A specific antigen-antibody interaction triggers the cellular pathophysiology of bullous pemphigoid

Monocytes enhance neutrophil-induced blister formation in an ex vivo model of bullous pemphigoid

BACKGROUND

Lesions of bullous pemphigoid (BP), an autoimmune subepidermal blistering disease characterized by the presence of tissue-bound and circulating autoantibodies to hemidesmosomal antigens, harbor a mixed inflammatory cellular infiltrate. In various models, neutrophils, eosinophils, mast cells, monocytes as well as B and T cells have been shown to be involved in the pathogenesis of BP. However, their interactions with and effective role in blister formation remain uncertain. This study was aimed at investigating the effect of monocyte/neutrophil interaction on blister formation in an ex vivo BP model.

METHODS

Skin cryosections were incubated with purified human neutrophils and monocytes, in the presence or absence of BP autoantibodies. Production of reactive oxygen species (ROS), degranulation, mediator release (neutrophil elastase [NE], myeloperoxidase [MPO], matrix metalloproteinase-9 [MMP-9]), binding of Fcγ receptor (CD16, CD32, CD64), and cell adhesion (CD18, ICAM-1) was investigated using appropriate inhibitors. Dermal-epidermal separation (DES) was assessed by light microscopy and quantified by Fiji software.

RESULTS

Monocytes and neutrophils synergistically interact resulting in a significantly higher DES compared to either monocytes or neutrophils separately (P < .0001). Monocyte/neutrophil-induced DES was associated with increased ROS production and was dependent on adhesion and FcγRIII binding. Upon stimulation by the granule-poor fraction of monocyte supernatants, neutrophils increased their release of MMP-9, thereby also DES at the dermal-epidermal junction of skin cryosections.

CONCLUSION

Our observations suggest that the interaction of cells, as shown here for monocytes and neutrophils, enhances mediator release resulting in an increased subepidermal blister formation. Thus, blocking intercellular cross talk promises a new therapeutic approach for blocking tissue damage in BP.

Evidence for a role of eosinophils in blister formation in bullous pemphigoid

BACKGROUND

Bullous pemphigoid (BP) is an autoimmune bullous disease of the skin characterized by subepidermal blister formation due to tissue-bound and circulating autoantibodies to the hemidesmosomal antigens BP180 and BP230. Although eosinophils and their toxic mediators are found abundantly in BP lesions, their role in blister formation has remained unclear.

OBJECTIVE

To investigate the role of eosinophils in the pathogenesis of BP with a specific focus on blister formation and to define conditions inducing dermal-epidermal separation (DES).

METHODS

In an ex vivo human model of BP, normal human skin cryosections were incubated with purified human peripheral blood eosinophils with or without activation in the presence or absence of BP autoantibodies, brefeldin A, diphenyleneiodonium, DNase or blocking F(ab')₂ fragments to CD16, CD18, CD32 and CD64. Dermal-epidermal separation was assessed by light microscopy studies and quantified using Fiji software.

RESULTS

Following activation with IL-5 and in the presence of BP autoantibodies, eosinophils induced separation along the dermal-epidermal junction of ex vivo skin. Dermal-epidermal separation was significantly reduced by blocking any of the following: Fcγ receptor binding (P = 0.048), eosinophil adhesion (P = 0.046), reactive oxygen species (ROS) production (P = 0.002), degranulation (P < 0.0001) or eosinophil extracellular trap (EET) formation (P = 0.048).

CONCLUSIONS

Our results provide evidence that IL-5-activated eosinophils directly contribute to BP blister formation in the presence of BP autoantibodies. Dermal-epidermal separation by IL-5-activated eosinophils depends on adhesion and Fcγ receptor activation, requires elevated ROS production and degranulation and involves EET formation. Thus, targeting eosinophils may be a promising therapeutic approach for BP.

