CA2421960A1

CA2421960A1 - Heavy chain libraries

Info

Publication number: CA2421960A1
Application number: CA002421960A
Authority: CA
Inventors: Ton Logtenberg; Erwin Houtzager
Original assignee: Individual
Current assignee: Janssen Vaccines and Prevention BV
Priority date: 2000-09-13
Filing date: 2001-09-12
Publication date: 2002-04-11
Also published as: WO2002028903A2; WO2002028903A3; AU2001294399A1

Abstract

The invention provides libraries which comprise binding molecules that are adapted to expression in an expression organism, but which are also transferable to a human context without undergoing a change in conformation and/or build up. The invention therefore provides a method for producing a human monoclonal antibody, said method comprising: providing a library of binding molecules, the binding domain of which consists essentially of human heavy chain variable fragments in a functional format, selecting from said library at least one heavy chain variable fragment having a desired binding affinity, and inserting a nucleic acid encoding said heavy chain variable fragment into a nucleic acid encoding the complementary part of at least a heavy chain of said human monoclonal antibody, allowing for expression of th e resulting heavy chain and for assembly of said heavy chain with a desired light chain, and producing a human monoclonal antibody. The conformation of said heavy chain variable fragment is such that it retains its binding affinity whether it is in phage display or in its normal heavy chain environment. A method for making a library for use in said method is also herewith provided, as well as methods of keeping heavy chain variable fragments in said conformation. The present inventionallows for producing larger libraries than the known ones. Besides, loss of specificities and affinities because of expression problems are reduced.

