Bullous Pemphigoid (BP) represents the most common autoimmune bullous disorder and prevalently occurs in the elderly (1). The disease is characterized by circulating IgG autoantibodies which mainly target the non-collagenous (NC)16A domain of Collagen XVII, also referred to as BP180, a main component of the hemidesmosomes, which maintains the integrity of the dermal-epidermal junction (DEJ) (2). Patients with BP also develop antibodies against BP230, a cytoplasmic protein of the hemidesmosomal plaque that cross-link BP180 to keratin (K) 5 and K14 (3, 4).

Bullous Pemphigoid encompasses a heterogeneous spectrum of manifestations. The classic type is characterized by diffuse tense blisters arising on a background of erythematous-edematous skin (5). Pruritus is always present and, in some patients, may precede for years the appearance of manifest lesions (6). Further, several non-bullous forms have been described (7).

Over the recent years, the incidence of BP is raising significantly (2, 8). This phenomenon is partly explained by an increasing aging population in western countries, and easier access to serological diagnostic kits. Epidemiologic studies showed that BP is associated to some neurological disorders (9–11), particularly dementia and Parkinson, drugs, such as dipeptidyl peptidase IV inhibitors (DPP4i) and PD-1 inhibitors (12, 13) and malignancies (14, 15). The pandemic showed that BP might occur after either SARS-CoV-2 infection or related vaccines (16, 17).

Pathogenically, IgG binding to either BP180 or BP230 activates a cascade of inflammatory mediators resulting in the loss of dermal-epidermal adhesion (18). The increasing knowledge of this complex inflammatory cascade is pivotal for developing new therapeutic strategies for the disease, as its therapeutic management is still largely based on long-term immunosuppressive treatments (10). Indeed, the purpose of this review is to provide a concise overview of the cytokine milieu of BP, with a special focus on molecules currently under-investigation as potential therapeutic targets.

IL-17/IL-23 axis has a supportive role in the pathogenesis of BP. In a IL17^–/– mice model, passive transfer of anti-BP180 antibodies led to reduced skin inflammation, whereas in wild-type mice levels of IL-17 production following anti-BP180 IgG passive transfer correlated to disease severity (48). Collectively, this findings suggest that IL-17 is pathogenically relevant in murine BP. In human BP IL-17 levels increase in lesional skin and blister fluid (49) and its production is mainly sustained by neutrophils (Figure 1). In the peripheral blood of BP patients, CD3+ lymphocytes appear to be the main source of IL-17 (24). Interestingly, longitudinal measurement of IL-17 and IL-23 serum concentration was found to predict relapse in BP patients, as relapsing patients were shown to have persisting increased levels of serum IL-17 as well as increasing serum concentration of IL-23 during the first month of treatment (50). The IL-17/IL23 axis promotes various pathological processes, including DNA extracellular trap formation (51), stimulation of IL-1β production in macrophages and production of matrix-metalloprotease (MMP)-9 and neutrophil elastase, enzymes involved in blister formation (52, 53).

Interestingly, only a subset of BP patients shows up-regulation of IL-17 at the baseline, without correlation with disease severity, suggesting that not all the patients could effectively benefit from therapeutic targeting of the IL-17/IL23 axis (30–32). Interestingly, targeting IL17/IL23 demonstrated efficacy in patients with coexisting BP and psoriasis, a prototype of IL17/IL23 driven disease (33, 54–57). Results from ongoing clinical trials are thereby necessary to understand the impact of this treatment approach also in BP patients without coexisting psoriasis (58). Inhibition of janus kinase may worth being investigated in BP, since this could allow simultaneous targeting of both IL-4 and IL-23 (59–62).

Complement activation plays a pivotal role in the pathogenesis of BP (63, 64). This assumption is supported by several evidences: (i) the complement component C3 is disposed in a linear fashion along the DEJ in perilesional BP skin, and is even stronger than IgG, or sometimes found in the absence of IgG, in direct immunofluorescence (65–68); (ii) the capacity of autoantibodies to activate complement ex vivo correlates with disease activity and levels of autoantibodies in BP patients (69); (iii) milder clinical phenotype of BP230-type seems to correlate with weaker complement deposition at DEJ (70) (iv) genetic deficiency and/or pharmacological depletion of various complement components reduce pathogenicity of anti-BP180 IgG and dampen skin inflammation in experimental mouse models of BP (71, 72).

In BP patients, the classical pathway activation, which occurs following antibody/antigen binding, is the major pathway (73). In vitro and in vivo evidence suggests that blockage of C1q prevents both complement activation and skin blistering. Likewise, genetic absence of C4 in experimental BP mouse models is enough to abolish mast cell (MC) degranulation and attraction of neutrophils (71) (Figure 1). The alternative pathway of complement activation plays a supportive role in the pathogenesis of BP (71). Accordingly, passive transfer of anti-BP180 antibodies in mice with genetic deficiency of alternative pathway component FB developed a delayed and mild disease phenotype (71).

