novel therapeutic candidates but also validates a unique general strategy to create broadly neutralizing antibodies to current and future SARS-CoV-2 variants. Statement of Signiﬁcance: We report engineering and characterization of SARS-CoV-2 therapeutic anti- bodies that neutralize all Omicron variants to date. Surrogate virus neutralization assay shows low ng/mL IC values and biolayer interferometry shows pM afﬁnity. The use of a proprietary platform technology, STage-Enhanced Maturation, is detailed. KEYWORDS: antibody engineering; neutralizing antibody; XBB.1.5; COVID-19; SARS-CoV-2 INTRODUCTION increasing the likelihood of breakthrough infections [1, 3]. At the time of writing, March 2023, XBB.1.5 and During the 3 years since the outbreak of coronavirus BQ.1.1 are dominant circulating strains, comprising 90% disease 2019 (COVID-19), the severe acute respiratory of all infections (cdc.gov). As a master of immune evasion, syndrome coronavirus 2 (SARS-CoV-2) virus has proven XBB.1.5 has been deemed the most transmissible variant exceptionally adept at mutating to evade the immune yet [4] and has the potential to dominate other variants response [1]. The emergence of the first Omicron variant worldwide. BA.1, having far more mutations in the spike receptor Antibody therapeutics are the standard of care for binding domain (RBD) compared to earlier variants, at-risk populations who are immunocompromised and dramatically increased the susceptibility of even previously thus susceptible to adverse COVID infection. Yet anti- vaccinated or infected individuals [2]. Later Omicron body therapeutics approved for treatment of COVID- sublineages BA.2 (from which XBB.1.5 is derived) and 19 have lost efficacy soon after new strains emerged. BA.4/5 (from which BQ.1.1 is derived) have continued BA.1 significantly reduced the potency of first-generation to evolve additional immune-evading mutations, further To whom correspondence should be addressed. Toshiaki Maruyama, 9823 Pacific Heights Blvd Ste J, Antibody Discovery, Abwiz Bio, San Diego, CA 92121, USA. Tel: 858-352-6911; Email: toshi@abwizbio.com © The Author(s) 2023. Published by Oxford University Press on behalf of Antibody Therapeutics. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Antibody Therapeutics, 2023 109 antibody therapeutics such as REGEN-COV (casirivimab– Spike trimers used in the study include Wuhan-Hu- imdevimab) [1]. BQ.1.1 resists neutralization by single or 1 (GenBank MN908947; Abwiz Bio Cat. #2720) as cocktail monoclonal antibody (mAb) therapies including well as variants Beta (Abwiz Bio Cat. #2652), Delta sotrovimab, bebtelovimab, bamlanivimab–etesevimab, and (Abwiz Bio Cat. #2611), BA.1 (Abwiz Bio Cat. #2672), evusheld (cilgavimab–tixagevimab) [5, 6]. XBB.1.5 has been BA.2 (Abwiz Bio Cat. #2460), BA.2.75 (Abwiz Bio Cat. shown to similarly evade neutralizing antibodies [3]. There #2664), BA.2.75.2 (Abwiz Bio Cat. #2676), BA.2.3.20 has been a call for novel, broadly active mAbs urgently (Abwiz Bio Cat. #2660), BN.1 (Abwiz Bio Cat. #2696), needed for prophylactic and/or therapeutic treatment XBB (Abwiz Bio Cat. #2700), XBB.1/XBB.1.9 (Abwiz in patients at high risk [5, 7, 8], especially given the Bio Cat. #2648), XBB.1.5/XBB.1.9.1 (Abwiz Bio Cat. danger associated with re-infection [9]. No currently U.S. #2712), XBB.1.16 (Abwiz Bio Cat. #2747), BA.4.6 (Abwiz Food & Drug Administration (FDA) approved antibody Bio Cat. #2692), BA.5 (Abwiz Bio Cat. #2688), BQ.1 therapeutics neutralize the latest Omicron variants BQ.1.1 (Abwiz Bio Cat. #2704), BQ.1.1 (Abwiz Bio Cat. #2668), and XBB.1.5. BF.7/BA.5.2.6/BF.11 (Abwiz Bio Cat. #2708), CH.1.1 Existing strategies to engineer neutralizing therapeutic (Abwiz Bio Cat. #2724), CA.3.1 (Abwiz Bio Cat. #2728), mAbs typically rely on antibody isolation from infected or and BR.2.1 (Abwiz Bio Cat. #2736). vaccinated individuals or from immunized humanized mice [10–12]. However, this process is poorly adapted for rapidly Immunization of rabbits evolving targets such as SARS-CoV-2, where the lead discovery phase must be repeated each time a new variant Three New Zealand White rabbits were immunized with emerges. Instead, we devised an approach that evolves anti- a recombinant protein including the receptor-binding body neutralization breadth in real time as the virus itself domain (RBD) of SARS-CoV-2 (Wuhan-Hu-1 strain). The evolves. Starting from a humanized lead candidate that immune sera were tested by ELISA using spike trimer, and showed strong neutralization of the Wuhan-Hu-1 strain, we the rabbit that showed the best titer was selected for library created six separate complementarity determining region construction. (CDR)-targeted libraries and selected the libraries on Wuhan-Hu-1 spike trimer to create diverse CDR pools. Library construction These CDRs were then randomly paired and iteratively selected on SARS-CoV-2 variants as they emerged, result- The bone marrow and spleen cells were collected and ing in a final panel of seven clones that strongly neutralize homogenized in TRI reagent (Molecular Research Center). all Omicron variants including XBB.1.5 and BQ.1.1. This Total ribonucleic acid (RNA) was isolated from the study not only details potential high-value therapeutic homogenate according to the manufacturer’s protocol. mAbs but also validates a general strategy to elicit broadly Messenger RNA was purified using messenger ribonucleic neutralizing mAbs for SARS-CoV-2 or other viruses. acid (mRNA) purification kit (Macherey-Nagel) according to the manufacturer’s protocol. First strand complemen- tary deoxyribonucleic acid (cDNA) was synthesized using PowerScribe MMLV Reverse Transcriptase (Monserate MATERIALS AND METHODS Biotechnology Group). cDNA was engineered using a method described in the US patent 9,890,414 and amplified Preparation of recombinant proteins with a single non-gene-specific primer. Amplified products were purified, digested, and sequentially ligated into a Spike trimers were prepared in-house for use in all proprietary rabbit Fab phagemid vector. assays. SARS-CoV-2 variants were