GB2616707A

GB2616707A - Methods

Info

Publication number: GB2616707A
Application number: GB2300352.8A
Authority: GB
Inventors: James William Finlay Jonathan; Louise Wall Hannah; Joy Rossant Christine; Yun Sanders Yun
Original assignee: Rxbiologics Ltd
Current assignee: Rxbiologics Ltd
Priority date: 2023-01-10
Filing date: 2023-01-10
Publication date: 2023-09-20
Anticipated expiration: 2043-01-10
Also published as: GB2616707B

Abstract

A method of identifying and/or selecting an antibody molecule which binds specifically to a target antigen. The method comprises the steps of: (1) providing a single-chain variant fragment (scFv) phage display library; wherein each scFv nucleic acid construct comprises: (a) a first nucleic acid sequence encoding either a human germline VH scaffold amino acid sequence including an HCDR1 and HCDR2, or a modified version thereof having at least 80% sequence identity to the human germline VH scaffold nucleic amino acid 10 sequence; (b) a second nucleic acid sequence encoding either an scFv human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof having at least 80% sequence identity to the human germline VL scaffold nucleic amino acid sequence; and (c) a third nucleic acid sequence encoding either a human B-cell donor-derived amino acid sequence including an HCDR3, or modified version thereof having at least 15 80% sequence identity to the human B-cell donor-derived amino acid sequence; (2) screening the scFv phage display library against the target antigen and selecting one or more binding scFv molecules; (3) identifying an enriched nucleic acid sequence encoding an HCDR3 from the one or more binding scFv molecules; (4) generating a fragment antigen-binding (Fab) display library of a population of engineered Fab nucleic acid constructs encoding Fab molecules including the enriched nucleic acid sequence or a modified version thereof; and (5) screening the Fab display library against the target antigen and selecting one or more binding Fab molecules.

Description

Methods

Field of the invention

The invention relates to methods of identifying and/or selecting antibody molecules which binds specifically to a target antigen, as well as libraries obtained or obtained according to the methods, and related products.

Background of the invention

Antibody discovery campaigns have employed antibody phage display (APD) technologies or immunisation approaches, including the use of transgenic animals and B cell cloning. APD technologies in particular have become powerful and effective in producing antibody molecules for diagnostic and therapeutic purposes. These technologies include methods for generating and screening single-chain variant fragment (scFv) phage display libraries and fragment antigen-binding (Fab) phage display libraries (see for example: US10,829,541; US8,921,281; US10,954,598; "Antibody Phage Display, Methods and Protocols", 2009, Aitken [Ed], Springer Protocols; "Phage Display, Methods and Protocols", 2018, Hust & Lim [Eds], Humana Press).

Key to the quality of APD libraries are their diversity (also called complexity). Various approaches have been adopted to introduce diversity into such libraries, falling into the following three 20 categories: (1) natural, using full length variable heavy (VH) / variable light (VL) chain repertoires amplified from donors (either naïve or immune libraries); (2) synthetic, by generating diversified complementary determining regions (CDRs) according to design; and (3) semi-synthetic, which relies on a combination of natural and designed CDRs.

The variable regions used in synthetic and semi-synthetic APD libraries have generally been chosen based on H and L chain germline gene frequency in the natural human repertoire and/or have been shown to have favourable properties. For example, Van Blarcom et al. (2018; mAbs 10: 256-268) took a total of four H chain and three L chain human germline frameworks as the backbone for the synthetic library, chosen based on their frequency in the memory compartment of the immune repertoire obtained from 10 donors. Similarly, Venet et al. (2012, PLoS ONE 7(8): e43471) used seven human H chain variable germline genes, five human kappa L chain variable germline genes and two human lambda L chain variable germline genes on the basis of their stability and frequency in the human antibody repertoire, to construct their libraries.

However, the importance of HCDR3 in driving antigen contact has become apparent. As early as 1992, Barbas et al. (1992, Proc. Natl Acad. Sci. USA 89: 4457-4461) generated a semi-synthetic Fab phage display library with synthetic diversity introduced only in the HCDR3. Subsequently, APD libraries have incorporated the natural diversity of the HCDR3 regions derived from the immunoglobulin genes from B cells (see for example Hoet et al. [2005, Nat. Biotechnol. 23, 344348]; Venet et al. [2012, supra]; Valadon et al. [2019, mAbs 11: 516-531]; Teixeira et al. [2021, mAbs 13: e1980942-1 to -16]; US8,921,281; US10,829,541; US10,954,508). In those libraries, HCDR3 regions are usually flanked by conserved framework (FW) sequences that allow the isolation of naïve or immune repertoire HCDR3 sequences.

Despite the advances in APD libraries, there remains a need for improvements for example to reduce processing time to generate lead antibodies that have ideal affinity, specificity, developability and/or lack of immunogenicity.

Summary of the invention

The present invention provides inter alia a method of identifying and/or selecting an antibody molecule which binds specifically to a target antigen, the method comprising the steps of: (1) providing a single-chain variant fragment (scFv) phage display library of a population of engineered scFv nucleic acid constructs encoding scFv molecules, wherein each scFv nucleic acid construct comprises: (a) a first nucleic acid sequence encoding either a human germline VH scaffold amino acid sequence including an HCDR1 and HCDR2, or a modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the human germline VH scaffold nucleic amino acid sequence; (b) a second nucleic acid sequence encoding either an scFv human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the human germline VL scaffold nucleic amino acid sequence; and (c) a third nucleic acid sequence encoding either a human B-cell donor-derived amino acid sequence including an HCDR3, or modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the human B-cell donor-derived amino acid sequence; (2) screening the scFv phage display library against the target antigen and selecting one or more binding scFv molecules; (3) identifying an enriched nucleic acid sequence encoding an HCDR3 from the one or more binding scFv molecules; (4) generating a fragment antigen-binding (Fab) display library of a population of engineered Fab nucleic acid constructs encoding Fab molecules, wherein each Fab nucleic acid construct comprises: (a) a fourth nucleic acid sequence encoding either a VH scaffold amino acid sequence including an HCDR1 and HCDR2, each or both of which are human germline and/or diversified, or a modified version having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, thereto; (b) a fifth nucleic acid sequence encoding either a Fab human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the Fab human germline VL scaffold nucleic amino acid sequence; and (c) an enriched nucleic acid sequence encoding either an HCDR3 identified in step (3), or a modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the HCDR3 identified in step (3); (5) screening the Fab display library against the target antigen and selecting one or more binding Fab molecules; and (6) optionally, reformatting the one or more binding Fab molecules into an alternative antibody molecule format, thereby identifying and/or selecting an antibody molecule which binds specifically to the target antigen.

Also provided according to the invention is an scFv phage display library and/or a Fab phage display library obtained or obtainable according to the method of the invention.

Further provided according to the invention is a library of vectors containing the nucleic acid constructs as defined here.

Additionally provided a host cell or library of host cells containing the library of the invention.

A kit containing a library of the invention is also encompassed.

Further provided according to the invention is an antibody molecule obtained or obtainable according to the method of the present invention.

Brief description of the drawings

Figure 1 is a schematic representation of an embodiment of the invention. An illustration of the region encompassing an scFv and gene 3 is shown with donor derived HCDR3 highlighted in black. The LacZ promoter is labelled and identified by a right angled arrow. The position of the VH gene is labelled and shaded in the grey and the VL gene in white. The GS represents the glycine/serine linker region that joins the VH to the VL. After an initial two or three rounds of selections the HCDR3s are amplified out and shuffled with diversified HCDR1 and HCDR2 fragments. This VH fragment is then cloned into the Fab phagemid display vector. A second wave of selections are carried out to identify Fab binders.

