Fig 1.
Humanization of 1E11 by CDR grafting.
Heavy chain (A) and light chain (B) amino acid sequence alignment of murine, human germline, and humanized 1E11. The CDR residues, defined according to the Kabat definition, are shown in bold. Amino acids are numbered according to the Kabat numbering scheme. Colons represent common residues between murine and human germline sequences.
Fig 2.
Humanized 1E11 shows equivalent binding properties and biological activity to the parental murine antibody.
A, The binding activities of hz1E11 and ch1E11 to HER2-ECD protein were analyzed by ELISA. Trastuzumab (TRA) was used as a positive control antibody against HER2 protein. B, The binding activity of the antibodies was analyzed by ELISA using recombinant HER2 sub-domain proteins. C, NCI-N87 cells were treated with antibodies for 4 days in the complete growth media and the cell viability was measured in duplicates (mean ± SD) using WST-1 reagent. The 100% viability was defined as the viability of the antibody-untreated wells. D, Mice bearing NCI-N87 xenograft tumors were treated with indicated dose of control antibody, trastuzumab, hz1E11, or trastuzumab plus hz1E11. Palivizumab was used as the isotype control antibody. Tumor volume (mm3) was expressed as mean ± SD (n = 6 mice/group). For clarity, only positive error bars are shown. **, P < 0.01; n.s., not significant as determined by two-way repeated measures ANOVA followed by Bonferroni post test.
Fig 3.
Three residues in CDR-H3 are important for antigen-antibody binding.
The binding activities of alanine scanning mutants of CDR-H3 (A) and CDR-L3 (B) were analyzed by ELISA. C, Quantitative dissociation kinetics of indicated mutants were analyzed by surface plasmon resonance. HER2-ECD protein was immobilized on a CM5 sensor chip followed by exposure to indicated mutant Fab antibodies. The 100% association was defined as the response unit (RU) at 200 seconds from the beginning of Fab injection. koff values were calculated using the BIAevaluation software.
Fig 4.
Library design, panning and binding properties of affinity-matured clones.
A, Design of the CDR-L3 and CDR-H3 libraries for affinity maturation. See text for detailed explanation. B, Sequence logo was generated using WebLogo program about unique CDR-L3 sequences from 4th round panning-2 (low-stringency panning) output of the CDR-L3 library (n = 39). C, The hz1E11 was immobilized on a CM5 sensor chip followed by sequential exposure to HER2-ECD-His and indicated antibodies using BIAcore2000. D, Trastuzumab was immobilized on a CM5 sensor chip followed by sequential exposeure to HER2-ECD-His and indicated antibodies using BIAcore 3000. E, The binding activities of the antibodies were analyzed by ELISA using recombinant HER2 sub-domain proteins.
Table 1.
Panning results of CDR-L3 and–H3 libraries.
Table 2.
koff measurement of selected clones.
Table 3.
Binding kinetics of the anti-HER2 antibodies. 1)
Fig 5.
Affinity-matured hz1E11 clones shows antitumor activity in HER2-overexpressing gastric cancer models.
NCI-N87 cells (A) and OE-19 cells (B) were treated with antibodies for 4 days in the complete growth media. The cell viability was measured in duplicates (mean ± SD) using WST-1 reagent. The 100% viability was defined as the viability of the antibody-untreated wells. C, NCI-N87 (n = 5 mice/group) cells were inoculated into mice and antibody treatments started when tumor volumes reached approximately 200 mm3. Mice received a dose of 20 mg/kg for single agent treatment and 10 mg/kg of each antibody for combination treatment. Administration days are indicated by arrows. Tumor volume (mm3) was expressed as the mean ± SD. For clarity, only positive error bars are shown.