made by overlap PCR and cloned to pCAGGS vector for production. Spike trimers include polybasic cleavage site mutation/deletion Phage panning and solubilizing mutations K986P/V987P as well as a trimerization motif and polyhistidine tag [13]. Following Library DNA was transformed into XL1-Blue cells large-scale purification, 300 μg of deoxyribonucleic acid (Agilent) for overnight growth and phage production (DNA) was transfected to HEK293T cells, with media following addition of M13KO7 Helper Phage (New Eng- harvested by centrifugation and filtration ∼7 days post land Biolabs). Phage were precipitated the following day transfection. His-tagged spike trimer was purified using from culture supernatant by addition of 4% PEG-8000/3% a Nickel Nitriloacetic acid (Ni-NTA) column: the spike NaCl on ice, followed by pelleting and resuspension in 1% trimer was washed with 50 mM imidazole then eluted with BSA/Dulbecco’s phosphate-buffered saline (DPBS) (block- 100 mM and 250 mM imidazole, followed by overnight ing buffer). High-bind microtiter wells (Immulon 4HBX) dialysis in DPBS. Quality and purity were assessed by were coated with recombinant spike trimer overnight, SDS-PAGE (sodium dodecyl sulfate–polyacrylamide gel washed with PBS, and blocked with blocking buffer for ◦ ◦ electrophoresis) gel and activity was assessed by testing for 1hat 37 C. Phage were added for incubation at 37 Cfor binding to angiotensin converting enzyme 2-human IgG 1 h, followed by washing with PBS, elution with 0.1 M Fc fusion (ACE2-Fc) by ELISA. In-house biotinylated HCl, and neutralization with 2 M Tris. Neutralized phage trimers were prepared for biolayer interferometry (BLI) were used to infect ER2738 (New England Biolabs) cells studies using EX-Link™ NHS-LC-Biotin (ThermoFisher for overnight propagation and precipitation as described A39257) followed by purification with 7 K MWCO above. Panning was performed for four or more rounds, Zeba™ Spin Desalting Columns (ThermoFisher 89882). with increasing washing stringency in later rounds to 110 Antibody Therapeutics, 2023 enrich specific binders. During phage panning of STage- ImmunoResearch 300-035-240). All assays were performed Enhanced Maturation (STEM) libraries, [1] the library was in triplicate. Raw ELISA values were normalized to six first transiently heated to 65 C then cooled to improve replicate positive wells omitting IgG and to three negative thermoresistance as described previously [14], followed by wells omitting both IgG and spike trimer, with negative [2] subtraction of the library on baculovirus particle (BVP) values forced to zero. Graphpad Prism was used to prepare coated microtiter wells to remove polyreactive clones before graphs and to derive inhibition of spike trimer binding to selection on spike protein trimer. ACE2 half maximal inhibitory concentration (IC ) values by fitting to a four-parameter logistic curve. Bebtelovimab, tixagevimab, cilgavimab, and S2X324 were purchased from Fab screening Proteogenix. Following the final round of phage panning, single colonies were prepared from the eluted page. Colonies were picked STage-Enhanced Maturation antibody engineering to 96-well culture plates with 1 mL of Super Broth medium and 50 μg/mL of carbenicillin. The culture In Stage 1, Six single CDR mutant libraries were designed plates were incubated at 37 C and 1 mM Isopropyl ß- based on amino acid usage by structurally characterized D-1-thiogalactopyranoside (IPTG) was added to induce human antibodies sharing similar canonical CDR struc- overnight soluble Fab production at 30 C. For Fab ELISA, tures among a subset that possessed identical germline microtiter wells were coated with 100 μL of spike trimer genes. This design strategy further humanizes the amino at 1–2 μg/mL in PBS. Following washing and blocking, acid usage in the CDRs and helps to avoid unwanted culture supernatants containing soluble Fab were added to mutations that could negatively affect the structure and the wells for incubation at 37 C for 1 h. After washing, folding of the antibodies while constraining library size to bound Fab was detected with peroxidase conjugated goat a diversity that could be sampled using our phage libraries. anti-rabbit IgG F(ab’) (Thermo Fisher Scientific) or Mutagenic oligos were synthesized and used to construct anti-human IgG F(ab’) secondary antibodies (Jackson separate CDR libraries by fragment amplification and over- ImmunoResearch). Positive clones were Polymerase Chain lap PCR using hN2Y gene as a template. Following diges- Reaction (PCR) amplified and sequenced by Sanger tion, ligation was performed at large scale using 10 μg sequencing (Eton Biosciences). digested vector DNA. These single CDR mutant libraries were panned on Wuhan-Hu-1 spike trimer to remove non- productive clones and to enrich all possible CDR mutants IgG production and humanization of rabbit antibody that retain binding to the spike trimer. Phage panning and Light and heavy chains of rabbit clone C-A11 were PCR Fab screening was performed as described above. In Stage 2, round four output phage from Stage 1 were amplified, digested, and cloned into a bi-cistronic rabbit used to amplify targeted CDR regions and combined by IgG vector for expression. All antibodies and recombinant overlap PCR into separate light chain (Lx; paired with the ACE2-Fc (Abwiz Bio Cat. #2566) were produced by tran- wild type heavy chain) or heavy chain (Hx; paired with the sient transfection of HEK293T cells with media harvested wild type light chain) libraries. The Lx library was panned by centrifugation and filtration ∼7 days post transfection. on the Beta spike trimer, and the Hx library was panned IgGs were purified by protein A column (Cytiva) followed separately on Wuhan-Hu-1, Alpha, Beta, and Epsilon spike by overnight dialysis in DPBS. CDRs of C-A11 rabbit antibody were grafted into selected human germline genes trimers. Phage panning and Fab screening was performed (IGHV3-23∗04 and IGKV-39∗01) to create a humanized as described above. LxC1-G10 clone was identified from the clone hN2Y. An expanded structural