Figure 2 is a schematic illustration of the pAMP-scFv phagemid vector used in an embodiment of the invention (see Example 1). The donor derived HCDR3s/FW4 pools are cloned into the Pstl/Xhol sites, replacing the tetracycline resistance gene, to generate a scFv which is fused to a 6xHis tag for purification and a HA tag for detection. There is an amber stop codon (not shown) between the tags and a gene III. In an E. coli amber suppressor strain the scFv is fused to the gene III protein and displayed on the phage. Expression is under control of a LacZ promoter and a pelB leader directs expression to the periplasm. An ampicillin marker and origins of replication are also shown.

Figure 3 is a schematic illustration of the cloning of the donor derived HCDR3 into the scFv region of the pAMP-scFv vector.

Figures 4A-C depicts HCDR3 length, IGKV-KJ and IGHJ chain usage in a VH1-46 kappa library embodiment of the invention.

Figures 5A&B show phage output titres after each round of selection (A depicts selection strategy A; B depicts selection strategy B; see Example 2 below) against PD-1.

Figure 6 is a schematic illustration of the isolation of HCDR3s from outputs of scFv phage display libraries and subsequent combination with diversified HCDR1 and HCDR2 by SOE PCR.

The internal arrows represent positions of primers used for amplifying the HCDR3 region.

Figure 7 is a schematic illustration of a pCAT-Fab phagemid vector used in an embodiment of the invention. VH gene inserts generated by HCDR3 shuffling are cloned into the Bgll/Xhol sites, replacing the tetracycline resistance gene, to enable Fab display.

Figure 8 is a graph showing a monoclonal phage ELISA against PD-1 for clones from the scFv library round 2 selection output and all subsequent rounds from the Fab libraries. An anti-PD-1 positive control and anti-TIGIT negative control are included.

Figure 9A&B are graphs showing the binding of Fabs derived from periplasmic extract samples for the Kappa library and Lambda library R3 to R5 selection outputs to human (A) and cyno (B) PD-1 in a TR-FRET format.

Figures 10A&B are graphs showing correlations between human and cyno PD1 binding levels by Fabs derived from periplasmic extract samples for the Kappa library (A) and Lambda library (B) R3 to R5 selection outputs in a TR-FRET format.

Figure 11 is a graph showing the extent of human PD-1 receptor-ligand inhibition by Fabs derived from periplasmic extract samples for the Kappa library and Lambda library IRS to R5 selection outputs in a TR-FRET format.

Figure 12 is a graph showing a binding ELISA for clones in IgG1 format to human PD-1.

Figure 13 is a graph showing the effect of the anti-PD-1 binders in a PDL-1/ PD-1 receptor blocking receptor cell assay.

Detailed description of the invention

According to one aspect of the invention, there is provided a method of identifying and/or selecting an antibody molecule which binds specifically to a target antigen, the method comprising the steps of: (1) providing a single-chain variant fragment (scFv) phage display library of a population of engineered scFv nucleic acid constructs encoding scFv molecules, wherein each scFv nucleic acid construct comprises: (a) a first nucleic acid sequence encoding either a human germline VH scaffold amino acid sequence including an HCDR1 and HCDR2, or a modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the human germline VH scaffold nucleic amino acid sequence; (b) a second nucleic acid sequence encoding either an scFv human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the human germline VL scaffold nucleic amino acid sequence; and (c) a third nucleic acid sequence encoding either a human B-cell donor-derived amino acid sequence including an HCDR3, or modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the human B-cell donor-derived amino acid sequence; (2) screening the scFv phage display library against the target antigen and selecting one or more binding scFv molecules; (3) identifying an enriched nucleic acid sequence encoding an HCDR3 from the one or more binding scFv molecules; (4) generating a fragment antigen-binding (Fab) display library of a population of engineered Fab nucleic acid constructs encoding Fab molecules, wherein each Fab nucleic acid construct comprises: (a) a fourth nucleic acid sequence encoding either a VH scaffold amino acid sequence including an HCDR1 and HCDR2, each or both of which are human germline and/or diversified, or a modified version having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, thereto; (b) a fifth nucleic acid sequence encoding either a Fab human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the Fab human germline VL scaffold nucleic amino acid sequence; and (c) an enriched nucleic acid sequence encoding either an HCDR3 identified in step (3), or a modified version thereof having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the HCDR3 identified in step (3); (5) screening the Fab display library against the target antigen and selecting one or more binding Fab molecules; and (6) optionally, reformatting the one or more binding Fab molecules into an alternative antibody molecule format, thereby identifying and/or selecting an antibody molecule which binds specifically to the target antigen.

In the method, the scFv phage display library comprises a distinct combination of germline or close to germline sequence H and L chains apart from the HCDR3 which is highly diverse and derived from human B-cell repertoires. The method uses enriched HCDR3 from the binding scFv molecules ("outputs") in step (3) and combines them with germline or diversified VH HCDR1 and/or HCDR2 via "HCDR3 shuffling" in step (4). These VH fragments are then paired with a restricted repertoire of germline sequence L chains as a Fab display library as part of an in-line maturation process. This increases the probability of identifying high affinity antibody molecules during the screening in step (5). Further rounds of selection and screening of the Fab library may be carried out to isolate antibody molecules ("clones") with improved affinities over the original library.

Thus the method of the invention in part mimics natural B cell antibody evolution whereby initial binding is driven primarily through the diversification present in HCDR3 (the scFv library), H and L chain pairing is established, which is then followed by antigen-driven affinity maturation through somatic hypermutation in the HCDR loops (the binding Fab molecules).

The method allows for the identification and/or selecting of suitable antibody molecules as close to the human germline sequence as possible with minimal passenger mutations that may confer sequence liabilities and increase developability risks. This streamlined process reduces the need for multiple iterative cycles of optimisation and liability correction that is required in many existing antibody discovery platforms.

A schematic showing features of a method of the invention is depicted in Figure 1.

In the method of the invention, the scFv molecules encoded by the population of engineered scFv nucleic acid constructs may (collectively) include at least two, for example at least three, four, five, six, seven, eight, nine or ten, human germline VH scaffold amino acid sequences including HCDR1 s and HCDR2s, or modified versions thereof as defined above. Thus the scFv molecules may include at least four human germline VH scaffold amino acid sequences including HCDR1 s and HCDR2s, or modified versions thereof as defined above. The scFv molecules may include at least five human germline VH scaffold amino acid sequences including HCDR1s and HCDR2s, or modified versions thereof as defined above. The scFv molecules may include at least six human germline VH scaffold amino acid sequences including HCDR1 s and HCDR2s, or modified versions thereof as defined above.

In the method of the invention, the Fab molecules encoded by the population of engineered Fab nucleic acid constructs may (collectively) include at least two, for example at least three, four, five, six, seven, eight, nine or ten, human germline VH scaffold amino acid sequences including diversified or modified versions thereof as defined above. Thus the Fab molecules may include at least four human germline VH scaffold amino acid sequences including diversified or modified versions thereof as defined above. The scFv molecules and/or Fab molecules may include at least five human germline VH scaffold amino acid sequences including diversified or modified versions thereof as defined above. The scFv molecules and/or Fab molecules may include at least six human germline VH scaffold amino acid sequences including diversified or modified versions thereof as defined above.

