by Yuji Ito
Graduate School of Science and Engineering, Kagoshima University, Japan
In recent decades, many antibody therapeutics have been developed for the clinical treatment of cancer, auto immune diseases, and infectious diseases. Initially, these antibodies were mainly developed by the humanization of murine antibodies produced by hybridomas. Although new technologies for producing human antibodies, such as trans-chromo mice with human immunoglobulin loci and human B cells that have been immortalized with EB virus, have recently emerged, antibody phage libraries remain an important standard technology for generating human antibodies. This technique has a proven record of success and was used to generate several of the antibodies currently on the medical market.
Antibody phage library technology makes it possible to select antibodies that are specific for a target antigen. It uses an antibody phage library constructed by the phage display technique, and the specific antibodies are identified with an in vitro selection method called biopanning. One advantage of this technology is the ability to obtain antibodies with particular characteristics by modifying the biopanning process.
High-throughput DNA sequencing using a next generation sequencer (NGS) was previously used mainly for genome analysis and transcriptome analysis, but some groups have also begun using it to analyze the immunoglobulin sequence repertoire in animals immunized with antigen or in patients with infectious diseases [1-3]. More recently, this NGS analysis has been applied to the selection of specific antibodies from an antibody library through biopanning [4, 5]. In this article, we introduce our research on the isolation of antigen-specific antibodies from an antigen-immunized alpaca VHH phage library  and we discuss the usefulness of combining antibody selection with NGS analysis.
An early application of NGS analysis to a single chain-variable fragment (scFv) antibody phage library was published by Ravn et al [5, 7]. To check the quality of the constructed phage library, the heavy chain variable domain (VH) and light chain variable domain (VL) sequences of scFv antibodies in the phage library were comprehensively analyzed by NGS, and the results confirmed the existence of a huge variety of VH and VL genes harboring different lengths of Complementarity Determining Region 3 (CDR3) in the constructed library . Ravn et al analyzed the antibody repertoire of the phages enriched by biopanning on NGS and identified the specific antibodies against several antigens . This approach is also applicable for isolating the specific binders from other phage libraries.
The VHH known as nanobody is an antigen-binding domain of heavy chain antibody that comprise only two heavy chains, and in mammals, only camelids produce it . This VHH is attractive for pharmaceutical and industrial uses because large amounts can be readily produced in bacteria  and it is easy to manipulate for protein engineering . We constructed a VHH antibody phage library from antigen-immunized alpaca and compared the usefulness and efficiencies of two different methods for isolating antibodies specific for VHH: biopannings performed following the conventional cloning/screening method (conventional method) and those performed using a NGS analysis/in silico data mining method (NGS method). The main steps in these two methods are depicted in Figure 1.
In our protocol, cDNA produced from the blood lymphocytes of alpaca immunized with antigen is used in a PCR with DNA primers specific to the hinge regions of the heavy chain antibodies (IgG2 and IgG3) to amplify the VHH genes. After construction of the phage library, the specific binders are enriched by a single round of biopanning. In the conventional method, 50–100 phages are cloned from the phage pool after biopanning and subjected to an enzyme-linked immunosorbent assay (ELISA) to screen for binding. Using this process, 50–90% of the separated phages generate positive clones. However, in this conventional method, there is a limit to the number of clones that can be screened for binding (maximum: 1,000) because the cloning step is laborious. Furthermore, the obtained positive clones often share common or homologous sequences, including within the CDR3. Sometimes, this overlap poses a serious problem because the main purpose of this step is to obtain various VHHs that recognize different epitopes with different CDR3 sequences. As a means for solving these problems, we added NGS analysis to our biopanning.
In the NGS method depicted in Figure 2, VHH genes are amplified by PCR from the pooled phage DNA before and after biopanning. The adaptor and index sequences are added by PCR and supplied for NGS analysis on a MiSeq (Illumina) using MiSeq Reagent Kits v3 (600-cycle). Low-quality sequencing data are first trimmed based on their quality scores, and reads <200 bases in length are discarded. The cleaned-up sequencing data are processed by the ‘Merge Overlapping Pairs’ tool to obtain complete sequences of the alpaca VHHs. These DNA sequence data are translated into amino acid sequences, and the frequencies (%) of the individual sequences in the library are calculated. Finally, the frequencies for identical sequences before and after biopannings are compared, and the ratio of the frequencies is calculated as the amplification factor. Because the antigen-specific VHH sequences enriched by biopanning have higher amplification factor scores, the best candidates out of the specific binders can be lined up according to this factor.
