Neonatal Fc Receptor Blockade by Fc Engineering Ameliorates Arthritis in a Murine Model

Although the focus of therapeutic approaches for autoimmunity has in the past been on targeting cellular immunity, considerable recent interest has been directed toward the humoral component for diseases in which Abs play a role in pathogenesis (1, 2). In many cases, current therapies for Ab-mediated diseases, such as systemic lupus erythematosus and myasthenia gravis, involve the use of immunosuppressive drugs or steroids, which have undesirable side effects (3, 4). Although high doses of intravenous immunoglobulin (IVIG) can be effective in ameliorating inflammatory diseases (5, 6), the use of this reagent can result in adverse events, such as immune complex-mediated damage (7, 8). Currently, there is also a worldwide shortage of IVIG (9). These shortcomings motivate the use of recombinant or synthetic approaches to develop new treatments.

A strategy that has recently been advocated to treat Ab-mediated autoimmunity is to design reagents that can lower Ab levels in vivo (10–12). This is of particular relevance because B cell depletion using Abs to target CD20 does not lower the levels of autoreactive IgGs sufficiently to modulate autoantibody-mediated disease due to the lack of CD20 expression by long-lived plasma cells (1, 13). It is well established that the MHC class I-related receptor, neonatal FcR (FcRn), regulates the levels and transport of Abs throughout the body (14–16). The inhibition of this receptor therefore provides a possible target for the therapy of IgG-mediated autoimmune diseases (10–12, 17–19). However, the ability of relatively low and single doses of FcRn blockers to ameliorate such diseases is uncertain. In addition, whether FcRn blockade alone is effective in treating Ab-mediated disease is a major issue regarding the feasibility of this strategy. For example, it is unclear whether other anti-inflammatory pathways, such as those involving FcγRIIB upregulation (20, 21), are necessary for therapeutic benefit. This question has been prompted by studies using high-dose IVIG to treat inflammatory disease: some studies support a role for FcRn blockade in the reduction of disease activity (19, 22, 23), whereas others have reported that the beneficial effects of IVIG are solely due to upregulation of the inhibitory FcγR, FcγRIIB (20, 21), or a combination of pathways involving both FcRn and FcγRIIB (24). Resolution of this issue is critical for the design of therapeutics for targeting Ab-mediated inflammatory disease.

Toward the goal of inhibiting FcRn function, we have recently described a class of Abs called Abdegs (Abs that enhance IgG degradation) that are engineered to bind with increased affinity to FcRn through their Fc region at both acidic and near neutral pH (10, 25). To date, however, the activity of Abdegs in treating Ab-mediated disease is untested. Wild-type Abs bind with very low affinity to FcRn at near neutral pH and are dependent on fluid-phase uptake for entry into cells, whereas Abdegs enter cells primarily by receptor-mediated processes (10, 26). Consequently, Abdegs compete very effectively with wild-type IgGs for FcRn interactions because they not only bind more strongly to this receptor at endosomal pH, but also accumulate within cells to much higher concentrations (10, 25). As such, these engineered Abs have a major competitive advantage for FcRn binding over existing lower-affinity, wild-type inhibitors such as IVIG. The binding properties of an Ab for FcRn also impact its in vivo t_1/2, with gain of binding at pH 7.4 resulting in lysosomal accumulation and lower in vivo persistence (25, 27, 28). Consequently, the competitive ability and in vivo t_1/2 of an Abdeg can be tuned for specific applications (16).

In this study, we have used a murine model of rheumatoid arthritis [via transfer of K/B×N serum (29)] to both analyze the therapeutic activity of Abdegs and address the fundamental question as to whether FcRn blockade without other contributing anti-inflammatory effects can alleviate ongoing, IgG-mediated disease. In this model, arthritis develops in normal mice following the transfer of anti–glucose-6-phosphate isomerase (anti-GPI) Abs that complex with endogenous GPI (30). These immune complexes infiltrate joints, where they initiate an inflammatory cascade within minutes following transfer (31–33). This model is therefore instructive for the analysis of therapies that target the humoral response. We demonstrate that Abdegs can effectively treat arthritis through a mechanism involving FcRn inhibition. Collectively, our observations indicate that the use of Fc-engineered Abs of the Abdeg class holds considerable promise for the treatment of arthritis, and by extension, other diseases in which Abs play a role in pathology.

