Association of the R92Q TNFRSF1A mutation and extracranial deep vein thrombosis in patients with Behçet's disease
Behçet's disease is a chronic, relapsing, multisystemic inflammatory disorder characterized by recurrent oral and genital ulcers and by ocular, articular, vascular, and central nervous system involvement. The tumor necrosis factor α (TNFα) pathway is likely involved in the pathophysiology of Behçet's disease. One of the 2 TNFα receptors is TNF receptor superfamily 1A (TNFRSF1A). We searched for R92Q TNFRSF1A mutations in patients with Behçet's disease.
A search for TNFRSF1A mutations was performed by polymerase chain reaction amplification of the TNFRSF1A gene, followed by denaturing high-performance liquid chromatography scanning.
Among the 74 unrelated European patients with Behçet's disease, 5 (6.8%) carried the R92Q TNFRSF1A mutation. The frequency of the R92Q mutation in patients with Behçet's disease was significantly higher than that in controls (P = 0.006 by Fisher's exact test). Deep vein thrombosis was significantly associated with the R92Q mutation (P = 0.001 [with Bonferroni adjustment for multiple comparisons]). Among the 30 patients with thrombosis, 10 had cerebral thrombophlebitis. None of these patients had the R92Q mutation. Among the 20 patients with Behçet's disease who had extracranial deep vein thrombosis, 6 had the R92Q mutation, whereas 14 did not (P < 0.0001)
The R92Q mutation in patients with Behçet's disease is associated with an increased risk of extracranial venous thrombosis. This new finding may help in understanding the complex prothrombotic state in patients with Behçet's disease.
Behçet's disease is a chronic, relapsing, multisystemic inflammatory disorder affecting both men and women. Although recurrent mucocutaneous lesions (mainly oral and genital ulcers) are the main symptoms, ocular (prevalence of 50–70%), articular (50%), vascular (30%), and central nervous system (CNS) involvement may occur. Behçet's disease is most common along the “Old Silk Route,” which spans the region from Japan and China in the Far East to the Mediterranean Sea, including countries such as Turkey and Iran. The diagnosis of Behçet's disease is based on clinical criteria defined by the International Study Group for Behçet's Disease in 1990 (1).
The etiology of Behçet's disease remains obscure, but vascular injury, enhanced production of proinflammatory cytokines, and neutrophil hyperfunction (2, 3) are characteristic of Behçet's disease. Tumor necrosis factor (TNF) seems to play a major role in the pathophysiology of Behçet's disease (3). TNF is a multifunctional cytokine secreted by monocytes or lymphocytes. TNF acts through the 2 distinct members of the TNF receptor superfamily, TNFRSF1A (also referred to as p55, TNFRI, and CD120a) and TNFRSF1B (TNFRII, p75, p80, CD120b). Increased levels of circulating TNF in Behçet's disease have been reported, as have high levels of TNF in aqueous humor from patients with Behçet's disease–associated uveitis (3). Increased numbers of monocytes and T lymphocytes expressing the γ/δ T cell receptor that overproduce TNF are found in patients with active disease (3). The presence of the TNF-1031C allele is independently associated with susceptibility to Behçet's disease in Caucasian patients, although the functional relevance of this polymorphism in disease pathogenesis remains unclear (3). The central pathogenic role of TNF in Behçet's disease is further supported by the beneficial effect of therapeutic agents that down-regulate TNF production, such as thalidomide (4), which selectively inhibits the production of TNF in human monocytes, and pentoxifylline, which selectively inhibits TNF transcription (5). More recently, dramatic effects of monoclonal antibody therapy directed to TNF (infliximab) or to its soluble receptor, p75 (etanercept), in refractory Behçet's disease have been reported (6).
TNF receptors are part of the TNF pathway and might be implicated in the pathogenesis of Behçet's disease. Increased levels of soluble TNF receptor (p75) have been reported in the blood of patients with Behçet's disease (3). The TNFR2 polymorphism has been recently assessed in Korean patients with Behçet's disease, but no difference between these patients and healthy individuals was found (7). TNFRSF1A mutations have been reported in TNF receptor–associated periodic syndrome (TRAPS; OMIM no. 134610) (8, 9). This autosomal-dominant disorder shares clinical features with Behçet's disease (8). The course of TRAPS, like that of Behçet's disease, is characterized by remitting/relapsing exacerbations of inflammation. The inflammatory episodes of both Behçet's disease and TRAPS involve skin, eye, joint, and scrotal involvement (8). Like Behçet's disease, TRAPS is responsive to treatment with steroids and etanercept. Additionally, in a previous study on the clinical spectrum on TRAPS, we identified the R92Q TNFRSF1A mutation in one patient with Behçet's disease (10).
