If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Bullous pemphigoid (BP) is an autoimmune blistering disease characterized by autoantibodies to COL17. Currently, systemic corticosteroids are used as first-line treatments for BP; alternatively, intravenous administration of high-dose IgG (IVIG) has been shown to be effective for patients with steroid-resistant BP in clinical practice. However, the effect of IVIG on BP has not fully been investigated. To examine the effects and mechanisms of action of IVIG against BP, we performed IVIG experiments using two experimental BP mouse models. One is a passive-transfer BP model that reproduces subepidermal separation in neonatal mice by the passive transfer of IgGs against COL17, such as polyclonal or monoclonal mouse IgG or IgG from BP patients. The other is an active BP model that continuously develops a disease phenotype in adult mice. IVIG decreased pathogenic IgG and the disease scores in both models. Injected IVIG distributed throughout the dermis and the intercellular space of the lower epidermis. Notably, IVIG inhibited the increase of IL-6 in both models, possibly by suppressing the production of IL-6 by keratinocytes. These results suggest that the inhibitory effects of IVIG on BP are associated with the reduction of pathogenic IgG and the modulation of cytokine production.
Bullous pemphigoid (BP) is characterized by tense blisters with itchy urticarial plaques and erythema that develop on the entire body. It is induced by autoantibodies to COL17 (also called BP180), a hemidesmosomal transmembrane protein at the dermal−epidermal junction (DEJ). The juxtamembranous extracellular noncollagenous 16A (NC16A) domain is preferentially recognized by autoantibodies in the sera of BP patients (
). The pathogenicity of IgG from BP patients (BP-IgG) was directly proven by our previous studies using COL17-humanized (COL17m−/−,h+) mice, which lack murine COL17 but express human COL17 in the skin (
). Thus, several complement-independent pathomechanisms also play important roles in blister formation in BP.
Currently, in clinical practice, systemic corticosteroids are used as the first-line treatment for moderate to severe BP, but long-term corticosteroid treatment may cause many dose-related adverse effects. Intravenous administration of high-dose IgG (IVIG) has also been reported as a safe, beneficial therapy for BP (
). In general, several mechanisms of anti-inflammatory effects for IVIG have already been proposed, such as (i) the inhibition of the recycling of pathogenic autoantibodies via neonatal Fc receptor saturation; (ii) the suppression or neutralization of pathogenic autoantibodies by anti-idiotypic antibody action; (iii) the non-specific blocking of the Fcγ receptor via the Fc portion derived from IgG preparations; and (iv) the modulation of cytokine production, such as decreases in TNF-α, IL-1α, IL-4, IL-6, IL-13, and IL-33, and an increase in IL-10 (
). However, the modes of action and the impact of IVIG on BP have not been fully investigated in vivo.
In this study, we used two experimental BP mouse models. One is a passive-transfer BP model that reproduces dermal−epidermal separation by the passive-transfer of IgG antibodies to COL17 into neonatal COL17-humanized mice (
). The other is an active BP model that was generated by the adoptive transfer of human COL17-immunized spleen cells into adult immunodeficient COL17-humanized mice. This model continuously produces anti-human COL17 IgG in a CD4+ T-cell−dependent manner and reproduces the BP phenotype (
). We demonstrate that IVIG reduces pathogenic antibodies to COL17, including polyclonal mouse IgG, monoclonal mouse IgG, and BP-IgG, and suppresses BP phenotype in experimental models. We also demonstrate the distribution of administered IVIG in the skin. Furthermore, we show that IVIG modulates the production of cytokines and chemokines, and that, of these, IL-6 is suppressed by IVIG both in vivo and in vitro. These findings clarify the mechanisms of action of IVIG in BP.
