Modified recombinant human IgG1‐Fc is superior to natural intravenous immunoglobulin at inhibiting immune‐mediated demyelination

Abstract Intravenous immunoglobulin (IVIG) is an established treatment for numerous autoimmune conditions. Although Fc fragments derived from IVIG have shown efficacy in controlling immune thrombocytopenia in children, the mechanisms of action are unclear and controversial. The aim of this study was to dissect IVIG effector mechanisms using further adapted Fc fragments on demyelination in an ex vivo model of the central nervous system–immune interface. Using organotypic cerebellar slice cultures (OSCs) from transgenic mice, we induced extensive immune‐mediated demyelination and oligodendrocyte loss with an antibody specific for myelin oligodendrocyte glycoprotein (MOG) and complement. Protective effects of adapted Fc fragments were assessed by live imaging of green fluorescent protein expression, immunohistochemistry and confocal microscopy. Cysteine‐ and glycan‐adapted Fc fragments protected OSC from demyelination in a dose‐dependent manner where equimolar concentrations of either IVIG or control Fc were ineffective. The protective effects of the adapted Fc fragments are partly attributed to interference with complement‐mediated oligodendroglia damage. Transcriptome analysis ruled out signatures associated with inflammatory or innate immune responses. Taken together, our findings show that recombinant biomimetics can be made that are at least two hundred‐fold more effective than IVIG in controlling demyelination by anti‐MOG antibodies.