Taken together, these findings prompted us to search for R92Q TNFRSF1A mutations in patients with Behçet's disease. We assessed 74 unrelated European patients with Behçet's disease and found that 6.8% had the R92Q TNFRSF1A mutation. The presence of the R92Q TNFRSF1A mutation in patients with Behçet's disease was significantly associated with extracranial deep vein thrombosis.
PATIENTS AND METHODS
One hundred twelve unrelated patients (40 women and 72 men) with Behçet's disease were included in the study. The diagnosis of Behçet's disease was made according to international criteria for Behçet's disease (1). Thirty-eight of the patients were Maghrebian (Maghreb is the area comprising the countries of Morocco, Algeria, and Tunisia), and 74 patients were European. All 112 of the patients had recurrent oral ulcers. The other main clinical features of Behçet's disease were as follows: genital ulcers (n = 81 patients), eye involvement (n = 70), skin involvement (n = 70), articular involvement (n = 47), deep venous thrombosis (n = 30), positive pathergy test (n = 22), and inflammatory CNS involvement (excluding strokes and cerebral thrombophlebitis) (n = 14).
A European control group comprised 909 healthy persons, 831 of whom were from Glasglow, Belfast, and Strasbourg and had been previously genotyped for the R92Q mutation (11). Because most patients with Behçet's disease were from Paris and its area, 78 controls from the Paris area were additionally genotyped. For all of the European controls, all 4 of their grandparents had to have been born in Europe. A Maghrebian control group was composed of 89 healthy Maghrebians.
The search for TNFRSF1A mutations was performed by polymerase chain reaction amplification of the TNFRSF1A gene followed by denaturing high-performance liquid chromatography scanning, as described previously (10, 11)
Fisher's exact test was applied to compare the R92Q TNFRSF1A mutation frequencies and the frequencies of the main clinical features of Behçet's disease. A Bonferroni adjustment was applied for multiple comparisons (corrected P).
Increased frequency of the R92Q TNFRSF1A mutation in patients with Behçet's disease.
The R92Q mutation was found in 12 of 909 European controls and in 5 of 74 unrelated European patients with Behçet's disease (1.3% versus 6.8%; P = 0.006 by Fisher's exact test). All patients were heterozygous for the R92Q mutation. No R92Q mutation was observed in the group of 89 Maghrebian controls. This frequency was not significantly different from that among the European controls (P > 0.05 by Fisher's exact test). One of the 38 Maghrebian patients with Behçet's disease had the R92Q mutation.
Genotype/phenotype correlation in patients with Behçet's disease and the TNFRSF1A mutation.
We next studied whether there was a correlation between the presence of the R92Q TNFRSF1A mutation and the main clinical manifestations of the 112 patients with Behçet's disease. Among the 6 patients with Behçet's disease and the R92Q mutation, all 6 had genital ulcers, 1 had eye involvement, 2 had skin involvement, 1 had articular involvement, and all 6 had venous thrombosis. The absence of eye involvement and the presence of venous thrombosis were significantly associated with the R92Q mutation (P = 0.027 and P = 0.0002, respectively, versus patients without the R92Q mutation, by Fisher's exact test). After correction for multiple comparisons, only the association with venous thrombosis was significant (P = 0.001) (Table 1). Among the 30 patients with venous thrombosis, 10 had cerebral thrombophlebitis; none of these 10 patients had the R92Q mutation. Among the 20 patients with Behçet's disease who had peripheral (i.e., extracranial location) venous thrombosis, 6 had the R92Q mutation, whereas 14 did not (P < 0.0001 by Fisher's exact test).
Table 1. Correlation between the presence of the R92Q TNFRSF1A mutation and the main clinical manifestations of BD*
Features of TRAPS in patients with Behçet's disease and TNFRSF1A mutations.