definition of CDRs Lx library and was cloned to IgG format and characterized was used, based on observed diversity from an internal as described above. database of rabbit antibodies. Following gene synthesis, In Stage 3, round four output phage from Stage 2 hN2Y light and heavy chain genes were cloned into a were used to amplify light and heavy chain libraries for bi-cistronic human IgG vector for production as above. sequential cloning and pairing to create a single combined S728-1157 light and heavy chain genes were synthesized by library. This library was panned for six rounds on BA.5 Twist Bioscience and cloned into our IgG expression vec- followed by two rounds on BA.2.75 and one round on tor, CR3022 expression plasmids were obtained from BEI either BQ.1.1 or XBB spike trimers. Fabs were screened Resources, NIAID, NIH for expression and purification as similarly as described above and tested for binding to the described above. respective spike trimer used for selection. Phage panning and Fab screening was performed as described above. Clones possessing 6R8 or 6R9 prefixes were identified Surrogate virus neutralization test of purified IgG from this screening and were converted to IgG format for Surrogate virus neutralization tests (sVNTs) were per- characterization as described above. formed according to the protocol described previously [15] with some modifications. Microtiter wells were coated with Biolayer interferometry 100 μL of ACE2-Fc (Abwiz Cat. #2566) at 2 μg/mL in PBS and blocked with 200 μL of 1% BSA/PBS. Purified Biotinylated spike trimers were prepared in-house as IgG were mixed 1:1 with a recombinant spike trimer at RT described above except BA.1 (ACROBiosystems SPN- and added to the wells. The bound spike trimer was detected C82Ee). BLI was performed using Octet Red96e with with peroxidase conjugated anti-his tag antibody (Jackson data collection at 30 C and sample shaking at 1000 rpm Antibody Therapeutics, 2023 111 Figure 1. STEM antibody engineering strategy. (A) Rabbit immunization with Wuhan-Hu-1 SARS-CoV-2 RBD was paired with phage display to identify a potent neutralizing clone; this mAb was humanized to create the lead candidate hN2Y. Single CDR libraries were constructed and selected on Wuhan- Hu-1 spike trimer to create a pool of functional CDR sequences. hN2Y predicted structure is shown, with all CDRs colored blue and targeted CDR(s) for a given library colored red. (B) Pre-selected CDR pools were amplified from phage and combined by overlap PCR to create separate combined light and combined heavy chain libraries. The light chain library was further selected on the Beta variant, resulting in the engineered clone LxC1-G10, which showed potent neutralization of BA.1 and was thoroughly characterized by the CoVIC at La Jolla Institute of Immunology [25]. The heavy chain library was selected on a combination of Alpha, Beta, Gamma, and Epsilon variants to broaden neutralization activity. (C) Pre-selected light and heavy chain libraries were then paired to create a single combined library. This library was selected on late-stage Omicron variants including BA.2.75.2, BA.5, BQ.1.1, and XBB to create broadly neutralizing 6R8/6R9 series antibodies. during all steps. All IgGs and spike trimers were prepared in Wuhan-Hu-1 RBD containing protein, and phage libraries running buffer containing 2% BSA, 0.002% Tween-20, PBS, were selected on Wuhan-Hu-1 spike protein trimer. Fab pH 7.4. Following 4 min baseline, 3 μg/mL of biotinylated neutralization assay was used to identify the potent trimer was captured to streptavidin sensors (Sartorius 18- clone C-A11. We then grafted CDRs onto a human 5019) for 4 min followed by an additional 4 min baseline. framework to create the humanized clone hN2Y, which IgG binding was monitored for 4 min followed by 30 min strongly neutralized the Wuhan-Hu-1 strain. As evasive dissociation in fresh buffer only wells. Each IgG was tested SARS-CoV-2 variants emerged, many groups attempted in 1:2 titration series from 10 to 0.16 nM. A reference to identify novel antibodies from vaccinated or infected sensor omitting IgG was included in each assay and used patient samples [16, 17]. We instead employed our STEM for subtraction. All data were globally fit to 1:1 binding platform to broaden neutralization potency of hN2Y to model to derive kinetic and affinity values, only omitting SARS-CoV-2 variants as they emerged in real time (Fig. 1). individual traces or times due to sensor error or low signal. Single CDR mutant libraries of hN2Y were first created For the in-tandem epitope binning assay, following the (Fig. 1A). CDR libraries were designed based on observed same baseline and antigen loading steps described above, amino acid frequencies for characterized human antibodies an initial 10 min 10 nM IgG-binding step was performed to possessing a similar germline and predicted CDR canonical allow for signal saturation, followed by immediate transfer structure usage to both further humanize CDRs and to to wells containing 10 nM IgGs to monitor binding for avoid introduction of sequence liabilities. By using phage an additional 10 min. The final nm shift values after the display, each CDR library can be sampled at sizes up to second IgG incubation step were used to prepare a heatmap 2E + 10. These separate CDR libraries were first selected of results. on Wuhan-Hu-1 spike trimer to eliminate non-binders and to isolate a large pool of diverse clones. Then, CDR pools from the selected phage populations were amplified and randomly paired by overlap PCR to create secondary RESULTS combined light chain and combined heavy chain libraries Therapeutic mAbs were engineered by STEM in real time as (Fig. 1B). To increase neutralization breadth, the combined SARS-CoV-2 variants emerged light chain library was selected on the Beta variant, and the Soon after the onset of the COVID-19 pandemic, we combined heavy chain library was selected separately on used our rabbit discovery platform, which pairs rabbit Alpha, Beta, Gamma, and Epsilon variants. We identified immunization with Fab-phage display, to create a thera- an exceptional candidate from the combined light chain peutic lead candidate mAb. Rabbits were immunized with library, LxC1-G10, which showed good neutralization 112 Antibody Therapeutics, 2023 Figure 2. Broad SARS-CoV-2 neutralization for engineered mAbs by sVNT. (A) Purified mAbs were tested by sVNT [15]. Wells were coated with recombinant ACE2-Fc, and His-tagged spike trimer protein was mixed with serially diluted mAb prior to addition. Bound spike trimer was detected by anti-His tag antibody. All experiments were performed in triplicates. 