key: cord-1006476-job2rehk
authors: Schlör, A.; Hirschberg, S.; Ben Amor, G.; Meister, T.L.; Arora, P.; Pöhlmann, S.; Hoffmann, M.; Pfaender, S.; Kamal Eddin, O.; Kamhieh-Milz, J.; Hanack, K.
title: SARS-CoV-2 neutralizing camelid heavy-chain-only antibodies as powerful tools for diagnostic and therapeutic applications
date: 2022-03-24
journal: bioRxiv
DOI: 10.1101/2022.03.24.485614
sha: bde9dd0f8cd601b4d357fa759e8ad115327696a6
doc_id: 1006476
cord_uid: job2rehk

Introduction The ongoing COVID-19 pandemic situation caused by SARS-CoV-2 and variants of concern such as B.1.617.2 (Delta) and recently, B.1.1.529 (Omicron) is posing multiple challenges to humanity. The rapid evolution of the virus requires adaptation of diagnostic and therapeutic applications. Objectives In this study, we describe camelid heavy-chain-only antibodies (hcAb) as useful for novel in vitro diagnostic assays and for therapeutic applications due to their neutralizing capacity. Methods Five antibody candidates were selected out of a naïve camelid library by phage display and expressed as full-length IgG2 antibodies. The antibodies were characterized by Western blot, enzyme-linked immunosorbent assays, surface plasmon resonance with regard to their specificity to the recombinant SARS-CoV-2 Spike protein and to SARS-CoV-2 virus-like particles. Neutralization assays were performed with authentic SARS-CoV-2 and pseudotyped viruses (wildtype and Omicron). Results All antibodies efficiently detect recombinant SARS-CoV-2 Spike protein and SARS-CoV-2 virus-like particles in different ELISA setups. The best combination was shown with hcAb B10 as catcher antibody and HRP-conjugated hcAb A7.2 as the detection antibody. Further, four out of five antibodies potently neutralized authentic wildtype SARS-CoV-2 and particles pseudotyped with the SARS-CoV-2 Spike proteins of the wildtype and Omicron variant, sublineage BA.1 at concentrations between 0.1 and 0.35 ng/mL (ND50). Conclusion Collectively, we report novel camelid hcAbs suitable for diagnostics and potential therapy. Graphical Abstract

Camelid heavy-chain-only antibodies (hcAbs) are described as a novel class of binding molecules with highly promising advantages for diagnostic and therapeutic applications.

Compared to mouse or human antibodies, hcAbs recognize their corresponding target only with one chain instead of a heavy and a light chain. The variable element of these heavy-chain-only antibodies is defined as camelid single domain antibody (VHH) or briefly as nanobody [1] [2] . Nanobodies are generated by phage display from naïve or immune libraries and represent the smallest recognition unit in the class of antibody molecules with a molecular size of 15 kDa. They are characterized by a high thermal stability, a very good tissue penetration, and the recognition of hidden or difficult-to-reach epitopes [3] . Therefore, these antibodies represent a new class of nextgeneration antibodies and are of high interest for biomedical applications. For therapeutic use it is favorable to use the 15 kDa fragment compared to full length antibodies to avoid antibody-dependent enhancement (ADE) which is often seen for human or humanized therapeutic antibodies [4] . Further, nanobodies can be expressed recombinantly in prokaryotic expression systems which is compared to the production of full length antibodies a cheap and easy to scale up alternative [5] . The first nanobodybased drug Caplacizumab was approved recently by the FDA and EMA [6] .

For in vitro diagnostic applications such as in ELISA systems, flow cytometry or immunofluorescence, nanobodies have to be expressed with a corresponding tag in order to provide a sufficient detection [7] . Beside different tags, it is advantageous to produce these nanobodies with a corresponding Fc part to create flexible opportunities for detection and to allow a system which is compatible with standard procedures such as biotin/streptavidin or enzyme-linked reactions.

There are already several nanobodies described which also neutralize SARS-CoV-2 [8] [9] [10] . The possibility to design multimeric formats was shown to significantly increase neutralizing effects. Nanobodies have entered the antibody field and gain more and more attention due to their unique properties. As camelid full length antibodies they represent a new opportunity for the creation of novel in vitro diagnostic assay formats.