Internalization of the 180 kDa bullous pemphigoid antigen as immune complexes in basal keratinocytes: an important early event in blister formation in bullous pemphigoid

We have previously shown that the 180 kDa bullous pemphigoid antigen (BPAG2) is distributed on the lateral-apical (as a pool) and ventral (as hemidesmosomes) cell membranes of monolayer cultured keratinocytes and that addition of IgG purified from bullous pemphigoid (BP) patients (BP-IgG) causes the internalization of immune complexes of BPAG2 and BP-IgG from the lateral-apical cell membrane. This internalization of BPAG2 is believed to inhibit the Ca2+ induced reformation of hemidesmosomes on the ventral cell membrane, possibly by inhibiting the supply of the antigen from the lateral-apical to the ventral membranes to form hemidesmosomes. The purpose of this paper is to examine, by using biopsy specimens from BP patients (12 cases), whether or not this internalization of BPAG2 is generated in situ. The fates of BPAG2, 230 kDa BPA (BPAG1) and bound BP-IgG were traced by immunofluorescence microscopy using monoclonal antibodies to BPAG1, BPAG2 and human IgG. In more than half of the lesional and perilesional biopsy specimens, internalization of BPAG2 into the basal cells was observed, while in normal skin BPAG2 was detected on the whole surface of the basal cells without its internalization. No internalization of BPAG1 was detected in BP and normal epidermis. These results suggest that binding of BP-IgG on the lateral-apical cell surface of basal cells causes internalization of BPAG2 in situ in the epidermis of BP patients similar to that seen in cultured cell systems, and that this internalization of immune complexes of BPAG2 and BP-IgG may play an important part in blister formation in BP.

Antibody-binding to the 180-kD bullous pemphigoid antigens at the lateral cell surface causes their internalization and inhibits their assembly at the basal cell surface in cultured keratinocytes

We demonstrated the effects of monoclonal antibodies to the 180-kD and 230-kD BP antigens (BPA) and of BP sera on Ca(++)-induced formation of hemidesmosomes in cultured human keratinocytes (a cell line, DJM-1) by immunofluorescence microscopy. Under low Ca++ (0.07 mM) conditions, the 180-kD and 230-kD BPAs were distributed homogeneously on the basal plasma membrane, while they formed a peculiar concentric ring or arch (ring/arch) arrangement in high-Ca++ (1.87 mM) medium. On the other hand, the apical-lateral cell membrane was stained homogeneously with antibodies to the 180-kD BPA, but not to the 230-kD BPA, both in low and high Ca++ media. The low-high Ca++ switch at first caused disappearance of the antigen from the basal plasma membrane and then formed the high-Ca++ ring/arch pattern within 3 hrs. In this system, monoclonal antibodies to the 180-kD and 230-kD BPAs and the sera from 5 BP patients, 2 pemphigus vulgaris (PV) patients, and 4 normal volunteers were added into the culture media. The addition of anti-180-kD BPA antibodies or any BP serum caused the internalization of the 180-kD BPA from the apical-lateral cell membrane and inhibited the Ca(++)-induced formation of the ring/arch pattern on the basal membrane, possibly by inhibiting the movement of the antigen from the lateral to the basal membrane to form hemidesmosomes.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparative study of bullous pemphigoid antigens among Japanese, British, and U.S. patients indicates similar antigen profiles with the 170-kD antigen present both in the basement membrane and on the keratinocyte cell membrane

There are a number of controversies relating to studies of bullous pemphigoid (BP) antigens from different institutions, mainly regarding the detection of the 230-kD and 170-kD BP antigens. In this study, in an attempt to resolve the discrepancies, we have examined and compared the reactivity by immunofluorescence, immunoblotting, and immunoprecipitation among the sera from Japanese, British and U.S. BP patients. Both the 230-kD and 170-kD BP antigens were detected by various sera from all populations with immunoblotting, whereas immunoprecipitation detected only the 230-kD BP antigen but not the 170-kD BP antigen. Immunoprecipitation was more sensitive than immunoblotting to detect the 230-kD antigen. These results indicate that both the 230-kD and 170-kD proteins are BP antigens found in all three populations. By immunofluorescence cell surface staining in the lower epidermis in addition to basement membrane zone staining was shown by a considerable number of patients' sera in all populations. Comparison between the results of immunofluorescence and immunoblotting revealed a clear correlation of this cell surface staining with the presence of antibodies against the 170-kD BP antigen. That the affinity-purified antibodies specific to the 170-kD BP antigen showed this cell surface staining further confirmed this correlation. These results may indicate a different nature of the 170-kD BP antigen from that of the 230-kD BP antigen.