K/B×N mice were generated by intercrossing KRN TCR transgenic mice (provided by C. Benoist and D. Mathis) with NOD mice (29). Offspring were typed by flow cytometric analysis of lymphocytes in the blood using an anti-Vβ6 Ab that recognizes the TCRβ chain of the transgene (29). Sera were harvested from 8- to 10-wk-old Vβ6⁺ K/B×N mice for use in transfer experiments. BALB/c and FcRn knockout mice (34) were purchased from The Jackson Laboratory (Bar Harbor, ME). All animal protocols have been approved by the Institutional Animal Care and Use Committee at the University of Texas Southwestern Medical Center (Dallas, TX).

Wild-type human IgG1 and the mutant MST-HN (Met²⁵² to Tyr, Ser²⁵⁴ to Thr, Thr²⁵⁶ to Glu, His⁴³³ to Lys, and Asn⁴³⁴ to Phe) are both specific for hen egg lysozyme and were purified using lysozyme-Sepharose (10, 35). To generate MST-HN + N297A, an Ab H chain expression construct encoding the MST-HN mutant in combination with N297A (to express in aglycosylated form) was transfected into a L chain-expressing NSO cell line (35). Ab-expressing clones were identified and expanded for protein expression, as described (10, 35). Binding properties of purified human IgG1 and MST-HN for FcRn were confirmed using surface plasmon resonance (BIAcore) (36) prior to use in vivo. IVIG (Gammunex) was obtained at 100 mg/ml from Talecris. The mouse anti-lysozyme Ab, D1.3 (37), was purified using lysozyme-Sepharose.

The effect of the MST-HN mutant on the clearance of ¹²⁵I-labeled wild-type mouse IgG1 (D1.3) was determined, as described (10). Drinking water was supplemented with 0.1% Lugol 72 h before radiolabeled mouse IgG1 was injected i.p. into BALB/c mice, and radioactivity was monitored at the indicated times by whole-body counting (25) (Atom Lab 100 Dose Calibrator). Seventy-two hours later, mice were i.v. injected with either PBS, 0.5, 1, or 2 mg MST-HN, and whole-body radioactivity was determined at the indicated times. Similarly, the effect of IVIG on the clearance of D1.3 was compared with that of MST-HN using an analogous approach, except that 72 h postdelivery of radiolabeled D1.3, mice were i.v. injected with either 1 mg wild-type IgG1 or MST-HN, or 25 or 50 mg IVIG, and whole-body counts were determined at the indicated times.

To analyze the effect of MST-HN on the clearance of mouse IgG1 in FcRn-deficient mice (34), mice were treated as above, except that a dose of 1 mg MST-HN or wild-type IgG1 was injected into FcRn-deficient mice 15–20 min following the i.v. delivery of radiolabeled mouse IgG1 (D1.3). β-phase t_1/2 values of radiolabeled D1.3 were determined by fitting the data to a decaying biexponential model using MATLAB (Mathworks) and custom written software (38).

Mouse IgG1 (D1.3) was radiolabeled, as described (25). Following i.p. injection of the radiolabeled Ab, counts in 10 μl blood were determined at the indicated times by scintillation counting.

Sera harvested from arthritic K/B×N mice were analyzed for anti-GPI titers using ELISAs. On day 0, arthritis was induced in BALB/c recipients by i.p. delivery of 150 μl K/B×N serum. Six hours following serum delivery, mice were treated with i.v. injections (200–250 μl per mouse) of MST-HN, wild-type human IgG1, IVIG, or PBS at the doses indicated. In cases in which 50 mg IVIG was used, i.v. injections (250 μl each) were delivered twice, first at 6 h postserum transfer and then after an additional 3–4 h. Mice were also treated either 72 h following, or 12 h prior to, K/B×N serum transfer with i.v. injections of 1 mg/mouse MST-HN, wild-type IgG1, or PBS. Ankle and paw thicknesses were determined using an analog caliper (Dyer 313 series) by two independent observers.

Inflammatory cell infiltration into ankle joints was visualized in hindlimbs isolated from mice 6 d following serum delivery. BALB/c mice were injected i.p. with 150 μl K/B×N serum and after 6 h treated i.v. with either 1 mg/mouse MST-HN, wild-type human IgG1, or PBS. Mice were euthanized, and hindlimbs were skinned and fixed in 10% neutral buffered formalin for ∼60 h. Following decalcification with Cal-Rite solution, tissue samples were embedded in paraffin wax, sagitally sectioned, deparaffinized, and stained with H&E. Slides were imaged and acquired using a Leica DM2000 upright photomicroscope (Leica Microsystems, Wetzlar, Germany) equipped with bright-field and epifluorescence optics and an Optronics Microfire color CCD camera (Optronics, Goleta, CA). A G4 Macintosh (OS 10.4.11) with Pictureframe 2.0 acquisition software was used to acquire and store images.

