Patent Grant
11629434
Monoclonal antibodies are hugely important as research tools, diagnostics and therapeutics. This is, in large part, due to the fact that monoclonal antibodies can be selected to bind with high specificity and affinity to almost any structural epitope.
Classical methods of immunizing animals to obtain antibodies are slow and cumbersome and, as a result, many in vitro selection techniques have been now developed. Examples of the techniques include nucleic acid display, phage display, retroviral display, and cell surface display (e.g., yeast, mammalian, and bacterial cells). In spite of these technological developments, it is still relatively difficult to obtain antibodies that possess the desired kinetic properties, selectivity, biophysical properties, and immunogenicity necessary for therapeutic use.
Accordingly, there is a need in the art for improved methods for the selection of antibodies against a desired target.
The invention provides methods and compositions for the production of novel antibodies that bind specifically to a target antigen. The invention is particularly useful for producing novel antibodies having the antigen binding specificity of a reference antibody but with improved properties (e.g., binding affinity, immunogenicity, and thermodynamic stability) relative to the reference antibody. In particular, the methods disclosed herein make it possible to rapidly generate an entirely novel antibody molecule starting only from a CDR3 of known antigen-binding specificity.
The disclosed methods also allow for the rapid identification of novel pairs of VH and VL domains having a high intrinsic thermostability. Prior art methods for selecting VH/VL binding pairs generally involve the enforced covalent linkage of VH and VL domains (e.g., as Fab or scFv fragments) and selection of VH/VL pairs based solely upon binding affinity of the linked VH and VL domains to a target antigen, without any being paid attention to the strength of non-covalent interaction between VH and VL domains. In contrast, the methods disclosed herein select stable VH/VL pairs based upon the strength of the non-covalent interaction between unpaired VH and VL domains (in addition to the binding affinity for a target antigen). As a result, the novel methods disclosed herein yield VH/VL pairs with greater intrinsic thermostability than those obtained using prior art methods.
Accordingly, in one aspect the invention provides a method for producing a V domain that binds specifically to a target antigen. The method generally comprises: (a) providing a library of chimeric, unpaired VH or VL domains wherein diversity lies in the FR1-FR3 regions of each domain, and wherein each member of the library comprises the CDR3 region sequence from a reference antibody that binds specifically to the antigen; (b) contacting the library with the antigen; and (c) selecting from the library at least one chimeric, unpaired VH or VL domain that binds specifically to the antigen, thereby producing a V domain that binds specifically to the antigen.
In certain embodiments, the method further comprises introducing additional amino acid sequence diversity into the library of step (a). In one embodiment, additional amino acid sequence diversity is introduced by random mutagenesis.
In certain embodiments, the method further comprises the step of (d) introducing additional amino acid sequence diversity into the VH or VL domain(s) selected in step (c).
In certain embodiments, the CDR3 region sequence is from a rodent, lagomorph, avian, camelid, shark, or human antibody.
In certain embodiments, each member of the library comprises an identical CDR3 region sequence.
In certain embodiments, the FR4 region sequences of said domains are human sequences.
In certain embodiments, the FR1-FR3 region sequences of the VH and VL domains are human sequences.
In certain embodiments, each member of the library comprises FR1-FR3 sequences encoded by a single human antibody VH or VL gene.
In certain embodiments, the library is a nucleic acid display library e.g., a dsDNA display library.
In another aspect, the invention provides library of chimeric, unpaired VH or VL domains wherein diversity lies in the FR1-FR3 regions of said domains, and wherein each member of the library comprises the CDR3 region sequence from the VH or VL domain of a reference antibody.
In certain embodiments, the CDR3 region sequence is from a rodent, lagomorph, avian, camelid, shark, or human antibody.
In certain embodiments, each member of the library comprises an identical CDR3 region sequence.
In certain embodiments, the FR4 region sequences of said domains are human sequences.
In certain embodiments, the FR1-FR3 region sequences of the VH and VL domains are human sequences.
In certain embodiments, each member of the library comprises FR1-FR3 sequences encoded by a single human antibody VH or VL gene.
In certain embodiments, the library is a nucleic acid display library e.g., a dsDNA display library.
In another aspect, the invention provides a method for selecting a stable VH/VL pair. The method generally comprises: (a) providing a VH domain that binds specifically to an antigen; (b) contacting the VH domain with a library of VL domains such that a library of VH/VL pairs is formed; (c) contacting the library of VH/VL pairs with the antigen; and (d) selecting from the library of VH/VL pairs at least one VH/VL pair that binds specifically to the antigen, thereby selecting a stable VH/VL pair.