IVIG reduced circulating polyclonal antibodies to COL17 and skin fragility in the passive-transfer BP model
To examine whether IVIG suppresses skin fragility that is simply induced by IgGs to COL17, we performed an IVIG experiment using an IgG passive-transfer BP model. First, antibodies to COL17 were isolated from the pooled sera of wild-type mice that had been immunized by grafting human COL17-expressing transgenic mice skin to their back (SG-IgG). We treated the neonatal mice with saline at 400 mg/kg/d (IVIG-400) or at 2,000 mg/kg/d (IVIG-2,000) at 2 hours prior to the injection of 150 μg/g of SG-IgG, and we examined them at 18 hours after injection (n = 15, n = 10, and n = 15, respectively). All of the saline-treated or IVIG-400−treated mice showed skin fragility and dermal−epidermal separation histologically (Figure 1a). Notably, 40% of the mice treated with IVIG-2,000 prior to SG-IgG were resistant to skin fragility (positive/total = 9/15). The intensities of IgG and C3 deposition at the DEJ of the skin were similar in all groups (Figure 1a). Inflammatory cell infiltration was not obvious in any of the mice. IVIG significantly decreased circulating antibodies to the NC16A domain of COL17 in a dose-dependent manner (Figure 1b). Titers of injected human IgG remained elevated for 48 hours after the IVIG-2,000 (Figure 1c). These findings show that passively transferred human IgGs persist for at least 2 days in mice, reduce circulating antibodies to the NC16A domain and mitigate skin fragility in the passive-transfer BP model.
IVIG reduces the binding of monoclonal antibodies against the NC16A domain of COL17 to the antigen
Next, we performed IgG passive-transfer experiments by using monoclonal mouse IgG1 to the NC16A domain of COL17 (TS39-3) (
) instead of SG-IgG, which consists of polyclonal antibodies to various epitopes on COL17. In this setting, we were able to directly evaluate the effect of IVIG on the BP model induced by an antibody to a single pathogenic epitope. Neonatal mice were treated with saline or IVIG-2,000 at 2 hours prior to the injection of 50 μg/g of TS39-3 and were examined at 24 hours after injection. All the saline-treated mice showed skin fragility (Figure 1d). Meanwhile, 86% of the mice treated with IVIG-2,000 prior to TS39-3 were resistant to skin fragility (positive/total = 1/7) (Figure 1d). Notably, the intensities of IgG deposition at the DEJ were significantly lower in the IVIG-2,000−treated mice than in the saline-treated mice (Figure 1e). Interestingly, human IgGs injected as IVIG distributed diffusely throughout the dermis, the subcutaneous tissues, and the intercellular space of the lower epidermis, but not at the DEJ (Figure 1d). IVIG reduced circulating TS39-3 as determined by ELISA (Figure 1f), although it failed to reduce the amount of total mouse IgG in the plasma (Supplementary Figure S1 online). Thus, IVIG eliminates the pathogenic antibodies and reduces the skin fragility in a passive-transfer BP model.
IVIG suppresses the skin fragility induced by BP-IgG
To examine the effect of IVIG in a more clinically relevant condition, we examined the effect of IVIG on a passive-transfer BP model induced by BP-IgG. Neonatal mice were treated with saline or IVIG-2,000 at 2 hours prior to the injection of 750 μg/g of BP-IgG and were examined at 24 hours after injection (n = 7 and 8, respectively). All of the saline-treated mice showed skin fragility, whereas 75% of the mice treated with IVIG-2,000 prior to BP-IgG were resistant to skin fragility (positive/total = 2/8) (Figure 2a). The deposition of BP-IgG at the DEJ was difficult to evaluate because both the BP-IgG and the IVIG were human IgGs (Figure 2a). Notably, the intensities of C3 deposition at the DEJ were significantly diminished in the IVIG-2,000−treated mice (Figure 2a, 2b). IVIG reduced circulating BP-IgG as determined by ELISA (Figure 2c). Thus, IVIG reduced BP-IgG, suppressed C3 deposition, and prevented skin fragility in the passive-transfer BP model.
IVIG modulates inflammatory cytokines and chemokines in the passive-transfer BP model
To further examine the mechanism of action of IVIG, we measured plasma cytokine and chemokine levels at different time points in the SG-IgG passive-transfer BP model. Most of the examined cytokines and chemokines, including IFN-γ, TNF-α, IL-1β, IL-2, IL-6, IL-10, IL-13, IL-17, eotaxin, G-CSF, GM-CSF, MIP-1α, MIP-1β, and MCP-1, increased after SG-IgG injection, whereas IL-1α, IL-12p40, and RANTES did not. IVIG-2,000 significantly decreased IL-6, IL-17, and MIP-1β at the early phase (Figure 3). Eotaxin also tended to be decreased at the early phase, and IFN-γ, MIP-1α, and MCP-1 tended to be decreased at the late phase by IVIG-2,000, although not significantly. Conversely, IL-10, an anti-inflammatory cytokine, was elevated in the IVIG-2,000−treated mice, although not significantly (Figure 3).