6R8/6R9 clones (blue) showed highly potent and broad neutralization of all Omicron variants tested including XBB.1.5 and BQ.1.1. LxC1-G10 (purple) light chain engineered clone showed improved binding compared to lead humanized candidate hN2Y (green) or parental rabbit clone C-A11 (red), though it lost potency for later Omicron variants. Bebtelovimab (orange) showed no neutralization of later Omicron variants. (B) IC values for all variants, with non-neutralizing clones reported at 1000 ng/mL. (C) RBD residues modified in the different SARS-CoV-2 variants compared to the Wuhan-Hu-1 strain. Residues are colored by mutation. Colored side bars for each mAb note variants that are neutralized with IC <100 ng/mL. of BA.1 and BA.2. After the emergence of later BA.2 to recombinant ACE2 protein by ELISA, shows strong lineage Omicron variants, we again used our platform correlation with results from pseudovirus or live virus to further improve potency. Light and heavy chains were neutralization tests [15]. We thus used high-throughput separately amplified from the pre-selected phage pools sVNT to quantify neutralization potency of mAbs across and further combined to create a single library (Fig. 1C). SARS-CoV-2 variants. In this assay, recombinant human The library was selected on BA.5, BA.2.75.2, BQ.1.1, and Fcγ -tagged ACE2 receptor is immobilized while IgGs XBB variants, all of which possess mutation at F486. A are pre-incubated with spike trimer before addition to diverse set of seven highly potent, broadly neutralizing the blocked ACE2 coated wells. ACE2-bound spike IgGs (6R8/6R9 clones) was identified following eight or trimer is detected by affinity tag using ELISA, and IC nine rounds of panning and is characterized below. values are derived. mAbs tested include the original neutralizing rabbit clone C-A11, the parent humanized In vitro neutralization assay shows broadly neutralizing clone hN2Y, LxC1-G10 possessing mutations across light mAbs chain CDRs, 6R8/6R9 clones possessing mutations across Neutralization titers for sera or mAbs measured using both heavy and light chain CDRs, and therapeutic mAb sVNT, which assays blocking of recombinant spike trimer bebtelovimab. Antibody Therapeutics, 2023 113 Notably, humanization of the rabbit antibody did not significantly reduce efficacy, as hN2Y retained similar neu- BQ.1.1 tralization potency for Wuhan-Hu-1 and Delta variants compared to C-A11 (Fig. 2A and B and Table 1). However, both IgGs showed significantly reduced neutralization of BQ.1 the original Omicron variant BA.1. LxC1-G10 contains mutations within all three light chain CDRs compared to BA.4.6 hN2Y. Neutralization of early Omicron variants by LxC1- G10 was significantly improved to low ng/mL IC values for BA.1, BA.2, BA.2.3.20, BA2.75, and BN.1. However, all three mAbs lost efficacy against variant BA.2.75.2, which BF.7/BA.5.2.6/BF.11 is notably different compared to these others due to F486S (Fig. 2C). LxC1-G10 showed effective neutralization of BA.2.12.1 lacking F486 mutation but weak neutralization BA.5 of BA.4/5 (possessing F486V) in a pseudovirus neutraliza- tion test by CoVIC at La Jolla Institute [18] in concordance XBB.1.16 with our observation in sVNT. To recover efficacy, as described above, we engineered a panel of seven mAbs (6R8A-A2, 6R8A-E1, 6R8A-E11, XBB.1/XBB.1.9 6R8C-D10, 6R8C-G8, 6R9A-A2, and 6R9E-A8), which show strong neutralization of all Omicron variants to date including BA.1, BA.2, BA.2.3.20, BA.2.75, BN.1, XBB.1.5/XBB.1.9.1 BA.2.75.2, CH.1.1, CA.3.1, BR.2.1, XBF, XBB, XBB.1 (XBB.1.9), XBB.1.5 (XBB.1.9.1), XBB.1.16, BA.5, BF.7 XBB (BA.5.2.6/BF.11), BA.4.6, BQ.1, and BQ.1.1. To our knowledge, these are the most broadly neutralizing ACE2- blocking antibodies characterized to date against the latest XBF Omicron variants of SARS-CoV-2. Compared to both hN2Y and LxC1-G10, these IgG clones possess mutations across CDRs L1, L2, L3, H1, and H3. BR.2.1 Bebtelovimab is a therapeutic SARS-CoV-2 antibody developed by Eli Lilly and first given Emergency Use CA.3.1 Authorization by the US FDA in February 2022, though the agency revoked authorization in late November 2022 due to an inability to neutralize variants BQ.1 and BA.2.75.2 BQ.1.1. Bebtelovimab showed no efficacy against these variants by sVNT or against CH.1.1, CA.3.1, XBB, XBB.1, XBB.1.16, or XBB.1.5 (Fig. 2A and B), matching CH.1.1 expected results. S728-1157 clone was developed as a broadly neutralizing antibody against Delta and BA.1 [19]. S728-1157 additionally neutralized BA.2, BN.1, and BN.1 BA.5 by sVNT but failed to neutralize BA.2.75.2, CH.1.1, CA.3.1, BR.2.1, XBF, XBB, XBB.1.5, XBB.1.16, BQ.1, BA.2.75 or BQ.1.1 (Fig. S1). Another previously characterized, broadly neutralizing antibody S2X324, identified from human plasma [17, 20], showed strong neutralization of BA.2.3.20 Wuhan-Hu-1, Delta, BA.2, BA.2.75, BR.2.1, XBF, and BA.5 in our sVNT but failed to neutralize CH.1.1, CA.3.1, XBB.1.5, XBB.1.16, BQ.1, or BQ.1.1 (Fig. S2). In contrast, BA.2 our engineered IgGs neutralized all SARS-CoV-2 variants tested and all currently circulating globally dominant BA.1 variants to date. Binding affinity correlates with neutralization potency Delta BLI was performed by capturing biotinylated spike trimer to a streptavidin-coated sensor surface to monitor bind- Wuhan-Hu-1 ing of serially diluted mAbs, with traces globally fit to 1:1 binding models to derive kinetic and affinity values. hN2Y showed slightly reduced affinity for the Beta variant IC50 (ng/mL) compared to C-A11 (Fig. S3), but affinity was increased to Table 1. IC values derived from sVNT data. Calculated IC values from sVNT assay are provided including 95% confidence intervals. 