Serum samples of BALB/c mice prior to induction of arthritis and after treatment were obtained via retro-orbital bleeding at the indicated times. These samples were used to determine anti-GPI Ab levels, total IgG levels, and albumin levels using sandwich ELISAs and serum dilutions of 1:500, 1:4,000, and 1:50,000, respectively. For measuring anti-GPI Ab levels, 1.25 μg/ml recombinant GPI (Roche) was used to coat 96-well plates, and bound Ab was detected using HRP rabbit anti-mouse IgG (H and L chain specific; Dako). IgG and albumin levels were determined using ELISA quantitation kits for mouse IgG and albumin (Bethyl Laboratories).

A total of 1 μg each of recombinant Ab or IVIG was electrophoresed on 12% polyacrylamide gels with BenchMark Prestained protein ladder (Invitrogen). Electrophoresed proteins were either transferred onto a polyvinylidene difluoride membrane (Millipore; Immobilon-P) or stained with Coomassie brilliant blue. The membrane was incubated with 1% Western blocking reagent (Roche) for 1 h, followed by 1 μg/ml biotinylated Elderberry Bark Lectin (Sambucus nigra agglutinin [SNA]; Vector Laboratories) in Western blocking reagent for 1 h, and subsequently washed twice with TBS/0.1% Tween and once with TBS (pH 7.4). The membrane was then incubated with 2.25 μg/ml Extravidin-peroxidase conjugate (Sigma-Aldrich) in Western blocking reagent for 30 min prior to washing, as above. Bound SNA was detected by incubation with 0.6 mg/ml 3,3-diaminobenzidine tetrahydrochloride with 0.01% (v/v) hydrogen peroxide.

The n for each data set is indicated, and statistical significance for each time point was determined by ANOVA using MATLAB (Mathworks). ANOVA was carried out by paired comparisons between different mouse groups using the Tukey-Kramer test at a 95% confidence interval (α = 0.05). Two-tailed Student t test was used to compare two treatment groups. The p values <0.05 were considered to be significant, and actual values are indicated. Outliers due to experimental artifacts were removed from the analyses. Data shown are mean values of treatment groups with error bars indicating SEs (SEM).

The current study is directed toward analyzing the therapeutic efficacy of a mutated human IgG1 (MST-HN, Met²⁵² to Tyr, Ser²⁵⁴ to Thr, Thr²⁵⁶ to Glu, His⁴³³ to Lys, and Asn⁴³⁴ to Phe) (10) that is designed to inhibit FcRn function in a mouse model of Ab-mediated arthritis. We initially investigated how the dose of MST-HN Abdeg impacts the levels of competing wild-type tracer IgG in vivo. BALB/c mice were used throughout these studies because this mouse strain is highly susceptible to arthritis following the transfer of arthritogenic serum derived from K/B×N mice (39). Mice were injected i.p. with radiolabeled mouse IgG1 and, 3 d later, 0.5, 1, or 2 mg MST-HN per mouse was delivered by i.v. injection. A dose of 0.5 mg per mouse (∼25 mg/kg) induces a lower decrease in tracer IgG1 levels relative to 1 mg (∼50 mg/kg) (Fig. 1). By contrast, comparison of doses of 1 and 2 mg MST-HN per mouse indicates no significant differences, demonstrating that for FcRn inhibition in BALB/c mice there is no advantage in using doses >1 mg per mouse (Fig. 1).

Anti-GPI Abs, primarily of the IgG1 isotype, are known to be the mediators of disease in the K/B×N model (32). In the current experiments, we induced arthritis by transferring serum from arthritic K/B×N mice by i.p. delivery into BALB/c recipient mice. Prior to the delivery of potential therapeutics, initial analyses were therefore carried out to determine the time at which i.p. injected mouse IgG1 could be detected in the serum of BALB/c mice. The levels of radiolabeled IgG1 in the blood approach maximal levels at the earliest sampling point of 2 h, but continue to increase slowly to peak levels until 6 h postdelivery (Fig. 2). Positron emission tomography imaging has shown that anti-GPI Abs in the serum localize to joints at detectable levels within several minutes of delivery (33), indicating that extravasation to the target site is very rapid.