Further analyses of epitopes for human monoclonal anti-basement membrane zone antibodies produced by stable human hybridoma cell lines constructed with Epstein-Barr virus transformants

We previously established Epstein-Barr virus (EBV)-transformed bullous pemphigoid (BP) patient lymphoblastoid cell lines, which produced human monoclonal anti-basement membrane zone antibodies. In the present study, we established two independent human-human hybridomas by fusion of these EBV transformants with a human B-cell line. These hybridomas, designated 5E-HY-4B and 10D-HY-8B, were very stable and showed a high yield of monoclonal antibody (MoAb) secretion. Each cell line was tetraploid and showed combined rearranged segments of immunoglobulin heavy-chain gene derived from both an EBV transformant and a parent cell. Immunoblot analysis showed that the 5E-HY-4B MoAb recognized the 230-kDa BP antigen but that the 10D-HY-8B MoAb did not show any reactivity. In contrast, both MoAbs precipitated the 230-kDa BP antigen with immunoprecipitation. These results indicate that the two MoAbs reacted with different epitopes on the 230-kDa BP antigen: a continuous epitope for the 5E-HY-4B MoAb and a conformation-dependent epitope for the 10D-HY-8B MoAb. This speculation was confirmed at the molecular level by the result that the fusion protein produced by a partial cDNA for the 230-kDa mouse BP antigen reacted with the 5E-HY-4B MoAb but not with the 10D-HY-8B MoAb. Furthermore, the study of the reactivity with fusion proteins of a series of deleted clones restricted the epitope for the 5E-HY-4B MoAb within the region with 114 amino acid residues in the C-terminal domain of the 230-kDa BP antigen.

Detection of the 170-kDa bullous pemphigoid antigen by immunoprecipitation

There has been controversy concerning the nature of the bullous pemphigoid (BP) antigen: immunoprecipitation identified BP antigen as a single, unique 230-kDa protein, whereas immunoblot analysis showed multiple antigen molecules, mainly 230- and 170-kDa proteins. In this study, to further characterize the 170-kDa protein, we have examined whether the 170-kDa protein is detected by immunoprecipitation. Extracts of human squamous cell carcinoma cells revealed the 170-kDa protein with immunoblot analysis. Although the conventional immunoprecipitation detected only the 230-kDa protein, some BP sera that detected the 170-kDa protein with immunoblotting also precipitated the 170-kDa protein with our modified immunoprecipitation, in which the cells were extracted with 1% sodium dodecylsulfate (SDS) buffer and reacted with the sera under reduced SDS concentration. The 170-kDa protein-specific BP sera clearly showed hemidesmosomal plaque staining with immunofluorescence of cultured cells. These results indicate that the 170-kDa protein is indeed one of the BP antigens and that the 230- and 170-kDa BP antigens are integrated in different ways in hemidesmosomes.

Differences in direct immunofluorescence staining patterns in epidermolysis bullosa acquisita and bullous pemphigoid

BACKGROUND

Both bullous pemphigoid (BP) and epidermolysis bullosa acquisita (EBA) are characterized by linear IgG deposits along the basement membrane zone (BMZ). Although these diseases can be distinguished by special tests, the staining pattern on direct immunofluorescence (DIF) is identical.

OBJECTIVE

The purpose of the present study was to reevaluate DIF as a diagnostic tool in differentiating EBA from BP.

METHODS

We performed DIF studies on biopsy specimens from eight consecutive patients with EBA and 18 consecutive patients with BP.

RESULTS

In five of eight cases of EBA, C3 deposition was essentially absent. IgG was the predominant class of immunoglobulins in the deposits and was present along the BMZ in all eight cases. C3 was present in 17 of 18 cases of BP. IgG was the predominant immunoglobulin present in 15 of these 18.

CONCLUSION

Patients with EBA are more likely to have linear IgG staining along the BMZ without concomitant C3 deposition than are patients with BP. This difference may be a function of the ability of the autoantibodies to fix complement.

