Several high affinity antibodies for InsR selected from each library were also shown Dasatinib ic50 to be functionally active as IgG2s. Examples of three InsR agonist antibodies are shown in Fig. 6A. scFv20 activates InsR 27% relative to activation by insulin and has an affinity of 230 pM. Fab20 and Fab21, as IgG2s, activate InsR by greater than 50% with affinities of 24 pM and 27 pM, respectively. In addition to agonists, positive and negative allosteric modulating antibodies were identified

(Bhaskar et al., 2012) (Fig. 6B). The positive allosteric modulator, scFv21 as an IgG2, did not significantly activate InsR without insulin, however, in the presence of insulin there was a significant increase in AKT phosphorylation with the addition of antibody. As an IgG2, the affinity of scFv21 for InsR in the absence of insulin could not be determined because it bound so poorly, however, in the presence of insulin, the KD was 1.6 nM (data not shown). For the negative allosteric modulator, scFv22, a decrease in AKT phosphorylation was seen with the addition of antibody (IgG2) in the presence of insulin.

The binding affinity of scFv22 as an IgG2 (KD = 50 pM) to selleck InsR was the same in the presence or absence of insulin. The VH-CDR3 confers the majority of the contacts between the antibody and antigen, and is therefore a major contributor to the specificity and affinity of the antibody (Amit et al., 1986 and Kabat and Wu, 1991). To analyze the amino acid usage of the two libraries, the VH-CDR3 sequences of naïve and selected clones were compared to VH-CDR3 sequences from 6580 productive human antibody sequences from the IMGT database (Giudicelli et al., 2006). While the amino acid distributions of the naïve and selected clones were, statistically, not the same as the IMGT distribution (χ2 test, Oxalosuccinic acid p < 0.003), the same general trends were observed as have been previously observed for human VH-CDR3s (Zemlin et al., 2003) (Fig. 7). Notably, the selected clones appeared to have a greater percentage of

tyrosine and glycine, but fewer cysteines than the naïve libraries or the antibodies in IMGT. The XFab1 and XscFv2 libraries were constructed to capture the diversity of the human antibody repertoire. Each library was generated from secondary lymphoid tissue from thirty healthy donors, and both these libraries have more than 250 billion antibody fragments, making both libraries larger than any of the previously published libraries (Sblattero and Bradbury, 2000, Lloyd et al., 2009 and Urlinger, 2011). These libraries encompass the full spectrum of V-gene families from IgM, IgG, IgE, IgA and IgD classes of immunoglobulins. The random pairing of VH and VL domains within and between donors increases the diversity by producing novel antibodies with heavy and light chain combinations that do not exist within the donor pool.