The therapeutic effect of treatment with 0.5 or 1 mg MST-HN (per mouse) on arthritis was initially assessed by transferring arthritogenic serum into BALB/c mice and treating 6 h later. Mice were treated with MST-HN (0.5 or 1 mg), or as controls with wild-type IgG1 (0.5 or 1 mg; same binding specificity as MST-HN) or PBS vehicle. Disease severity was monitored by determining ankle and paw thickness. The data shown in Fig. 3A and 3B clearly demonstrate that MST-HN treatment at a time point when substantial accumulation of anti-GPI:GPI immune complexes in joints is expected (Fig. 2) (33) results in significantly lower disease activity relative to controls (wild-type IgG1 or PBS). Although both 0.5 mg (∼25 mg/kg) and 1 mg (∼50 mg/kg) MST-HN induce therapeutic effects, the higher dose is more effective in treating disease. In addition, there was no significant difference between wild-type IgG1 and PBS treatment groups (Fig. 3A, 3B). Similar results were observed when paw swelling was used to assess disease activity (data not shown). Delivery of MST-HN to mice following injection of control (nonarthritogenic) serum did not have any impact on joint size (data not shown), indicating the specificity of the effects on autoantibody-mediated disease.

Histopathological analysis of ankle joints showed markedly less infiltration of immune cells in mice treated with MST-HN relative to mice in control treatment groups (Fig. 3C). As a result, no obvious bone damage was observed in MST-HN–treated mice, whereas mice treated with either wild-type IgG1 or PBS showed substantial inflammatory cell invasion and bone/tissue damage (Fig. 3C). We also determined the anti-GPI levels in sera of mice immediately prior to and following delivery of MST-HN (∼50 mg/kg), wild-type IgG1, or PBS (Fig. 3D). Disease activity was found to correlate with anti-GPI Ab levels, with MST-HN treatment resulting in a rapid decrease in these Ab levels within several hours of delivery.

To further investigate the therapeutic effects of MST-HN, we analyzed whether MST-HN delivery during established disease (3 d postserum transfer) ameliorated arthritis. Relative to control groups, MST-HN treatment resulted in a rapid attenuation of disease (Fig. 4A). In addition, we determined the effect of administration of MST-HN prior to K/B×N serum transfer. Prophylactic treatment with MST-HN 12 h prior to K/B×N serum transfer resulted in less severe disease relative to that observed in wild-type IgG1-treated mice (Fig. 4B). However, the reduction was not as pronounced compared with the effect of Abdeg delivery 6 h posttransfer of arthritogenic serum (Fig. 3B).

Earlier studies have indicated that sialylation through α2,6 linkage to the N-linked carbohydrate, present on 1–3% of IgG molecules in IVIG preparations, mediates anti-inflammatory effects by inducing the upregulation of FcγRIIB (20, 21, 40). It was therefore important to determine whether similar effects could contribute to disease amelioration induced by MST-HN. Immunoblotting with a lectin (SNA) specific for α2,6-linked sialic acid indicated that MST-HN, wild-type IgG1, and IVIG all bear similar levels of this type of sialic acid (Fig. 5). A possible contribution of this sialic acid form to the anti-inflammatory effects of MST-HN was therefore directly assessed by comparing the activity of this Ab with its parent, wild-type counterpart. Significantly, our data demonstrate that at a dose of 0.5 or 1 mg per mouse, wild-type IgG1 has no therapeutic activity (Fig. 3A, 3B), suggesting that the disease amelioration induced by the same doses of MST-HN involves FcRn inhibition.

To further investigate the FcRn dependence of MST-HN activity, the impact of MST-HN delivery on the clearance of ¹²⁵I-radiolabeled mouse IgG1 in FcRn knockout mice (34) was examined. No significant difference in the t_1/2 of radiolabeled IgG1 was observed in mice treated with either 1 mg MST-HN (30.4 h ± 1.15, SEM, n = 5 mice) or 1 mg wild-type IgG1 (32.2 h ± 1.05, SEM, n = 5 mice). These results are therefore consistent with the concept that a primary mechanism of Abdeg action is through FcRn inhibition.

