Behçet’s disease (BD) is an inflammatory disorder of unknown aetiology characterised by recurrent attacks affecting the mucocutaneous tissues, eyes, joints, blood vessels, brain and gastrointestinal tract. It is a multifactorial disease, and several triggering factors including oral cavity infections and viruses as described by Hulusi Behçet himself [1] may induce inflammatory attacks in genetically susceptible individuals. Investigations of inflamed tissues suggest a vasculitis or vasculopathy with mixed-cellular perivascular infiltrates and thrombotic tendency as the underlying pathology, and because of involvement of both venous and arterial vessels of all sizes, BD has been classified as “variable vessel vasculitis” in the 2012 International Chapel Hill Consensus Conference on the Nomenclature of Vasculitides [2].

BD has unique immune-mediated inflammatory characteristics, and because of the absence of pathogenic autoimmune T and B cell responses as well as the similarities of its recurrent manifestations to various hereditary autoinflammatory disorders (Fig. 1), BD had already been suggested as an example of acquired or multifactorial autoinflammatory diseases [3]. However, during an attempt to classify inflammatory disorders, BD was positioned in between monogenic autoinflammatory and monogenic autoimmune disorders, as a “mixed-pattern disease”, because of its strong association with a major histocompatibility complex (MHC) Class I antigen, human leukocyte antigen (HLA)-B*51, as an evidence of adaptive immunity component [4].

Fig. 1
figure 1

Clinical features of Behçet’s disease overlapping with monogenic autoinflammatory disorders. Abbreviations: MKD mevalonate kinase deficiency, TRAPS tumour necrosis factor receptor-associated periodic syndrome, CAPS cryopyrin-associated periodic syndrome, FMF familial Mediterranean fever, PAPA pyogenic arthritis, pyoderma gangrenosum and acne syndrome, DIRA deficiency of interleukin-1 receptor antagonist

Full size image

Monogenic disorders are naturally occurring models linking genetic variations to functional abnormalities, and all of these classification efforts in common inflammatory disorders including BD actually aim to better understand the observed pathologies rather than to discuss the nosology by giving them another name or lumping the splits into larger but more heterogeneous groups [5]. Hopefully, at the end, all immunogenetic and functional studies are expected to provide new data to translate into clinical practice by elucidation of pathogenic mechanisms and critical inflammatory pathways to target in patients with BD as well as with other common inflammatory conditions [6].

Definition of autoinflammatory disorders

Autoinflammatory disorders were originally described in 1999 as rare hereditary diseases associated with inborn errors in the innate immune system genes and characterised by seemingly unprovoked episodes of inflammation developing without pathogenic high-titre autoantibodies or antigen-specific T cells [7]. With the improved understanding of their pathogenic basis, the definition of autoinflammatory disorders was revised to encompass a larger spectrum of clinical entities with a significant host predisposition and characterised by abnormally increased inflammatory response, which is mediated predominantly by the cells and molecules of the innate immune system, without ruling out possible contribution of adaptive response [7, 8]. This updated concept has led to understanding autoinflammatory components within the mild to moderate inflammatory findings of several common traits including atherosclerosis and diabetes; randomised clinical trials are currently testing the potential of treatments targeting interleukin (IL)-1-driven inflammation in their management [9–12].

Clinical findings of BD

BD has several clinical and inflammatory features fitting well to the revised definition of autoinflammatory disorders [3, 8]. Most of the clinical findings of BD are self-limited with a variable recurrence rate, and some of them such as oral aphthous ulcers or papulopustular skin lesions may heal spontaneously without any scarring. But some manifestations, such as retinal, vascular and parenchymal neurologic findings, can cause permanent tissue damage during each inflammatory episode. Several typical clinical findings of BD overlap with the manifestations of monogenic autoinflammatory conditions such as mevalonate kinase deficiency (Fig. 1) [3], which may even result in diagnostic confusions especially in countries where BD is not prevalent [13]. Challenges associated with similar manifestations are also true for the differential diagnosis of BD with complex autoinflammatory conditions such as Crohn’s disease [14].