Here, we describe the identification and recombinant production of camelid full length antibodies from a naïve camelid library that allow detection of SARS-CoV-2 with a high sensitivity and neutralize SARS-CoV-2 with a high potency.

A naïve camelid VHH library was generated as previously described by Schlör et. al [11] . Briefly, RNA was isolated from peripheral blood mononuclear cells (PBMC;

NucleoSpin RNA Mini Plus, Macherey-Nagel GmbH & Co.KG, Düren, Germany). For cDNA first strand synthesis RevertAid kit (Thermo Scientific) was used with a combination of random hexamer and oligo dT primers. Following a 2-step PCR protocol for VHH amplification with the primer combination P12/13 and NEM01-03_PhD. Primer sequences are listed in Table 1 . Amplified VHHs were cloned into phagemid vector pADL-22c (Antibody Design Labs). The library was transformed into E. coli XL1 blue (Agilent) by electroporation.

For biopanning, transformed XL1 blue were cultured and inoculated with M13KO7 helper phages. The resulting phages were enriched in five rounds of biopanning. Full length SARS-CoV-2 Spike protein (SPN-C52H4; AcroBiosystems) was used as antigen in decreasing amounts starting from 15 to 0.1 µg. Post-panning analyses were performed as described by Coomber [12] . Monoclonal phages were tested in phage enzyme-linked immunosorbent assays (ELISA) for antigen specific binding to recombinant SARS-CoV-2 Spike protein. Positive clones were sequenced and become heavy chain only antibodies by cloning into the expression plasmid pNEM_camAb.

For the generation of a full length heavy-chain-only antibody, the different VHHs were amplified with SL13/14 and cloned into the vector pNEM_camhAb. pNEM_camAb was designed in our lab as vector for eukaryotic protein expression with a camelid Fc fragment (CH2-CH3 domain) resulting in hcAbs.

HcAbs were expressed using the Expi293 expression system (Thermo Fisher Scientific). Culturing and transfection was performed according to the manufacturer's protocol. After 5 to 6 days after transfection, the cell culture supernatants were harvested and filtered, followed by protein A purification (ProSep® Ultra Plus; Merck).

Elution of hcAbs was achieved by a pH shift with glycine buffer as described previously [13] . Finally, the hcAbs were dialyzed against PBS (pH 7.4). The integrity and purity were analyzed using SDS PAGE, Western Blot and ELISA.

To investigate the different hcAb candidates in Western Blot, 1 µg SARS-CoV-2 Spike protein (antibodies-online, ABIN6952734) per lane was applied onto a 4-12% SDS polyacrylamide gradient gel. The gel was blotted using the NuPage system and and Bis-Tris buffer as transfer buffer. To block unspecific binding, Rotiblock was used as blocking agent and for the dilution of the first and the second antibody. Five different hcAb candidates were applied in a concentration of 1 µg/mL (diluted in Rotiblock) and 

HRP-conjugation of hcAbs was done using a modified version of the classical periodate method as published in 1985 [14] . For activation of HRP, 0.5 mg/mL HRP in 100 mM NaHCO3 (pH 8.1) were mixed 1:1 with 12.5 mM NaIO4 and incubated at RT for 2 h in the dark. In the next step, the same volume of activated HRP and hcAbs (1 mg/mL in NaHCO3; pH 9.2) were incubated in a glas-wool plugged Pasteur pipet filled with Sephadex G-25 (GE Healthcare) for 3 h at RT in the dark. The elution of conjugated hcAbs from Sephadex was performed with 100 mM NaHCO3 (pH 9.2). To stop the reaction, 1/20th volume of 5 mg/mL NaBH4 was added and incubated at 4 °C. After 30 min another volume (1/10) of freshly prepared NaBH4 solution was added and incubated at 4 °C for 1 h. Finally, after an ammonium sulfate precipitation overnight HRP conjugated hcAbs were dissolved in PBS (pH 7.4).

The hcAb-binding properties of B10 were analyzed at 25 °C on a Biacore T200 instrument (GE Healthcare) using 10 mM HEPES pH 7.4, 300 mM NaCl, 3 mM EDTA, 0.05% Tween 20, 0.25 mg/ml BSA as running buffer. SARS-CoV-2 Spike protein-RBD-mFc (Acrobiosystem; SPD-C5259) was captured by a covalently immobilized antimouse IgG on a C1 sensor chip. Increasing concentrations of hcAb B10 (0.23 -60 nM) were injected. Analyte responses were corrected for unspecific binding and buffer responses. The sensorgrams was fitted with a 1:1 binding model using Biacore evaluation software.

