Prothrombotic state and impaired fibrinolysis in bullous pemphigoid, the most frequent autoimmune blistering disease

Authors

  • A. V. Marzano,

    1. Unità Operativa di Dermatologia, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Dipartimento di Fisiopatologia Medico-Chirurgica e dei Trapianti, Università degli Studi di Milano, Milano, Italy
    Search for more papers by this author
  • A. Tedeschi,

    1. Unità Operativa di Allergologia e Immunologia Clinica, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milano, Italy
    Search for more papers by this author
  • I. Polloni,

    1. Unità Operativa di Dermatologia, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Dipartimento di Fisiopatologia Medico-Chirurgica e dei Trapianti, Università degli Studi di Milano, Milano, Italy
    Search for more papers by this author
  • C. Crosti,

    1. Unità Operativa di Dermatologia, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Dipartimento di Fisiopatologia Medico-Chirurgica e dei Trapianti, Università degli Studi di Milano, Milano, Italy
    Search for more papers by this author
  • M. Cugno

    Corresponding author
    • Medicina Interna, Dipartimento di Fisiopatologia Medico-Chirurgica e dei Trapianti, Università degli Studi di Milano, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milano, Italy
    Search for more papers by this author

Correspondence: M. Cugno, Internal Medicine, Department of Pathophysiology and Transplantation, University of Milan, Via Pace 9, 20122 Milano, Italy.

E-mail massimo.cugno@unimi.it

Summary

Bullous pemphigoid (BP) is a potentially life-threatening autoimmune blistering disease that is burdened with an increased risk of cardiovascular events. In BP, there is an interplay between inflammation and coagulation both locally, which contributes to skin damage, and systemically, which leads to a prothrombotic state. Fibrinolysis is an important defence mechanism against thrombosis, but has only been studied locally in BP and no systemic data are available. The aim of this observational study was to evaluate systemic fibrinolysis and coagulation activation in patients with BP. We measured parameters of fibrinolysis and coagulation by immunoenzymatic methods in plasma from 20 patients with BP in an active phase and during remission after corticosteroid treatment. The controls were 20 age- and sex-matched healthy subjects. Plasma levels of plasminogen activator inhibitor type 1 (PAI-1) antigen, PAI-1 activity and tissue plasminogen activator (t-PA) antigen were significantly higher in the BP patients with active disease than in healthy controls (P = 0·0001 for all), as were the plasma levels of the fibrin fragment d-dimer and prothrombin fragment F1+2 (P = 0·0001 for both). During remission after treatment, levels of PAI-1 antigen and PAI-1 activity decreased significantly (P = 0·008 and P = 0·006, respectively), and there was also a significant decrease in plasma levels of d-dimer (P = 0·0001) and F1+2 (P = 0·0001). Fibrinolysis is inhibited in patients with active BP, due mainly to an increase in plasma levels of PAI-1. Corticosteroids not only induce the regression of BP lesions, but also reduce the inhibition of fibrinolysis, which may contribute to decreasing thrombotic risk.

Introduction

Bullous pemphigoid (BP) is an autoimmune blistering disease that occurs typically in the elderly [1] and is burdened with a high risk of death, due mainly to sepsis and cardiovascular events [2]. It involves the skin and rarely the mucous membranes, and is characterized by the presence of blisters usually surrounded by erythematous–oedematous lesions. The diagnosis is supported by histology showing a subepidermal blister with a dermal mixed inflammatory cell infiltrate usually rich in eosinophils, and a direct immunofluorescence examination of perilesional skin revealing the linear deposition of immunoglobulin (Ig)G and/or C3 in the basement membrane zone (BMZ). Circulating anti-BMZ autoantibodies can be detected by means of an enzyme-linked immunosorbent assay (ELISA) for two hemidesmosomal antigens, BP180 and BP230 [3, 4]. Autoantibodies against these antigens play an important role in the pathogenesis of BP, as well as complement activation and leucocyte infiltration [1, 5]. Inflammatory cells (particularly autoreactive T cells and eosinophils) participate in blister formation by producing and releasing a number of cytokines and soluble factors that amplify and maintain tissue damage [6-8]. The inflammatory response induces an activation of blood coagulation which is involved both locally, by amplifying the inflammatory network in lesional skin, and systemically, by leading to a prothrombotic state [4, 9-12]. Indeed, it has been reported that the risk of thrombosis is increased in patients with BP [2, 13, 14] and we have found an annual incidence of venous thrombosis of 8% [4], which is clearly higher than that observed in the general elderly population (0·28–0·41% per year) [15]. It is well known that the inflammatory response inhibits fibrinolysis, which contributes to the prothrombotic state seen in conditions such as sepsis [16], inflammatory bowel diseases [17] and rheumatoid arthritis [18]. However, to the best of our knowledge, no data are available concerning systemic fibrinolysis in BP patients, although it has been shown to be involved at local level in lesional skin in humans and experimental BP models [19-23].