Description

Title: Heavy chain libraries Technical Field: The present invention relates to the fields of molecular biology and immunology and in particular to the field of designing, for example, human antibodies having a desired binding affinity through display and selection techniques.
Background: The exposure to a highly diverse and continuously changing environment requires a dynamic immune system that is able to rapidly adapt in order to adequately respond to potentially harmful micro-organisms. Higher organisms have evolved specialized molecular mechanisms to ensure the implementation of clonally-distributed, highly diverse repertoires of antigen-receptor molecules expressed by cells of the immune system: immunoglobulin (Ig) molecules on B lymphocytes and T cell receptors on T lymphocytes. For B lymphocytes, a primary repertoire of (generally low affinity) Ig receptors is established during B cell differentiation in the bone marrow as a result of rearrangement of gerrnline-encoded gene segments.
Further refinement of Ig receptor specificity and affinity takes place in peripheral lymphoid organs where antigen-stimulated B lymphocytes activate a somatic hypermutation machinery that specifically targets the immunoglobuhn variable (~ regions. During this process, B cell clones with mutant Ig receptors of higher affinity for the inciting antzgen are stimulated into clonal proliferation and maturation into antibody-secreting plasma cells (renewed m 1).
In recent years, recombinant DNA technology has been used to mimic many aspects of the processes that govern the generation and selection of ~5 natural human antibody repertoires (renewed m 2,3) For instance, the construction of large repertoires of antibody fragments expressed on the surface of filamentous phage particles and the selection of phages by "pannrng"
on antigens has been developed as a versatile and rapid method to obtain antibodies of desired specificities (reviewed in 4,5). Further optimization of the affinity of individual phage antibodies has been achieved by creating mutant antibody repertoires that are expressed on bacteriophage particles and sampled for higher affinity mutants by selection for binding to antigen under stringent conditions (reviewed m 6). Various approaches have been used to create mutated antibody repertoires, including chain shuffling (7,8), error prone PCR (9), use of E, coli mutator strains (10) or approaches more specifically directed to the complementarity determining regions ("CDRs") of the antibody molecule, like CDR "walking" and parsimonious mutagenesis (11-13).
Libraries created so far have a more limited span of specificities than possible. This is in large part due to the fact that many specificities present are not expressed or exposed properly by the organism, for example, chosen for expression of the library components. This is most likely due to a lack of adaptation of the expression products to the expression environment.
Furthermore, the libraries created so far, if they contain a desired specificity, require engineering of the nucleic acid encoding the specificity m order to be able to create a fully human monoclonal antibody. For instance, in single chain Fv molecules, the light chain encoding sequence and the heavy chain encoding sequence are separated from the linker sequence and 20. separately inserted into a complementary part of a heavy chain encoding sequence and a light chain encoding sequence. Upon this rearranging of the variable parts, specificity and affinity may change The present invention solves these problems at least in part. Other advantages and embodiments of the present invention will be clear from the detailed description below.
Disclosure of the Invention The invention now provides libraries which comprise binding molecules that are adapted to expression in the expression organism, but which are also transferable to a human context without undergoing a change in conformation and/or build up.
Thus, the present invention provides a method for producing a human monoclonal antibody, said method comprising: providing a library of binding molecules, the binding domain of which consists essentially of human heavy chain variable fragments in a functional format, selecting from said library at least one heavy chain variable fragment having a desired binding affinity, and inserting a nucleic acid encoding said heavy chain variable fragment into a ' nucleic acid encoding the complementary part of at least a, heavy chain of said human monoclonal antibody, allowing for expression of the resulting heavy chain and for assembly of said heavy chain with a desired light them, and producing a human monoclonal antibody. The present inventors have found that having only a heavy chain derived variable fragment determining the binding affinity of the binding molecules in the library, that, as long as they are presented. m a functional format, this will suffice for creating a library at least as large as the known ones, but typically will allow for producing even larger libraries. Also, the loss of specificities and affinities because of expression problems can be reduced, especially according to the preferred embodiments as disclosed herein below. A heavy chain variable fragment is defined as anything based on a fragment the size of a CDR (complementarity determining region) of a heavy chain (e.g., CDR 3) to a heavy chain variable fragment as usually defined m the art. Also, the way the heavy chain variable fragments are encoded, allows for the direct insertion into a (preferably) standard complementary part of the heavy chain encoding nucleic acid without '~5 significantly altering its conformation, affinity and/or specificity The resulting heavy chain (upon expression) can then be assembled with a (preferably standard) fight chain. However, this light chain will typically not have any significant binding affinity for the molecule recognized by the heavy chain varrable fragment ' The nucleic acids encoding the heavy and light chains of the resulting human monoclonal antibody may be the same or different. They typically are expressed in a eukaryotic cell, preferably a human cell, preferably a cell like PER.C6. It may be either transient expression or from insertions in the host cell's genome; the latter being preferred.
Detailed Description of the Invention In a preferred embodiment, the methods of the invention are carried out in a manner wherein the heavy chain variable fragment is in a functional format through fusion to a structural protein designed for that purpose. A
functional format means that its conformation is such that it retains it binding affinity whether it is in phage display, or in its normal heavy chain environment. Methods of keeping heavy chain variable fragments in such a conformation are an important aspect of the present invention. It is disclosed herein how to provide amino acid sequences capable of simulating the conformation of the heavy chain variable fragment in phage display surroundings the way they are in the natural surroundings. One way is fusyng~
a variable fragment with a known affinity to random sequences, expressing the resulting nucleic acids and selecting for the known affinity.
In another preferred embodiment, the equality of the conformation of the phage display fragment and the fragment m the heavy chain environment is removal of at least one sequence which is responsible for associating with a light chain. In this format, an indifferent light chain variable fragment can be used as a structural amino acid sequence. According to the invention, the 2~ heavy chain variable fragment is preferably inserted into a standard human heavy chain encoding nucleic acid, derived from a human antibody backbone which is prevalent in the population, these include, but are not limited to members of the VHI, VH3 or VH4 gene families.
The same is true for the light chain. These include, but are not limited to members of the Vkappal, Vkappa3 and Vlambda3 gene familes WO 02/28903 ~ PCT/NLO1/00670 This way, the invention provides a kit of parts consisting of heavy chain variable fragments having the desired binding affinity to cut from the library and a set of ready to use monoclonal antibody encoding nucleic acids to insert them m.
Thus, the invention also provides a human monoclonal antibody obtainable by a method according to the invention as disclosed above. As explained previously herein, the invention provides a method for producing a structural amino acid sequence or a nucleic acid sequence encoding such an amino acid sequence for keeping a human heavy chain variable fragment in a functional format upon expression of a nucleic acid encoding such a fragment in a fusion with a nucleic acid encoding a protein expressed associated with the surface of a phage particle, comprising fusing a nucleic acid sequence encoding a possible structural amino acid' sequence to a nucleic acid which is a fusion of a human heavy chain variable fragment with a known binding affinity and the nucleic acid encoding a protein expressed associated with the surface of a phage particle and expressing said' nucleic acid in the context of a suitable phage expression system and selecting fusions which expose the desired binding affinity. The fusions in functional alignment basically mean that the order in which the sequences are present can be different and be functional. The heavy chain variable fragment and the structural amino acid sequence encoding parts should be next to each other, in either direction. The phage surface protein encoding nucleic acid can be on either side. The linkage may be direct or indirect The ammo acid sequence designed for keeping a heavy chain variable fragment in the proper conformation mll work for other heavy chain variable fragments as well. The invention thus also includes these ammo acid sequences (proteinaceous substances) and their encoding nucleic acids. Thus, one can make a library of heavy chain variable fragments in proper conformation, because of the presence of the novel structural sequence.
The invention further comprises a method for making a library for use m a method according to the invention, comprising cloning a number of randomized nucleic acids derived from a heavy chain variable fragment in functional alignment with a nucleic acid encoding a proteinaceous substance as disclosed hereinabove, and providing the resulting nucleic acid in functional alignment with a nucleic acid encoding a protein expressed associated with the surface of a phage particle and expressing the resulting nucleic acids comprising said heavy chain variable fragment, the proteinaceous substance encoding acid and said surface protein encoding nucleic acid in the context of a suitable phage expression system, thus producing said library. The invention also provides a phage display library obtainable by a method disclosed above.
Examples Example 1 Generation of a library of heavy chain variable regiozis using a soluble variable heavy chain 3 domain (sVH3).
The phagemid PDV UO3 is the basis vector for generating a library of binding molecules consisting of variable heavy chain 3 domains. A nucleic acid ~0 sequence of the phagemid PDV U03 is given m figure I. Instead of gVIIIp protein m the PDV U03 vector gIIIp can also be used.
The core of the soluble VH3 domain is given in figure 2. The dots indicate places, representing CDR1 and CDR2 m an unaltered VH domain, where through varying the ammo acid sequence, VH domains of various ~5 binding specificities can be obtained. The place marked "CDR3" in the figure, also indicates a place where through varying ammo acids, VH domains comprising varrous binding specificities can be obtained. Of course said CDR3 regions may vary m size, at least according to the natural VH3 size variation m CDR3 By varying the amino acid sequence in the CDR regions it is possible 30 to generate VH3 domains with varying specificities. The solubility of sVH3 versus an unmodified VH3 is due to mutations m framework 2 and framework 3, said mutations leading to a change in the hydxophobicity of the VH3 domain such that the hydrophilicity of the mutated VH3 domain increases. The solubility of sVH3 allows the generation of a phage comprising a binding molecule consisting of a VH domain in the absence of a light chain.
Libraries of binding specificities based on sVH3 domains can be generated by methods known in the art as long as the basic ammo-acid sequence given in figure 2 is used. Other amino-acid sequences then given in figure 2 can also be used provided that they result in a sufficiently soluble VH3 domain. A person skilled m the art can arrive at the library by for instance chosen primers with at least partial overlap and building an ever larger part of the sVH3 domain by consecutively amplifying resulting product with a further partially overlapping primer. The CDR3 domain being located at the extreme end of the VH domain requires attention in the amplification procedure. Preferably, one or more (partially overlapping) primers are used that result in a restriction site being present at the extreme end of the amplified product such that the resulting sVH3 library can easily be cloned into PDV U03. A preferred combination of enzymes to clone the library into PDV U03 is NcoI and XhoI, wherein NcoI is located near the leader in PDV
U03 that is fused to the start of the sVH3 domain.
The resulting phagemids are electroporated into E. coli TG1 or XL1-blueTEN. The bacteria are plated onto smtable culture plates that include 5% glucose. The next day the resulting colonies are collected and stored. Several of these collections are inoculated m liquid medium and helper ',5 phages. After 1 night at degrees 30 C the phages are harvested. The resulting phages are selected for the appropriate target and amplified using said bacteria. The amplified phages were sequenced and shown to be as expected.