To determine the neutralization capacity of hcAbs against SARS-CoV-2, a neutralization assay with a propagation-incompetent VSV*ΔG pseudovirus system or full length virus was performed as previously described [15] . The expression vector for the Omicron spike (based on isolate hCoV19/Botswana/R40B58_BHP_3321001245/2021; GISAID Accession ID: EPI_ISL_6640919) was generated by Gibson assembly using five overlapping DNA strings (Thermo Fisher Scientific, sequences available upon request), The cells were incubated for 72 h before they were stained with crystal violet to visualize cytopathic effects.

Human codon optimized sequences of genes encoding the S and E structural proteins 

Camelid single domain antibodies were selected by phage display out of a naïve camelid library as previously described [11] . We could identify around 15 monophages with a specific ELISA signal of more than 0.5 when recombinant full length SARS-CoV-2

Spike protein was used for coating ( Fig. 1) . After sequencing, we chose 5 different candidates -A7.2, B10, D3, D12, G10 -for recombinant expression of camelid full length IgG2 antibodies. Antibodies were purified by protein A chromatography [12] and characterized by Western blot analyses for their specific recognition ( Fig. 2A) . All hcAbs specifically recognize the recombinant Spike protein in Western blot, whereas A7.2 and B10 showed a more intense signal than the other candidates.

The purified hcAbs were further characterized in different ELISA setups, to verify the optimal combinations and concentrations. First, the antibodies were serially diluted on recombinant SARS-CoV-2 Spike protein (Fig. 2B) where they showed a specific recognition till a working concentration of 0.5 µg/mL. All hcAbs showed a similar performance in this assay format.

In the next step, we tested pairwise combinations to find the optimal catcher and detector antibodies for a sandwich immunoassay principle (Fig. 3) . To ensure a direct detection, we conjugated hcAb B10 and A7.2 to HRP and tested their performance in a direct ELISA (Fig. 3A) . A7.2 showed a more succinct labeling with HRP and a better performance in the assay compared to B10. Therefore, we used HRP-conjugated hcAb A7.2 in a dilution of 1:1600 as detection antibody. Figure 3B shows the results for the sandwich immunoassay. We coated all hcAb candidates in different concentrations of 10, 5 and 2.5 µg/mL on the solid phase. Recombinant SARS-CoV-2 Spike protein was used as antigen and HRP-conjugated hcAb A7.2 as detection antibody. The usage of B10 as catcher hcAb and HRP-conjugated hcAb A7.2 as detector seemed to be the most sensitive combination in this sandwich immunoassay format (Fig. 3B, 2nd ).

In Figure 4 we have summarized the results for our generated SARS-CoV-2 VLPs.We could prove that the generated VLPs have the expected particle size (Fig. 4A ) and carry the Spike (S) as well as the nucleoprotein (N) which were detectable with human reconvalescent patient sera in Western blot experiments (Fig. 4B) . The presence of E and M protein was not specifically tested because their presence is necessary for the release of VLPs from the host cell [17] [18] . The VLPs were used as antigen coated on the solid phase and detected by either hcAb B10 or A7.2 followed by a HRP-conjugated secondary antibody (Fig. 4C) . The dilution series revealed a working concentration of 1.25 µg/mL for hcAb B10 and 0.6 µg/mL for hcAb A7.2.

Next, we characterized the hcAb candidate B10 in surface plasmon resonance spectrometry (SPR). SPR measurements were performed with immobilized SARS-CoV-2 Spike protein (RBD domain) recombinantly expressed with a murine Fc part. For hcAb B10 highly reproducible kinetic data could be received, revealing a KD value of 0.39 nM (Fig. 5) .

Neutralization assays were performed with full length SARS-CoV-2 wildtype and SARS-CoV-2 pseudotyped viruses for the wildtype, and the Omicron variant, respectively. hcAbs showed a similar capacity as shown for the full length wildtype virus. Surprisingly, we found that also the new variant of concern B1.1.529 (Omicron) was neutralized by the hcAbs B10, D3 and D12 in a range of 0.1 -0.35 ng/mL (ND50). For hcAb A7.2 there was no neutralization detected (data not shown).