Characterization of bullous pemphigoid antibodies by use of recombinant bullous pemphigoid antigen proteins

The seroreactivity of patients with bullous pemphigoid (BP) to recombinant proteins representing sequences in the carboxyl domain of the murine 230-kD BP antigen (BPA) was determined. Sera from 133 patients with BP, 20 patients with pemphigus, and 21 normal subjects were examined by Western blotting by using two recombinant proteins: RP120 (MW = 120 kD), representing the C-terminal half of the 230-kD BPA, and RP60 (MW = 60 kD), representing the C-terminal quarter. These RP120 and RP60 were recognized by 84% and 61%, respectively, of the BP sera that reacted with the 230-kD BPA in epidermal extract, and not by any of pemphigus or normal sera. Furthermore, these RP120 and RP60 were not recognized by any BP sera that reacted only with the 170-kD BPA, which is known to be another major BPA. These findings indicate that one or more of the major antigenic regions localizes in the carboxyl-half domain of the 230-kD BPA, and also suggest that the 230-kD BPA may be distinct from the 170-kD BPA.

Alpha 6 beta 4 integrin heterodimer is a component of hemidesmosomes

Antisera that recognize the alpha 6 and beta 4 subunits of integrins were found by immunoelectron microscopy to localize to hemidesmosomes in the basal cells of mouse corneal epithelium. Immunoprecipitation experiments using extracts of metabolically labeled corneal epithelial cells indicate that the primary alpha 6-subunit-containing integrin heterodimer present is alpha 6 beta 4 and not alpha 6 beta 1. Here we extend previous studies to report that by immunofluorescence microscopy the alpha 6 integrin subunit colocalizes with bullous pemphigoid antigen and type VII collagen in newly forming hemidesmosomes in the developing 17-day fetal rabbit eye. Neither the composition of the anchoring filaments, which span the region between the hemidesmosomal plaque and the lamina densa of basement membrane where the globular domain of type VII collagen is located, nor the extracellular ligand of alpha 6 beta 4 is known. Once anchoring filament proteins are identified, it will be of interest to determine whether any bind to alpha 6 beta 4.

Partial cDNA cloning of the 230-kD mouse bullous pemphigoid antigen by use of a human monoclonal anti-basement membrane zone antibody

A cDNA clone with the coding sequence for the 230-kD bullous pemphioid (BP) antigen was isolated from a mouse cDNA expression library by using an anti-basement membrane zone human monoclonal antibody (MoAb-5E). The lambda gt11 cDNA expression library was constructed from poly(A)+RNA from the mouse epidermal cell line, Pam cells, by random priming. 1.5 X 10(5) recombinant clones were screened by immunostaining with MoAb-5E and one positive clone (BPM1) was obtained. All of the ten BP sera but none of the five normal or seven pemphigus sera tested reacted with the fusion protein produced by BPM1. The size of the cDNA was 3.2 kb. Northern blot analysis indicated that BPM1 cDNA hybridized to a mRNA of about 9 kb, which is large enough to encode for a 230-kD protein. DNA sequencing demonstrated a 2,991-bp open reading frame encoding a peptide of 115 kD. Sequence comparison between mouse and human cDNA clones revealed that the 230-kD BP antigen was well conserved during evolution except for the carboxyl terminus. Highly conserved and hydrophilic regions in the molecule were considered to be good candidates for antigenic determinants. This cDNA clone will be useful not only for diagnosis of BP, e.g., enzyme-linked immunosorbent assay using recombinant proteins or synthetic peptides as antigens, but also for pathophysiologic study in which mouse models of BP might be used.

Production of human monoclonal anti-basement membrane zone (BMZ) antibodies from a patient with bullous pemphigoid (BP) by Epstein-Barr virus transformation. Analyses of the heterogeneity of anti-BMZ antibodies in BP sera using them

We established three lymphoblastoid cell lines from a bullous pemphigoid (BP) patient's peripheral blood by means of EBV transformation, which produced human monoclonal anti-basement membrane zone (BMZ) IgG antibodies. A blocking immunofluorescence test using these MAbs, designated 5A, 5E, and 10D, revealed that 5A and 5E recognized the same or a closely associated epitope, but the epitope for 10D was completely different, 18 of 30 BP sera blocked the reactivity of 10D MAb and 17 sera blocked 5E, while 9 sera did not block the staining of either antibody. Immunoblot analysis demonstrated that both 5A and 5E MAbs reacted exclusively with a protein band of approximately 230 kD in normal human epidermal extracts. However, 10D did not show any protein band. 22 of 30 BP sera strongly reacted with the same 230-kD protein, while none of control sera showed such reactivity. These results clearly demonstrated the heterogeneity of anti-BMZ antibodies in terms of epitopes. These MAbs should be useful in future investigations concerning not only the immunopathology but also the biochemial and molecular analyses of the BP antigen.