In certain embodiments, the method further comprises the step of introducing additional amino acid sequence diversity into library of VL domains of step (b). In one embodiment, the additional amino acid sequence diversity is introduced by random mutagenesis.
In certain embodiments, the library of VL domains of step (b) comprises human VL domains.
In certain embodiments, the library of VL domains or VH/VL pairs is a nucleic acid display library, e.g., a dsDNA display library.
In certain embodiments, the complementary VH domain of step (a) is produced by the methods disclosed herein.
In another aspect, the invention provides a method for selecting a bispecific, stable VH/VL pair. The method generally comprises: (a) providing a VH domain that binds specifically to a first antigen; (b) contacting the VH domain with a library of VL domains such that a library of VH/VL pairs is formed; (c) contacting the library of VH/VL pairs with a second antigen; (d) selecting from the library of VH/VL pairs at least one VH/VL pair that binds specifically to the second antigen; and (e) contacting the VH/VL pair(s) selected in step (d) with the first antigen; and (f) selecting at least one VH/VL pair that binds specifically to the first antigen, thereby selecting a bispecific, stable VH/VL pair.
In certain embodiments, the method further comprises the step of introducing additional amino acid sequence diversity into library of VL domains of step (b). In one embodiment, the additional amino acid sequence diversity is introduced by random mutagenesis.
In certain embodiments, the library of VL domains of step (b) comprises human VL domains.
In certain embodiments, the library of VL domains or VH/VL pairs is a nucleic acid display library, e.g., a dsDNA display library.
In certain embodiments, the complementary VH domain of step (a) is produced by the methods disclosed herein.
The invention provides methods and compositions for the production of novel antibodies that bind specifically to a target antigen. The invention is particularly useful for producing antibodies having the antigen binding specificity of a reference antibody but with improved properties (e.g., binding affinity, immunogenicity, and thermodynamic stability) relative to the reference antibody.
As used herein, the terms “VH domain” and “VL domain” refer to single antibody variable heavy and light domains, respectively, comprising FR (Framework Regions) 1, 2, 3 and 4 and CDR (Complementary Determinant Regions) 1, 2 and 3 (see e.g., Kabat et al. (1991) Sequences of Proteins of Immunological Interest. (NIH Publication No. 91-3242, Bethesda). The CDR boundaries can be defined using any art recognized numbering system.
As used herein, the term “V domain” refers to a single polypeptide comprising a VH domain or VL domain that is devoid of constant region sequences that facilitate the covalent pairing of said VH domain or VL domain with a complementary VL domain or VH domain, respectively.
As used herein, the term “FR1-FR3” refers to the region of a VH encompassing FR1, CDR2, FR2, CDR2 and CDR3, but excluding the CDR3 and FR4 regions.
As used herein, the term “chimeric” refers to an antibody variable domain comprising amino acid sequences from two or more different antibody variable domain, e.g., a variable domain with CDR3 sequences from a reference antibody and FR1-FR3 sequences from one or more different antibodies.
As used herein, the term “unpaired” refers to VH or VL that are not linked (either covalently or non-covalently) to a complementary VL or VH domain, respectively.
As used herein, the term “complementary VL or VH domain” refers to a VL or VH domain that associates with a VH or VL domain, respectively, to form a VH/VL pair.
As used herein, the term “VH/VL pair” refers to a non-covalent dimer of a single VH and a single VL domain, wherein the VL and VH domains are associated through the natural VH/VL dimer interface in a similar manner to that observed in a complete, tetrameric immunogobulin molecule, and the dimer can bind specifically to at least one target antigen. A “stable VH/VL pair” is a VH/VL pair that does not exhibit significant dissociation of the substituent VH and VL domains under physiological conditions.
As used herein, the term “bispecific” refers to a VH/VL pair, wherein the VH and the VL domains bind to different antigens.
As used herein, the term “nucleic acid display library” refers to any art recognized in vitro cell-free phenotype-genotype linked display, including, without limitation those set forth in, for example, U.S. Pat. Nos. 7,195,880; 6,951,725; 7,078,197; 7,022,479; 6,518,018; 7,125,669; 6,846,655; 6,281,344; 6,207,446; 6,214,553; 6,258,558; 6,261,804; 6,429,300; 6,489,116; 6,436,665; 6,537,749; 6,602,685; 6,623,926; 6,416,950; 6,660,473; 6,312,927; 5,922,545; and 6,348,315, and in WO2010/011944, which are all hereby incorporated by reference in their entirety.