6R8/6R9 clones show low ng/mL IC 50 50 50 values for all SARS-CoV-2 variants tested. NA, no significant neutralization activity (>1000 ng/mL); UD, undefined C-A11 10.1 11.6 NA 67.1 23.4 10.2 10.5 NA NA NA NA NA NA NA NA NA NA NA NA NA NA hN2Y 19.4 19.8 211 91.9 63.6 38.5 32.8 NA NA NA NA NA NA NA NA NA NA NA NA NA NA LxC1-G10 7.0 18.3 5.7 4.6 9.1 14.4 7.1 NA NA NA NA NA NA NA NA NA 636 NA 309.7 NA NA 6R8A-A2 11.7 12.7 6.2 5.5 11.5 9.2 6.7 31.8 16.5 15.4 18.7 31.8 17.1 19.0 20.9 37.4 10.1 13.1 12.4 15.2 10.3 6R8A-E1 11.9 8.9 5.4 5.2 11.5 8.9 6.9 26.6 14.9 13.3 12.6 18.8 20.3 22.8 20.6 33.1 10.9 14.4 12.6 13.2 9.3 6R8A-E11 11.4 11.5 5.7 6.2 12.5 9.8 7.6 33.1 20.0 17.9 15.4 27.5 29.0 21.2 28.5 39.3 10.9 17.8 13.3 14.3 15.4 6R8C-D10 12.7 11.6 5.8 5.0 12.3 8.1 6.6 25.1 13.3 19.5 19.1 16.3 11.7 14.9 16.9 20.2 10.2 14.1 12.7 11.0 12.0 6R8C-G8 12.0 9.7 6.0 5.6 11.6 8.8 7.2 34.9 14.0 15.8 15.9 26.1 19.0 15.7 21.7 39.3 11.5 16.2 11.2 12.5 11.1 6R9A-A2 9.8 10.5 6.7 6.2 8.5 10.5 7.3 33.5 8.7 24.9 20.9 33.3 10.1 14.7 17.8 39.7 9.4 10.0 8.6 9.3 14.1 6R9E-A8 13.1 12.6 5.6 5.6 13.5 8.9 5.6 18.8 13.9 17.9 11.8 12.8 19.0 15.9 19.6 28.5 9.8 16.5 12.9 12.5 7.4 bebtelovimab 8.1 9.5 55.9 7.2 407 171 48.4 NA 38.6 NA 149 79.5 NA NA NA NA 5.3 9.9 8.8 NA NA 114 Antibody Therapeutics, 2023 Figure 3. BLI shows high affinity to all SARS-CoV-2 variants for engineered mAbs. In each assay, biotinylated spike protein trimer was captured to the streptavidin-coated sensor surface and dipped into a 7-fold dilution series of IgG at a starting concentration of 10 nM. Binding was monitored for 4 min, followed by dissociation monitored for 30 min. Traces were referenced to a sensor omitting IgG. (A) For each variant tested, each experiment is shown with the same scale y-axis. C-A11, hN2Y, and LxC1-G10 lost binding activity for BQ.1.1 and XBB.1.5, but all seven 6R8/6R9 clones showed strong binding. Therapeutic mAbs tested under the same conditions showed weak or no binding. (B) BLI traces were globally fit to a 1:1 binding model to derive K 6 −1 −1 −6 −1 affinity values (pM), (C) k , on rates (× 10 M ∗s ), and (D) k , off rates (×10 s ). Some mAb affinities were below the detection limit of the assay (<1 pM) due to exceptionally slow off rates even when measured for 30 min. For some weakly binding clones, affinity, and/or kinetic values could not be determined. N.B., non-binding; n.d., not derived. Antibody Therapeutics, 2023 115 low pM binding K for LxC1-G10. LxC1-G10 possessed 24 pM K affinity for BA.1, 5-fold improved over hN2Y. However, LxC1-G10 binding affinity was itself decreased nearly 100-fold for BQ.1.1 (2.1 nM) compared to BA.1 (24 pM) (Fig. 3A and B). The seven 6R8/6R9 mAbs possessed 7 pM or lower K for BQ.1.1 and below 70 pM K for XBB.1.5. Binding affinity for these mAbs was increased nearly 10 000-fold (for BQ.1.1) or above undetectable levels (for XBB.1.5) compared to parent clone hN2Y. While on-rates were simi- lar (Fig. 3C), the difference in binding affinity for BQ.1.1 is primarily due to a nearly 1000-fold reduction in off- −3 −1 rate, with hN2Y possessing 10 s off-rate while 6R8/6R9 −6 −1 clones showed 10 s off rates (Fig. 3D). 6R8C-G8 and 6R9E-A8 were also tested on XBF, CH.1.1, and CA.3.1 variants; these mAbs showed strong binding to all three, though weaker relative binding to CA.3.1. Bebtelovimab showed weak binding to BQ.1.1 by BLI and no binding at all to XBB.1.5. Evusheld is an antibody cocktail of Tixagevimab and Cilgavimab first given Emer- Figure 4. BLI epitope binning shows no change in binding site for gency Use Authorization by the US FDA in December engineered mAbs. In an in-tandem assay, 10 nM of primary mAb was first saturated on the spike protein trimer for 10 min, followed by dipping into 2021, though the agency revoked authorization January fresh 10 nM solution of secondary mAb. The nm shift values obtained 2023 due to the inability to neutralize the latest Omi- at the end of the secondary mAb incubation are reported and colored cron variants. Both tixagevimab and cilgavimab showed by heatmap. hN2Y, LxC1-G10, 6R8C-G8, and 6R9E-A8 all binned into no binding to BQ.1.1 or to XBB.1.5 (Fig. 3A and B). In the same cluster. CR3022 did not cross-compete with any other mAb as contrast, our engineered IgGs show low pM binding K for expected, while Bebtelovimab and Cilgavimab clustered together. these dominant SARS-CoV-2 Omicron variants currently circulating in the US. engineered a single IgG, originally developed for neutral- ization of the Wuhan-Hu-1 strain. Engineered mAbs recognize the same epitope as the parent Our humanized clone hN2Y was derived from a rabbit antibody hN2Y lead candidate. Rabbits are an excellent source of potent antibodies due to the exceptionally high diversity of their We then tested whether broader neutralization potency immune repertoire by use of somatic gene conversion and was achieved by significantly altering the epitope. BLI was somatic hypermutation [22]. C-A11 rabbit mAb possesses used for in-tandem epitope binning. In total, 10 nM of a 19 amino acid CDR H3 and 9 amino acid CDR L3. IgG was first saturated on the sensor for 600 s, followed Long CDR H3 loops, typically defined as ≥24 amino acids, by dipping the sensor into new solutions of 10 nM IgGs. are a common feature of broadly neutralizing antiviral The results are presented as a heatmap using the final nm mAbs [23]. Though not technically classified as ‘long,’ the shift values obtained after incubation with the second IgG longer CDR H3 may have contributed to C-A11’s potent (Fig. 4). hN2Y, LxC1-G10, 6R8C-G8, and 6R9E-A8 all neutralization of early variants through Delta. Despite high competed with each other, suggesting that they bind to CDR diversity, humanization of rabbit clones is relatively identical or overlapping epitopes. CR3022, a neutralizing straightforward due to their restricted use of a limited antibody that does not