Because the R92Q mutation has been described in TRAPS (9, 10), we looked for clinical features of TRAPS in the 6 patients with Behçet's disease and TNFRSF1A mutations. Of these 6 patients, none had typical features of TRAPS. One patient with the R92Q mutation had a history of pericarditis, a finding that is unusual in Behçet's disease but that has been reported as a clinical manifestation in TRAPS (10). The family histories of all patients with Behçet's disease were negative for cases suggestive of TRAPS or Behçet's disease.
The R92Q carrier frequency in the group of European patients with Behçet's disease was significantly higher than that in the European control group (6.8% and 1.3%, respectively; P = 0.006). The frequency of the R92Q mutation in the European control group was in good accordance with that observed in a previous North American study (9). The R92Q mutation has been detected in 1 (1.4%) of 71 subjects of Irish origin and in 3 (0.95%) of 317 subjects of Caucasian origin (9).
All patients with Behçet's disease had a diagnosis fulfilling the criteria of the International Study Group for Behçet's Disease (1). None of these patients had typical features of TRAPS, an autosomal-dominant disorder associated with mutations of the TNFRSF1A gene (9, 10). We can conclude that the high frequency of the R92Q TNFRSF1A mutation was associated with Behçet's disease and was not associated with underdiagnosed TRAPS. A high frequency (5.1%) of the R92Q mutation has been previously reported in a cohort of 135 patients with rheumatoid arthritis (9). These findings suggest that this low penetrance of the TNFRSF1A mutation contributes to complex genetic disorders characterized by an autoinflammatory process. Although we did not find a significant association between the presence of the R92Q mutation and other genetic factors known to contribute to Behçet's disease, such as B51 or MICA A6 (data not shown), it is likely that the phenotypic manifestations of R92Q in Behçet's disease depend on other linked or unlinked modifying genes and/or on modifying environmental factors.
In patients with Behçet's disease, the presence of the R92Q mutation was associated with an increased risk of developing venous thrombosis, especially extracranial venous thrombosis. Cambien et al (12) have shown that under inflammatory conditions, TNFα exerted an antithrombotic effect that is mediated through TNF receptors located in the blood wall. TNFα, through its TNF receptors, stimulates inducible nitric oxide synthase (iNOS), leading to the generation of NO, which in turn inhibited thrombi formation (12). In the absence of TNFα, the process of thrombosis was not significantly different between mice knocked-out for TNF receptors and their corresponding wild-type, indicating that in the absence of inflammation, TNF receptors were not involved in thrombosis (12).
Interestingly, there is a body of work suggesting that NOS activity is impaired in the setting of Behçet's disease. Reduced plasma nitrate and nitrite levels have been observed in patients with active Behçet's disease (13). Plasma from patients with Behçet's disease decreased the level of NO production in cultured human umbilical vein endothelial cells (14). Endothelium-dependent arterial dilatation of the brachial artery, which is mediated by the release of endothelial NO, is impaired in patients with Behçet's disease (15). Consistent with these results and our findings, we propose that the R92Q mutation may interfere with the local antithrombotic effect of TNFα by impairing the endothelial TNF receptor–mediated iNOS stimulation and could not counterbalance the procoagulant activity induced by the vascular wall inflammation. This last point is of particular importance, because in the in vivo model described by Cambien et al (12), thrombosis was always initiated by a vascular wall injury. Immune or inflammatory vascular cell damage is an essential element of Behçet's disease but not of TRAPS. This may explain why there is no increased frequency of thrombosis in TRAPS even though this disease exhibits high levels of TNF and mutation of TNF receptor. The R92Q mutation has also recently been associated with an increased risk of premature myocardial infarction and carotid intima-media thickness (11).
In conclusion, we observed a high frequency of the R92Q TNFRSF1A mutation in patients with Behçet's disease. In patients with Behçet's disease, the R92Q mutation is associated with an increased risk of extracranial venous thrombosis. Although the exact contribution of mutation to the genesis of thrombosis is still to be determined, this new finding may help to better understand the complex prothrombotic state in patients with Behçet's disease. Because our study lacked statistical power for the Maghrebian population, we cannot reach definite conclusions regarding Maghrebian patients with Behçet's disease. However, it is likely that the conclusions drawn from the analysis of European patients with Behçet's disease can be extended to the Maghrebian population inasmuch as we found no R92Q mutation among 89 Maghrebian controls.
We thank François Poirier, who provided data from the Etude Cas-Témoins de l'Infarctus du Myocarde (ECTIM) study.