Characterization of bullous pemphigoid antigen synthesized by cultured human squamous cell carcinoma cells

This report concerns the characterization of bullous pemphigoid antigen (BPA) in cultured cells derived from a human squamous cell carcinoma (SCC cells) as identified by sera of 9 patients with bullous pemphigoid (BP). Immunoglobulin G from 5 of 9 BP sera bound the cell surfaces of SCC cells with a punctate staining pattern by immunofluorescence. The other four BP sera and all 12 controls sera showed no specific staining. To characterize the antigen for pemphigoid antibody, we immunoprecipitated NP-40 extracts of cells labeled with [14C] amino acids using nine BP sera and 12 controls sera. These immunoprecipitates were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis and by fluorography. The five BP sera that gave positive immunofluorescence results also precipitated a protein with molecular weight (mol wt) of 110 kD, whereas the other four BP sera negative by immunofluorescence and all 12 controls sera did not precipitate this protein. These results indicate that the sera of patients with BP contain antibodies against the protein with mol wt 110 kD in the SCC cells used in this study. With use of four BP sera, two of which were positive to this mol wt 110 kD BPA and two of which were negative, immunoprecipitation was carried out in Pam cells. One 110 kD-positive serum and two 110 kD-negative sera, precipitated a BPA of mol wt 220 kD from Pam cells labeled with [35S] methionine. These results suggest that BP sera may recognize more than one antigen, in addition to possible BPA heterogeneity in different individuals.

A 230-kD basic protein is the major bullous pemphigoid antigen

Recent studies have demonstrated that bullous pemphigoid (BP) antigen, in extracts of cultured human keratinocytes and normal human epidermis, is a 230-kD polypeptide. However, other recent studies have suggested that there is heterogeneity of BP antigen at the molecular level. To demonstrate that this 230-kD polypeptide is the major BP antigen and to further characterize it, we tested multiple BP sera by both immunoprecipitation of extracts of radiolabeled cultured human keratinocytes and immunoblotting of extracts of normal human epidermis. Thirty-six of 37 BP sera immunoprecipitated the 230-kD polypeptide, and 11 of 13 BP sera identified the 230-kD polypeptide by immunoblotting. Eighty-six disease and normal control sera did not bind the 230-kD BP antigen. Immunoprecipitation was a more sensitive assay than was immunoblotting to detect the 230-kD antigen; BP sera that were weak or negative by immunoblotting were strongly positive by immunoprecipitation. To demonstrate that this 230-kD polypeptide shared epitopes, defined by BP sera, with the epidermal basement membrane zone, we showed that BP IgG affinity purified on this polypeptide bound the epidermal basement membrane zone by immunofluorescence. To further characterize the BP antigen and to compare the antigen synthesized by cultured human keratinocytes to the antigen extracted from epidermis, we analyzed the 230-kD BP antigen by two-dimensional gel electrophoresis. BP antigen immunoprecipitated from culture extracts or BP antigen identified by immunoblots of epidermal extracts was a single spot on two-dimensional gels, with a pI of about 8. These data indicate that, although other polypeptides are sometimes identified (e.g., a 166 kD band by immunoprecipitation or a 180 kD band on immunoblots), the major BP antigen identified in cultured human keratinocytes is similar to the antigen found in normal human epidermis and is a basic 230-kD polypeptide.

Isolation of complementary DNA for bullous pemphigoid antigen by use of patients' autoantibodies

Autoantibodies from bullous pemphigoid (BP) patients define a 230-kD protein found in the basement membrane of stratified squamous epithelia. The purpose of this study was to isolate and characterize a cDNA clone with coding sequences for BP antigen. Poly(A+) RNA derived from total RNA of cultured keratinocytes was used, with oligo-dT priming, to construct a cDNA library in the lambda gt11 expression vector, which was screened by the immunoperoxidase method with one BP serum. One darkly stained clone, called here the BP clone, was further characterized. 9 of 9 BP sera, but none of 6 normal and 11 pemphigus sera, bound the plaques of this BP clone. Furthermore, BP IgG affinity purified on plaques of this clone, but not unrelated clones, bound the epidermal basement membrane by immunofluorescence and immunoprecipitated the 230-kD BP antigen from extracts of cultured keratinocytes. Eco RI digestion of the BP clone's cDNA insert demonstrated a 680- and 1,500-bp fragment. Northern blots of total keratinocyte RNA showed that complementary riboprobes transcribed from both fragments hybridized to a 9-kb RNA. Dideoxy DNA sequencing from the 5' end of the BP cDNA demonstrated a 1,992-bp open reading frame, encoding a peptide of 76 kD. This BP cDNA clone will be valuable for understanding the protein structure, expression, and gene organization of BP antigen.