WO 02/28903 $ PCT/NLO1/00670 Generation of a structural protein capable of supporting proper VH3 function.
The phagemid PDV U02 is the basis vector for generating a library of binding molecules consisting of variable heavy chain 3 domains further comprising a structural protein (SP) capable of supporting VH3 function. (SP
does not comprise intrinsic antigen binding capacity). The sequence of a first SP (SP1) is obtained by shortening the binding loops of CDR1 and CDR2 in the light chain VK3 such that the binding properties are destroyed but the heavy 1'0 chain supporting function of the light chain is essentially left intact.
This is achieved by deleting amino acid from CDR1 and CDR2 such that these CDRs do not contain antigen binding capacity. In this Example, the 4 amino acids representing amino acid 28-31 are omitted from CDRl. These amino acids represent the most variable region in the CDR1 region of VK1 (012). From CDR2, 3 amino acids, representing amino acid 53-55 in VKl (012) are omitted.
VKl CDR3 is replaced by a VSV-tag. The VSV-tag used contains the amino acid sequence YTDIEMNRLGK. A nucleic acid encoding SP1 was generated synthetically using assembly PCR and the correctness of the nucleic acid sequence was verified by sequencing. The nucleic acid contains a Notl site and a SaeI site such that cloning of SP1 into PDV U02 does not disrupt the reading frame of the gIII protein The Notl site is located near the putative N-termmal part of SP1.
SP2 was generated based on VK3 (A27) by omitting the 5 amino acids representing ammo acid 28-31A are omitted from CDR1. These amino acids represent the most variable region in the CDR1 region of Vtc3 (A27). From CDR2, 3 amino acids, representing ammo acid 53-55 in VK3 (A27) are omitted.
The CDR3 of VK3 (A27) is replaced by a VSV-tag. The VSV-tag used contains the ammo acid sequence YTDIEMNRLGK. A nucleic acid encoding SP2 was generated synthetically using assembly PCR and the correctness of the nucleic acid sequence was verified by sequencing The nucleic acid contains a NotI site and a SacI site such that cloning of SP2 into PDV LT02 does not disrupt the reading frame of the gIII protein. The NotI site is located near the putative N-terminal part of SP2.
A VH3 framework and CDR1 and CDR2 randomized region used in this example is depicted in figure 4. The nucleic acid sequence encoding this VH3 framework is also given in figure 4. This nucleic sequence is optimized fox codon usage in both E. coli and human cells. Table I depicts nucleic acid sequences that are optimized for codon usage in E. coli and human cells. The nucleic acid sequences encoding the framework are flanked by restriction sites x0 NcoI and Xh.oI such that the reading frame of the gIII protein is left intact.
The framework is cloned into PDV U02 using the sites indicated.
The resulting phagemids containing either SPI together with the framework or SP2 together with the frame work are electroporated into E. cola TGI or XLI-blueTEN. The bacteria are plated onto suitable culture plates that I5 include 5% glucose. The next day the resulting colonies are collected and stored. Several of these collections are inoculated in liquid medium and helper phages. After I night at 30 degrees C, the phages are harvested. The resulting phages are selected for the appropriate target and amplified using said bacteria. The amplified phages were sequenced and shown to be as expected.