As used herein, the term “specifically binds to” refers to the ability of a binding molecule (e.g., a VH or VL domain) to bind to an antigen with an affinity of at least about 1×10−6 M, 1×10−7 M, 1×10−8 M, 1×10−9 M, 1×10−1° M, 1×10−11M, 1×10−12 M, or more, and/or bind to a target with an affinity that is at least two-fold greater than its affinity for a nonspecific antigen.
As used herein, the term “antigen” refers to the binding site or epitope recognized by a binding molecule (e.g., a VH or VL domain).
As used herein, the term “reference antibody” refers to any antibody that binds specifically to an antigen of interest.
As used herein, the term “antibody” refers to IgG, IgM, IgA, IgD or IgE or an antigen-binding fragment thereof (e.g. VH and/or VL), whether derived from any species naturally producing an antibody, or created by recombinant DNA technology.
In one aspect, the invention provides a method for producing a V domain that binds specifically to a target antigen. This method allows for the rapid generation of an entirely novel antibody molecule starting only from a CDR3 of known antigen-binding specificity. The method generally comprises the steps of: (a) providing a library of chimeric, unpaired VH or VL domains wherein diversity lies in the FR1-FR3 regions of each domain, and wherein each member of the library comprises a CDR3 region sequence that binds specifically to the antigen; (b) contacting the library with the antigen; and (c) selecting from the library at least one chimeric, unpaired VH or VL domain that binds specifically to the antigen.
The CDR3 region sequence can be artificial, naturally derived, or a combination thereof. Naturally derived CDR3 region sequences can be from any organism that produces an antibody including, but not limited to, rodent, lagomorph, avian, camelid, shark, or human.
The CDR3 sequence can be from an antibody heavy chain or a light chain. However, the skilled worker will appreciate that the CDR3 must match the V-domain context into which it is placed e.g., a heavy chain CDR3 should be used in a VH domain library and a light chain CDR3 should be used in a VL domain library.
In certain embodiments, the library contains a single species of CDR3 sequence. In a particular embodiment, the CDR3 sequence is from a single reference antibody that binds to the target antigen. In other embodiments, the library contains multiple species of CDR3 sequence. In a particular embodiment, the multiple species of CDR3 sequence are variants of a single CDR3 sequence from a single reference antibody. Such variants can be produced by alteration (e.g. substitution, addition, deletion and/or modification) of at least one amino acid in the CDR3 sequence from the reference antibody. Alterations can be generated randomly (e.g., by random mutagenesis) or in a site-directed fashion using any art recognized means.
In general, the library comprises a diversity plurality of diverse FR1-FR3 region amino acid sequences. The plurality of FR1-FR3 region amino acid sequences can be from any source including, without limitation, naturally-occurring variable regions (e.g, from the antibody VH or VL gene repertoire of an animal), artificial antibody variable regions, or a combination thereof. In certain embodiments, each member of the library comprises FR1-FR3 sequences encoded by a single antibody VH or VL gene (e.g., a human VH or VL gene). In other embodiments, the FR1-FR3 region sequences are human sequences. In a particular embodiment, the FR1-FR3 region sequences are naturally-occurring human antibody variable region sequences from the B-cells of healthy donors.
In certain embodiments, the V domain further comprises an FR4 region. The FR4 region can be from the reference antibody or from another source. Suitable sources include, without limitation, naturally occurring human antibody variable regions, artificial antibody variable regions, or a combination thereof.
The V-domain library can be generated using any art recognized methods including, without limitation, ab initio nucleic acid synthesis and DNA or RNA polymerase-mediated assembly. In a preferred embodiment, the library is assembled by PCR using overlapping oligonucleotides. Suitable oligos for assembling a VH domain library include those set forth in SEQ ID No: 4-21. Suitable oligos for assembling a VL domain library include those set forth in SEQ ID No: 70-102.
The chimeric, unpaired VH or VL domain(s), selected using the methods of invention, can be paired with a complementary VL or VH, respectively, to generate a VH/VL pair using the methods disclosed herein.
In another aspect, the invention provides a method for selecting stable VH/VL pairs. This method selects stable VH/VL binding pairs based upon the strength of the non-covalent interaction between unpaired VH and VL domains and allows for the rapid identification of novel pairs of VH and VL domains having a high intrinsic thermostability. The method generally comprises the steps of: (a) providing a VH domain that binds specifically to an antigen; (b) contacting the VH domain with a library of VL domains such that a library of VH/VL pairs is formed; (c) contacting the library of VH/VL pairs with the antigen; and (d) selecting from the library of VH/VL pairs at least one VH/VL pair that binds specifically to the antigen.
