Invented by Laura Simmons, Genentech Inc
The Genentech Inc invention works as followsThe present invention provides methods of producing humanized antigens and increasing the yields of antibodies or antigen-binding fragments in cell culture. In one aspect, an amino acid residue from the framework region of the variable is replaced by a corresponding one from a subgroup of the variable sequence consensus that shares the greatest sequence identity with HVR1 or HVR2 of the variable sequence. In another aspect, a position adjacent to a cys that is involved in an intrachain disulfide bonds of the variabledomain is occupied by an amino that corresponds with an amino found in that position within a variabledomain consensus sequence subgroup which has the highest sequence identity to the HVR1 or HVR2 amino acids sequences.
Background for Methods of producing humanized antibodies and increasing yield of antibodies or Antigen Binding Fragments in Cell Culture
Antibodies and humanized antibodies have proven to be very useful in diagnostics and therapy.” Humanized antibodies are antibodies where CDRs (or hypervariable regions) from non-human antibodies are combined with framework regions of humans to form antigen binding molecules. The exchange is also known as a “CDR swap”. There are several ways to select human framework sequences. One method is to select a human variable-domain sequence with a framework sequence very similar to the nonhuman antibody used as the source for the CDRs. A human variable domain consensus is used as a source for the human framework region. A straight CDR exchange does not always result in molecules with high affinity for antigens. Therefore, additional modifications or changes are needed to increase the affinity of humanized antibodies. Humanization can be a time-consuming process due to the need for additional modifications. Humanization can also result in antibodies that are not produced at high yields in cell culture.
Some uses of antibodies require large amounts of fully assembled, full-length antibodies. There are many techniques available to produce antibodies recombinantly. These include E. coli and yeast as well as plant, insect, or mammalian cell systems. In large-scale antibody production, both eukaryotic and protokaryotic systems were used. E. coli is a particularly useful organism for the production of engineered antibodies such as humanized ones. E.coli expression systems have several advantages, including a convenient and well-studied gene technology that allows constructs to easily be expressed directly, as well as the relatively easy and inexpensive large-scale production made possible by E.coli’s fast growth and fermentation.
Full length antibodies are composed of two heavy chains that are linked by disulfide and two light chains. Each light chain is linked by a disulfide to one heavy chain. Each chain contains a variable domain at its N-terminus (VH or LV) and a constant domain or domains near the C-terminus. The constant domain in the light chains is disulfide-bonded to the heavy chains first constant domain, while the light-chain variable domain is aligned and disulfide-bonded with the heavy-chain variable domain. Each variable domain of the heavy chain and light chains includes framework regions and hypervariable region (HVRs), as well as an intrachain disulfide. (See e.g. Chothia et al., J. Mol. Biol. Biol. 186:651-663, (1985); Novotny & Haber, Proc. Natl. Acad. Sci. USA 82:45924596 (1985); Padlar et al., Mol. Mol. Immunol. 23(9) 951-960, 1986; Padlar et al. J. Mol. Biol., 216:965-973 (1990). Antibody fragments can also be produced, and often include combinations of heavy chain and light-chain variable domains to form an antibody binding site. Antibody fragments can include, for instance, Fab, F(ab?) and Fab? Antibody fragments include, for example, Fab, Fab?, F(ab? fragments.
Generally, prokaryotic antibody production involves the synthesis of light and heavy chain in the cytoplasm. This is followed by the secretion of the chains into the periplasm for processing. The heavy and light chain can also be directed to form inclusion bodies in the cytoplasm. The folding of light and heavier chains takes place in tandem with the assembly of folded light and heavier chains into an antibody molecule. During these folding and assembling processes, multiple covalent and noncovalent interactions take place between and within heavy and light chain. The efficiency and fidelity in these processes can have a significant impact on the yield of antibodies. After synthesis of heavy and light chain, protein aggregation can occur or even proteolysis. This will reduce the yield of antibody.
The production and stability of fragments of antibodies have been studied in greater depth than full-length antibodies. The stability and/or yields of scFv fragments or Fabs of natural antibody produced in host cell have often been found to not be sufficient. Honneger et al., J. Mol. Biol., 309:687-699 (2001). The stability of the antibody or fragment of an antibody when incubated in physiological conditions is critical for therapeutic efficacy. To increase the yield of therapeutic antibodies or fragments, it is important to improve folding efficiency and production yields. It is not always the case that expression yield of scFvs in bacterial periplasm correlates with their stability. Worn et al., J. Mol. Biol., 305:989-1010 (2001). Some stable scFv fractions only show poor expression yields within bacterial periplasm, and some mutations may affect folding efficiency in vivo but not stability. Worn et al., supra. “The many factors that influence the periplasmic yield of expression and/or the stability of scFv have not been fully understood.
The structural features that are thought to be responsible for the stability or in vivo fold of antibody fragments were described previously. In bacteria, for example, it was found that the FR1 of antibodies fragments influences in vivo fold of antibody fragments. de Haard et al., Prot. Eng., 11: 1267-1276 (1998). De Haard et. al. It was suggested that mutations in residue 6 of the heavy chain interfered the folding of the scFv. de Haard et al. supra. Jung et al. The amino acids at positions H6,H7, and H10 (H9) have been used to describe four different conformations for the FR1. Jung et al. J. Mol. Biol., 309:701 (2001). Mutations in these residues can disrupt the FR1 structure, particularly at residue 6. This can negatively affect folding yields and stability. Jung et al. supra. It is believed that residue 6 in the heavy chains contributes to the stability Fab, Fv and ScFv fragments without disulfide bond. Langdyk et al., J. Mol. Biol., 283:95 (1998). Disulfide bonding also contributes to the stability antibody domains. The H6 residue stabilized the scFv when disulfide was removed but not restored. Other point mutations in residues were described as either stabilizing or destabilizing specific scFvs. Worn et al., supra. The effect of a particular mutation in an antibody or in the antibody framework can be unpredictable, and it may or may affect in vivo fold efficiency. “A mutation in a specific residue of an antibody or antibody fragment may be beneficial for folding and yield in one case, but not in another.
Producing high affinity humanized antibody can be time-consuming, and the resultant antibodies are not ideal for cell culture production. When antibodies or antibody fragments are produced in cell cultures, multiple factors can affect their yield and/or stabilities. These factors are unpredictable and not well understood. “There is still a need to improve the process for producing humanized antibody and the yield of antibodies in cell culture. This is especially true for bacterial cell cultures.
The present invention relates to methods of improving the humanization of an antibody or antigen-binding fragment, and improving the yield of antigen-binding fragments or antibodies in cell culture. This is especially true for bacterial cell cultures. The invention is based upon the discovery that the sequence of the variable primary domains of antibodies can be modified or designed to improve the folding, assembly and yield of antigen binding fragments or antibodies. “The invention involves not only identifying which residues are to be substituted, but also identifying the substitutions that should be made at these residues in a predictable manner in order to improve yield of antibodies.
One method of humanizing antibodies is to combine HVRs from a human antibody with framework regions derived from human consensus sequences that are derived from the subgroups of heavy and light chains most frequently occurring in the sequence compilation by Kabat et al., Sequences Of Proteins Of Immunological Interest NIH, 1991. The most common heavy and light chains consensus sequences do not always produce an antibody which can be produced with high yields in cell culture. In one embodiment of the invention, the invention provides for a method of selecting a human subgroup sequence for atleast one of the framework regions or regions by identifying the subgroup sequence that shares the most sequence similarity with the HVR1 sequence and/or HVR2 of the nonhuman antibody. The method can reduce the time required to produce a high-yield humanized antigen or antibody binding fragment in cell culture.
The method for producing a humanized or recombinant antibody or antigen-binding fragment may include expressing in a cell host a variable-domain containing at least one FR-sequence from a human subgroup-consensus sequence and an HVR1 or HVR2-sequence from a nonhuman antibody, where the human subgroup-consensus sequence is that human subgroup-consensus sequence which has the HVR1 or HVR2 a
In other embodiments, the method includes comparing the HVR1 or HVR2 sequences of the nonhuman antibody to the HVR1 or HVR2 sequences of the human consensus subgroups that corresponds to the heavy or the light chain. Then, selecting the subgroup consensus human variable domain sequence that is most similar to the HVR1 or HVR2 sequences of the nonhuman antibody and using at least one FR from that sequence as the sequence for the antigen-binding fragment or recombinant The recombinant antibodies can be prepared using a single FR, or a combination of several FRs selected from FR1, 2, FR3, and FR4 as well as mixtures. In one embodiment, the humanized antibody or fragment of antigen binding includes the selected consensus sequence of the heavy chain subgroup of human that shares the highest sequence identity with the HVR1 and/or the HVR2 of non-human antibodies. This method of improving the humanization process may result in a humanized antigen-binding fragment or antibody that can be produced with high yields in cell culture and in less time.
In one aspect, the invention provides a method for producing an antigen or antibody binding fragment with high yields in cell culture. The method consists of expressing the variable domain of an antibody or binding fragment that contains at least one modified FR, in which the modified amino acid has been substituted for at least one position in the human subgroup consensus sequence, the different amino acids being the amino acids found at the corresponding position in the HVR1 or HVR2 amino sequences with the highest sequence identity, and the recovered antibody or binding fragment variable comprising modified framework is obtained from the host cells.
In one aspect, the invention provides a method for producing an antibody that has a higher yield in cell culture. The method involves expressing in a cell host a variable-domain of an antibody or antigen-binding fragment that contains at least one modified FRS. This modified FR is formed by replacing at least one amino in the corresponding position of a consensus human subgroup variable sequence, which has the HVR1 or HVR2 sequence with the highest sequence identity. The modified FR of the antibody or antigen-binding fragment has a higher yield in a culture compared to the parent variable domain. It is also recovered from the host cells.
The method includes: “In one embodiment the method comprises comparing the HVR1 or HVR2 of a parental antibody or antigen-binding fragment variable domain to the corresponding HVR1 or HVR2 of each of the human subgroup consensus sequences, and selecting the consensus that has the highest sequence identity; b), identifying at the least one different amino acid from the amino at the corresponding place of the human subgroup consensus sequence selected; and c), substituting the identified at least amino in step (b)
At lease one FR from a heavy chain or light domain variable or both may be modified or chosen for use according to the invention’s methods. Modifications may be made to 1 FR, or multiple FRs selected from FR1,FR2,FR3,FR4 and their mixtures. In a FR at least one amino acid is substituted, preferably multiples. In one embodiment, the amino acids at the positions of the selected consensus subgroup sequence in the selected framework region residues are substituted for all the framework regions residues that are different from the parent antibody.
The method can also include expressing the heavy-chain variable domain of an antibody or antigen-binding fragment that contains at least one modified framework in a cell host, in which the modified framework has at least one amino acids position replaced with a different one, in which the different amino is the one found at the corresponding position in the human heavy chain subgroup consensus sequence, which has the HVR1 or HVR2 amino sequence with the highest sequence identity to the corresponding HVR1 or HVR2 amino sequence of the heavy-
Another aspect of the invention is a method to improve the yield of an antibody or antigen-binding fragment in culture. The method consists of modifying the at least one FR of a variable subgroup of an antibody or binding fragment to be at least 50% identical to the corresponding FR of a subgroup consensus sequence. This modified FR contains a substitution at at least one position of amino acids with a different one, the different amino is found at the corresponding FR of a subgroup consensus variable domain sequence.
Another aspect of the invention is a method to produce an antibody or antigen-binding fragment with high yields in cell culture. The method consists of expressing in a cell host a modified version of an antibody or binding fragment’s variable, which has been modified to replace at least one amino-acid position that is part of a disulfide link in the intrachain of the variable, with another amino-acid, the other amino-acid being the amino-acid found in the corresponding position in the human subgroup consensus variable sequence, that has the HVR1 or HVR2 sequence with the highest sequence identity, and the recovery of the modified variant.
Another aspect of the invention is a method to improve the yield of antigen or antibody binding fragments in culture. The method includes: a. identifying an amino-acid position within a first variabl domain of an antigen or antibody binding fragment parent that is adjacent to a cys forming an intrachain disulfide link in the first variabl domain parent; b. selecting a consensus variable domain sequence that has the highest sequence identity to the HVR1 or HVR2 amino-acid sequences of the firstvariabl domain parent; c. placing a different acid at that position, the different amino In one embodiment, the amino acid at the position proximal of a cys is substituted by the amino acid at the corresponding place in the human subgroup consensus. In one embodiment, the amino acid closest to a cys can be modified before being incorporated into an antibody or antigen-binding fragment.
Another aspect of the invention is expressing a polynucleotide in a cell host. The method includes expressing in a cell a polynucleotide that encodes a variable region of an antibody or binding fragment, which comprises at least a modified FR, in which the modified amino acid has been substituted in at most one amino-acid position with a different one, in which the different amino-acid is the amino-acid found in the corresponding FR positions of a consensus sequence for the human subgroup variable, with the highest sequence identity to the HVR1 or HVR2 corresponding HVR2 FR FR FR FR FR FR, the at least FR, the at least FR, the at least a modified FR, the HVR, corresponding HVR1 or HVR2 corresponding to the HVR, and/or HVR2 corresponding corresponding HVR2 corresponding HVR1 or HVR2 corresponding HVR1 or HVR2 amino corresponding HVR2 corresponding variable domain.
In one embodiment, the method includes expressing in a cell host a polynucleotide that encodes a modified variation domain of the antigen or antibody binding fragment, wherein a different amino is substituted at a position adjacent to a Cys residue that participates an intrachain disulfide bonds, and wherein this different amino is the amino found at the position of a consensus human subgroup variable sequence that has an HVR1 or HVR2 sequence with the highest sequence identity to the corresponding HVR1 or HVR1 or HVR1 or HVR1 or HVR1 or HVR1 or HVR1 or HVR1 or H VR1 or HVR1 or HVR1 or HVR2 corresponding to the HVR1 or HVR1 or HVR1 or HVR2 HVR1 or HVR1 or HVR2 HVR1 or HVR1 or HVR2 HVR1 or corresponding HVR1 or HVR1 or HVR2 aClick here to view the patent on Google Patents.