The activation of complement generates the attachment of component C3 and release of anaphylatoxins C3a and C5a, of which C5a can mediate polymorphonuclear leukocytes chemotaxis and both C3a and C5a can mediate MC degranulation (74). Of relevance in the generation of experimental BP in mice is the interaction between C5a and C5aR1 (75). C5aR2 conversely plays an anti-inflammatory and protective role in BP (76, 77). C5a receptor 1 (C5aR1) on MC was shown to be critical for the formation of skin lesions (74). Further, C5a/C5aR1 interaction on the surface of neutrophils activates the release of LTB4 via 5-lipoxygenase (78). Collectively, these molecules are indispensable to an efficient recruitment of neutrophils into the interstitial skin tissue (79, 80) (Figure 1 and Table 1).

A last consequence of complement activation in BP is the formation of the terminal membrane attack complex, which exerts direct cytotoxic effects in the epidermal basal cells.

It is worth noting that loss of dermal-epidermal adhesion in BP may also occur via complement-independent mechanisms (76, 81, 82), which are thought to be preponderant during early and non-blistering phases of BP, where non-fixing complement IgG4 subclasses are predominant (83–85). These mechanisms include (i) internalization of BP180 from the surface of keratinocytes after IgG binding to BP180 (86); (ii) direct release of cytokines, e.g., IL-6 and IL-8, from keratinocytes (87); (iii) induction of MC degranulation and eosinophil activation by IgE autoantibodies (88).

It is thus possible that both complement-dependent and complement-independent mechanisms work together in inducing and perpetuating BP inflammation and blistering (85).

Loss of dermal-epidermal adhesion in BP is critically associated with skin infiltration of neutrophil and eosinophil granulocytes (89–91). In passive transfer models, pathogenicity of antibodies is significantly reduced in the absence of myeloid granulocytes (89, 92, 93). Moreover, BP models induced by genetic deletion of BP180 pathogenic domains are characterized by spontaneous infiltration of granulocytes (94). Granulocytes induce blistering by different mechanisms such as the release of proteases, e.g., MMP-9, reactive oxygen species (ROS) and either neutrophil or eosinophil extracellular traps (51, 92, 95–97) (Figure 1). In humans, neutrophils and eosinophils localize differently in BP skin, with the first predominating in the blister fluids and the second in the dermal skin (98). The blister fluid, as well as sera, of BP patients over-express several chemoattractant and pro-inflammatory molecules, including IL-1, IL-8, IL-10, IL-5, Tumor necrosis factor-alpha, IL-6, CCL17, CCL-1, galectin-9, periostin, CCL11, CCL26, thymic stromal lymphopoietin and eotaxins (99–107) (Table 1). Dermal-epidermal separation induced by eosinophils is mainly orchestrated by IL-5 and depends on adhesion and Fcγ receptor activation (108). Neutrophils and eosinophils also regulate different aspects of BP inflammation. As an example, NET release from neutrophils acts systemically by inducing B-cell differentiation into plasma cells via activation of the MAPK/p38 cascade (109). Eosinophils mediate specific IgE pathogenicity, release cytokines which enhance Th2 cell recruitment and stimulate peripheral nerve terminals, e.g., by releasing IL-31 (88, 110–112).

Cross talk between immune cells is likely to potentiate the effector functions of granulocytes. Accordingly, neutrophils in BP release significantly more ROS and MMP-9 when stimulated with monocyte supernatants in vitro (113). Recently, Granzyme B was shown to critically regulate monocyte-dependent neutrophil recruitment in BP, and its inhibition significantly ameliorated pemphigoid disease induced by immunization with anti-COL7 antibodies in mice (114).

Monocytes and neutrophils are also activated by CXCL10, whose levels are increased in early-relapsing patients and is produced by keratinocytes, fibroblasts, and infiltrating immune cells (115).

Tissue resident macrophages are increased in BP skin and are mainly polarized toward the M2 phenotype, which express CD163 and CD206 (116, 117). M2 macrophages produce large amounts of Th2-type cytokines and stimulate T-cell and eosinophil recruitment by releasing CCL18, CCL22, CCL24, and CCL2 (117, 118).

Furthermore, CD163+ M2 macrophages stimulated by LL37 in vitro produced CXCL10 and CCL20 as well CXCR5+, CXCL13+, which contribute to recruitment of TFH cells (43, 117).

Finally, BP skin is enriched in both basophils and MC. Basophils are implicated in BP-associated itch (119). The role of MCs is still matter of debate. Experimental evidence suggests that MC degranulation in mice with BP occur after different stimuli, including complement fractions and specific IgE antibodies (74). Finally, macrophage-mediated neutrophil infiltration depends on MC activation (120, 121) (Figure 1). However, while studies using KIT-dependent MC knock-out mice demonstrated that MC activation trigger BP (121), Kit-independent MC-deficient mice still develop the disease, without significant changes in immune cells infiltration. Collectively, these findings raise the hypothesis that MC activation could be a bystander effect of BP inflammation (122).

Until now, several targeted therapies for BP have been developed, including (i) cell-depleting therapies; (ii) autoantibody-targeting therapies and (iii) single cytokine/molecule-directed therapies (Table 1).