Hemidesmosomes and anchoring fibril collagen appear synchronously during development and wound healing

Bullous pemphigoid antisera and monoclonal antibodies to type VII collagen were used to localize hemidesmosomes and anchoring fibrils, respectively, in tissues of developing eyes and healing corneal wounds of New Zealand white rabbits. In the 17-day fetal rabbit eye, both antibodies colocalize to the epithelial-stromal junction of the lid and conjunctival region, but neither binds to the cornea, and electron microscopy demonstrates hemidesmosomes only where the antibodies bind. By 20 days of fetal development, the antibodies colocalize in cornea, and, by electron microscopy, hemidesmosomes are shown to be present as well. In healing 7-mm corneal wounds, both antibodies colocalize at the wound periphery within 66 h. By electron microscopy, hemidesmosomes along small segments of basal lamina are also shown to be present at the wound periphery at this time. These demonstrations of the synchronous assembly of hemidesmosomes and anchoring fibrils support the hypothesis of linkage of hemidesmosomes through the basement membrane to anchoring fibrils.

Identification of pemphigus vulgaris antigen extracted from normal human epidermis and comparison with pemphigus foliaceus antigen

Immunoprecipitations of cultured keratinocyte extracts have shown that pemphigus vulgaris (PV) sera bind a polypeptide of 210,000 mol wt with disulfide-linked chains of 130,000 and 85,000 mol wt. To identify proteins in normal human skin recognized by PV antibodies, we performed immunoprecipitations of normal human epidermal extracts. All 22 PV sera tested immunoprecipitated a complex of polypeptides (PV complex) of 210,000, 130,000, and 85,000 mol wt, after reduction. One- and two-dimensional gel electrophoresis showed that the 130,000- and 85,000-mol-wt polypeptides of the PV antigen from both cultured keratinocytes and epidermis have identical charges and sizes. In addition to precipitating the PV complex, 14 of 22 PV sera also have antibodies to a calcium-sensitive epitope on a different complex of polypeptides (PF complex) which has previously been shown to be precipitated by all pemphigus foliaceus (PF) sera. The PF complex consists of polypeptides of 260,000, 160,000, 110,000, and 85,000 mol wt. Although the majority of PV sera also precipitate the PF complex, no PF sera precipitate the PV complex. Thus, PV and PF can be absolutely distinguished on a molecular level using the patients' autoantibodies. The PV and PF complexes, although distinct, have certain similarities. The 85,000-mol-wt polypeptide of each is identical. The 160,000-mol wt-peptide of the PF complex and the 130,000-mol-wt peptide of the PV complex have the same isoelectric point and both are capable of disulfide linkage to the 85,000-mol-wt polypeptide. The PV and PF complexes are closely related and may prove important in cell adhesion.

Bullous pemphigoid

Bullous pemphigoid is an autoimmune blistering dermatologic disease characterized clinically by tense bullae that may develop on normal or erythematous skin. The major histologic feature is a subepidermal blister with variable degrees of dermal inflammation. Other immunologically mediated blistering skin diseases may mimic bullous pemphigoid, including herpes gestationis, cicatricial pemphigoid, dermatitis herpetiformis, and epidermolysis bullosa acquisita. These diseases will be discussed. Most patients with bullous pemphigoid demonstrate circulating autoantibodies reactive with an antigen located in the lamina lucida region of the basement membrane zone. Complement activation by these autoantibodies initiates influx and activation of mast cells along with other inflammatory cells. Tissue injury with damage and eventual destruction of the basement membrane occurs as a result of the release of inflammatory mediators. The final result is subepidermal blister formation. The course of bullous pemphigoid tends to be self-limited, and successful treatment is usually effected with systemic glucocorticosteroids.