IVIG reduces disease severity and circulating autoantibodies in the active BP model
We next tried to examine the effect of IVIG on the active BP model that mimics the human BP phenotype better than the passive-transfer BP model does (
). Immunized spleen cells reacting with COL17 were intravenously transferred into Rag-2−/−/COL17-humanized mice at day 0. This model produces antibodies to the NC16A domain, as well as to the other domains of COL17. The mice were intravenously administrated two different doses of IVIG from day 1 to day 21 daily at 400 mg/kg/d (IVIG-400, n = 10) or 2,000 mg/kg/d (IVIG-2,000, n = 14), or with saline (n = 13), and the mice were clinically evaluated during a 5-week period. The saline-treated mice developed the disease phenotype, including skin detachment, erythemas, erosions, blisters, and crusts, especially on the face, ears, and chest and back, at day 14, and the disease scores peaked at day 35 (Figure 4a, 4b). Of note, the IVIG-2,000−treated mice showed significantly lower disease scores than the saline-treated mice showed at every observation point from day 14 to day 35. Histopathologic analysis of the back skins at day 35 revealed that both IVIG-400 and IVIG-2,000 reduced crust formation (Figure 4c). Notably, the IVIG-2,000−treated mice showed significantly lower titers at day 11 than the saline-treated mice showed. The effect became unclear at day 21 (Figure 4d). The intensities of IgG and C3 deposition at the DEJ were similar between the saline-treated and IVIG-2,000−treated mice 5 weeks after transfer (Figure 4e). The concentration of human IgGs administered as IVIG in the plasma was higher in the IVIG-2,000−treated mice than that in the IVIG-400−treated mice or the saline-treated mice, and the injected human IgGs were mostly eliminated within a week after the final injection of IVIG (Figure 4f). Neither IVIG-400 nor IVIG-2,000 reduced the concentration of total mouse IgG in the plasma at day 11 (Supplementary Figure S2 online).
Next, we examined the therapeutic effect of IVIG-2,000 by administering it for 5 consecutive days at two different timings: from days 1 to 5 (IVIG-2,000-1) or from day 6 to 10 (IVIG-2,000-2). In the early phase (2 to 4 weeks after transfer), the IVIG-2,000-2 tended to reduce the disease severity, whereas the saline or IVIG-2,000-1 did not, but in the late phase, there were no differences between the regimens (Figure 5a). Neither regimen reduced the titers of circulating autoantibodies (Figure 5b).
IVIG decreases IL-6 and increases IL-10 in the active BP model
We also measured cytokine levels in the active BP model at day 11. We focused on the four cytokines that were altered by IVIG in the passive-transfer BP model. IVIG-2,000 significantly decreased IL-6 and increased IL-10 in the active BP model, whereas IL-17 and MIP-1β showed no changes (Figure 6a). The changes could be associated with the reduction of disease severity by IVIG-2,000 in the active BP model.
IVIG decreases IL-6 release from cultured keratinocytes
IVIG reduced serum IL-6 levels in both the passive-transfer BP model and the active BP model. IL-6 is known to be produced by keratinocytes in response to stimulation with antibodies to COL17 (
). In addition, we found that administered IVIG distributes in the intercellular space of the lower epidermis (Figure 6b). Therefore, we examined the effect of IVIG on IL-6 production by keratinocytes. HaCaT cells, which are immortalized human keratinocytes, were stimulated with BP-IgG in the presence or absence of IVIG. The release of IL-6 from the HaCaT cells was significantly reduced by IVIG (Figure 6c), suggesting that IVIG may directly act on keratinocytes and inhibit the production of IL-6.