block ACE2 receptor binding [21], number of germline genes [24]. After humanization, hN2Y does not cross-compete with any antibody as expected. showed almost no change in affinity or neutralization Interestingly, both Bebtelovimab and Cilgavimab, which activity compared to C-A11 and showed even stronger block ACE2 receptor binding but do not bind BQ.1.1 or neutralization of some variants including BA.1 and XBB.1.5, were binned into a distinct cluster that does not BA.2. compete with our engineered mAbs. This demonstrates that Omicron variant BA.1 reduced the neutralization broader neutralization activity of the engineered IgGs was potency of hN2Y, as observed for many other therapeutic likely achieved by optimizing binding to conserved SARS- antibodies developed prior to December 2021. As detailed CoV-2 residues within the same epitope as the parent mAb, above, LxC1-G10 was selected by panning a combined light rather than by significantly changing the binding site. chain library on the Beta variant, which possesses RBD mutations K417N/E484K/N501Y (Fig. 2C). LxC1-G10 showed sub-pM affinity for the Beta variant, >100-fold DISCUSSION increased compared to hN2Y (Fig. 3B). This increased We have detailed the engineering and activity of broadly binding was sufficient to improve neutralization of early neutralizing mAbs that show potent efficacy against all Omicron subvariants BA.1, BA.2, BA.2.3.20, BA.2.75, and dominant Omicron circulating variants including XBB.1.5 BN.1, despite those variants possessing many additional and BQ.1.1. Uniquely, instead of identifying a novel clone mutations within the spike protein. LxC1-G10 shares an from infected or immunized samples, we rather iteratively identical heavy chain as hN2Y but possesses over 14 amino 116 Antibody Therapeutics, 2023 acid differences distributed across all light chain CDRs. is designed to sample the broadest sequence diversity Light chain CDR mutations alone were thus sufficient including inter-CDR cooperative mutations and directly to confer significant improvement in binding affinity and incorporates developability filters early during phage neutralization potency. panning. However, LxC1-G10 lost activity against all Omicron Comparing hN2Y to 6R8/6R9 mAbs, there are four variants possessing a mutation at F486. F486 mutations are to five amino acid differences in CDR L1, four in CDR found in all later Omicron variants and have been linked L2, four in CDR L3, six to nine in CDR H1, and two to immune evasion [1]. The further engineered 6R8/6R9 in CDR H3, totaling 21–24 amino acid mutations accu- series mAbs, selected from panning a combined light mulated during affinity maturation. This is a very large and heavy chain library on Omicron variants possessing number of changes which would be difficult to impossi- F486S (BA.2.75.2 and XBB) and F486V (BA.5 and ble to obtain by any other established affinity maturation BQ.1.1) mutations, showed exceptional neutralization method. Notably, none of the CDR sequences from the potency and binding for all SARS-CoV-2 variants tested, 6R8/6R9 clones were identified during screening of single including those with alternate amino acids at this site CDR libraries (panned on Wuhan-Hu-1) or separate com- such as F486I (BR.2.1) and F486P (XBF and XBB.1.5). bined CDR libraries (separately panned on Alpha, Beta, Compared to both hN2Y and LxC1-G10, 6R8/6R9 mAbs Gamma, Delta, or Epsilon), making it unlikely that a CDR possess two mutations in CDR H3 (which is identical walking [33] or parallel CDR optimization [26] approach among all 6R8/6R9 clones) and 6–9 mutations in CDR would be successful. Alternative methods combine satu- H1. Compared to LxC1-G10, 6R8/6R9 mAbs possess rated mutagenesis libraries by DNA shuffling [34], but the 11–13 mutations across all light chain CDRs. Extensive likelihood of these methods resulting in two or more tan- mutagenesis across all CDRs except H2 was thus required dem mutations within an individual CDR, as observed in to confer the increased neutralization activity and binding all 6R8/6R9 clones, is very low due to the improbability of affinity observed. recombination within such close proximity. Computational Both our early lead candidate hN2Y (CoVIC-359), and methods used for affinity maturation have shown promise, LxC1-G10 (CoVIC-362) were submitted to the CoVIC but the quality of prediction begins to fall and the complex- at La Jolla Institute of Immunology (https://covicdb. ity in modeling begins to increase dramatically beyond five lji.org/) for analysis and testing against over 350 other amino acid mutations, making it impossible to predict the final 6R8/6R9 sequences in silico [35]. Taken together, only therapeutic antibody candidates [18, 25]. CoVIC is an by applying our STEM platform, which samples excep- international effort to conduct side-by-side analyses of tionally large diversity across all CDRs, could these potent candidate antibody therapeutics targeting the SARS-CoV- neutralizing antibodies be created. The absence of competi- 2 spike protein in standardized assays. Both antibodies tor mAbs showing similar potency further underscores our showed nearly 100% protection against cell infection in approach. pseudovirus neutralization assay for Beta, Delta, BA.1, We have characterized an antibody panel that outper- BA.1.1, and BA.2 variants, and showed efficacy with live forms all existing FDA approved antibody therapeutics Wuhan-Hu-1 virus in cell-based neutralization assay and against all globally dominant circulating SARS-CoV-2 during in vivo mouse challenge [25]. Both were binned into strains to date. This study validates our unique STEM the same epitope class, RBD-2a, characterized by bivalent platform. Though our lead humanized clone showed no intra-spike binding, where both arms of the IgG bind to binding to or neutralization of XBB.1.5, we successfully two spike units within a single trimer. This binding motif recovered potency by engineering through repeated sam- was further validated by cryo-EM in the CoVIC study, pling and combination of CDRs progressively against