In the method of the invention, the scFv molecules encoded by the population of engineered scFv nucleic acid constructs may (collectively) include at least two, for example at least three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen or fifteen, human germline VL scaffold amino acid sequences including LCDR1 s, LCDR2s and LCDR3s, or modified versions thereof as defined above. Thus the scFv molecules may include at least three human germline VL scaffold amino acid sequences including LCDR1 s, LCDR2s and LCDR3s, or modified versions thereof as defined above. The scFv molecules may include at least four human germline VL scaffold amino acid sequences including LCDR1 s, LCDR2s and LCDR3s, or modified versions thereof as defined above. The scFv molecules may include at least five human germline VL scaffold amino acid sequences including LCDR1 s, LCDR2s and LCDR3s, or modified versions thereof as defined above.

In the method of the invention, the Fab molecules encoded by the population of engineered Fab nucleic acid constructs may (collectively) include at least two, for example at least three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen or fifteen, human germline VL scaffold amino acid sequences including LCDR1s, LCDR2s and LCDR3s, or modified versions thereof as defined above. Thus the Fab molecules may include at least three human germline VL scaffold amino acid sequences including LCDR1s, LCDR2s and LCDR3s, or modified versions thereof as defined above. The Fab molecules may include at least four human germline VL scaffold amino acid sequences including LCDR1s, LCDR2s and LCDR3s, or modified versions thereof as defined above. The Fab molecules may include at least five human germline VL scaffold amino acid sequences including LCDR1s, LCDR2s and LCDR3s, or modified versions thereof as defined above.

According to the method, the scFv phage display library may comprise: (i) a kappa scFv phage display library (also referred to herein as the "VHVK" phage display library) in which each second nucleic acid sequence is a kappa ("K") light chain sequence or a modified version thereof; and (ii) a lambda scFv phage display library (also referred to herein as the "VHVA" phage display library) in which each second nucleic acid sequence is a lambda ("A") light chain sequence or a modified version thereof.

The kappa scFv phage display library may comprise a first set of vectors having first nucleic acid sequences combined with second nucleic acid sequences containing kappa light chain sequences or derivatives thereof. The lambda scFv phage display library may comprise a second set of vectors having first nucleic acid sequences combined with second nucleic acid sequences containing lambda light chain sequences or derivatives thereof.

In the method of the invention, the Fab phage display library may comprise: (i) a kappa Fab phage display library in which each fifth nucleic acid sequence is a kappa light chain sequence or a derivative thereof; and (ii) a lambda Fab phage display library in which each fifth nucleic acid sequence is a lambda light chain sequence of a derivative thereof.

The kappa Fab phage display library may comprise a third set of vectors having fourth nucleic acid sequences combined with fifth nucleic acid sequences containing kappa light chain sequences or derivatives thereof. The lambda Fab phage display library may comprise a fourth set of vectors having fourth nucleic acid sequences combined with fifth nucleic acid sequences containing lambda light chain sequences or derivatives thereof.

The second nucleic acid sequences may be the same or substantially the same as the fifth nucleic acid sequences.

The enriched nucleic acid sequences encoding the HCDR3s may be combined for the Fab display library with either: the third set of vectors if the enriched nucleic acid sequences are derived from binding scFv molecules obtained using the first set of vectors; or the fourth set of vectors if the enriched nucleic acid sequences are derived from binding scFv molecules obtained using the second set of vectors.

Depending on which scFv library the enriched HCDR3 were originally extracted from, the VH fragments may be cloned into a pool of Fab phagemid display vectors encoding either kappa or lambda germline L chain genes as used in the scFv library (referred to as a "light chain shuffle").

In other words, if one or more enriched HCDR3s are identified from a kappa scFv phage display library, then a kappa Fab phage display library may be generated for further screening against the target antigen. Additionally or alternatively, if one or more enriched HCDR3s are identified from a lambda scFv phage display library, then a lambda Fab phage display library may be generated for further screening against the target antigen.

In this way, the VH fragment can find an optimal pairing as a Fab format. Accordingly, the natural HCDR3 sequences are combined with sequences as close to the human germline sequence as possible with minimal sequence liabilities.

Switching from scFv to Fab format offers additional advantages. A Fab format is generally considered to be more stable and less prone to form dimers or multimers than scFv.

Furthermore, the monomeric nature of Fab molecules, combined with the typical lower display levels on phage and lower soluble expression levels, acts in favour of selecting and identifying high affinity binders.

Also, because a restricted or defined pool of common light chains may be used, the method allows for the straightforward and rapid generation of bispecific antibodies.

The method of the invention has a particular advantage of being able to isolate and identify by sequence analysis distinct families of antibody molecules derived from a common or related HCDR3. This opens up the possibility to interrogate the sequence space and properties of clones with positive attributes.

In the method of the invention, the population of scFv and/or Fab nucleic acid constructs may include a plurality of first nucleic acid sequences and/or fourth nucleic acid sequences which are one or more (for example, at least two, three, four, five, six, seven, eight, nine or ten) of the human germline VH gene segment-derived nucleic acid sequences selected from the group consisting of: VH1-2, VH1-46, VH1-58, VH1-69, VH3-7, VH3-11, VH3-13, VH3-23, VH3-30, VH5-51, a modified version of any of these segments encoding HCDR1 and/or HCDR2 and having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the corresponding unmodified segment, and a combination thereof.

The human germline VH gene segments used for the scFv and/or Fab nucleic acid constructs may exclude the following specific segments or combination of segments: (1) VH3-23 (as used in Hoet et al. [2015, supra], and Valadon et al. [2019, supra]); and/or (2) VH1-46, VH1-69 and VH3-23; and/or (3) IGHV1-24, IGHV3-7, IGHV4-30-4 and IGHV1-18 (as used in Teixeira et al. [2021, supra] and US11,254,931); and/or (4) VH1-69, VH3-15, VH3-23 and VH5-51 (as used in Van Blarcom et al. [2018, supra]); and/or (5) VH1-2, VH1-69, VH1-18, VH3-23, VH3-30, VH3-48 and VH5-51 (as used in Venet et al. [2012, supra]).

In the method of the invention, the population of scFv and/or Fab nucleic acid constructs may include a plurality of second and/or fifth nucleic acid sequences which are: (i) one or more (for example, at least two, three, four, five, six, seven, eight or nine) human kappa L chain gene segment-derived nucleic acid sequences selected from the group consisting of: K1-5, K1-6, K1- 16, K1-39, K2-28, K3-15, K3-11, K3-20, K3-D7, a modified version of any of these segments encoding LCDR1, LCDR2 and/or LCDR3 and having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the corresponding unmodified segment, and a combination thereof, combined with a kappa J gene segment-derived nucleic acid sequence selected from the group consisting of: J2, J4, a modified version of J2 or J4 having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the corresponding unmodified segment, and a combination thereof; or (H) one or more (for example, at least two, three, four, five or six) human lambda L chain gene segment-derived nucleic acid sequences selected from the group consisting of: L2-8, L2-14, L2- 18, L3-27, L7-43, L4-46 a modified version of any of these segments encoding LCDR1, LCDR2 and/or LCDR3 and having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the corresponding unmodified segment, and a combination thereof, combined with a lambda J gene segment-derived nucleic acid sequence selected from the group consisting of: J1, J2, a modified version of J1 or J2 having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, to the corresponding unmodified segment and a combination thereof; or a combination of (i) and (ii).