Using this NGS method, we identified variable VHH antibodies specific for several different antigens. Although many of the identified VHHs for a given antigen were identical or highly homologous to the VHHs selected by the conventional method, we also found several VHHs that could not be found by the conventional method, probably because their frequencies in the library were too low. Most, but not all, of the VHHs that were produced by bacteria showed a clear antigen binding activity with a high level of specificity.
The NGS method can incorporate a large number (104-6) of target sequences (or clones) for analysis/screening compared with the conventional method, and is therefore very effective at finding a wider variety of specific VHH sequences. However, it is essential to check the function of the candidates produced by in silico data mining through the regeneration of the VHH gene, the production of VHHs, and the analysis of antigen binding. In the reconstruction of the gene of interest for a given antibody, several papers described the recovery of antibody genes by inverse PCR using the sequence information of heavy chain CDR3  or CDR3/FR-4 .
Analysis using a combination of NGS and a phage library is spreading, not only for VHH but also for scFv, non-immunoglobulin scaffolds, and further random peptide libraries . By employing biopanning methods that can differentiate the affinity, stability, and folding rate of binding molecules, application of this technique could spread to further characterizations of the identified binders.
This research is a collaborative work with Dr. N. Miyazaki and Dr. S. Inoue in ARK resource (Kumamoto, Japan) and Dr. Y. Hagiwara (AIST Health Research Institute, Japan).
About the Author
Dr. Yuji Ito is a professor in the Department of Chemistry and Bioscience, Graduate School of Science and Engineering, Kagoshima University (Japan). He received PhD in Pharmaceutical Science from Kyushu University in 1993. Dr. Ito is interested in the design of antibody fragments and small peptides that mimic protein function, specifically aimed at pharmaceutical uses, for which he mainly employs phage display techniques. He is currently a project leader for the “Basic Science and Platform Technology Program for Innovative Biological Medicine” project (2014-2018) that is sponsored by the Japan Agency for Medical Research and Development (AMED) and aims to develop next-generation therapeutics.
- DeKosky BJ, Ippolito GC, Deschner RP, et al. High-throughput sequencing of the paired human immunoglobulin heavy and light chain repertoire. Nat Biotechnol 2013; 31(2):166-9.
- Georgiou G, Ippolito GC, Beausang J, et al. The promise and challenge of high-throughput sequencing of the antibody repertoire. Nat Biotechnol 2014; 32(2):158-68.
- Zhu J, Ofeka G, Yang Y, et al. Mining the antibodyome for HIV-1-neutralizing antibodies with next-generation sequencing and phylogenetic pairing of heavy/light chains. Proc Natl Acad Sci USA 2013; 110(16):6470-5.
- D’Angelo S, Kumar S, Naranjo L, et al. From deep sequencing to actual clones. Protein Eng Des Sel 2014; 27(10):301-7.
- Ravn U, Gueneau F, Baerlocher L, et al. By-passing in vitro screening—next generation sequencing technologies applied to antibody display and in silico candidate selection. Nucleic Acids Res 2010; 38(21):e193.
- Miyazaki N, Kiyose N, Akazawa Y, et al. Isolation and characterization of antigen-specific alpaca (Lama pacos) VHH antibodies by biopanning followed by high-throughput sequencing. J Biochem 2015; 158(3):205-15.
- Ravn U, Didelot G, Venet S, et al. Deep sequencing of phage display libraries to support antibody discovery. Methods 2013; 60(1):99-110.
- Hamers-Casterman C, Atarhouch T, Muyldermans S, et al. Naturally occurring antibodies devoid of light chains. Nature 1993; 363(6428):446-8.
- Mizukami M, Tokunagab H, Onishia H, et al. Highly efficient production of VHH antibody fragments in Brevibacillus choshinensis expression system. Protein Expr Purif 2014; 105C:23-32.
- Zhang J, Tanha J, Hirama T, et al. Pentamerization of single-domain antibodies from phage libraries: a novel strategy for the rapid generation of high-avidity antibody reagents. J Mol Biol 2004; 335(1):49-56.
- Spiliotopoulosa A, Owen JP, Maddison BC, et al. Sensitive recovery of recombinant antibody clones after their in silico identification within NGS datasets. J Immunol Methods 2015; 420:50-55.
- Rentero Rebollo I, Sabisz M, Baeriswyl V, et al. Identification of target-binding peptide motifs by high-throughput sequencing of phage-selected peptides. Nucleic Acids Res 2014; 42(22):e169.