WO 02/28903 I~ PCT/NLO1/00670 References 1 Berek, C., & Milstein, C. 1987. Mutation drift and repertoire shift in the maturation of the immune response. Immunol. Rev. 96:23.
2 Winter, G. & Milstein, C. 1991. Man-made antibodies. Nature. 349:293.
3 Vaughan, T.J., Osbourn, J.K., & Tempest, P.R. 1998. Human antibodies by design. Nat. Biotechnol. 16,535.
4 Winter, G., Griffiths, A.D., Hawkins, R.E., & Hoogenboom, H.R. 1994.
I0 Making antibodies by phage display technology. Annu. Rev. Immunol. 12:433.

5 Burton, D.R., ~ Barbas, C.F. 1994. Human antibodies from combinatorial libraries. Adv.Immunol. 57:I91.

6 Hoogenboom, H.R. 1994. Designing and optimizing library selection strategies for generating high-affinity antibodies. Trends in Biotechnol.
I5 15:62.

7 Marks, J.D., Griffiths, A.D., Malmqvist, M., Clackson, T., Bye, J.M., &
Winter, G. 1992. Bypassing immunisation: high affinity human antibodies by chain shuffling. Bio/Technology. 10:779.

8 Clackson, T., Hoogenboom, H.R., Griffiths, A.D., & Winter, G. 199I.
20 Making antibody fragments using phage display libraries. Nature.. 352:624.

9 Hawkins, R.E., Russel, S.J., & Winter.G. 1992. Selection of phage antibodies by binding affinity: mimicking affinity maturation. J.Mol.Biol.
226:889.
Low, N.M., Holliger, P.H., & Winter, G. 1996. Mimicking somatic 25 hypermutation: affinity maturation of antibodies displayed on bacteriophage using a bacterial mutator strain. J.Mol.Biol. 260,359.
I1 Barbas, C.F., Hu, D., Dunlop, N., Sawyer, L., Cababa, D., Hendry, R.M., Nara, P.L., & Burton, D.R. I994. In vitro evolution of a neutralizing human antibody to human immunodeficiency virus type 1 to enhance affinity and 30 broaden strain cross-reactivity. Proc. Natl. Acad. Sci. USA. 91:3809.