Camelid antibodies represent a promising antibody format for therapeutic as well as in vitro diagnostic applications. These types of antibodies provide favourable properties such as a high thermostability, easy and reproducible expression in prokaryotic and eukaryotic expression systems. With 15 kDa in size, nanobodies serve as the smallest known antigen binding unit and are used mainly for therapeutic applications [19] . For in vitro diagnostic assay systems nanobodies are not practicable due to the absent Fc chain which is needed to build up a sensitive detection. Current in vitro diagnostic systems use polyclonal or monoclonal full length antibodies as detector antibodies coupled to HRP or biotin to provide a measurable and sensitive detection of the target.

So far, only few assay systems based on camelid nanobodies are published [20] [21] [22] but none of these are using camelid full length antibodies for detection. This study identifies suitable camelid antibody candidates specific for SARS-CoV-2 Spike protein selected by phage display from a camelid naïve library. We have expressed the nanobodies as full length heavy-chain-only antibody formats of an IgG2 isotype to establish tools for an ELISA based detection system. Normally, we would expect lower affinities when selecting candidates from a naïve library because the immune system was not specifically challenged against the target [23] . But the five candidates chosen from the selection rounds in this study showed good binding properties against the target. The hcAb B10 was characterized further with an affinity constant of 0.39 nM which is compared to other SARS-CoV-2 nanobodies from immune or synthetic libraries, a highly affine binder. This very good performance was also mirrored by the usability of hcAb B10 as catcher antibody in the developed ELISA format. Due to the high specificity it was possible to reduce the coating concentration to less than 1 µg/mL. Together with hcAb A7.2 as a detector antibody, a very potent antibody pair could be established. The successful conjugation of these hcAb candidates with routinely used enzymes such as HRP offers a standardized use of hcAbs also for in vitro diagnostic applications. Until now, there is no similar assay system described in literature. Our results prove the feasibility of assay systems completely designed and coordinated with camelid full length heavy-chain-only antibodies. These opportunities offer new opportunities and provide flexible tools for future applications.

Further, we could show that these hcAb candidates are not only useful as in vitro diagnostic tests components. The neutralization data for the full length wildtype virus as well as the pseudotyped wildtype and the new variant of concern B1.1.529 (Omicron) reveal a possible therapeutic application too. With four out of five candidates which are able to effectively neutralize the SARS-CoV-2 wildtype virus as well as the newly emerged Omicron variant we can provide highly efficient tools useful for a range of different applications. As recently described, the therapeutic potential of current human antibodies against the newly emerged Omicron variant is reduced caused by the increased number of mutations in the RBD domain and the resulting loss of specificity [24] . Our hcAb portfolio can therefore provide an alternative and improve the current availability of antibodies for both, a diagnostic as well as a therapeutic use.

Camelid antibodies have a high potential for diagnostic and therapeutic applications because of their exclusive properties compared to routinely used human or mouse antibodies. The approval of the first camelid therapeutic antibody is an important step.

Also for in vitro diagnostic assay systems camelid antibodies provide substantial advantages. They can easily be adapted to a routine detection by using corresponding Fc parts and specific secondary antibodies. To create novel tools and establish applications based on camelid antibodies will significantly impact future developments in the antibody field. 

Characterization of SARS-CoV-2 Spike protein specific VHHs

Recombinant Spike protein was applied to a 4-12% SDS polyacrlyamide gradient gel and blotted using the NuPage system. HcAb candidates (1 µg/mL) were incubated for 18 h. Detection took place using a HRP-conjugated secondary antibody and WesternBright Sirius as substrate solution.

Wells were coated with antigen (SARS-CoV-2 Spike protein, 3 µg/mL) and blocked with 100 µL/well PBS/ 1% casein. HcAbs were added in different dilutions. Detection was performed with a HRP-conjugated secondary antibody (ABIN1981272, 1 µg/mL) and optical density was measured at 450 nm with a reference wavelength of 620 nm. Mean and standard deviation of three independent measurements is shown. HcAbs B10 and A7.2 were conjugated with HRP to serve as secondary antibody in a sandwich ELISA. To detect the HRP conjugation, SARS-CoV-2 Spike protein was coated on the solid phase (1 µg/mL). HRPconjugated hcAb B10 and A7.2 were added in different dilutions. Binding was detected by adding TMB substrate solution and optical density was measured at 450 nm with a reference wavelength of 620 nm.