With this background, we evaluated systemic fibrinolysis by measuring the plasma parameters of 20 patients with BP in an active phase and in clinical remission after systemic corticosteroid treatment, and correlated the results with coagulation markers and the parameters of disease activity.

Materials and methods

Patients

We conducted an observational study enrolling 20 consecutive patients with previously untreated active BP (10 males and 10 females; mean age 76 years, range 53–99) who were admitted to our Dermatology Department from January 2010 to June 2011. The diagnosis of BP was established on the basis of clinical and immunopathological criteria. All the patients had a clinical picture of generalized BP without any mucous membrane involvement (mean disease duration: 1 month, range 0–2); the skin lesions (vesiculobullous and/or erythematous–oedematous lesions) covered a median 40% of total body area (range 20–60%). Direct immunofluorescence examinations of the perilesional skin revealed the linear deposition of IgG and/or C3 in the BMZ in all cases, circulating anti-BP180 autoantibodies were detected by means of an ELISA. Concomitant neoplastic or inflammatory diseases were excluded on the basis of clinical and instrumental examinations. None of the patients had thyroid dysfunction or atrial fibrillation and were taking drugs affecting coagulation. Three of the 20 BP patients had type 2 diabetes and were receiving treatment with oral anti-diabetic drugs with an acceptable disease control (haemoglobin A1c values 6·5, 6·7 and 7·0, respectively).

After taking the blood samples, patients with active disease were treated with methylprednisolone at an initial dose of 0·5–0·75 mg/kg/day. When either new lesions or pruritic symptoms have not occurred for at least 2 weeks, the tapering of steroid was started until reaching the minimal dose of 0·05–0·1 mg/kg/day.

All the patients were also studied during clinical remission, defined as the absence of any new BP lesions with the complete healing of the previous lesions for a minimum of 4 weeks. At the time of sampling, they were being treated with low-dose corticosteroids (methylprednisolone 4 mg daily).

Controls

The control group consisted of 20 age- and sex-matched apparently healthy subjects with no history of thrombosis (10 males and 10 females; mean age 75 years, range 55–94).

The protocol was approved by our Institutional Review Board, and all the subjects gave their informed consent before participating in the study.

Blood sampling

Morning fasting blood samples were taken from all the BP patients and the 20 normal subjects in vacutainer tubes (Beckton & Dickinson, Rutherford, NJ, USA) by means of the clean puncture of an antecubital vein with minimal stasis, using sodium citrate 3·8% as anti-coagulant. The samples were centrifuged at 2000 g at 4°C to obtain plasma, which was then divided into aliquots, frozen and stored at −80°C until testing.

Plasminogen activator inhibitor type 1 (PAI-1) antigen

Plasminogen activator inhibitor type 1 (PAI-1) antigen was measured using a commercially available ELISA (Innotest PAI-1; Byk Gulden, Konstanz, Germany). The intra- and interassay coefficients of variation (CVs) were, respectively, 8 and 13%.

PAI-1 activity

PAI-1 activity was measured using a commercially available bioimmunoassay (Zymutest PAI-1 activity; Hyphen BioMed, Neuville-sur-Oise, France) with intra- and interassay CVs of 3·5 and 5·6%.

Thrombin activatable fibrinolysis inhibitor (TAFI) antigen

TAFI antigen was measured using a commercially available ELISA (Zymutest TAFI antigen; Hyphen BioMed) with intra- and interassay CVs of 7 and 14%.