in which LxC1-G10 was observed to bind to both spike SARS-CoV-2 variants. The ability to recover neutral- units in the ‘up’ conformation, a conformational state that ization activity from a single lead mAb demonstrates exposes the ACE2 receptor-binding site [25]. Our epitope the exceptional CDR sequence diversity that can be binning assay (Fig. 4) demonstrated that the engineered obtained by applying an iterative selection and library 6R8/6R9 mAbs likely bind in the same RBD-2a epitope creation strategy. By continuing to engineer a single class. The increased avidity conferred by bivalent binding mAb, lead time was significantly reduced and significant likely contributes to the strong neutralization activity and resources were saved by avoiding the need to repeat the high affinity for 6R8/6R9 mAbs. discovery process by immunization or screening. In vitro In traditional humanization, animal-derived CDRs are affinity maturation of a single neutralizing antibody has grafted into human framework germline genes. Then, the potential to be far more efficient and cost-effective further engineering is performed using various methods to compared to discovery campaigns reliant on in vivo affinity obtain the best lead candidate for therapeutic mAb devel- maturation by successive vaccination and/or infection. In opment [26–32]. These methods suffer from notable set- patient samples, broadly neutralizing antibodies must be backs including [1] the inability to easily sample mutations identified from a large background of non-neutralizing that work cooperatively within or across CDRs such as in anti-spike protein antibodies, while affinity maturation of saturation mutagenesis and CDR walking, [2] introduction a neutralizing antibody allows for fine tuning around a of unwanted framework mutations such as in error prone known neutralizing epitope. As in this study, filters can be PCR, [3] lack of secondary selection pressure when combin- included during phage panning, including heat treatment ing beneficial mutations such as in parallel CDR optimiza- to remove unstable clones [14] and negative subtraction on tion [26], and [4] almost all omit selection pressure for well- BVP to remove polyreactive clones, to ensure that the final behaved clones. In contrast, our in vitro STEM platform Antibody Therapeutics, 2023 117 resulting candidates retain good developability profiles. project, and K.C.E. and T.M. wrote the manuscript, with This approach can be adapted for future therapeutic mAb all authors providing comments and edits. development against SARS-CoV-2 or other viruses, even Kevin C. Entzminger (Conceptualization, Data curation, to allow for rapid development of broadly neutralizing Formal analysis, Investigation, Methodology, Project therapeutic mAbs in real time. administration, Supervision, Validation, Writing—original draft, Writing—review & editing [Equal]), Jonathan K. Fleming (Data curation, Formal analysis, Investigation, ACKNOWLEDGMENTS Methodology, Writing—review & editing [Equal]), Paul D. Entzminger (Data curation, Formal analysis, Investigation, We thank Drs Erica Ollmann Saphire and Sharon Schendel Methodology, Writing—review & editing [Equal]), Lisa at CoVIC of La Jolla Institute for testing and providing Yuko Espinosa (Data curation, Investigation [Equal]), us the data for our antibodies. CoVIC is an international Alex Samadi (Data curation, Investigation, Methodology, effort to conduct side-by-side analyses of candidate anti- Writing—review & editing [Equal]), Yuko Hiramoto body therapeutics targeting the SARS-CoV-2 spike protein (Data curation, Investigation, Writing—review & editing in standardized assays. [Equal]), C.J. Okumura (Data curation, Funding acquisi- tion, Project administration, Writing—review & editing [Equal]), and Toshiaki Maruyama (Conceptualization, SUPPLEMENTARY DATA Data curation, Formal analysis, Investigation, Method- Supplementary data are available at ABT online. ology, Project administration, Supervision, Writing— original draft, Writing—review & editing [Equal]) FUNDING Funding for this study was provided by Abwiz Bio Inc. ETHICS AND CONSENT STATEMENT The CoVIC was supported by the COVID-19 Therapeutics Accelerator (INV-006133), the Bill and Melinda Gates Not applicable. Foundation (OPP1210938), a supplement to National Institutes of Health/National Institute of Allergy and Infectious Diseases (grant U19 AI142790-S1), and the ANIMAL RESEARCH STATEMENT GHR Foundation. Rabbit immunization was performed by ProSci (San Diego, CA), following all Institutional Animal Care and Use Com- mittee guidelines. Care was taken throughout the study to CONFLICTS OF INTEREST STATEMENT prevent any undue animal suffering. All authors are employees of Abwiz Bio Inc. Abwiz Bio has filed patent applications relating to anti-SARS-CoV- 2 antibodies and STEM technology with K.C.E., C.O., and T.M. as inventors. Some mAbs in this study are being REFERENCES considered for development of therapeutic antibodies. 1. Cao, Y, Yisimayi, A, Jian, F et al. BA.2.12.1, BA.4 and BA.5 escape antibodies elicited by Omicron infection. Nature 2022; 608: 593–602. 2. Willett, BJ, Grove, J, MacLean, OA et al. SARS-CoV-2 Omicron is an immune escape variant with an altered cell entry pathway. Nat DATA AVAILABILITY Microbiol 2022; 7: 1161–79. The datasets used and/or analyzed during the current study 3. Qu, P, Faraone, JN, Evans, JP et al. Extraordinary evasion of are available from the corresponding author on reasonable neutralizing antibody response by Omicron XBB.1.5, CH.1.1 and request. CA.3.1 variants. 2023. https://doi.org/10.1101/2023.01.16.524244. 4. Callaway, E. Is coronavirus variant XBB.1.5 a global threat? Nature 2023; 613: 222–3. 5. Arora, P, Kempf, A, Nehlmeier, I et al. Omicron sublineage BQ.1.1 AUTHORS’ CONTRIBUTIONS resistance to monoclonal antibodies. Lancet Infect Dis 2022; 23: K.C.E. isolated the neutralizing clone C-A11 from the 22–3. 6. Imai,M,Ito,M,Kiso, M et al. Efficacy of antiviral agents against rabbit library, designed CDR mutant libraries, performed omicron subvariants BQ.1.1 and XBB. New Engl J Med 2023; 388: selection of clones that bind multiple variants of COVID- 89–91. 