Mean and standard deviation of two independent measurements is shown.

Purified hcAbs were coated on the solid phase in different concentrations of 10, 5 and 2.5 µg/mL. SARS-CoV-2 Spike protein was added (2.5 µg/mL). HcAb A7.2 was applied as HRP-conjugated secondary antibody in a dilution of 1:1600. Detection of optical density was performed at 450 nm with a reference wavelength of 620 nm. Cross-hatched areas represent the value for the unspecific background. 

Perspective on therapeutic and diagnostic potential of camel nanobodies for coronavirus

Nanobodies: natural single-domain antibodies

RESPIRE study group , Nebulised ALX-0171 for respiratory syncytial virus lower respiratory tract infection in hospitalised children: A double-blind, randomised, placebo-controlled

The potential danger of suboptimal antibody responses in COVID-19

Exploiting nanobodies' singular traits

Nanobody approval gives domain antibodies a boost

Nanobody Technology: A Versatile Toolkit for Microscopic Imaging, Protein-Protein Interaction Analysis, and Protein Function Exploration, Front Immunol

Structure-guided multivalent nanobodies block SARS-CoV-2 infection and suppress mutational escape

Nanobodies from camelid mice and llamas neutralize SARS-CoV-2 variants

Camel nanobodies broadly neutralize SARS-CoV-2 variants, bioRxiv

Generation and validation of murine monoclonal and camelid recombinant single domain antibodies specific for human pancreatic glycoprotein 2

Panning of Antibody Phage-Display Libraries

Generation of murine monoclonal antibodies with specificity against conventional camelid IgG1 and heavy-chain only IgG2/3

Chapter 11 Preparation of enzyme-antibody or other enzyme-macromolecule conjugates, Editor(s): P. Tijssen

Rapid Quantification of SARS-CoV-2-Neutralizing Antibodies Using Propagation-Defective Vesicular Stomatitis Virus Pseudotypes, Vaccines (Basel)

Comparable Environmental Stability and Disinfection Profiles of the Currently Circulating SARS-CoV-2 Variants of Concern B.1.1.7 and B.1.351

Construction of SARS-CoV-2 Virus-Like Particles by Mammalian Expression System

Minimal system for assembly of SARS-CoV-2 virus like particles

The Therapeutic Potential of Nanobodies

Development and Evaluation of a Novel VHH-Based Immunocapture Assay for High-Sensitivity Detection of Shiga Toxin Type 2 (Stx2) in Stool Samples

Use of VHH antibodies for the development of antigen detection test for visceral leishmaniasis

Highly Sensitive Detection of Zika Virus Nonstructural Protein 1 in Serum Samples by a Two-Site Nanobody ELISA

Phage Display Libraries for Antibody Therapeutic Discovery and Development

Omicron escapes the majority of existing SARS-CoV-2 neutralizing antibodies

Morphometry of SARS-CoV and SARS-CoV-2 particles in ultrathin plastic sections of infected Vero cell cultures

Samples were diluted 1:5000 prior to measuring in NanoSight LM10. The size distribution of individual particles and the particle concentration of the samples is shown

B) Confirmation of the presence of Spike-and nucleoprotein in SARS-CoV-2 VLPs For the detection of Spike-and nucleoprotein on SARS-CoV-2 VLPs human reconvalescent serum was used in a dilution of 1:100 in western blot anlayses. Reference proteins served as controls. For the ELISA, VLPs (5 µg/mL) were coated on the solid phase and detected with commercial antibodies specific for the Spike-and the nucleoprotein. Reference proteins were used in a concentration of 0.25 µg/mL to serve as control

C) Dose-dependent detection of SARS-CoV-2 VLPs using hcAbs B10 and A7

HcAbs B10 and A7.2 were applied in different dilutions. Detection was performed using a HRP-conjugated secondary antibody. The optical density was measured at 450 nm with a reference wavelength of 620 nm. Mean and standard deviation of three independent measurements is shown. ELISA: 2-log dilution of B10 and A7.2. Kruskal Wallis test with Benjamini, Krieger and Yekuteli correction for multiple comparison against 0 µg/ml ***

We would like to thank Dr.