The human germline VL gene segments used for the scFv and/or Fab nucleic acid constructs may exclude the following specific segments or combination of segments: (1) VK3-15; and/or (2) VK3-20 and VK4-1 (as used in Valadon et al. [2019, supra]); and/or (3) VK1-39, VK3-20 and VL1-47 (as used in Van Blarcom et al. [2018, supra]); and/or (4) IGKV1-12, IGKV2D-29, IGKV3-11 and IGLV2-14 (as used in Teixeira et al. [2021, supra] and US11,254,931).

In the method of the invention, the fourth nucleic acid sequence may encode either a diversified HCDR1 and/or diversified HCDR2 modified to reflect natural amino acid abundance and variation in VH1 and VH3 human germline families, optionally excluding HCDR1 and/or HCDR2 sequences with sequence liabilities, or a modified version having at least 80% sequence identity, for example at least 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity, thereto.

Sequence liabilities to be excluded may include: deamidation motif NA, NG and/or NS; and/or isomerisation motif DG and/or DS; and/or hydrolysis motif DP and/or DQ; and/or unpaired cysteine C; and/or N-glycosylation motif NXT and/or NXS, where X is any amino acid except P. Additional sequence liabilities which may be excluded are found in Table 1 of Teixeira et al. (2021, supra).

In another aspect of the invention there is provided an scFv phage display library, such as a lambda scFv phage display library or kappa scFv phage display library, obtained or obtainable according to the method of the invention.

Additionally according to the invention there is provided a Fab library, such as a lambda Fab phage display library or kappa Fab phage display library, obtained or obtainable according to the method of the invention.

The present invention also provides a library of vectors containing a nucleic acid construct of the invention as defined herein.

Further provided in an aspect of the invention is a host cell or library of host cells containing a library of the invention.

The invention also encompasses a kit containing the library of the invention.

An antibody molecule obtained or obtainable according to the method of the invention is further encompassed.

An "antibody molecule" is an immunoglobulin molecule capable of specific binding to a target, such as a carbohydrate, polynucleotide, lipid, polypeptide, etc., through at least one antigen recognition site, located in the variable region of the immunoglobulin molecule.

As used herein, the term "antibody molecule" encompasses not only whole or intact polyclonal or monoclonal antibodies, but also any antigen binding fragment (for example, an "antigen-binding portion") or single chain thereof, fusion proteins comprising an antibody, and any other modified configuration of the immunoglobulin molecule that comprises an antigen recognition site including, for example without limitation, scFv, single domain antibodies (for example, shark and camelid antibodies), maxibodies, minibodies, intrabodies, diabodies, triabodies, tetrabodies, v-NAR and bis-scFv. The antibody molecule may be monovalent, bivalent or multivalent.

An "antibody molecule" encompasses an antibody of any class, such as IgG, IgA, or IgM (or sub-class thereof), and the antibody need not be of any particular class. Depending on the antibody amino acid sequence of the constant region of its heavy chains, immunoglobulins can be assigned to different classes. There are five major classes of immunoglobulins: IgA, IgD, IgE, IgG, and IgM, and several of these may be further divided into subclasses (isotypes), for example Igal IgG2, IgG3, IgG4, IgA1 and IgA2. The heavy-chain constant regions that correspond to the different classes of immunoglobulins are called alpha, delta, epsilon, gamma, and mu, respectively. The subunit structures and three-dimensional configurations of different classes of immunoglobulins are well known.

The term "antigen binding portion" of an antibody molecule, as used herein, refers to one or more fragments of an intact antibody that retain the ability to specifically bind to a target antigen. Antigen binding functions of an antibody molecule can be performed by fragments of an intact antibody. Examples of binding fragments encompassed within the term "antigen binding portion" of an antibody molecule include Fab; Fab'; F(ab')2; an Fd fragment consisting of the VH and CH1 domains; an Fv fragment consisting of the VL and VH domains of a single arm of an antibody; a single domain antibody (dAb) fragment, and an isolated complementarity determining region (CDR).

The term "Fc" or "Fc region" is used to define a C-terminal region of an immunoglobulin heavy chain. The Fc region may be a native sequence Fc region or a variant Fc region. Although the boundaries of the Fc region of an immunoglobulin heavy chain might vary, the human IgG heavy chain Fc region is usually defined to stretch from an amino acid residue at position Cys226, or from Pro230, to the carboxyl-terminus thereof. The numbering of the residues in the Fc region is that of the EU index as in Kabat. The Fc region of an immunoglobulin generally comprises two constant domains, CH2 and CH3. As is known in the art, an Fc region can be present in dimer or monomeric form.

A "variable region" of an antibody refers to the variable region of the antibody light chain or the variable region of the antibody heavy chain, either alone or in combination. As known in the art, the variable regions of the heavy and light chain each consist of four framework regions (FRs) connected by three complementarity determining regions (CDRs) also known as hypervariable regions, contribute to the formation of the antigen binding site of antibodies. When choosing FR to flank CDRs, for example when humanizing or optimizing an antibody, FRs from antibodies which contain CDR sequences in the same canonical class are preferred. Such FRs may nevertheless be modified for example by mutation or deletion, including by increasing or shortening their length if required.

Sequence identity between nucleotide or amino acid sequences can be determined by comparing an alignment of the sequences. When an equivalent position in the compared sequences is occupied by the same base or amino acid, then the molecules are identical at that position. Scoring an alignment as a percentage of identity is a function of the number of identical amino acids or bases at positions shared by the compared sequences. When comparing sequences, optimal alignments may require gaps to be introduced into one or more of the sequences to take into consideration possible insertions and deletions in the sequences. Sequence comparison methods may employ gap penalties so that, for the same number of identical molecules in sequences being compared, a sequence alignment with as few gaps as possible, reflecting higher relatedness between the two compared sequences, will achieve a higher score than one with many gaps. Calculation of maximum percent identity involves the production of an optimal alignment, taking into consideration gap penalties.

Suitable computer programs for carrying out sequence comparisons are widely available in the commercial and public sector. Examples include MatGat (Campanella et al., 2003, BMC Bioinformatics 4: 29), Gap (Needleman & Wunsch, 1970, J. Mol. Biol. 48: 443-453), FASTA (Altschul et al., 1990, J. Mol. Biol. 215: 403-410), Clustal W 2.0 and X 2.0 (Larkin et al., 2007, Bioinformatics 23: 2947-2948) and EMBOSS Pairwise Alignment Algorithms (Needleman & Wunsch, 1970, supra; Kruskal, 1983, In: Time warps, string edits and macromolecules: the theory and practice of sequence comparison, Sankoff & Kruskal (eds), pp 1-44, Addison Wesley). All programs may be run using default parameters.

In a specific example, sequence comparisons may be undertaken using the EMBOSS Needle "Pairwise Sequence Alignment" (see https://www.ebi.ac.uk/Tools/psa/emboss needle/), which determines an optimum alignment (including gaps) of two sequences when considered over their entire length and provides a percentage identity score. Default parameters for amino acid sequence comparisons ("Protein" option) may be Output format: pair; Matrix: Blosum62; Gap Open: 10; Gap Extend: 0.5; End Gap Penalty: false; End Gap Open: 10; End Gap Extend: 0.5. Default parameters for nucleotide sequence comparisons ("DNA" option) may be Output format: pair; Matrix: DNAfull; Gap Open: 10; Gap Extend: 0.5; End Gap Penalty: false; End Gap Open: 10; End Gap Extend: 0.5.