In addition to the well-established role of FcRn in regulating IgG levels (14, 15), it also functions to transport and recycle albumin (41). The lack of overlap between the IgG and albumin binding sites on FcRn (42) has enabled us to use albumin levels to assess whether the MST-HN mutant impacts FcRn expression and/or downregulation. For example, it is conceivable that through high-affinity, bivalent binding to FcRn, MST-HN might promote lysosomal trafficking of this receptor. Importantly, albumin levels did not vary significantly between different treatment groups, whereas total IgG levels decrease (Fig. 6). These data demonstrate that the MST-HN mutant acts through competition with endogenous IgGs for binding to FcRn in the absence of significant effects on FcRn downregulation/degradation.

Comparison of the impact of 1 mg MST-HN and 1 mg IVIG on arthritis indicated that, by contrast with MST-HN, 1 mg IVIG had no significant therapeutic effect (data not shown). This result is not unexpected, because IVIG comprises primarily human IgG1, which at this low dose also does not alter disease incidence and severity (Fig. 3B). We therefore determined the dose of IVIG necessary to induce similar therapeutic effects as 1 mg MST-HN. Mice were treated with 1 mg MST-HN, 25 mg IVIG, 50 mg IVIG, or, as control, 1 mg wild-type IgG1, and disease activity was monitored (Fig. 7A). The reduction in ankle swelling in mice treated with 1 mg MST-HN was between that induced by 25 and 50 mg IVIG (∼1250 and 2500 mg/kg, respectively), although differences between MST-HN and IVIG treatment groups were not significant. Similar results were observed for paw swelling (data not shown). Consequently, ∼25- to 50-fold more IVIG is needed to induce the therapeutic benefit that is achievable with 1 mg MST-HN. Consistent with the observations in Fig. 3B and 3D, disease severity correlated with the amount of anti-GPI Abs present (Fig. 7B). In addition, MST-HN and IVIG lowered levels of competing mouse IgG1 under noninflammatory conditions in vivo (Fig. 7C).

Recent interest has focused on targeting the humoral arm of the immune response for the treatment of many autoimmune diseases (1, 2). In the current study, we demonstrate that recombinant Abs of the Abdeg class can be used as an effective therapeutic to treat ongoing disease in the K/B×N serum transfer model of Ab-mediated arthritis. Consistent with the ability of Abdegs to inhibit FcRn, disease blockade is accompanied by a reduction in serum anti-GPI levels. Abdegs therefore have considerable promise as therapeutic reagents for IgG-mediated autoimmunity.

Although IVIG is currently used to treat a number of autoimmune diseases, high doses (∼1–2 g/kg) are needed for efficacy (6, 43). This poses a challenge for use in the clinic that is exacerbated by the worldwide shortage of IVIG (9). In addition, the delivery of IVIG can result in adverse side effects due to the polyclonal nature of this therapeutic and immune complex-mediated damage (7, 8). These shortcomings have prompted the development of alternative approaches, such as the use of Fc-engineered Abs that bind with increased affinity to FcRn at endosomal pH and have longer in vivo t_1/2 (11). Delivery of these Abs alleviates autoimmunity in a murine model involving the passive transfer of human Abs from arthritic patients (11). Abs of this class compete poorly for FcRn due to the need for fluid-phase uptake into cells, and consequently, they require high and repeated dosing to show efficacy. Another strategy that has shown beneficial effects in a rat model of myasthenia gravis is to use Abs that block FcRn by binding through their variable domains (12). However, such Abs bind to FcRn with similarly high affinity at both pH 6.0 and 7.4, resulting in short in vivo t_1/2 and the need for repeat and frequent dosing (12).

An alternative approach that is described in this work is to use Fc engineering to generate Abdegs that bind through their Fc region with high affinity to FcRn in the pH range 6.0–7.4 (10). Consequently, relative to their wild-type counterparts, Abdegs accumulate in cells to higher concentrations through receptor-mediated uptake, where they outcompete wild-type IgGs for FcRn binding in endosomes (10, 25, 28). Nevertheless, due to the intrinsic pH dependence of FcRn–IgG interactions (27, 44), Abs of the Abdeg class such as MST-HN retain significantly lower affinity at pH 7.4 relative to pH 6.0 (10). Consequently, MST-HN persists with a t_1/2 in mice of ∼40 h (38), providing a significant advantage over the therapeutic use of shorter-lived Abs that bind to FcRn through their V regions (12, 17). Indeed, the available knowledge concerning how FcRn–IgG interaction properties impact the t_1/2 and FcRn-blocking activity of an Abdeg enables the optimization of Abdeg behavior for different applications (16, 25).

