Table 1.
Functional characterization of primary humanized 4A11 variants.
| Variants | Human germline acceptor framework | TGFβ binding by Biacore SPR (KD: pM)d | TGFβ inhibition by HEK-Blue cells (IC50: nM)e | Blockingc | |||||
|---|---|---|---|---|---|---|---|---|---|
| LC | HC | TGFb1 | TGFb2 | TGFb3 | TGFb1 | TGFb2 | TGFb3 | ||
| rbt4A11 | – | – | > 5000a | 1.6 | 1.2 | > 667b | 0.9 | 0.1 | Complete |
| h4A11.v1 | IGKV4-1*01 | IGHV3-48*01 | > 5000a | 4.6 | 14.4 | > 667b | > 667b | > 667b | Incomplete |
| h4A11.v2 | IGKV1D-39*01 | IGHV3-48*01 | > 5000a | 1.4 | 13.3 | > 667b | > 667b | > 667b | Incomplete |
| h4A11.v3 | IGKV4-1*01 | IGHV4-59*06 | > 5000a | 9.6 | 12.1 | > 667b | 3.9 | 0.2 | Complete |
| h4A11.v4 | IGKV1D-39*01 | IGHV4-59*06 | > 5000a | 10 | 10.1 | > 667b | 1.9 | 0.1 | Complete |
| h4A11.v5 | IGKV3-20*01 | IGHV3-48*01 | > 5000a | 12.1 | 13.9 | > 667b | > 667b | > 667b | Incomplete |
| h4A11.v6 | IGKV2-24*01 | IGHV3-48*01 | > 5000a | 8.4 | 12.5 | > 667b | > 667b | > 667b | Incomplete |
| h4A11.v7 | IGKV3-20*01 | IGHV4-59*06 | > 5000a | 6.3 | 9.7 | > 667b | 0.9 | 0.2 | Complete |
| h4A11.v8 | IGKV2-24*01 | IGHV4-59*06 | > 5000a | 8.2 | 14.1 | > 667b | 4.8 | 0.7 | Complete |
aThe binding was not observed at the highest TGFβ1 concentration (5 nM); therefore, the reported KD are lower limits.
bThe maximun inhibition did not achieve over 90% at the highest antibody concentration (667 nM); therefore, the reported IC50 are lower limits.
cThe complete blocking or incomplete blocking was defined by TGFβ2 and TGFβ3 inhibition greater than 90% or less than 70%, respectively, at the highest antibody concentration.
dThe affinity constant (KD) for each variant is calculated using a global fitting model and reported as a single value for the whole data set with an average fitting standard error of 10%.
eThe half-maximal inhibitory concentration (IC50) for each variant from quadruplicate experiments is calculated using nonlinear regression with a four-parameter variable slope curve fit and reported as a single value with an average fitting standard error of 10%.