Rituximab, a B-cell depleting therapy, is still not approved for BP, but often applied off-label to patients who fail conventional therapies. BP patients receiving one or more RTX cycles experience high rate of complete remission (CR) (approximately 75–92% of CR and 40% of CR off therapy), with significant drop of autoantibody titer (27–29, 123). Notably, RTX was shown to halve the 5-year mortality rate of BP in one study (124). Although RTX is mostly applied according to either the lymphoma protocol (375 mg/m2 weekly for 4 weeks) or the rheumatoid arthritis protocol (2 infusions of 1,000 mg 2 weeks apart), increasing number of studies addressed the efficacy of low (2 infusions of 500 mg 2 weeks apart) or ultra low (100 mg weekly for 4 weeks) doses of RTX in BP (125, 126).

Intravenous immunoglobulins showed pleiotropic anti-inflammatory effects (26), including increased autoantibody catabolism, and demonstrated meaningful positive effects in several cases and studies in BP both as monotherapy or combined with RTX (127–130). More recently, efgartigimod, a monoclonal antibody targeting the neonatal Fc receptor and thereby hastening the internalization and degradation of immunoglobulins, has entered clinical trials in BP after promising results in pemphigus and myasthenia gravis patients (131, 132). Several studies reported disease improvement with omalizumab, a monoclonal antibody targeting the IgE-specific Fc epsilon receptor III (133–137) (Table 1). Omalizumab offers a favorable safety profile, making it suitable for patients with contraindication to prolonged corticosteroid/immunosuppressive regimens (138, 139). Intriguingly, it showed efficacy also in patients without detectable serum anti-BP180/BP230 IgE. Responder patients show a decrease in IgE skin deposition, circulating IgG autoantibody levels and circulating basophils, which suggests immune-modulatory effects beyond IgE inhibition (137, 138, 140).

Over the recent years, complement activation has served as an attractive target in BP, owing to the established role in BP pathogenesis demonstrated in animal studies.

In one study blockage of C1s by the specific inhibitor, TNT003, successfully blocked the complement activating capacity of BP sera. Likewise, Gutjahr et al. (141) found that tinzaparin sodium inhibited autoantibody-induced complement activation in BP sera.

More recently, Sadik et al. (142) reported the results of a phase IIa non-randomized clinical trial of BP patients treated with nomacopan (NCT05061771), an inhibitor of leukotriene B4 and complement C5. Seven of the nine patients recruited demonstrated remarkable reduction of Bullous Pemphigoid Disease Area Index (BPDAI) and pruritus after approximately 1.5 months. No serious adverse events were reported (Table 1).

Since a first report in 2018 (143), dupilumab, an IL-4 alpha subunit receptor inhibitor has been used increasingly in BP. Dupilumab mainly acts by suppressing IL-4- and IL-13-producing CD4+ T cells (144). In a case series of 13 patients published in 2020, dupilumab demonstrated a satisfactory response in 92.3% (145). BP patients receiving dupilumab in combination with conventional therapies achieve faster clinical response with a reduction of the cumulative steroid dose compared to those treated with conventional therapies (146, 147). Interestingly, combination of omalizumab and dupilumab demonstrated efficacy in one case of recalcitrant BP (148). Finally, the drug holds promise as a treatment for special settings of patients, including highly recalcitrant, rituximab-resistant, or immune checkpoint inhibitor-induced BP (149–151). So far, attempts to block IL-5 (mepolizumab) failed to show an impact on BP activity (NCT01705795), while a study with benralizumab, a monoclonal antibody targeting the IL-5 receptor is currently ongoing (NCT04612790) (Table 1).

Finally, it will be intriguing to evaluate the efficacy of nemolizumab, a monoclonal antibody targeting IL-31, on disease activity and pruritus of BP.

An intriguing aspect of the pathogenesis of BP is that antibody/antigen binding activates different pathways, which seem to act in parallel rather than as a single cascade. Hence, combining different target therapies will represent a feasible way to reduce the cumulative exposure of patients to systemic steroids. In a merely speculative manner, combination of rituximab and dupilumab might effectively target the T-B-cell cross-talk involved in the loss of tolerance against BP autoantigens; while, combination of anti-complement drugs and either neutrophil-or eosinophil-targeting therapies might be best suited to impair the effector phase of BP inflammation and pruritus. Indeed, with the number of available therapeutic options rapidly increasing, clinicians should focus on identifying comorbidities, clinical variables (e.g., bullous vs. non-bullous phenotypes and pruritus intensity), laboratory [e.g., neutrophil-rich vs. eosinophil-rich infiltrates at histopathology, or the intensity of complement deposition at direct immunofluorescence (DIF)] and serological findings (e.g., titer of IgG and IgE antibodies against BP180/BP230) or molecular factors (e.g., cytokine concentration) which may influence therapy-response and decision-making.

RM contributed to the design, writing, and speculation. FS, DD, and CP contributed to the writing and speculation. EA contributed to the design and writing. GD contributed to the design, writing, and supervision. All authors contributed to the article and approved the submitted version.