Cross-species analysis of Fc engineered anti-Lewis-Y human IgG1 variants in human neonatal receptor transgenic mice reveal importance of S254 and Y436 in binding human neonatal Fc receptor

Abstract

IgG has a long half-life through engagement of its Fc region with the neonatal Fc receptor (FcRn). The FcRn binding site on IgG1 has been shown to contain I253 and H310 in the CH2 domain and H435 in the CH3 domain. Altering the half-life of IgG has been pursued with the aim to prolong or reduce the half-life of therapeutic IgGs. More recent studies have shown that IgGs bind differently to mouse and human FcRn. In this study we characterize a set of hu3S193 IgG1 variants with mutations in the FcRn binding site. A double mutation in the binding site is necessary to abrogate binding to murine FcRn, whereas a single mutation in the FcRn binding site is sufficient to no longer detect binding to human FcRn and create hu3S193 IgG1 variants with a half-life similar to previously studied hu3S193 F(ab')2 (t1/2β, I253A, 12.23 h; H310A, 12.94; H435A, 12.57; F(ab')2, 12.6 h). Alanine substitutions in S254 in the CH2 domain and Y436 in the CH3 domain showed reduced binding in vitro to human FcRn and reduced elimination half-lives in huFcRn transgenic mice (t1/2β, S254A, 37.43 h; Y436A, 39.53 h; wild-type, 83.15 h). These variants had minimal effect on half-life in BALB/c nu/nu mice (t1/2β, S254A, 119.9 h; Y436A, 162.1 h; wild-type, 163.1 h). These results provide insight into the interaction of human Fc by human FcRn, and are important for antibody-based therapeutics with optimal pharmacokinetics for payload strategies used in the clinic.

Keywords: Antibody engineering; Fc receptors; molecular biology; neonatal Fc receptor; transgenic mice.

Figure 1.

Computational analysis of the FcRn interaction with hu3S193 Fc. (A) Ribbons-style representation of the model of the huFcRn complex with hu3S193 Fc. FcRn is composed of a MHC-like binding chain (red) and β-2-microglobulin (white). A single heavy chain of hu3S193 Fc (blue) with the residues involved in the interface with FcRn displayed as CPK spheres. (B) Free energy contribution plot for the residues of hu3S193 Fc interacting with huFcRn and muFcRn, respectively. Data were fitted to a linear relationship with an R² value of 0.82.

Figure 2.

ELISA analysis at pH 6.0 of binding of hu3S193 wild-type and variants to FcRn. Each panel compares binding of FcRn to wild-type hu3S193 with a series of different hu3S193 variants at different concentrations. (A-D) Binding obtained from hu3S193 wild-type and variants to muFcRn. (E-H) Binding obtained from hu3S193 wild-type and variants to huFcRn. n = 2; bars, SD.

Figure 3.

BIAcore analysis of the binding of immobilized biotinylated muFcRn and huFcRn to mutant and wild-type hu3S193 antibodies. (A) BIAcore analysis of the binding of immobilized biotinylated muFcRn at pH 6.0 to hu3S193 mutant antibodies (667 nM). Data shown is average maximum biosensor responses observed of 2 independent batches of antibody. (B) BIAcore analysis of the binding of immobilized biotinylated muFcRn at pH 6.0 to hu3S193 wild type antibodies (667 nM). Data shown is a typical BIAcore sensorgram of each antibody. (C) BIAcore analysis of the binding of immobilized biotinylated huFcRn at pH 6.0 to hu3S193 mutant antibodies (667 nM). Data shown is average maximum biosensor responses observed of 2 independent batches of antibody. (D) BIAcore analysis of the binding of immobilized biotinylated huFcRn at pH 6.0 to hu3S193 mutant antibodies (667 nM). Data shown is a typical BIAcore sensorgram of each antibody. 1, wild-type; 2, N434A; 3, Y436A; 4, H433A; 5, S254A; 6, S254A/Y436A; 7, I253A; 8, H435A; 9, H310Q; 10, H310A. n = 2; bars, SD.

Figure 4.

Blood clearance studies of ¹¹¹In-CHX-A″ DTPA-labeled hu3S193 wild-type and mutants in mice. (A) BALB/c nu/nu mice: H310A, H435A and I253A/H310A; (B) BALB/c nu/nu mice: S254A, S254A/Y436A, N434A, Y436A; (C) BALB/c nu/nu mice: I253D, I253P, H310D, H310E, H310Q; (D) huFcRn Tg mice: H310A, H435A and I253A/H310A; (E) huFcRn Tg mice: S254A, S254A/Y436A, N434A, Y436A; (F) huFcRn Tg mice: I253D, I253P, H310D, H310E, H310Q; n = 4–5; bars, SD.

Figure 5.

Key pharmacokinetic parameters calculated from single non-linear regressions (2-phase decay) from individual blood clearance curves obtained from BALB/c nu/nu mice and huFcRn transgenic mice. (A) BALB/c nu/nu mice: t_1/2β; (B) huFcRn Tg mice: t_1/2β; (C) BALB/c nu/nu mice: AUC; (D) huFcRn Tg mice: AUC. n = 3–6; bars, SEM.

Figure 6.

Correlation between RUmax values obtained by BIAcore in vitro, and in vivo parameters (elimination half-lives (t_1/2β) and AUC) obtained by blood clearance studies of hu3S193 IgG1 wild-type and variants in BALB/c nu/nu and huFcRn transgenic mice. (A) BALB/c nu/nu mice: RUmax and t_1/2β; (B) huFcRn Tg mice: RUmax and t_1/2β; (C) BALB/c nu/nu mice: RUmax and AUC; (D) huFcRn Tg mice: RUmax and AUC.