IVIG suppressed disease severity in both the passive-transfer BP model and the active BP model with the reduction of circulating antibodies to the NC16A domain of COL17. First, we examined the effect of IVIG on a neonatal passive-transfer BP model in which the disease is induced by SG-IgG, polyclonal antibodies to COL17. IVIG significantly reduced circulating antibodies to the NC16A domain and partially suppressed skin fragility in a dose-dependent manner. However, direct immunofluorescence failed to detect the reduction of the deposition of SG-IgG at the DEJ. We previously reported that antibodies to the NC16A domain but not to other domains on COL17 are relevant to the skin fragility in the BP model (
). Therefore, we speculated that the amount of SG-IgG binding to the NC16A domain was decreased by IVIG, but that the reduction was masked by the deposition of a large amount of antibodies to other domains on COL17. To address this issue, we next performed a passive-transfer experiment by using TS39-3, a pathogenic monoclonal antibody to the NC16A domain. As expected, IVIG reduced the circulating as well as the skin-binding TS39-3 antibody and suppressed skin fragility. However, we were unable to evaluate the C3 deposition in this model because TS39-3 is a mouse IgG1 antibody that has no complement activation ability. We next performed IVIG experiments using a BP-IgG passive-transfer model. Notably, IVIG significantly diminished C3 deposition at the DEJ induced by BP-IgG. Taken together, these findings suggest that IVIG reduces circulating antibodies to COL17 and diminishes the deposition of IgG and C3 at the DEJ, although the results depend on experimental system.
How does IVIG reduce the pathogenic antibodies in vivo? Previous studies have demonstrated that the IVIG saturates the neonatal Fc receptor, a protective receptor that prevents IgG degradation by lysosomes and returns intact IgG to circulation, and enhances the clearance of the free pathogenic autoantibody from circulation (
). The reduction of the pathogenic antibody by IVIG in our model might be explained by this mechanism.
A passive-transfer BP model using neonatal mice is a simple and easy method for assessing the efficacy of IVIG, but it has limitations: (i) IVIG was administered intraperitoneally and not intravenously due to technical difficulties; (ii) IVIG was administered in a single dose, and (iii) the observation period was short. To overcome these issues, we also utilized an active BP model. IVIG decreased circulating antibodies to the NC16A domain as well as disease severity in the active BP model in a dose-dependent manner. Furthermore, delayed short-term administration of IVIG temporally showed a reduction of disease severity, suggesting that IVIG may have not only a preventive effect but also a therapeutic effect on the active BP model.
To explore the other factors that contribute to diminished disease severity, we further examined plasma cytokine and chemokine levels. In the passive-transfer BP model, the administration of SG-IgG induced various cytokines and chemokines, especially at the early phase (6 hours after SG-IgG injection) and, interestingly, IVIG decreased IL-6, IL-17, and MIP-1β (CCL4). The skin fragility in a passive-transfer BP model is considered to be associated with the reduction of COL17 due to the binding of antibodies to the NC16A domain rather than due to the inflammatory process (
). Therefore, we next examined several cytokines in the active model that were changed in the passive-transfer model. Interestingly, IL-6 was again decreased in IVIG-treated active BP model. IVIG was reported to suppress the release of IL-6 from blood cells of normal children and reduce serum IL-6 in children with Kawasaki disease (
). Therefore, we assumed that IVIG may suppress IL-6 production by keratinocytes. As expected, IVIG reduced the release of IL-6 by cultured keratinocytes that were stimulated with BP-IgG, suggesting a novel mechanism of action of IVIG in BP.
Conversely, IL-10, an anti-inflammatory cytokine, was elevated by IVIG in the active BP model. Previous studies demonstrated that serum IL-10 levels (
). Although little is known about how IVIG affects the differentiation of IL-10−producing cells, our results suggest that IVIG may regulate the inflammatory responses of BP by promoting the increase of IL-10−producing cells.
In conclusion, IVIG decreases antibodies to the NC16A domain and modulates the production of cytokines and chemokines, resulting in the reduction of disease severity in experimental BP models. IVIG might act directly on keratinocytes to suppress the production of IL-6. This study provides significant evidence that IVIG is effective for BP through several mechanisms, although the details should be further investigated.