Dysregulated activation of IL-1 beta due to inborn errors in the inflammasome components is the main finding in several monogenic autoinflammatory conditions [15]. However, other proinflammatory cytokines may also take part in the development of disease-specific features of other hereditary and acquired forms of autoinflammatory disorders. Mucocutaneous lesions are common in almost all of these diseases; however, variability of skin lesions may provide some hints for the differential diagnosis as well as for the involved pathogenic pathways in those disorders. In NLRP3-inflammasome-related disorders such as cryopyrin-associated periodic syndrome (CAPS), increased IL-1 beta production results in erythematous or urticarial rash, associated with dermal neutrophilic infiltrates [15]. However, pustular or acne-like skin findings of BD resemble more to that of deficiency of IL-1 receptor antagonist (DIRA) syndrome or pyogenic arthritis, pyoderma gangrenosum and acne (PAPA) syndrome, and orogenital aphthous ulcers of BD resemble more to mevalonate kinase deficiency; all of which suggest involvement of other cytokines, such as IL-1 alpha, in addition to IL-1 beta, in the pathogenesis of mucocutaneous BD lesions. Manifestation-specific investigations are expected to identify regulatory mechanisms involved in the biology of IL-1 alpha, IL-1 beta, IL-18, IL-36 and other innate pro-inflammatory cytokines in different organs, which may help explain BD pathogenesis further.

Significant host predisposition: HLA-B*51 and beyond

BD is not a Mendelian disease, but it has a significant genetic component documented by a familial aggregation with a high λs value [16]. HLA-B*51 is the strongest genetic susceptibility factor for BD described so far, and this association has been replicated in different ethnic groups at different strengths [17, 18]. Presence of HLA-B*51 is important in the definition of typical BD-phenotype, since it is mainly associated with classical disease manifestations, but not with some rare findings such as gastrointestinal disease [19, 20]. Recent genomewide association studies (GWASs) confirmed the dominant role of HLA-B region in the genetic tendency to BD [21–23]. Detailed analysis of the MHC showed additional weaker but independent associations in the Class I region, especially around the HLA-A locus [21, 22, 24].

HLA-B*51 may contribute to BD pathogenesis by several different mechanisms, involving both adaptive (by presentation of certain pathogenic peptides to CD8 T cells) and innate (by interacting with natural killer cell receptors as well as activating intracellular inflammatory pathways associated with heavy chain folding problems and endoplasmic reticulum stress) immune responses (Fig. 2) [18]. Identification of endoplasmic reticulum aminopeptidase 1 (ERAP1) polymorphisms as a recessively inherited risk factor only in HLA-B*51 positive individuals revealed the important role of peptides loaded onto antigen-binding groove of HLA-B*51 [25]. ERAP1 polymorphisms may affect the enzymatic activity and possibly peptide specificity of this endoplasmic reticulum aminopeptidase, and organ-specific peptidome produced by ERAP1 and possibly other peptidases may affect BD susceptibility. Moreover, investigation of HLA-A and HLA-B alleles in a large cohort of patients and controls provided further information about the pathogenic mechanisms associated with HLA-B*51 by defining weaker new risk alleles such as HLA-B*15, HLA-B*27 and HLA-A*26 as well as protective alleles such as HLA-B*49 and HLA-A*03 [26]. These HLA alleles have different binding affinities for the second position of the peptides (P2), and co-inheritance of HLA-B*49 or HLA-A*03 may neutralise the HLA-B*51-associated risk for BD. Therefore, MHC region-associated risk for BD is determined by the combination of HLA-B and HLA-A alleles as well as ERAP1 alleles-defined peptidome, which may indicate the important role of altered regulation of cellular cytotoxicity in BD pathogenesis [26]. However, it is still not clear yet whether tissue-specific peptidome and stimulation of HLA-B*51 restricted cytotoxic T cell response or antigen-independent innate inflammatory response induced by folding problems associated with inappropriate matches between slow-folding HLA-B*51 molecules and available peptides is the main pathogenic mechanism in BD pathogenesis. Elucidation of MHC Class I-mediated inflammatory mechanisms is expected to help in the development of novel therapeutic approaches such as interactions with peptide production [27], peptide loading or folding properties, and it may also make the induction of long-term remission possible.