Tissue plasminogen activator antigen (t-PA)

t-PA antigen was measured using a commercially available ELISA (Imunolyse tPA; Biopool, Umea, Sweden), in accordance with the manufacturer's instructions. The intra- and interassay CVs were, respectively, 6·5 and 8%.

d-dimer

d-dimer levels were measured by means of an ELISA (Zymutest d-dimer; Hyphen BioMed), in accordance with the manufacturer's instructions. The intra- and inter-assay CVs were, respectively, 10 and 15%.

Prothrombin F1 + 2

Prothrombin fragment F1+2 levels were measured using a sandwich ELISA (Enzygnost F1+2; Behring Diagnostic GmbH, Frankfurt, Germany), with intra- and interassay CVs of, respectively, 5 and 8%.

C-reactive protein (CRP)

CRP was measured by means of an ELISA (Zymutest CRP; Hyphen BioMed, Andresy, France) with intra- and inter-assay coefficients of variation (CVs) of 7–11%.

Statistics

As the data were positively skewed, they were log-transformed before analysis and are given as the anti-log values of the mean values and standard deviations (SDs). Student's t-test for unpaired data was used to assess the statistical significance of the differences between the normal controls and the patients with active BP. The effect of treatment was analysed using Student's t-test for paired samples. Correlations were assessed by means of least-square linear regression. The significance level was set at P < 0·05. Data were analysed using the spss PC statistical package, version 17·00 (SPSS Inc., Chicago, IL, USA).

Results

Patients with active BP

Figure 1 shows that PAI-1 antigen and active PAI-1 levels were significantly higher in the 20 BP patients with active disease (25·06 ± 8·88 ng/ml and 15·65 ± 5·75 ng/ml) than in the 20 healthy controls (10·04 ± 7·80 ng/ml and 7·25 ± 5·49 ng/ml) (P = 0·0001 for both). Figure 2 shows that plasma t-PA levels were also significantly higher in the patients (34·70 ± 33·22 ng/ml versus 6·60 ± 6·78 ng/ml; P = 0·0001), whereas there was no significant between-group difference in TAFI levels (91·58 ± 23·93% versus 92·73 ± 20·61%).

Figure 1.

Plasma levels of plasminogen activator inhibitor type 1 (PAI-1) antigen and PAI-1 activity in 20 healthy controls and 20 patients with bullous pemphigoid during active disease (active BP) and during remission after corticosteroid treatment (remission BP).

Figure 2.

Plasma levels of thrombin activatable fibrinolysis inhibitor (TAFI) antigen and tissue plasminogen activator (t-PA) antigen in 20 healthy controls and 20 patients with bullous pemphigoid during active disease (active BP) and during remission after corticosteroid treatment (remission BP).

As shown in Fig. 3, plasma d-dimer and F1+2 levels were both markedly higher in the patients with active BP (2350 ± 3676 ng/ml and 551 ± 484 ng/ml) than in the controls (188 ± 107 ng/ml and 106 ± 42 ng/ml) (P = 0·0001 for both).

Figure 3.

Plasma levels of d-dimer and prothrombin fragment F1+2 levels in 20 healthy controls and 20 patients with bullous pemphigoid during active disease (active BP) and during remission after corticosteroid treatment (remission BP).

To rule out the possible influence of diabetes on our results, we have analysed differences in fibrinolysis and coagulation parameters between BP patients and normal controls after the exclusion of the three diabetic BP patients and their sex- and age-matched controls. In the 17 BP patients with active disease, PAI-1 antigen and active PAI-1 levels were significantly higher (22·13 ± 8·68 ng/ml and 16·76 ± 5·55 ng/ml) than in the 17 sex- and age-matched healthy controls (8·65 ± 6·29 ng/ml and 6·21 ± 4·37 ng/ml) (P = 0·0001 for both). Plasma t-PA levels were also significantly higher in the 17 patients (36·91 ± 32·02 ng/ml versus 6·09 ± 4·45 ng/ml; P = 0·0001). Finally, plasma d-dimer and F1+2 levels were both markedly higher in the 17 patients with active BP (2774 ± 3817 ng/ml and 631 ± 487 ng/ml) than in the 17 controls (183 ± 107 ng/ml and 106 ± 44 ng/ml) (P = 0·0001 for both).