Sequence comparisons may be performed over the full length of the reference sequence.

Experimental Particular non-limiting examples of the present invention will now be described with reference to the accompanying figures.

Example 1. Isolation of HCDR3s from Human B cells for the construction of scFv phage display libraries The isolation of naive HCDR3s for use in the construction of the "semi-synthetic" scFv antibody libraries was carried out using methods similar to those described previously (see Van Blarcom et al. [2018, supra]; Hoet et al. [2005, supra]; Venet et al. [2012, supra]; Valadon et al. [2019, supra]). Briefly, cDNA was synthesised from RNA isolated from PBMCs of 53 healthy donors, using a Superscript First Strand Synthesis kit. The HCDR3s were amplified from the cDNA using a universal forward primer (HCDR3-univrsl-F) designed on a conserved region in the VH framework 3 (FW3) and an equal mix of three reverse primers that can prime off all six human IGHJ germline genes in the framework 4 (FW4) region (Table 1).

Table 1.

Primer direction Sequence (5'-3') Forward GGACACTGCAGTGTATTACTGTGC (SEQ ID NO: 1) Reverse GACACTCGAGACGGTGACCATTGTCC (SEQ ID NO: 2) Reverse GACACTCGAGACGGTGACCGTGGTCC (SEQ ID NO: 3) Reverse GACACTCGAGACGGTGACCAGGG (SEQ ID NO: 4) The forward primer contained a Pstl restriction enzyme site and the reverse primers a Xhol site that enabled cloning of the amplified fragment into the scFv phagemid display precursor vector pAMP-scFv (see Figure 2). The phagemid vector pAMP-scFv was synthesised by gene synthesis for the display of scFv fragments. Methods for generating scFv and Fab phagemid display vectors are well known in the art since the first scFv and Fab display vectors were described and detailed (McCafferty et al. [1990, Nature 348: 552-554]; Barbas et al. [1991, Proc. Natl Acad. Sci. USA 88: 7978 -7982]; Hoogenboom et al. [1991, Nucleic Acids Res. 19: 4133-4137]). Most designs are derived from the pHEN1 phagemid vector (Hoogenboom et al. [1991, supra]). In brief, for the pAMP-scFv vector used here, the donor derived HCDR3s/FW4 pools were cloned into Pstl/Xhol sites, replacing the tetracycline resistance gene, to generate a scFv which is fused to a 6xHis tag for purification and a HA tag for detection. There is an amber stop codon between the tags and the gene III. In an E. coil amber suppressor strain the scFv is fused to the gene III protein and displayed on the phage. Expression is under control of the LacZ promoter, operator and Cap-binding site. A pelB leader directs expression to the periplasm. The vector has an ampicillin resistance gene and bacterial and phage origins of replication (see Figure 3).

A total of 300 unique scFv phagemid display precursor vectors were constructed by gene synthesis using established procedures. Each of 10 VH germline or germline-like scaffolds were individually combined with 18 VK/JK and 12 VA/JL germline or germline like sequences. In some cases, silent mutations were incorporated into the germline nucleic acid sequence. Specifically, silent mutations were incorporated into the VH germline nucleic acid sequences to introduce a Pstl restriction site at the corresponding amino acid sequence positions 99-101 (numbering according to!MGT), with the exception of the VH5-51 germline sequence. For the VH5-51 sequence an existing Pstl site, found at the corresponding amino acid sequence position 89 and 90, was removed through a silent mutation. Subsequent introduction of the Pstl site, corresponding amino acid sequence positions 99-101, resulted in methionine at position 101 being changed to a valine (numbering according to!MGT). Several silent mutations were introduced into the VH3-23 sequence so that it matched the sequence of the universal forward primer listed in Table 1. The VK2-28 germline sequence was modified in CDR1 to remove a sequence liability, asparagine at position 34 was substituted to serine. These 300 H and L chain combinations have the germline or germline like sequence throughout except for the region encompassing the HCDR3 which is replaced by a tetracycline resistance gene. The 10 VH scaffolds used were: VH1-2, VH1-46, VH1-58, VH1-69, VH3-7, VH3-11, VH3-13, VH3-23, VH3-30 and VH5-51. These VH germline scaffolds were reported to have good properties (see Tiller et al., 2013; mAbs 5: 445-470). They also offer structural and sequence variation to improve the chances of binding different epitopes on the target antigen. The L chains used are germline or germline like gene sequences. 9 VK scaffolds combined with either JK2 or JK4 to create a total of 18 different VK genes and 6 VA scaffolds combined with either JL1 or JL2 to create a total of 12 different VA genes. The H and L chain genes used are summarised in Table 2.

Table 2. H and L chain germline genes used in scFv library construction

VH VK JK VA JL

1-2 1-5 J2 2-8 J1 1-46 1-6 J4 2-14 J2 1-58 1-16 2-18 1-69 1-39 3-27 3-7 2-28 7-43 3-11 3-15 7-46 3-30 3-11 3-23 3-20 3-13 3D-7 5-51 In all, 180 VH/VK and 120 VH/VA precursor vectors were generated. The VH/VK and VH/VA pAMP-scFv precursor vectors were then pooled for each of the ten chosen VH germlines. Each vector pool was digested with Pstl/Xhol to remove the tetracycline stuffer, and gel purified. The HCDR3s/FW4 were prepared as 5 pools each containing the HCDR3s/FW4 sequences derived from 10 or 11 donors, digested with Pstl/Xhol and the fragment column purified. Each of these HCDR3s/FW4 pools were then cloned into the Pstl/Xhol sites of each set of VH germline pAMPscFv precursor vectors (see Figure 3).

As an example, pAMP-scFv containing VH1-46 fragment was paired with 18 VK germline sequences. These vectors were then combined, digested with PstI/Xhol and purified by gel extraction. This vector set was then ligated to the 5 pools of Pstl/Xhol HCDR3s/FW4 fragments derived from 10 or 11 donors using T4 DNA ligase, i.e. 5 scFv kappa sub-libraries are generated for each VH germline. The ligation reactions were column purified and then electroporated into TG1 cells. Table 3 gives the library size achieved from using the VH1-46/VK precursor vector set with the 5 pools of the donor derived HCDR3s/FW4 fragments.

Table 3. Library size of the VH1-46/VK vector set with each HCDR3/FW4 donor pool Library ID Library size (cfu) VH1-46/VK + HCDR3/FW4 pool 1 5.60E+09 VH1-46/VK + HCDR3/FW4 pool 2 5.40E+09 VH1-46/VK + HCDR3/FW4 pool 3 5.20E+09 VH1-46/VK + HCDR3/FW4 pool 4 5.60E+09 VH1-46/VK + HCDR3/FW4 pool 5 6.40E+09 TOTAL 2.82E+10 The transformants from the libraries were rescued from agar plates using 2xYT media supplemented with ampicillin (100 mg/ml) and 17% glycerol to make a library master stock stored at -80 °C. The quality of each library was assessed by Sanger sequencing. The HCDR3 length and IGHJ chain usage was found to mimic the natural human repertoire (Figure 4).