The VH domain that is used to screen for stable VH/VL pairs can be from any source (artificial, naturally derived, or a combination thereof). In certain embodiments, the VH domain is a from a reference antibody. In other embodiments, the VH domain is a chimeric VH domain selected using the CDR3/Framework methods disclosed herein.
The library of VL domains that is used to screen for stable VH/VL pairs can be from any source (artificial, naturally derived, or a combination thereof). In certain embodiments, the library of VL domains is a human VL domain library. In a particular embodiment, the human VL domain library comprises naturally-occurring human antibody VL region sequences from the B-cells of healthy donors. The VL domain libraries disclosed herein are particularly suitable for use in these methods.
Although it is preferred to screen a library of VL domains using a VH domain as the bait, the skilled worker will appreciate that it is entirely within the scope of the invention to perform the converse screen i.e., to screen a library of VH domains using a VL domain as the bait. This method general comprises that steps of: (a) providing a VL domain that binds specifically to an antigen; (b) contacting the VL domain with a library of VH domains such that a library of VH/VL pairs is formed; (c) contacting the library of VH/VL pairs with the antigen; and (d) selecting from the library of VH/VL pairs at least one VH/VL pair that binds specifically to the antigen, thereby selecting a stable VH/VL pair.
In another aspect, the invention provides, a method for selecting a stable VH/VL pair, wherein the method generally comprises: (a) providing a first nucleic acid display library of VH domains, wherein members of the library comprise a VH domain linked to its cognate nucleic acid coding sequence; (b) providing a second nucleic acid display library of VL domains, wherein members of the library comprise a VL domain linked to its cognate nucleic acid coding sequence; (c) contacting the first nucleic acid display library with the second nucleic acid display library such that a library of VH/VL pairs is formed; (d) contacting the library of VH/VL pairs with an antigen; and (e) selecting from the library of VH/VL pairs at least one VH/VL pair that binds specifically to the antigen, and, (f) ligating together the nucleic acid coding sequences of the VH and VL domains in the VH/VL pairs selected in step (e). This method is particularly advantageous in that it allows for the simultaneous screening of VH and VL domain libraries and the precise determination of the sequences of the VH and VL domains in each selected, stable VH/VL pair.
The nucleic acid coding sequences in each selected VH/VL pair can be ligated together using any appropriate art recognized means (e.g., chemical and/or enzymatic). RNA can be ligated with, for example, RNA ligase. DNA can be ligated, with, for example, DNA ligase. Ligation of the VH and VL domain nucleic acid coding sequences can be facilitated by the use of, for example, nucleic acid linkers, adaptors and/or restriction enzyme digestion. In certain embodiments, the nucleic acid coding sequences of the VH and VL domain in each selected VH/VL pair are ligated together to form a single, continuous nucleic acid sequence (linear or circular), and the VH and VL domain sequences present in the continuous nucleic acid sequence are determined. Nucleic acid sequence determination can achieved by any art-recognized means, including without limitation, single molecule DNA sequencing.
The above methods for identification of stable VH/VL pairs are particularly suitable for use in combination with the CDR3/framework shuffling methods disclosed herein. This combination of methods allows selection of a novel, stable VH/VL pair specific for a target antigen using only an antigen-specific CDR3 sequence as a starting point.
The methods of the invention are particularly useful for screening for stable VH/VL pairs. However, one skilled in the art will appreciate that the invention can be used more broadly to screen for any stable protein dimer, wherein at least one monomer of the dimer binds to a ligand. Suitable protein dimers include, without limitation, immunoglobulin superfamily members (e.g., T-cell receptors), hormones, cytokines, transcription factors, and the like.
In another aspect, the invention provides a method for selecting a bispecific, stable VH/VL pair, (e.g., a VH/VL pair in which the VH and VL domains bind to two different antigens). The method generally comprises the steps of: (a) providing a VH domain that binds specifically to a first antigen; (b) contacting the VH domain with a library of VL domains such that a library of VH/VL pairs is formed; (c) contacting the library of VH/VL pairs with a second antigen; (d) selecting from the library of VH/VL pairs at least one VH/VL pair that binds specifically to the second antigen; and (e) contacting the VH/VL pair(s) selected in step (d) with the first antigen; and (f) selecting at least one VH/VL pair that binds specifically to the first antigen.
In one embodiment, the first and second antigens are different epitopes present in a single molecule. In another embodiment, the first and second antigens are epitopes present in two separate molecules.