12 Yang, W.-P., Green, K., Pinz-Sweeney, S., Briones, A.T., Burton, D.R., &
Barbas, C.F. 1995. CDR walking mutagenesis for the affinity maturation of a potent human ant-HIV-1 antibody into the picomolar range. J. Mol. Biol.
254:392.
l3Balint, R.F., & Larrick, J.W. 1993. Antibody engineering by parsimonious mutagenesis. Gene. 137:109.

CODON USAGE ~IN E.COLI ANO H.SAPIENS
Amirioacid Preferential Alternative Classic . ~ ' Modern codon codons' Ala ~ A GCC GCT GCA

Cys C TGC TGT

Asp ~ D GAT ' GAC

Glu E GAA GAG

Phe F TTC TTT

Gly G GGC

His H CAC CAT

Ile I ATC ATT

Lys K AAA AAG

Leu L CTG

Met ' M ATG

Asn N AAC AAT

Pro P

Gln Q

Arg R CGC

Ser S AGC AGT TCC TCT

Thr T ACC

Val V GTG GTG

Trp W TGG

Tyr Y ~ TAC TAT

*: The codon usage in E.coli and H.sapiens does not correspond. The use of CCG is strongly preffered in E.coli while the proline codon in H.sapiens -is strongly biased for CCC. The codon advised for the Phage Display Technique is CCG as the E.eoli forms the basis of the selection. After maturation of the COR-regions the proline codons might be replaeed by CCC, This way both E.coli and human cell=lines may optimal synthesize the desired single chain or other antibody products.
19~5°FI°~'°lJ"i"~ ~ll~~~"~' ~Rl.ifL.E 26)

Claims

1. A method fox producing a human monoclonal antibody, said method comprising:
providing a library of binding molecules, the binding domain of which consists essentially of human heavy chain variable fragments in a functional format, selecting from said library at least one heavy chain variable fragment having a desired binding affinity, and inserting a nucleic acid encoding said heavy chain variable fragment into a nucleic acid encoding the complementary part of at least a heavy chain of said human monoclonal antibody, allowing for expression of the resulting heavy chain and for assembly of said heavy chain with a desired light chain, and producing a human monoclonal antibody.

2. The method according to claim 1 wherein said heavy chain variable fragment is in a functional format through fusion to a structural protein designed for that purpose.

3. The method according to claim l, wherein at least one sequence of said heavy chain variable fragment relevant only for association with a light chain is removed.

4. The method according to claim 1, wherein the complementary part of the heavy chain is derived from VH3, VH4 or VH1.

5. The method according to claim 1, wherein the light chain is derived from to a member of a Vkappa1, Vkappa3 and Vlambda3 gene family.

6. A human monoclonal antibody obtainable by a method according to claims 1-5.

7. A method for producing a structural amino acid sequence or a nucleic acid sequence encoding such an amino acid sequence for keeping a human heavy chain variable fragment in a functional format upon expression of a nucleic acid encoding such a fragment in a fusion with a nucleic acid encoding a protein expressed associated with the surface of a phage particle, said method comprising:
fusing a nucleic acid sequence encoding a possible structural amino acid sequence to a nucleic acid which is a fusion of a human heavy chain variable fragment with a known binding affinity and said nucleic acid encoding a protein expressed associated with the surface of a phage particle and expressing said nucleic acid in the context of a suitable phage expression system and selecting fusions which expose the desired binding affinity.

8. A proteinaceous substance or a nucleic acid encoding it, which substance is capable of keeping a heavy chain variable fragment in a functional conformation, obtainable by a method according to claim 7.

9. A method for making a library for use in a method according to claim 1, saia method comprising:
cloning a number of randomized nucleic acids derived from a heavy chain variable fragment in functional alignment with a nucleic acid encoding a proteinaceous substance according to claim 8, and providing the resulting nucleic acid in functional alignment with a nucleic acid encoding a protein expressed associated with the surface of a phage particle and expressing the resulting nucleic acids comprising said heavy chain variable fragment, said proteinaceous substance encoding acid and said surface protein encoding nucleic acid in the context of a suitable phage expression system, thus producing said library.

10. A phage display library obtainable by a method according to claim 9.