A total of 50 VHVK and VHVA sub-libraries were generated resulting in a final combined library size of 1011. For selections, phage were produced for all the VHVK sub-libraries. The phage were then combined to make a master VHVK library, also referred to as the kappa scFv library. Likewise, phage produced from the VHVA sub-libraries were combined to make a master VHVA library, also referred to as the lambda scFv library. The final VHVK and VHVA scFv libraries were kept as separate library pools. Phage from the VHVK and VHVA scFv libraries were produced by standard procedures, summarised as follows. Library stocks were used to inoculate 2xYT medium supplemented with ampicillin (100 pg/ml) and glucose 2% (w/v) in a sufficient volume to cover the overall library size. The culture was grown at at 37 °C with shaking at 250rpm until 0D600 was 0.5-0.7. Cells were then infected with M13K07 helper phage at a multiplicity of infection of approximately 10 and incubated for a further 60 mins at 37 °C shaking at 150 rpm. The culture was then centrifuged at 2600g for 10 minutes and the supernatant was completely removed. The cells resuspended in 2xYT media supplemented with ampicillin (100 pg/ml) and kanamycin (50 pg/ml) and grown overnight at 25 °C at 250 rpm. The virions were purified by PEG precipitation.

Example 2. Phage display selection using the scFv libraries Selections were performed against biotinylated human and cyno PD-1 (Acro Biosystems) by solution phase panning. 1012 -1013 phage from each of the VHVK and VHVA libraries were used in the first round of panning. Phage were blocked with PBS containing 3% (w/v) skimmed milk (MPBS) for one hour at room temperature. Prior to incubation with antigen, the phage were deselected on streptavidin coated magnetic beads (Dynal M-280). The biotinylated human or cyno PD-1 (100 nM) was pre-captured on streptavidin coated beads and washed prior to incubation with the blocked phage library. After an 1 hour incubation at room temperature beads were captured using a magnetic stand and washed with PBS/0.1% Tween 20 followed by PBS. The bound phage were eluted first with triethylamine (TEA) and then with 0.2 M glycine pH2.5.

The eluates were neutralized and then used to infect exponentially growing E. col i TG1 cells. An aliquot of the infected TG1 was serial diluted to titre the selection output. The remaining infected TG1 were plated on 2xYT agar plates containing 2% glucose and 100 pg/ml ampicillin and incubated overnight. After overnight incubation at 30 °C, the cells were scraped from the surface to generate a master glycerol stock and to produce phage for a second round of selection. Phage were purified essentially as described in Example 1 but scaled down in volume. A second round of selection was performed essentially as described above with the exception that the human and cyno antigen was alternated to enrich for clones specific to both antigens and neutravidin beads were used. Human PD-1 in round 1 followed by cyno PD-1 in round 2 was referred to as selection A and the vice versa as selection B. Phage outputs were titrated after each round to identify target specific enrichment (Figure 5).

Example 3. Isolation of HCDR3 from scFv library selection outputs and shuffling with synthetic HCDR1 and HCDR2 to generate a library of diversified VHs This process takes the HCDR3s from the selection outputs from the scFv libraries as described in the above examples and combines them with fragments containing diversified VH HCDR1 and HCDR2 by splice overlap extension (SOE) PCR. There is an overlap between the two fragments of 23 by that enabled this. This is referred to as 'HCDR3 shuffling' (see Van Blarcom et al. [2018; supra]), as depicted in Figure 6. The primers used to amplify the HCDR3s from the selection outputs are shown in Table 4.

Table 4. Primers for amplifying HCDR3 Sequence GACACTGCAGTGTATTACTGTGC (SEQ ID NO: 5) CCTTGGTCGACGCGCTCGAGACGGTGACC (SEQ ID NO: 6) The HCDR1 and HCDR2 sequences were designed to reflect natural amino acid abundance and variation found in the respective VH1 and VH3 germline families with sequence liabilities removed (see Tables 5 and 6). The diversity was generated in the HCDR1 and HCDR2 of d5DNA fragments for each of the nominated VH1 and VH3 germline genes up to the start of HCDR3, by gene synthesis. A Bgll site was appended to the 5' end of the fragment to allow subsequent cloning into the Fab vectors.

Table 5. HCDR1 and HCDR2 sequences of VH1 and VH3 germline families Germline family HCDR1 sequence (ORE aa 26-33) HCDR2 sequence (ORF aa 51-58) 1-2 GYIFIGYY (SEQ ID NO: 7) INPNSGGT (SEQ ID NO: 8) 1-46 GYTFTSYY (SEQ ID NO: 9) INPSGGST (SEQ ID NO: 10) 1-58 GE:TT= (SEQ ID NO: 11) IVVGSGNT (SEQ ID NO: 12) 1-69 GGIFSSYA (SEQ ID NO: 13) IIPIFGTA (SEQ ID NO: 14) 3-7 GFTFSSYW (SEQ ID NO: 15) IKQDGSEK (SEQ ID NO: 16) 3-11 GFIFSDYY (SEQ ID NO: 17) ISSSGSTI (SEQ ID NO: 18) 3-30 GFTFSSYG (SEQ ID NO: 19) ISYDGSNK (SEQ ID NO: 20) 3-23 GFTESSYA (SEQ ID NO: 21) ISGSGGST (SEQ ID NO: 22) 3-13 GFTFSSYD (SEQ ID NO: 23) IGTAGDT- (SEQ ID NO: 24) Table 6. The positions diversified in the HCDR1 and HCDR2 for each of the nominated germlines. The amino acid residues diversified at each position are shown in bold. The corresponding parental germline sequence for HCDR1 and HCDR2 are shaded. Any combinations that give rise to the following sequence liabilities are omitted: Deamidation motif NG and NS; the isomerisation motif DG, DS; the hydrolysis motif DP and DQ; and the N-glycosylation motif NXT and NXS, where X is any amino acid except for proline.

ORF as # Germline 1-2 3 -7 3-13 HCDR1 HCDR2 2 2 2 2 3 3 3 3 5 5 5 5 5 5 5 5 6 7 8 9 0 1 2 3 1 2 3 4 5 6 7 8

T GI N N

V S. V

F F A V V N

S T A

G F G S S

F L S S L A F A V E A

E V V V E F S

V S S S

F

S

F. T S S. D. N S S 1-

F F S S V A K S S S E K

V A V S S S

H V

S S V V

T

C F.. Dr A D

F F S S V A A S

V A V V S

H

S S

T

In Example 2 above, the HCDR3 fragments were isolated from the round 2 selection outputs A and B. Briefly, a 100 pl aliquot of the glycerol stock from each output was used. Cells were pelleted and the plasmid purified from them using QlAprep Spin Miniprep Kit according to the manufacturer's protocol. The primers shown in Table 4 were used to amplify the HCDR3s using the plasmid preparation. The gel purified HCDR3 fragments were then combined with the diversified VH HCDR1 and HCDR2 fragments by splice overlap extension (SOE) PCR. The SOE PCR reaction was cleaned up using a Qiagen PCR purification column.

Example 4. Construction of Fab phagemid display vectors for cloning the HCDR3 shuffled 10 fragment For the Fab phage display library, a pCAT-Fab phagemid display vector was designed and constructed as depicted in Figure 7. The backbone of the vector is similar to the previously described pAMP-scFv vector (see Example 1), but here the displayed antibody is a Fab which consists of separate H and L chains including the CH1 and L chain constant domains. The L chain is preceded by an OmpA leader and the resistance marker is chloramphenicol.

