An engineered immunomodulatory IgG1 Fc suppresses autoimmune inflammation through pathways shared with i.v. immunoglobulin

Abstract

Immunoglobulin G (IgG) antibodies in the form of high-dose intravenous immunoglobulin (IVIG) exert immunomodulatory activity and are used in this capacity to treat inflammatory and autoimmune diseases. Reductionist approaches have revealed that terminal sialylation of the single asparagine-linked (N-linked) glycan at position 297 of the IgG1 Fc bestows antiinflammatory activity, which can be recapitulated by introduction of an F241A point mutation in the IgG1 Fc (FcF241A). Here, we examined the antiinflammatory activity of CHO-K1 cell-produced FcF241A in vivo in models of autoimmune inflammation and found it to be independent of sialylation. Intriguingly, sialylation markedly improved the half-life and bioavailability of FcF241A via impaired interaction with the asialoglycoprotein receptor ASGPR. Further, FcF241A suppressed inflammation through the same molecular pathways as IVIG and sialylated IgG1 Fc and required the C-type lectin SIGN-R1 in vivo. This contrasted with FcAbdeg (efgartigimod), an engineered IgG1 Fc with enhanced neonatal Fc receptor (FcRn) binding, which reduced total serum IgG concentrations, independent of SIGN-R1. When coadministered, FcF241A and FcAbdeg exhibited combinatorial antiinflammatory activity. Together, these results demonstrated that the antiinflammatory activity of FcF241A requires SIGN-R1, similarly to that of high-dose IVIG and sialylated IgG1, and can be used in combination with other antiinflammatory therapeutics that rely on divergent pathways, including FcAbdeg.

Keywords: Autoimmune diseases; Immunoglobulins; Immunology; Immunotherapy.

Figure 1. Manipulation of Fc glycosylation of CHO-K1–produced Fc^F241A.

(A) Sequence alignment of WT human IgG1 Fc (Fc^WT) and Fc^F241A, with the gray box designating residue 241 and the black box designating N297 in both sequences. Graphic of human IgG Fc structure with positions of F241 (gray hexagon) and N297 (black hexagon) marked, showing their relative proximity and location within the interior of the Fc structure. (B) Schematic representing the different N-linked glycoforms that may be present on Fc N297 and their nomenclature. Blue squares, N-acetylglucosamine (GlcNAc); yellow circles, galactose; green circles, mannose; red triangles, fucose; purple diamonds, sialic acid. (C) CHO-K1 cells were transfected with plasmids for either Fc^WT or Fc^F241A, and then N297 glycoforms on the purified Fc products were analyzed via HPLC. (D) Percentages of N297 glycans from CHO-K1 cells expressing Fc^F241A (green), Fc^F241A transfected with ST6GAL1 (Fc^F241A/ST6, blue), Fc^F241A transfected with B4GALT1 and ST6GAL1 (Fc^F241A/B4ST6, purple), and Fc^F241A transfected with siRNA against SLC35A1 (Fc^F241A/siSLC, maroon). Fc glycoforms were analyzed via HPLC. Bar graphs are plotted as means with SDs in C and D.

Figure 2. Sialylation of Fc^F241A is not necessary for antiinflammatory activity but improves half-life and bioavailability in vivo.

(A) Female WT C57BL/6 mice (n = 5) were given K/BxN serum alongside PBS, IVIG, Fc^F241A, Fc^F241A/ST6, Fc^F241A/B4ST6, or Fc^{F241A/SLC35A1} in a preventative manner, and swelling was scored for 10 days. Day 8 clinical scores, representing maximum separation between PBS and Fc^F241A-based treatments, are plotted. (B) Female humanized FcRn mice (n = 6) were dosed with 20 mg/kg of Fc^F241A, Fc^F241A/ST6, Fc^F241A/B4ST6, or Fc^F241A/siSLC. Serum concentration of hIgG Fc was measured via ELISA. Half-life, area under the curve from the time of dosing to the last measurable concentration (AUC_last), and clearance rate were calculated. (C) Correlation between sialylation on Fc^F241A and clinical scores, half-life, AUC_last, and clearance. Corresponding R² and P values are shown. Clinical score correlation was generated from A; other correlations were generated from B. (D) Female WT mice (n = 4) were given control or ASGPR block before Fc^F241A/B4ST6 or Fc^F241A/siSLC. Serum concentration of hIgG Fc was measured via ELISA at 1, 3, and 7 days. Fc^F241A/B4ST6 and Fc^F241A/siSLC were also analyzed for ASGPR binding on Octet. (E) NTN was induced in female WT mice (n = 3–5) that were given PBS, IVIG, or varying doses of Fc^F241A. Day 7 serum was used to quantify anti-sheep mIgG via ELISA, represented as absorbance at 450 nm. Day 7 blood urea nitrogen (BUN) levels for NTN mice were also quantified. (F) Female WT C57BL/6 mice (n = 6–10) had EAE induced and were treated with PBS, IVIG, or 100 mg/kg Fc^F241A on days 5, 10, 15, and 20. EAE clinical scores were recorded daily. Data are plotted as means with SDs in A, B, and D–F. Statistics are ordinary 1-way ANOVA with Tukey’s multiple comparisons (A, B, and D–F) or simple linear regression (C).