In certain embodiments, it is desirable to select for a promiscuous light chain that can bind to at least two different VH domains (with different antigen specificities) but that cannot, itself, specifically bind to an antigen (e.g., has a low affinity for all antigens). Such light chains are particularly useful in that they can be used in the assembly of full-length, bispecific, heterotetrameric antibodies. These bispecific antibodies generally comprising a first heavy chain with a first antigen specificity, a second heavy chain with second antigen specificity, and two molecules of the selected promiscuous light, wherein each heavy chain is paired with a different promiscuous light chain molecule.
The skilled worker will appreciate that any type of VH or VL domain expression library can be employed in the methods of the invention. Suitable expression libraries include, without limitation, nucleic acid display, phage display, retroviral display, and cell surface display libraries (e.g., yeast, mammalian, and bacterial cells). In certain embodiments, the library is a nucleic acid display library. In a preferred embodiment, the nucleic acid display library is a DNA-display library constructed as exemplified herein or in WO2010/011944, which is hereby incorporated by reference in its entirety.
Methods for screening expression libraries are well known in the art. See, for example, Antibody Engineering: Methods and Protocols. Methods in Molecular Biology Volume 248, (B.K.C. Lo, Ed) Humana Press, 2004 (ISBN: 1-58829-092-1), which is hereby incorporated by reference in its entirety. Suitable methods of screening nucleic acid display libraries, include, without limitation those set forth in U.S. Pat. Nos. 7,195,880; 6,951,725; 7,078,197; 7,022,479; 6,518,018; 7,125,669; 6,846,655; 6,281,344; 6,207,446; 6,214,553; 6,258,558; 6,261,804; 6,429,300; 6,489,116; 6,436,665; 6,537,749; 6,602,685; 6,623,926; 6,416,950; 6,660,473; 6,312,927; 5,922,545; and 6,348,315, and in WO2010/011944, which are all hereby incorporated by reference in their entirety. In a preferred embodiment, the screening of nucleic acid-display libraries is performed as exemplified herein or in WO2010/011944.
The methods of the invention are particularly well suited to generating VH or VL domains, and/or VH/VL pairs that bind to a target antigen with high affinity. To enhance affinity for the antigen, multiple rounds of library screening can be performed, with additional amino acid sequence diversity being introduced at each screening round, if desired. If desired, the stringency of the method can be enhanced by altering the assay conditions to reduce the affinity of the VH or VL domains for the antigen, for example, by altering the pH and/or salt concentration of the antigen binding reaction. Such methods selectively enrich for VH or VL domains with the highest affinity and stability.
The methods of the invention also allow for selection of VH or VL domains, and/or VH/VL pairs having a different antigen specificity to that of a starting reference antibody. For example, the HCDR3 from an antibody that only binds to human PDGFRβ can be used to select for a VH domain, and/or VH/VL pair that binds to both human and mouse PDGFRβ. Such a selection is exemplified in Example 2 herein.
In certain embodiments, additional amino acid sequence diversity is introduced into the VH or VL domain library, or the VH or VL domain(s) selected from the library. Amino acid sequence diversity can be introduced by any art-recognized means. In certain embodiments, amino acid sequence diversity is introduced by alteration of the nucleic acid sequence(s) encoding the VH or VL domain library, or the VH or VL domain. In certain embodiments amino acid sequence diversity is introduced by random mutagenesis. Such random mutagenesis can be achieved, for example, by low-fidelity PCR amplification of the nucleic acid sequence(s) encoding the VH or VL domain library, or the VH or VL domain.
The VH and/or VL domains employed in the methods of the invention can be used in isolation or fused to additional amino acids (e.g., epitope tags) and/or domains. In certain embodiments, at least one VH domain in a library is fused to at least one CH1 domain, CH2 domain, CH3 domain or a combination thereof. In a particular embodiment, at least one VH domain in a library is fused to at least one heavy chain constant region comprising at least one CH1 domain, CH2 domain and CH3 domain. In certain embodiments, at least one VL domain in a library is fused to at least one light chain constant region.
VH or VL domains, and/or VH/VL pairs selected using the methods of the invention can be incorporated into another antibody format including, without limitation, scFv, Fab, and/or complete tetrameric antibody.
The present invention is further illustrated by the following examples which should not be construed as further limiting. The contents of Sequence Listing, figures and all references, patents and published patent applications cited throughout this application are expressly incorporated herein by reference.