Materials and Methods
Generation of the passive-transfer BP model and IVIG treatment
SG-IgG was passively transferred into neonatal COL17-humanized mice as described previously (
). Different doses of IVIG (NIHON Pharmaceutical Co, Ltd, Japan) (400 mg/kg or 2,000 mg/kg) or saline was intraperitoneally administered to neonatal COL17-humanized mice, and 150 μg/g SG-IgG was intraperitoneally transferred at the same time or 2 hours later. At 18 hours after the injection of SG-IgG, the fragility of back skin was evaluated by up to four incidences of gentle rubbing. The evaluation was carried out by a blinded investigator. Ear or back skin was obtained after the evaluation of skin fragility and processed for light microscopy (hematoxylin and eosin) and for direct immunofluorescence using FITC-conjugated antibodies to mouse IgG and C3. In the hematoxylin and eosin sections, the histopathologic results were scored as follows: 0, negative; 1, minimal; 2, mild; 3, moderate; 4, marked. In some experiments, 50 μg/g TS39-3 or 750 μg/g BP-IgG was intraperitoneally transferred instead of SG-IgG and the mice were evaluated 24 hours after the injection. All animal procedures were conducted according to the guidelines of the Hokkaido University Institutional Animal Care and Use Committee.
Generation of the active BP model
Wild-type mice were immunized by human COL17-expressing Tg mouse skin grafts as described. Spleen cells were isolated and pooled from several immunized wild-type mice and intravenously transferred into Rag-2−/−/COL17-humanized mice through a tail vain in 500 μl phosphate-buffered saline (
). To determine the appropriate number of spleen cells, we compared the disease severity of mice at three different numbers of spleen cells (1.0, 2.0, and 4.0 × 108 cells/mouse), and we found 1.0 × 108 cells to be the best for inducing the grade of BP appropriate for this study.
IVIG for the active BP model
Spleen cells (1.0 × 108) were transferred to Rag-2−/−−/COL17-humanized mice at day 0. The mice were treated with IVIG from day 1 to day 21 (for 3 weeks) at different doses (400 mg/kg/d or 2,000 mg/kg/d) or saline. In some of the experiments, the mice were treated with 2,000 mg/kg/d of IVIG from day 1 to day 5, or from day 6 to day 10. Then, the mice were examined for general condition and for percentage of body surface area affected by cutaneous lesions (i.e., erythema, hair loss, blisters, erosions, and crusts) at days 1, 6, 9, 11, 14, 18, 21, 24, 28, 32, and 35. We also collected plasma to examine circulating anti-human COL17 IgG levels at days 0, 6, 9, 11, 18, 21, 28, and 35. The animals were then sacrificed at day 35, and skin sections were taken for histologic examination. Back, neck, or ear skin was processed for light microscopy (hematoxylin and eosin) and direct immunofluorescence using FITC-conjugated antibodies to IgG and C3. In the hematoxylin and eosin sections, the histopathologic results were scored as follows: 0, negative; 1, minimal; 2, mild; 3, moderate; 4, marked.
Data were expressed as mean ± standard deviation. Statistical analyses were performed using SAS, Release 9.3 (SAS Institute Inc, Cary, NC) and GraphPad Prism (GraphPad Software, La Jolla, CA). In the SG-IgG passive-transfer BP model, an unpaired t test was performed for plasma IgG titer. In the TS39-3 and BP-IgG passive-transfer BP models, the statistical differences in fluorescence intensity and plasma IgG titer were determined by unpaired t test. Comparisons of cytokine and chemokine levels in the passive-transfer model were made by using Dunnett’s multiple comparison test. For the active BP model, the Wilcoxon rank-sum test was performed for histopathologic examinations and ANOVA was performed for all other examinations. Multiplicity adjustment was performed by the Holm method. The unpaired t test was used for a comparison of cytokine levels in the active model. For comparison of IL-6 from HaCaT cells, Tukey’s multiple comparisons test was used. P < 0.05 was considered significant compared with the control.
Conflict of Interest
This study was supported by NIHON Pharmaceutical Co, Ltd. H. Ujiie reports receiving consulting and lecture fees from NIHON Pharmaceutical Co, Ltd. The remaining authors state no conflict of interest.
We thank K. B. Yancey (Department of Dermatology, Medical College of Wisconsin, Milwaukee, WI) for providing human COL17 cDNA transgenic mice and Mika Tanabe for her technical assistance.
Bullous pemphigoid is an autoantibody-mediated skin blistering disease. Previous studies revealed that intravenous Ig is therapeutic in animal models of bullous pemphigoid by saturating the IgG-protective receptor FcRn, thereby accelerating degradation of pathogenic IgG. Sasaoka et al. demonstrate that the inhibitory effects of intravenous Ig on bullous pemphigoid are also associated with negative modulation of cytokine production by keratinocytes.