SOE-PCR generated VH fragments from Example 3 were cloned into a pool of linearised pCATFab phagemid display vectors (specifically into the Bgll/Xhol site of the pCAT-Fab vector, replacing the tetracycline resistance gene) encoding either kappa or lambda germline L chain genes as used in the initial scFv library. VH fragments from the VHVK library selection outputs were ligated into a pCAT-Fab kappa vector pool and VH fragments from the VHV% library selection outputs into a pCAT-Fab lambda vector pool, at the Bgll/Xhol sites. After clean up, the ligated DNA were electroporated into TG1 cells to generate bacterial libraries. In total, 30 pCATFab phagemid precursor vectors were synthesised. Eighteen of these vectors each contained one of 9 VK germline scaffolds combined with either JK2 or JK4 and twelve with 6 VA germline scaffolds combined with either JL1 or JL2. These are the same L chain germline combinations as shown in Table 2 above.

Example 5. Selections and screening against PD-1 using Fab libraries constructed with 30 the shuffled HCDR3.

The assembled VH genes as described in Example 3 were digested with Bgll and Xhol and ligated to the respective pool of pCAT-Fab phagemid display vectors encoding either kappa or lambda germline L chain genes. The ligation reactions, after clean-up using a mini-elute column (Qiagen), were electroporated into TG1 cells. Phage from these Fab display libraries were prepared and selected similarly to what was described in Examples 1 and 2 for the scFv libraries but under appropriate antibiotic selection. Three rounds of selections were carried out with decreasing concentration of the biotinylated PD-1 antigen at 100, 50 and 10 nM. The antigen was preloaded onto the beads and the human and cyno form alternated in sequential rounds.

A direct phage ELISA was performed on a small number of clones from chosen rounds of selection that confirmed that the selection had generated specific binders (Figure 8). To perform phage ELISA, PD-1 at 1 pg/mlwas coated by passive adsorption onto Maxisorp plates in lx PBS. Plates were blocked with 3% MPBS. The phage supernatant produced from single colonies was first blocked in 3% MPBS prior to addition to the antigen coated plate. This was subsequently washed three times with PBST, and once with lx PBS prior to detection with commercially available anti-M13 HRP antibody conjugate (Sino Biological). Colour development was carried out with TMB and after quenching with 0.5 M sulphuric acid the absorbance at 450 nm determined.

A total of 1052 clones were picked, 352 from each round, and expressed soluble Fab in periplasmic extracts tested in direct binding and competition assays. For periplasmic extracts a starter culture for each clone, in 96 well format, was grown overnight in 2xYT supplemented with chloramphenicol 25 pg/ml and 2% glucose. 2 pl from each well was transferred into 2 ml deep well plate containing 900 pi 2xYT supplemented with chloramphenicol 25 pg/ml and 0.1 glucose. This was incubated at 30 °C at 250 rpm until an OD600 between 0.5 -0.8) was reached. At this point soluble Fab expression was induced by the addition of IPTG to a final concentration of 0.5 mM. The plates were then incubated overnight at 30 °C at 250 rpm. Soluble Fab was released from the cell pellets of the overnight cultures by first resuspending the pellet in 120 pl ice cold lx MES (MOPS, EDTA, sucrose) buffer followed by the addition of 180 pl of 20% ice cold MES buffer. Cells were spun down at x 2800g for 15 min at 4 °C and 180 pl of supernatant containing the soluble expressed Fabs was retained.

The clones were tested in a direct binding TR-FRET assay. 10 pi of diluted periplasmic extracts in 1 xPBS/0.1 % BSA were added into the wells of a 384 well plate. 5 pl of donor detection mix containing 8 nM human PD-1 Fc (Sino Biological) and 4 nM anti human IgG-cryptate (Cisbio) in 2x KF buffer (0.8 M KF in 1x PBS/ 0.1% BSA) was added into each well. This was followed by the addition of 5 pl of the acceptor mix containing 40 nM Mab anti-HA-D2 (Cisbio) in 2x KF buffer. The plates were protected from light and incubated at 4 °C overnight. The plates were read on a BMG ClarioStar microplate reader using the TR-FRET module. The assay measured the binding of HA and His tagged Fab fragments in crude periplasmic preparations to huFc tagged human or cyno PD-1. In the assay human Fc (IgG1) tagged antigens were detected using an anti-human Fc specific antibody labelled with europium-cryptate. Fabs that bound to the target antigen were detected using a fluorophore labelled secondary antibody specific for the HA or His tag. A FRET complex was formed through the binding interaction of the PD1 with the Fab and the close proximity of donor and acceptor fluorophores lead to an emission at 665 nm. Binding was observed for clones across all 3 rounds of selections for both kappas and lambdas (Figures 9A and 9B). Cross-reactivity (binding to both human and cyno PD1) was also observed (Figures 10A and 10B).

The clones were also tested in a TR-FRET competition assay. Here, Fabs were tested to see if they competed with the PDL1 ligand for binding to PD1. 10 pl of diluted periplasmic extracts in 1xPBS/0.1% BSA were added into the wells of a 384 well plate. 5 pl of donor detection mix containing 8 nM human PD-1 Fc (Sino Biologicals) and 4 nM anti human IgG-cryptate (Cisbio) in 2x KF buffer was added into each well. This was followed by the addition of 5 pl of the acceptor mix containing 100nM PD-L1 twin strep tagged protein (ACRO Biosystems) and 40 nM Strop-Mab Immo-Dy649 (IBA Life Sciences) in 2x KF Buffer. The plates were protected from light and incubated at 4 °C overnight. The plates were read on the BMG ClarioStar microplate reader using the TR-FRET module. A FRET complex was formed through the close proximity binding interaction of the PD-1 with the PD-L1. Specific inhibition of the interaction between PD-1 and PD-L1 was achieved with a number of clones resulting in a reduced fluorescent signal at 665 nm (Figure 11). The number of clones reducing the percentage specific binding increased from selection round 3 onwards for the Lambda library and plateaued after R4 for the Kappa library.

Example 6. Properties of reformatted Fabs as IgG A subset of the Fab PD-1 inhibitors from Example 5 were produced as IgG in CHO cells using standard methodology. These were tested by ELISA for binding to PD-1 as a titration series.

Goat anti-human IgG HRP (Jackson ImmunoResearch) was used as the secondary detection antibody. TMB substrate was used for colorimetric detection and the absorbance at 450 nm measured (Figure 12). The clones were further tested in a PD-1/ PDL-1 Blockade Bioassay (Promega Cat: J1250) according to the manufacturer's instructions. In brief, the PD-1/PD-L1 Blockade Bioassay is a bioluminescent cell-based assay that can measure the potency and stability of antibodies and of biologics designed to block the PD-1/PD-L1 interaction. The bioassay consists of two genetically engineered cell lines: PD-1 effector cells that express human PD-1 and a luciferase reporter driven by an NFAT response element (NFAT-RE); and PD-L1 aAPC/CHO-K1 cells. When co-cultured, the PD-1/PD-L1 interaction inhibits TCR signalling and NFAT-RE mediated luminescence. When the PD-1/PD-L1 interaction is disrupted, for example by an anti-PD-1 antibody, TCR activation induces luminescence that can be detected.

As shown in Figure 13, Antibody 3 gave a response in both assays comparable to the clinically approved anti-PD-1 control antibody (Nivolumab).

Although the present invention has been described with reference to preferred or exemplary embodiments, those skilled in the art will recognise that various modifications and variations to the same can be accomplished without departing from the spirit and scope of the present invention and that such modifications are clearly contemplated herein. No limitation with respect to the specific embodiments disclosed herein and set forth in the appended claims is intended nor should any be inferred.