Figure 3. Divergent antiinflammatory pathways are elicited by Fc^F241A/B4ST6 and Fc^Abdeg.

(A) Sequence alignment and schematics of Fc^WT and Fc^Abdeg, with the gray boxes designating the locations of Abdeg mutations and the black box designating the location of N297. (B) Dissociation constants (K_D, μM) and association constants (K_A, M^–1 s^–1) determined by surface plasmon resonance (SPR) of Fc^WT, Fc^F241A/B4ST6, and Fc^Abdeg with mFcRn and hFcRn are plotted. (C) Female WT C57BL/6 (n = 3–9) and humanized homozygous FcRn (Tg32) mice (n = 3–4) were given 1 dose of 50 or 100 mg/kg of Fc^F241A/B4ST6 or 10 mg/kg of Fc^Abdeg, and serum mouse IgG was measured via ELISA out to day 7 post-dose. (D) Fold change of cell surface binding MFI of PBS, Fc^F241A/B4ST6, Fc^F241A/siSLC, or Fc^Abdeg to SIGN-R1^+/+, SIGN-R1^–/–, and hDC-SIGN⁺/SIGN-R1^–/– murine bone marrow–derived macrophages (BMDMs) as detected by FACS. Plot shows the fold change in binding, represented as MFI, in comparison with PBS. (E) Female WT C57BL/6 mice (n = 5 per group) and female and male SIGN-R1^–/– mice (n = 5) were given arthritogenic K/BxN serum alongside PBS, IVIG 1 g/kg, Fc^F241A/B4ST6 50 mg/kg, or Fc^Abdeg 10 mg/kg in a preventative manner, and joint swelling was clinically scored for 10 days. Day 7 clinical scores for each group, representing the maximum separation in clinical score between PBS and Fc^F241A-based treatments, are plotted. Bar graphs are plotted as means with SDs, and statistics are ordinary 1-way ANOVA with Tukey’s multiple comparisons in B, D, and E.

Figure 4. Combinatorial antiinflammatory activity of Fc^F241A/B4ST6 and Fc^Abdeg.

(A) Day 8 clinical scores of mice (n = 5) given arthritogenic K/BxN serum on day 0, then PBS, IVIG 1 g/kg, Fc^F241A/B4ST6 50 mg/kg, or Fc^Abdeg 10 mg/kg on day 2. (B) Day 10 serum concentrations of Fc^F241A/B4ST6 in mice (n = 4–5) given Fc^F241A/B4ST6 50 mg/kg alone or combined with increasing doses of Fc^Abdeg (1–10 mg/kg). (C) Relative serum IgG titers of mice (n = 3) given PBS or Fc^Abdeg 1 mg/kg, then serum mouse IgG levels days 1, 3, and 7 after administration. Data are normalized to PBS (100%). (D) Day 6 clinical scores of K/BxN-treated mice (n = 3–5) given PBS, IVIG 1 g/kg, Fc^F241A/B4ST6 50 mg/kg, Fc^Abdeg 1 mg/kg, or a combination of Fc^F241A/B4ST6 and Fc^Abdeg. (E) Day 9 clinical scores of K/BxN-treated mice (n = 5) given PBS, IVIG 1 g/kg, Fc^F241A/B4ST6 50 mg/kg, Fc^Abdeg 1 mg/kg, or a combination of Fc^F241A/B4ST6 and Fc^Abdeg on day 2. All bar graphs are plotted as means with SDs. Statistics are ordinary 1-way ANOVA with Tukey’s multiple comparisons (A, B, D, and E) or 1-tailed Mann-Whitney test (C).