In the current study, treatment of mice 6 h following K/B×N serum transfer with a single dose of either 0.5 or 1 mg Abdeg ameliorates arthritis, with the higher dose resulting in maximal competitive effects and almost complete ablation of disease symptoms. By comparison with Abdeg, our studies show that 25- to 50-fold higher concentrations of IVIG are required to induce a comparable reduction in disease activity. Thus, the binding properties of Abdegs translate into potent therapeutic effects. Importantly, MST-HN delivery 72 h postserum transfer reduces disease activity when joints are visibly swollen, demonstrating therapeutic benefit in established disease. The potency of the Abdeg in reducing arthritis, however, was not as pronounced when administered prophylactically (12 h prior to K/B×N serum transfer) compared with delivery at 6 h postserum transfer. This observation is consistent with the relatively short t_1/2 of MST-HN (38), which is expected to render the prophylactically delivered Abdeg less effective in competing with anti-GPI Abs for FcRn binding.

Although in earlier studies the stoichiometry of FcRn binding to Fc (IgG) was uncertain, it is now established that the interaction comprises an asymmetric 2:1 FcRn:Fc (or IgG) complex (45–47). Relative to their homodimeric counterparts, heterodimeric Fc fragments with only one functional FcRn interaction site have shorter in vivo t_1/2 and are recycled less efficiently (48, 49). This, combined with our recent observations demonstrating that Abdegs enter the lysosomal pathway in cells (28), raises the question as to whether the presence of two relatively high-affinity binding sites on an Abdeg for FcRn could decrease recycling activity by diverting this receptor into lysosomes. In addition to transporting IgG, FcRn recycles serum albumin (41), a protein that is critical in maintaining oncotic pressure, blood pH, and transporting a range of small molecules. Importantly, the binding sites for albumin and IgG on FcRn are distinct (42). We therefore reasoned that by monitoring albumin levels, we could determine whether FcRn expression and/or intracellular trafficking were affected by Abdeg treatment. Our studies demonstrate that Abdegs function specifically through blocking FcRn–IgG interactions, and do not alter FcRn expression or distribution to impact albumin levels. The specificity of Abdeg effects for IgG indicates that the application of these reagents in the clinic will not result in the undesirable side effect of altering albumin homeostasis.

A question of direct relevance to the design of therapeutics is whether FcRn blockade alone is sufficient to treat Ab-mediated disease. For example, sialylation (α2,6-linked sialic acid) present on 1–3% of IgG molecules in IVIG has been shown to induce anti-inflammatory effects by upregulating FcγRIIB (21, 40). This upregulation has been reported to be either an important, or the sole, contributing factor to the therapeutic benefit of IVIG in the K/B×N serum transfer model (20, 21, 24). Our immunoblotting data indicate that the levels of α2,6-linked sialic acid on MST-HN or wild-type IgG1 are analogous to those present on IVIG, although the possibility of cross-reactivity of SNA with different forms of sialic/neuraminic acid on the recombinant Abs cannot be excluded (50). We therefore reasoned that comparison of the effects of the delivery of MST-HN and wild-type IgG1 on arthritis would enable us to assess the role of FcRn inhibition versus other possible anti-inflammatory pathways on the therapeutic effects of the Abdeg. Significantly, at the low doses (∼25–50 mg/kg) of recombinant Abs used in this study, wild-type IgG1 delivery does not modulate arthritis, suggesting that a primary mode of action of the MST-HN mutant is through FcRn blockade. This is also consistent with the inactivity of MST-HN in further reducing the t_1/2 of IgG in FcRn knockout mice, combined with the relationship between therapeutic efficacy and the ability of MST-HN or relatively high doses of IVIG to decrease Ab levels in vivo.

In the current study, we show that the use of Abs of the Abdeg class offers advantages over alternative approaches, including the ability of relatively low, single doses to treat disease when delivered following the initiation of immune complex-mediated inflammation. Collectively, our observations demonstrate that Abdegs represent highly efficacious and specific therapeutics that could have broad applicability in the management of IgG-mediated autoimmunity.

We thank Drs. Christophe Benoist and Diane Mathis for generously providing KRN transgenic mice. We are indebted to Dr. James Richardson and John Shelton (Molecular Pathology Core, University of Texas Southwestern Medical Center) for histologic preparations. We are grateful to Dr. Sripad Ram for expert advice on statistical analysis, Dr. Wentao Mi for advice on mouse husbandry, and Tom Lo for assistance.

E.S.W. is an inventor on a pending patent describing the use of Abdegs as FcRn inhibitors. The other authors have no financial conflicts of interest.