All documents cited herein are incorporated by reference in their entirety.

Claims

Claims 1. A method of identifying and/or selecting an antibody molecule which binds specifically to a target antigen, the method comprising the steps of: (1) providing a single-chain variant fragment (scFv) phage display library of a population of engineered scFv nucleic acid constructs encoding scFv molecules, wherein each scFv nucleic acid construct comprises: (a) a first nucleic acid sequence encoding either a human germline VH scaffold amino acid sequence including an HCDR1 and HCDR2, or a modified version thereof having at least 80% sequence identity to the human germline VH scaffold nucleic amino acid sequence; (b) a second nucleic acid sequence encoding either an scFv human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof having at least 80% sequence identity to the human germline VL scaffold nucleic amino acid sequence; and (c) a third nucleic acid sequence encoding either a human B-cell donor-derived amino acid sequence including an HCDR3, or modified version thereof having at least 80% sequence identity to the human B-cell donor-derived amino acid sequence; (2) screening the scFv phage display library against the target antigen and selecting one or more binding scFv molecules; (3) identifying an enriched nucleic acid sequence encoding an HCDR3 from the one or more binding scFv molecules; (4) generating a fragment antigen-binding (Fab) display library of a population of engineered Fab nucleic acid constructs encoding Fab molecules, wherein each Fab nucleic acid construct comprises: (a) a fourth nucleic acid sequence encoding either a VH scaffold amino acid sequence including an HCDR1 and HCDR2, each or both of which are human germline and/or diversified, or a modified version having at least 80% sequence identity thereto; (b) a fifth nucleic acid sequence encoding either a Fab human germline VL scaffold amino acid sequence including an LCDR1, LCDR2 and LCDR3, or a modified version thereof version thereof having at least 80% sequence identity to the Fab human germline VL scaffold nucleic amino acid sequence; and (c) an enriched nucleic acid sequence encoding either an HCDR3 identified in step (3), or a modified version thereof having at least 80% sequence identity to the HCDR3 identified in step (3); (5) screening the Fab display library against the target antigen and selecting one or more binding Fab molecules; and (6) optionally, reformatting the one or more binding Fab molecules into an alternative antibody molecule format, thereby identifying and/or selecting an antibody molecule which binds specifically to the target antigen.
2. The method according to claim 1, wherein the scFv phage display library comprises: (i) a kappa scFv phage display library in which each second nucleic acid sequence is a kappa ("K") light chain sequence or a modified version thereof; and (ii) a lambda scFv phage display library in which each second nucleic acid sequence is a lambda ("A") light chain sequence or a modified version thereof. 5
3. The method according to claim 2, wherein the kappa scFv phage display library comprises a first set of vectors having first nucleic acid sequences combined with second nucleic acid sequences containing kappa light chain sequences or derivatives thereof, and the lambda scFv phage display library comprises a second set of vectors having first nucleic acid sequences combined with second nucleic acid sequences containing lambda light chain sequences or derivatives thereof.
4. The method according to any preceding claim, wherein the Fab phage display library comprises: (i) a kappa Fab phage display library in which each fifth nucleic acid sequence is a kappa light chain sequence or a derivative thereof; and (ii) a lambda Fab phage display library in which each fifth nucleic acid sequence is a lambda light chain sequence of a derivative thereof.
5. The method according to claim 4, wherein the kappa Fab phage display library comprises a third set of vectors having fourth nucleic acid sequences combined with fifth nucleic acid sequences containing kappa light chain sequences or derivatives thereof, and the lambda Fab phage display library comprises a fourth set of vectors having fourth nucleic acid sequences combined with fifth nucleic acid sequences containing lambda light chain sequences or derivatives thereof.
6. The method according to claims 3 and 5, in which the enriched nucleic acid sequences encoding the HCDR3s are combined for the Fab display library with either: the third set of vectors if the enriched nucleic acid sequences are derived from binding scFv molecules obtained using the first set of vectors; the fourth set of vectors if the enriched nucleic acid sequences are derived from binding scFv molecules obtained using the second set of vectors.
7. The method according to any preceding claim, in which the population of scFv and/or Fab nucleic acid constructs include a plurality of first nucleic acid sequences and/or fourth nucleic acid sequences which are one or more (for example, at least two, three, four, five, six, seven, eight, nine or ten) of the human germline VH gene segment-derived nucleic acid sequences selected from the group consisting of: VH1-2, VH1-46, VH1-58, VH1-69, VH3-7, VH3- 11, VH3-13, VH3-23, VH3-30, VH5-51, a modified version of any of these segments encoding HCDR1 and/or HCDR2 and having at least 80% sequence identity to the corresponding unmodified segment, and a combination thereof.
8. The method according to any preceding claim, in which the population of scFv and/or Fab nucleic acid constructs include a plurality of second nucleic acid sequences and/or fifth nucleic acid sequences which are: (i) one or more (for example, at least two, three, four, five, six, seven, eight or nine) human kappa L chain gene segment-derived nucleic acid sequences selected from the group consisting of: K1-5, K1-6, K1-16, K1-39, K2-28, K3-15, K3-11, K3-20, K3-D7, a modified version of any of these segments encoding LCDR1, LCDR2 and/or LCDR3 and having at least 80% sequence identity to the corresponding unmodified segment, and a combination thereof, combined with a kappa J gene segment-derived nucleic acid sequence selected from the group consisting of: J2, J4, a modified version of J2 or J4 having at least 80% sequence identity to the corresponding unmodified segment, and a combination thereof; or (ii) one or more (for example, at least two, three, four, five or six) human lambda L chain gene segment-derived nucleic acid sequences selected from the group consisting of: L2-8, L2-14, L218, L3-27, L7-43, L4-46 a modified version of any of these segments encoding LCDR1, LCDR2 and/or LCDR3 and having at least 80% sequence identity to the corresponding unmodified segment, and a combination thereof, combined with a lambda J gene segment-derived nucleic acid sequence selected from the group consisting of: J1, J2, a modified version of J1 or J2 having at least 80% sequence identity to the corresponding unmodified segment and a combination thereof; or a combination of (i) and (ii).
9. The method according to any preceding claim, in which the fourth nucleic acid sequence encodes either a diversified HCDR1 and/or diversified HCDR2 modified to reflect natural amino acid abundance and variation in VH1 and VH3 human germline families, optionally excluding HCDR1 and HCDR2 sequences with sequence liabilities, or a further modified version of having at least 80% sequence identity thereto.
10. The method according to claim 9, in which the sequence liabilities are: deamidation motif NA, NG and/or NS; and/or isomerisation motif DG and/or DS; and/or hydrolysis motif DP and/or DO; and/or unpaired cysteine C; and/or N-glycosylation motif NXT and/or NXS, where X is any amino acid except P.
11. An scFv phage display library obtained or obtainable according to the method of any of claims 1 -10.
12. A Fab phage display library obtained or obtainable according to the method of any of claims 1-10.
13. A library of vectors containing the nucleic acid constructs as defined in any of claims 1-10.
14. A host cell or library of host cells containing the library of any of claims 11-13.
15. A kit containing the library as defined in any of claims 11-14. 10
16. An antibody molecule obtained or obtainable according to the method of any of claims 1-10.
17. The method, library, kit or antibody molecule according to any preceding claim, in which the antibody molecule is a whole antibody, an antibody fragment or a smaller antibody molecule format, for example a single chain antibody molecule, an scFv molecule, a Fab fragment or a bivalent antibody molecule.