19, and performed BLI studies. T.M. designed the project, 7. Planas, D, Bruel, T, Staropoli, I et al. Resistance of omicron cloned RBD containing spike protein for immunization, subvariants BA.2.75.2, BA.4.6 and BQ.1.1 to neutralizing built the rabbit phage libraries, performed panning and antibodies. 2022. https://doi.org/10.1101/2022.11.17.516888. 8. Cao, Y, Jian, F, Wang, J et al. Imprinted SARS-CoV-2 humoral screening of the libraries, humanized the neutralizing immunity induces convergent Omicron RBD evolution. Nature 2022; clone C-A11, and performed sVNTs. J.K.F. produced 614: 521–9. https://doi.org/10.1038/s41586-022-05644-7. all spike trimers, ACE2 protein, and antibodies. P.D.E. 9. Bowe, B, Xie, Y, Al-Aly, Z. Acute and postacute sequelae associated performed sVNTs and analyzed the data. L.Y.E. performed with SARS-CoV-2 reinfection. Nat Med 2022; 28: 2398–24055. phage panning on Omicron variants. A.S. preformed 10. Brouwer, PJM, Caniels, TG, van der Straten, K et al. Potent neutralizing antibodies from COVID-19 patients define multiple manufacturability and biophysical risk assessment with targets of vulnerability. Science (1979) 2020, 1979; 369: 643–50. J.K.F., Y.H., and T.M. created spike trimer variant vectors, 11. Taylor, PC, Adams, AC, Hufford, MM et al. Neutralizing cloned antibodies into IgG expression vector, and prepared monoclonal antibodies for treatment of COVID-19. Nat Rev plasmid DNA for transfection. T.M. and C.O. oversaw the Immunol 2021; 21: 382–93. 118 Antibody Therapeutics, 2023 12. Morales-Núñez, JJ, Muñoz-Valle, JF, Torres-Hernández, PC et al. 24. Knight, KL. Restricted VH gene usage and generation of antibody Overview of neutralizing antibodies and their potential in diversity in rabbit. Annu Rev Immunol 1992; 10: 593–616. COVID-19. Vaccine 2021; 9: 1376. 25. Callaway, HM, Hastie, KM, Schendel, SL et al. Bivalent intra-spike 13. Keating, SM, Mizrahi, RA, Adams, MS et al. Generation of binding provides durability against emergent omicron lineages: recombinant hyperimmune globulins from diverse B-cell repertoires. results from a global consortium. Cell Rep 2023; 42: 112014. Nat Biotechnol 2021; 39: 989–99. 26. Barbas, CF III, Burton, DR. Selection and evolution of high-affinity 14. Entzminger, KC, Johnson, JL, Hyun, J et al. Increased fab human anti-viral antibodies. Trends Biotechnol 1991; 14:230–4. thermoresistance via VH-targeted directed evolution. Protein Eng 27. Simons, JF, Lim, YW, Carter, KP et al. Affinity maturation of Des Selection 2015; 28: 365–77. antibodies by combinatorial codon mutagenesis versus error-prone 15. Tan, CW, Chia, WN, Qin, X et al. A SARS-CoV-2 surrogate virus PCR. MAbs 2020; 12: e1803646. neutralization test based on antibody-mediated blockage of 28. Kennedy, PJ, Oliveira, C, Granja, PL et al. Monoclonal antibodies: ACE2–spike protein–protein interaction. Nat Biotechnol 2020; 38: technologies for early discovery and engineering. Crit Rev Biotechnol 1073–8. 2018; 38: 394–408. 16. Westendorf, K, Zentelis, S, Wang, L et al. LY-CoV1404 29. Rathore, AS, Sarker, A, Gupta, RD. Recent developments toward (bebtelovimab) potently neutralizes SARS-CoV-2 variants. Cell Rep antibody engineering and affinity maturation. Protein Pept Lett 2022; 39: 110812. 2018; 25: 886–96. 17. Cameroni, E, Bowen, JE, Rosen, LE et al. Broadly neutralizing 30. Tabasinezhad, M, Talebkhan, Y, Wenzel, W et al. Trends in antibodies overcome SARS-CoV-2 omicron antigenic shift. Nature therapeutic antibody affinity maturation: from in-vitro towards 2022; 602: 664–70. next-generation sequencing approaches. Immunol Lett 2019; 212: 18. Hastie, KM, Li, H, Bedinger, D et al. Defining variant-resistant 106–13. epitopes targeted by SARS-CoV-2 antibodies: a global consortium 31. Lim, CC, Choong, YS, Lim, TS. Cognizance of molecular methods study. Science 2021; 374: 472–8. for the generation of mutagenic phage display antibody libraries for 19. Changrob, S, Halfmann, PJ, Liu, H et al. Site of vulnerability on affinity maturation. Int J Mol Sci 2019; 20: 1861. SARS-CoV-2 spike induces broadly protective antibody to 32. Prassler, J, Steidl, S, Urlinger, S. In vitro affinity maturation of antigenically distinct omicron SARS-CoV-2 subvariants. 2022. HuCAL antibodies: complementarity determining region exchange https://doi.org/10.1101/2022.10.31.514592. and RapMAT technology. Immunotherapy 2009; 1: 571–83. 20. Park,Y,Pinto,D,Walls,AC et al. Imprinted antibody responses 33. Yang, W, Green, K, Pinz-Sweeney, S et al. CDR walking against SARS-CoV-2 Omicron sublineages. Science 2022; 378: mutagenesis for the affinity maturation of a potent human 619–27. anti-HIV-1 antibody into the picomolar range. J Mol Biol 1995; 254: 21. Tian, X, Li, C, Huang, A et al. Potent binding of 2019 novel 392–403. coronavirus spike protein by a SARS coronavirus-specific human 34. Banach, BB, Tripathi, P, Pereira, LDS et al. Highly protective monoclonal antibody. Emerg Microbes Infect 2020; 9: 382–5. antimalarial antibodies via precision library generation and yeast 22. Weber, J, Peng, H, Rader, C. From rabbit antibody repertoires to display screening. JExp Med 2022; 219: e20220323. rabbit monoclonal antibodies. Exp Mol Med 2017; 49: e305. 35. Bachas, S, Rakocevic, G, Spencer, D et al. Antibody optimization 23. Sok, D, Burton, DR. Recent progress in broadly neutralizing enabled by artificial intelligence predictions of binding affinity and antibodies to HIV. Nat Immunol 2018; 19: 1179–88. naturalness. 2022. https://doi.org/10.1101/2022.08.16.504181.

Journal

Antibody Therapeutics – Oxford University Press

Published: Apr 13, 2023

Keywords: antibody engineering; neutralizing antibody; XBB.1.5; COVID-19; SARS-CoV-2

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5

Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5

Rapid engineering of SARS-CoV-2 therapeutic antibodies to increase breadth of neutralization including BQ.1.1, CA.3.1, CH.1.1, XBB.1.16, and XBB.1.5

References (39)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies