Method for mass humanization of non-human antibodies
The present invention relates to a method for producing a population of nucleic acids encoding at least one protein comprising at least one immunoglobulin variable domain having a non-human-derived CDR3 amino acid sequence embedded in essentially human framework sequences, as well as to a population of nucleic acids and a population of proteins relates thereto and uses thereof.
1. A library that comprises a plurality of different humanized antibodies, wherein each humanized antibody of the plurality of humanized antibodies comprises:
at least one immunoglobulin variable domain having a non-human complementarity determining region 3 (CDR3); and
a human framework scaffold that comprises a first human framework region (FR1), a second human framework region (FR2), a third human framework region (FR3), a fourth human framework region (FR4), a complementarity determining region 1 (CDR1), and a complementarity determining region 2 (CDR2);
wherein:
the non-human CDR3 is embedded in the human framework scaffold;
the FR1 and FR2 are interspaced by the CDR1;
the FR2 and FR3 are interspaced by the CDR2;
the FR3 and FR4 are interspaced by the non-human CDR3;
each amino acid position of either the CDR1 or CDR2 comprises an amino acid residue from either a human species or a non-human species;
each amino acid at each amino acid position of the CDR1 is selected by blending a first positional weight matrix (PWM) of amino acid positional variability from naturally occurring, non-human CDR1s and a second PWM of amino acid positional variability from naturally occurring, human CDR1s; wherein the first PWM is generated on a computer by calculating a first relative frequency of each amino acid at each position of the naturally occurring, non-human CDR1s, the second PWM is generated on the computer by calculating a second relative frequency of each amino acid at each position of the naturally occurring, human CDR1s, blending the first PWM and the second PWM on the computer produces a blended PWM that provides for amino acid variation observed in both human and non-human CDR1s; and
each amino acid at each amino acid position of the CDR2 is selected by blending a first positional weight matrix (PWM) of amino acid positional variability from naturally occurring, non-human CDR2s and a second PWM of amino acid positional variability from naturally occurring, human CDR2s; wherein the first PWM is generated on a computer by calculating a first relative frequency of each amino acid at each position of the naturally occurring, non-human CDR2s, the second PWM is generated on the computer by calculating a second relative frequency of each amino acid at each position of the naturally occurring, human CDR2s, blending the first PWM and the second PWM on the computer produces a blended PWM that provides for amino acid variation observed in both human and non-human CDR2s.
2. The library of claim 1 , wherein each amino acid sequence of the CDR1 comprises at least one amino acid from a non-human CDR1 amino acid sequence and at least one amino acid from a human CDR1 amino acid sequence, and each amino acid sequence of the CDR2 comprises at least one amino acid from a non-human CDR2 amino acid sequence and at least one amino acid from a human CDR2 amino acid sequence.
3. The library of claim 1 , wherein each amino acid sequence of the CDR1 or the CDR2 exhibits at least 30% sequence identity to a non-human CDR1 or a non-human CDR2 sequence, respectively.
4. The library of claim 1 , wherein each amino acid sequence of the CDR1 or the CDR2 exhibits at least 30% sequence identity to a human CDR1 or a human CDR2, respectively.
5. The library of claim 1 , wherein a similarity between a CDR1's or a CDR2's immunoglobulin variable domain and a CDR3's native immunoglobulin variable domain is increased with respect to a feature selected from the group consisting of amino acid sequence, cosmology, length, canonical structure, heavy/light interface mount angle, and any combination thereof.
6. The library of claim 1 , wherein two C-terminal amino acids of the FR2 are from non-human sequences.
7. The library of claim 1 , wherein two C-terminal amino acids of the FR3 are from non-human sequences.
8. The library of claim 1 , wherein each humanized antibody comprises a peptide that allows display of a humanized antibody on a virus, a cell, or a surface.
9. The library of claim 8 , wherein the peptide is A-agglutinin-binding subunit (Aga2p).
10. The library of claim 8 , wherein the peptide is protein III (pIII).
11. The library of claim 1 , wherein each humanized antibody of the plurality of humanized antibodies comprises a scFv, a Fab, a Fab′, F(ab′) 2 , or a Fv.
12. The library of claim 1 , wherein the sequence of the CDR1 or the sequence of the CDR2 does not comprise cysteine, methionine, and tryptophan residues.