N563A/C575A. Briefly, the N glycans from 50 µg of each sample was released by incubation with NEB Rapid™ PNGase F and isolated from peptides using Sep-Pak C18 cartridges (Waters) as described previously 1 . The released N-glycans were permethylated, prior to Matrixassisted laser desorption ionization (MALDI) MS analysis. Data were acquired using a 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems) in the positive ion mode. The data were analyzed using Data Explorer (Applied Biosystems) and Glycoworkbench 2 . The proposed assignments for the selected peaks were based on composition together with knowledge of biosynthetic pathways.
The observed binding to CD22 was particularly surprising as this receptor prefers a-2,6 linked neuraminic acid and not a-2,3 linkages attached by CHO-K1 cells, although proximity-labeling experiments have recently shown that glycan-independent interactions of CD22/Siglec-2 with Ig in the BCR is possible (61).
The insertion of multiple glycan sites into the Fc, in particular at Asn 221 , enables new receptor interactions that are not possible with solely Asn 297 -directed approaches (Fig. 11A). For example, we generated the di-glycan D221N/C309L/N297A/C575A mutant that displayed marked binding to Siglec-1 and Siglec-4 (MAG), either FcgRs or complement proteins (Tables II, III) yet was highly effective at blocking hemagglutination by influenza A virus (Fig. 10C).
As glycan-mediated binding is essential for the influenza virus to infect cells of the respiratory tract, mutations in HA that lead to loss of receptor binding are unlikely to survive any neutralizing Abs induced during an immune response (Fig. 11B). Modeling of the D221N/C575A mutant shows that the distance from the N-terminal to the C-terminal tips of the Fc is ∼60Å (Fig. 11B), which is the same distance between the sialic acid-binding domains on the HA trimer (65). The Asn 221 and Asn 563 sugars located at the tips of the Fc are not constrained by their location within the Fc, as with Asn 297 , and would therefore be expected to be highly mobile and flexible with respect to searching out the HA-binding pocket.
Alternative anti-influenza therapeutic strategies are urgently needed. The use of IVIG during the 2009 and 1918 pandemics reduced mortality from influenza by 26 and 50%, respectively (66, 67), and a recent randomized, placebo-controlled study suggests these figures may be improved by enhancing influenza-specific Abs in IVIG (Flu-IVIG) preparations (36). As Flu-IVIG is manufactured in advance of future epidemics, there may be modest or no neutralizing activity against emerging strains. Combinations of Flu-IVIG or neuraminidase inhibitor drugs with Fc sialic acidbinding domain blockers may enhance the efficacy of Flu-IVIG or OSCs were demyelinated with 5 µg/mL anti-MOG antibody and either 6% rabbit (n = 8-17 OSC per group) or 11% human complement (n = 3-5 OSC per group) for three days in the absence or presence of 1, 3 or 10 µg/mL of three different glycosylated Fc leads (D221N/C309L/N563A/C575A; N563A/C575A and D221N/C57A) or the equimolar amount of WT-IVIG. D221N/C309L/N563A/C575A and N563A/C575A inhibit demyelination at 10 µg/ml independent of the use of human or rabbit complement. D221N/C57A only protected significantly when demyelination was induced with human complement. Demyelination was quantified by GFP fluorescence emitted from myelin of OSCs prepared from transgenic mice expressing GFP under regulatory elements of the myelin protein PLP. Significances were calculated in respect to the demyelinated control using one-way analysis of variance and Dunnett's post hoc test. *P ≤ 0.05. Values are depicted as mean ± SEM.   Complement binding and activation was assessed by ELISA on Fc coated wells as described previously (23-25). Serial dilutions of Fc from 10 µg/ml were coated onto maxi-sorb 96 ELISA plates at 100 µl/well (Fisher Scientific, NC9229197) in 0.05M carbonate buffer (Sigma C3041) overnight at 4 o C. Following washing with PBST (PBS-Tween-20 0.05%), plates were incubated with human serum diluted 1/100 in gelatin veronal buffer (Sigma G6514) for 1h at room temperature. After washing in PBST, plates were incubated with either a 1/250 dilution of sheep anti-human C1q-HRP (Serotec: 2221-5004P) or a 1/500 dilution of mouse anti-human C5b-9 mAb (Serotec: MCA2611) in PBST supplemented with 0.5% w/v BSA (Sigma) for 2h. For detection of the C5b-9 binding mAb, plates were washed as before prior to incubation for 2h with a 1:500 dilution of anti-mouse IgG2b-HRP (Southern Biotech: 1090-05). Plates were then washed ten times in PBST before developing for HRP substrate 3,3',5,5'-tetramethyl-benzidine dihydrochloride (Sigma T3405) in phosphate-citrate buffer (Sigma P4922). Reactions were stopped with 25µl 2M H2SO4 and plates read on a plate reader at OD450nm. Duplicate experiments are shown.  Receptors were coated down on ELISA plates (Nunc) in carbonate buffer pH 9 (Sigma-Aldrich) at 2 µg/ml overnight at 4°C, unless otherwise specified. The plates were blocked in PBS/0.1% Tween-20 (PBST) containing 5% dried skimmed milk. Plates were washed three times in PBST before adding Fc mutant proteins at the indicated concentrations and left at 4°C overnight. Plates were washed as above and incubated for 2h with 1:500 dilution of an alkaline phosphatase-conjugated goat F(abʹ)2 anti-human IgG (Jackson Laboratories). Binding of the secondary detecting Fab'2 anti-human Fc was checked by direct ELISA to every mutant to ensure there were no potential biases in the detection of binding of different mutants to different receptors. Plates were washed and developed with 100 µl/well of a Sigmafast p-nitrophenyl phosphate solution (Sigma-Aldrich). Plates were read at 405nm, and data plotted with GraphPad Prism. n=2 independent experiments.  Methods describing the binding of mutants to tetrameric human DC-SIGN (Elicityl), Siglec-1, Siglec-4, and Siglec-3 (Stratech Scientific), Siglec-2, CD23, dec-1, dec-2, clec-4a, clec-4d, MBL and MMR (Stratech Scientific or Bio-Techne) have been described previously (23-25). Error bars represent SD around the mean; n=2 independent experiments.

Fig. S8. Microglia depletion via clodronate strongly reduces immune-mediated demyelination.
Six days prior to demyelination, OSC was either pre-treated with 300 µg/ml of clodronate (dissolved in H2O; Sigma-Aldrich, Hamburg, Germany) or left untreated. In both groups, OSC were demyelinated with 5 µg/mL anti-MOG antibody and 8% rabbit complement (rbC) in the presence or absence of 10 µg/mL of the N563A/C575A mutant or were left untreated.
Quantification of the GFP+ area relative to day 0 (n = 4-12 OSC per group). Significant differences between treatment groups were calculated with the student's t-test.    Arvcf E030030I06Rik Atoh8