VH domains that bind specifically to human PDGFRβ were selected using DNA display as set forth in WO2010/011944, which is hereby incorporated by reference in its entirety. Specifically, a naïve, human VH domain DNA display library containing derived from 10 bone marrow donors was subject to six rounds of selection against human PDGFRβ. The selected binders were cloned and sequenced. From this screen VH domain clones A4, B4 and G2 were selected, the amino acid sequences of which are set forth in Table 1.
A. VH Library Construction
To screen for VH domains with improved binding characteristics, the HCDR3 sequence of clone A4 (designated XB1511) was shuffled into a naïve human VH library, which was further selected for binding to human and mouse PDGFRβ. Specifically, the DNA sequence coding for the HCDR3 of clone A4 (SEQ ID NO: 1) was synthesized and assembled into a library comprising framework regions 1-3 of naïve human VH domains amplified from bone marrow B cells and PBMCs using framework specific oligonucleotides. Human VH framework regions 1-3 were amplified using 5′ VH family-specific and 3′ generic FR3 reverse primers to create separate libraries of VH family framework regions. The VH family framework libraries and the XB1511 HCDR3 were shuffled by further PCR amplification using 5′ T7TMV and 3′ XB1511 FR3CDR3FR4 oligos. This also added a T7TMV promoter sequence at the 5′ end for in vitro transcription/translation. A C-terminal Cμ3 sequence and a FLAG tag (for purification after translation) were also added by PCR using FR4 Cu3 Reverse and Y109 primers, respectively, together with the 5′ T7TMV primer. The nucleic acid sequences of the oligonucleotides used for preparation of the HCDR3-shuffled VH library are set forth in Table 2. A schematic representation of the general concept of CDR3/framework shuffling is set forth is
B. Library Screening
The HCDR3 shuffled VH domain library was then transcribed into an mRNA library and subjected to selection with dsDNA display technology as set forth in WO2010/011944. A schematic representation of the screening method is set forth in
C. Binding Specificity of Selected HCDR3 Shuffled VH Domains
The R4 binding pool selected above was assessed for binding to both human and mouse PDGFβ using a 35S Met-labelled in vitro translated library. Specifically, binding of the pool to epoxy beads, 100 nM of human IgG, human PDGFβ and mouse PDGFRβ were assessed. As shown in
D. Binding Affinity of Selected HCDR3 Shuffled VH Domains
The R4 framework shuffled human and mouse PDGFβ enriched VH domain pool was cloned into E. coli expression vectors, produced and purified. The binding kinetics of the VH domains to human and mouse PDFGR was determined using surface Plasmon resonance on a Biacore T100. Briefly, human and mouse PDGFR-hIgG1-Fc chimeric fusion protein were separately immobilized using a Series CMS sensorchip (CMS) coupled to anti-hIgG1 Fc monoclonal antibody. For each cycle, the PDGFR fusion protein was first captured, followed by the injection of VH for 115 seconds at a flow rate of 100 uL/min (association). Immediately following the association phase is a dissociation phase of 600 seconds. The surface was regenerated at each cycle with a single injection 3M MgCl2 (10 uL/min, 60 seconds). Multiple concentrations of VH were injected (0.55 nM-40 nM) and the resulting sensorgram were analyzed with T100 Evaluation software. The binding kinetics was determined using 1:1 binding curve fitting. The binding kinetics of VH domain clones XB2202 and XB2708 to human and mouse PDGFR are shown in
A. Construction of VL DNA Libraries
Human VL libraries (Vkappa and Vlamda) were constructed from B cells of young healthy donors (Allcells) by RT-PCR. To ensure the diversity of the library, 300 million mononuclear cells from bone marrow of each donor and total of 10 donors and 100 million of peripheral blood mononuclear cells from each donor and total of 10 donors were obtained for naïve VH and VL library construction. A schematic of the library generation method is set forth in
Oligonucleotide primers for cDNA synthesis and subsequent PCR amplification of the Vkappa and Vlamda sequences were design as set forth in Table 4. Specifically, multiple sense primers were designed from the Vκ and Vλ FR1 regions of each family with an upstream UTR sequence. The anti-sense primers for κ and λ gene amplification were designed from the constant regions nested to Cκ1 (Cκ2) or Jλ with the same Cκ2 downstream (JλCκ2). The Vκ and Vλ libraries carry the same C-terminal sequence for PCR amplification during the selection cycles.
mRNA was prepared with individual donors using a FastTrack mRNA preparation kit (Invitrogen) following the protocol provided by the kit. First strand cDNA was synthesized from the isolated mRNA using primers specific for the light chain kappa and lambda constant regions (Cκ1 and Cλ1).