Fig. 2
figure 2

Possible HLA-B*51-related pathogenic mechanisms in Behçet’s disease

Full size image

Abnormally increased inflammatory response

Increased or exaggerated inflammatory response to non-specific stimuli is a known feature of BD, and it is even used as a diagnostic test in the skin pathergy reaction, as an erythematous papulopustular response to hypodermic needle trauma, which extends to 48 h [28–30]. Significantly larger eryhthematous skin reaction compared to healthy controls can also be elicited in BD patients by injecting monosodium urate crystals, a known NLRP3-inflammasome activator, intradermally [31]. A prominent innate immune response has consistently been reported in BD patients by increased expression of proinflammatory cytokines including IL-1, IL-6, IL-8 and tumour necrosis factor (TNF) [32, 33]. Lipopolysaccharide (LPS)-induced production of IL-1 by monocytes may even be higher in BD patients compared to patients with familial Mediterranean fever, known to be associated with MEditerranean FeVer gene (MEFV) variations causing increased IL-1 activation [32]. Similarly, activated neutrophils are frequently observed in pathological specimens, and BD has generally been classified amongst neutrophilic dermatoses [34].

Targeted deep re-sequencing of the selected innate immune genes identified associations with rare variants in toll-like receptor 4 (TLR4), MEFV, and nucleotide-binding oligomerization domain-containing protein 2 (NOD2) genes in BD patients [35]. The frequency of the most penetrant p.Met694Val mutation of MEFV gene, a variant not found in Japanese population, was found to be increased in Turkish patients [35]. Interestingly, despite all clinical genetic similarities, Crohn’s disease-associated risk alleles of the TLR4 and NOD2 genes were found to be protective for BD, suggesting a stronger inflammatory response to microbial triggers, which may explain relative protection from intestinal disease in BD compared to Crohn’s disease [35].

Along with MHC Class I and ERAP1 alleles, association with these innate response gene variations suggests a defect in sensing and processing of pathogen-associated and/or danger-associated molecular patterns leading to a stronger inflammatory response in BD pathogenesis [23].

Treatments targeting innate cytokines, especially TNF, have successfully been used in BD patients, who are refractory to standard of care [36]. Several case reports suggest that IL-1 and IL-6 are also promising targets in patients resistant or intolerant to other regimens including anti-TNFs [37]. Favourable responses to IL-1 blockade by anakinra, canakinumab or gevokizumab have been described for several BD manifestations including refractory uveitis and mucocutaneous lesions [38, 39], and it has been suggested that anti-IL-1 agents may provide a better tuberculosis-related safety profile in BD patients compared to anti-TNF agents [39]. Blocking IL-6 activity with tocilizumab has also been shown to be effective in BD patients with refractory neurologic or ocular disease [40, 41]. However, no favourable response or a paradoxical exacerbation was reported for mucocutaneous BD manifestations with tocilizumab treatment [42, 43]. Results of randomised controlled trials for agents targeting IL-1 and IL-6 have still been waited to establish their efficacy and safety in different manifestations of BD patients.