PCR amplification of the Vkappa and Vlamda sequences was performed with Cκ2 and Vκ family specific or JλCκ2 mix and Vλ family specific oligos using cDNA synthesized from mRNA prepared from the various source of B cells as template. The PCR was performed with individual families and individual donors for 18-20 cycles. After gel purification, Vκ and Vλ libraries from different sources were pooled to generate Vκ and Vλ libraries.
B. Generation of VL Fusion Libraries by dsDNA Display
Vk and VL DNA libraries generated using the methods set forth in this Example were transcribed into mRNA libraries using the T7 Megascript kit (Invitrogen, Cat #AM1334). The mRNA was purified with RNeasy MinElute Cleanup Kit (Qiagen, Cat #74204) following protocol provided by the kit. A total of 600 pmol of RNA (300 pmol of Vk and Vl libraries) was ligated and assembled with dsDNA display linkers and components as described in WO2010/011944. The assembled VL library was subjected to in vitro translation system to create a fusion library in which each VL domain (phenotype) is stably fused to its coding sequence (genotype). 35S Met was incorporated in the translation process to radiolabel the fusions. The library was then purified with oligo dT cellulose, converted into a dsDNA display library using the standard molecular biology techniques of reverse transcription, RNaseH digestion, 2nd strand DNA synthesis, followed by flag tag purification. A schematic of this process is set forth in
C. Identification of VL Pairs for XB1511, and XB2202 VH Domains
As a proof of concept, PDGFRβ VH binding domain XB1511 was translated as free protein (with incorporation of 35S Met in the translation reaction) and affinity purified through a c-terminal flag tag. The XB1511 and a purified VL fusion library (prepared as above) were then mixed at an equal molar ratio and incubated at 25 C overnight to allow for in vitro association of VH and VL fusion domains through their hydrophobic patch. The mixture was then contacted with PDGFRβ target pre-immobilized on Epoxy450 beads or in solution and captured by protein A beads. Complexes that bound to the immobilized PDGFRβ target were washed and eluted with 0.1N KOH. PCR was performed with VL specific primer sets to recover the VLs that bound to the PDGFRβ target, both as VH-VL pairs and as unpaired VL domains. The VL pairing was performed for 3 rounds, with low stringency (100 nM PDGFRβ) for the first 2 rounds and higher stringency (10 nM PDGFRβ) for the third round. PDGFRβ VH binding domain XB2202 was also paired with VL library similarly for 2 rounds. For each round of XB2202 VL pairing and selection, the stringency was increased by kinetic controlled on and off rate strategy to identify VLs that can pair with VH stably and enhance the VH binding. A schematic of this process is set forth in
VL pools identified above were then cloned into Blunt Zero TOPO vector (Invitrogen) and VL-encoding DNA sequences were amplified from the resultant bacterial colonies by PCR using M13 forward and reverse primers. The individual amplified VL-encoding DNA sequences were then sequenced. The sequence data obtained from VL pools showed that a diverse repertoire of VLs was enriched through the process. Multiple families and frameworks were present in the pool. Several VLs were present as replicates or families. Distinct VL families could be identified and several VLs were present more than once. Exemplary VL sequences identified using the methods of the invention that pair with the PDGFRβ-binding VH domains XB1511 and XB2202 are set forth in Table 4 herein.
D. Evaluation of Identified VH and VL Pairs
To evaluate the characteristics of the identified VH-VL pairs, 10-12 scFVs from each pool were constructed and produced by either in vitro translation or by E. coli expression, followed by affinity purification.
A PDGFRβ binding ELISA assay was performed to assess the binding of the scFv to immobilized PDGFRβ and to determine the EC50. Specifically, 2 ug/mL of human PDGFRβ and human Fc or IgG in PBS was immobilized on Maxisorp plates at 4° C. overnight. The plate was then washed and blocked with superblock. In vitro translated crude scFv lysate was diluted 1:3 in 1×PBST. 100 ul of the diluted scFv lysate was loaded into each well of Maxisorp plates and incubated for 1 hour at room temperature. scFv that bound to immobilized PDGFRβ was detected by anti-flag antibody-HRP at 1:5000 dilution and a TMB substrate. The plate was read on a Molecular Device plate reader with end point assay at OD 450 nm. As shown in
The affinity of several scFvs was determined by solution based equilibrium binding assay. Specifically, 120 pmol of scFv RNA was translated into free protein with 35S Met incorporated. The translated reaction mixture was 3-fold diluted in binding buffer containing 1×PBS with 0.025% triton, 1 mg/mL BSA and 0.1 mg/mL sssDNA. Human PDGFRβ was diluted in the same binding buffer to final concentrations from 100 nM to 0 nM. The diluted scFv mixture was incubated with hPDGFRβ in final volume of 100 ul on Kingfisher plates (Thermofisher Scientific, 97002084). Following incubation, 25 ul of protein A magnetic beads (Invitrogen) were used to capture the PDGFRβ from solution. The captured PDGFRβ was washed and eluted in kingfisher Reader (Thermofisher Scientific). The amount of scFv (labeled with 35S Met) bound to the magnetic bead-immobilized hPDGFRβ was counted using a scintillation counter and the Kd was calculated with Graph Pad Prism 5. For the XB1511-derived scFv tested, 2 scFv showed an 8-10 fold higher KD, 1 showed 2.5 fold higher KD, and 4 showed a similar KD when compared to XB1511 VH alone (
XB1511 VH and D8 VL were expressed as heterotetrameric IgG in 293T cells. Cell culture supernatant was collected after 48 hours and 96 hours and the expressed IgG was purified with protein A agarose beads. The IgG was produced at 8 mg/L without any optimization. To evaluate the biological activity of the XB1511/D8 IgG, HFF-1 human foreskin fibroblasts were seeded in 384-well BIND biosensors and allowed to attach overnight in serum-free media. The fibroblast cells were then stimulated with 5 ng/mL or 10 ng/mL of PDGFBB ligand and allowed to migrate for 18 hours. BIND Scanner images were captured every 15 minutes and software analysis tools used to measure the track lengths of individual cell migration responses. Track length is represented by a “heat map” from blue (no migration) to red (maximal migration). As shown in
The thermostability of the XB2202 VH and scFv XB2202-A4 were determined. Specifically, 1 mg/mL of XB2202 and XB2202-A4 were incubated at 4 C, 37 C, 60 C and 70 C for 12 hours and PDGFRβ binding ELISA was performed to test the binding activity of the protein after incubation. As shown in
This application is a division of U.S. patent application Ser. No. 15/847,203, filed Dec. 19, 2017, which is a division of U.S. patent application Ser. No. 14/005,085, filed Dec. 24, 2013, now U.S. Pat. No. 9,880,160, which is the U.S. national stage of International Patent Application Serial No. PCT/US2012/029086, filed Mar. 14, 2012, which claims priority to U.S. Provisional Patent Application Serial Nos. 61/566,778, filed Dec. 5, 2011; and 61/453,106, filed Mar. 15, 2011, the entire disclosures of which are hereby incorporated herein by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 20050255552 | Flynn et al. | Nov 2005 | A1 |
| 20060134098 | Bebbington et al. | Jun 2006 | A1 |
| 20100105569 | Hsieh et al. | Apr 2010 | A1 |
| Number | Date | Country |
|---|---|---|
| 2712298 | Jul 2009 | CA |
| WO 2008130704 | Oct 2008 | WO |
| WO 2013085972 | Jun 2013 | WO |
| Entry |
|---|
| Brezinschek et al. (1998) “Pairing of variable heavy and variable k chains in individual naive and memory B cells,” The Journal of Immunology. 160(10):4762-4767. |
| Chen et al. (2008) “Construction of a Large Phage-Displayed Human Antibody Domain Library with a Scaffold Based on a Newly Identified Highly Soluble, Stable Heavy Chain Variable Domain,” J. Mol. Biol. 382(3):779-789. |
| Hur et al. (2010) “Development of the dual-vector system-III (DVS-III), which facilitates affinity maturation of a Fab antibody via light chain shuffling,” Immunol. Let. 132(1-2):24-30. |
| International Search Report corresponding to International Patent Application No. PCT/US2012/029086, dated Mar. 28, 2013. |
| Kang et al. (1991) “Antibody redesign by chain shuffling from random combinatorial immunoglobulin libraries,” Proc Natl Acad Sci USA. 88(1):11120-11123. |
| Supplementary European Search Report with Written Opinion corresponding to European Patent Application No. 12757543.9, dated May 11, 2015. |
| Written Opinion corresponding to International Patent Application No. PCT/US2012/029086, dated Mar. 28, 2013. |
| Number | Date | Country | |
|---|---|---|---|
| 20200400657 A1 | Dec 2020 | US |
| Number | Date | Country | |
|---|---|---|---|
| 61566778 | Dec 2011 | US | |
| 61453106 | Mar 2011 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 15847203 | Dec 2017 | US |
| Child | 16862851 | US | |
| Parent | 14005085 | US | |
| Child | 15847203 | US |