Activation of adaptive immune response has also been documented in BD patients [33], with oligoclonal T cell expansions correlating with disease exacerbations as well as with findings of Th1 and to lesser extent Th17 type of polarizations [44–47]. Despite strong HLA-B*51 association, so far, no HLA-B*51 restricted cytotoxic response has been documented as the cause of any BD manifestations. On the other hand, associations with IL10, IL23R and STAT4 gene variations revealed by GWAS may define a response pattern affecting the regulation of innate and adaptive immune responses [23]. Documented associations with MHC Class I alleles (HLA-B*27 and HLA-C*06) and ERAP1 variations in ankylosing spondylitis (AS) and psoriasis, respectively, and association with IL10, NOD2 and TLR4 variations in Crohn’s disease and with IL23R variations in all three diseases suggest that BD shares several genetic susceptibility factors with spondyloarthritis, which affect certain inflammatory pathways [21–23, 25, 35]. Regarding the combined effects of MHC Class I and ERAP1 genes, protective ERAP1 alleles in HLA-B*27 positive AS and HLA-C*06 positive psoriasis patients, respectively, are associated with an increased risk for HLA-B*51 positive BD patients. Therefore, pathologic response may depend on the set of ERAP1 polymorphisms and their matching MHC Class I alleles, but whether or not the different combinations of ERAP1 polymorphisms with HLA-B*27/HLA-C*06 or HLA-B*51 results in a similar inflammatory response should be elucidated.

The pathogenic mechanisms associated with spondyloarthritis-related IL23R variations have not been clarified yet. Increased IL-23 signalling may be critical in the development of enthesitis, through activation of innate lymphoid cells [48]. On the other hand, its role in Th17 type polarisation is not clear, and pathogenic role of IL-17 and Th17 cells may show variation amongst spondyloarthritides. Although blocking IL-17 activity has been shown to be effective in AS, psoriasis and psoriatic arthritis [49], a randomised controlled trial could not show any efficacy of secukinumab, an anti-IL-17 monoclonal antibody, in BD patients with uveitis refractory to conventional treatments [50], very similar to the negative results obtained in patients with Crohn’s disease [51].

On the other hand, open-label or controlled observations showed beneficial effects with monoclonal antibodies targeting T cells such as anti-CD52, alemtuzumab or Campath [52], or B cells, such as rituximab [53]. Therefore, the possible contribution of adaptive immune response in different BD manifestations deserves further studies.

Lastly, a recent randomised controlled trial showed favourable effects of apremilast, a selective phosphodiesterase 4 (PDE4) inhibitor, compared to placebo in recurrent oral ulcers of BD patients [54]. Apremilast had already been shown to be effective in psoriasis and psoriatic arthritis by increasing intracellular cyclic AMP (cAMP) level, which results in an anti-inflammatory state by inhibiting the activity of TNF, IFN-gamma, IL-12 and IL-23 and by increasing IL-1Ra and IL-6 levels in inflammatory cells, which may also explain the observed efficacy in BD patients [55].

In summary, clinical trial data supports critical role of innate cytokines TNF, IL-1 and IL-6 in the development of inflammatory episodes of BD, and targeting T cells or B cells may also provide favourable results. Successful results obtained by apremilast also document the potential of small molecules by regulating several proinflammatory cytokines through modulating cAMP levels.

Conclusions

BD has several autoinflammatory features including recurrent self-limited manifestations overlapping with monogenic autoinflammatory disorders, significant host predisposition and abnormally increased inflammatory response, with a strong innate immune activation. HLA-B*51 is the strongest susceptibility factor affecting the disease risk and typical phenotype, and non-HLA genetic associations such as ERAP1, IL23R and IL10 variations suggest shared genetic susceptibility factors with spondyloarthritis. Combined effects of peptides produced by ERAP1 and other proteases and MHC Class I alleles may change the disease risk, and it is not clear yet whether peptide-HLA allele combination results in an adaptive response by self-antigen directed cytotoxic response or an innate response by affecting NK cell activity or causing unfolded protein response. Understanding of MHC Class I-driven inflammatory response is expected to provide insights for the development of better treatment and remission-induction options in BD as well as in AS and psoriasis.