In the patients with active BP, disease severity (expressed as the percentage of involved body surface area) correlated significantly with the number of blood eosinophils (r = 0·705, P = 0·01) and the plasma levels of d-dimer (r = 0·713, P = 0·0001) and F1+2 (r = 0·703, P = 0·001). Plasma CRP levels correlated directly with the levels of PAI-1 antigen (r = 0·722, P = 0·0001), PAI-1 activity (r = 0·514, P = 0·021), t-PA antigen (r = 0·547, P = 0·012) and F1+2 (r = 0·450, P = 0·047) and the number of blood eosinophils correlated with PAI-1 antigen (r = 0·585, P = 0·046), PAI-1 activity (r = 0·680, P = 0·015) and d-dimer (r = 0·710, P = 0·010). Anti-BP180 autoantibody levels only correlated with d-dimer (r = 0·495, P = 0·026) and F1+2 (r = 0·458, P = 0·042).

Patients with BP during treatment-induced remission

In the 20 BP patients during remission after treatment, the levels of PAI-1 antigen and active PAI-1 decreased significantly from 25·06 ± 8·88 ng/ml to 16·99 ± 7·05 ng/ml and from 15·65 ± 5·75 ng/ml to 11·19 ± 5·14 ng/ml (P = 0·008 and P = 0·006, respectively) (Fig. 1). The mean differences were 5·30 ng/ml [95% confidence interval (CI): 1·65–8·96 ng/ml] for PAI antigen and 4·00 ng/ml (95% CI: 1·66–6·35 ng/ml) for active PAI. There was an albeit not significant decrease in tPA levels (from 34·70 ± 33·22 to 32·74 ± 27·80 ng/ml). Plasma TAFI levels did not change significantly (Fig. 2), but there was a significant decrease in the plasma levels of d-dimer (from 2350 ± 3676 ng/ml to 571 ± 651; P = 0·0001) and F1+2 (from 551 ± 484 ng/ml to 188 ± 216; P = 0·0001). The mean differences were 2804 ng/ml (95% CI: 744–4865 ng/ml) for d-dimer and 414 ng/ml (95% CI: 191–638 ng/ml) for F1+2.

Discussion

The results of this study of a group of patients with active BP show that fibrinolysis is inhibited, due mainly to an increase in the plasma levels of PAI-1 activity and antigen, and that systemic corticosteroid treatment acts not only by inducing the regression of bullous and inflammatory BP lesions, but also by reducing the inhibition of fibrinolysis.

The homeostasis of the fibrinolytic system is finely regulated by plasminogen activators such as t-PA, and natural inhibitors such as PAI-1 and TAFI. It is thought that t-PA plays a relevant role in initiating fibrinolysis and thrombolysis. The high circulating levels of t-PA antigen in patients with active BP do not conflict with the reduction in fibrinolysis because the t-PA immunoassay largely measures circulating complexes of t-PA and PAI-1. Consequently, increased concentrations of t-PA antigen provide indirect information concerning PAI-1 expression and indicate reduced rather than increased fibrinolysis [24]. The inhibition of fibrinolysis may have important effects on systemic circulation; high plasma PAI-1 levels are generally considered to be a cardiovascular risk factor [25]. Clinical and experimental evidence suggests that the long-term effects of PAI-1 are crucial factors in the occurrence of thrombotic events. The increased risk of cardiovascular events in BP can therefore be attributed partially to the inhibition of fibrinolytic system, which may act synergistically with the previously demonstrated increased activation of blood coagulation associated with the disease [4, 9, 12]. Another factor possibly contributing to the increased risk of thrombosis in BP is the presence of anti-phospholipid antibodies, which have been detected in about 20% of cases in a series of 28 patients with this disease [26]. The possible influence on our results of comorbidities such as hyperthyroidism and diabetes, which may impair the fibrinolytic process [27, 28], has also been considered. None of our patients had thyroid dysfunction and the alterations of fibrinolysis and coagulation were evident even after the exclusion of the three diabetic patients.

Activation of the coagulation system has local effects on the skin (by contributing to inflammation, tissue damage and blister formation) and systemic effects on the blood stream that increase thrombotic risk [10, 11]. At local level, it has been demonstrated that the fibrinolytic system is activated in blister fluid taken from BP patients [21] and plays a critical role in blister formation in experimental BP by mediating the physiological activation of metalloproteinase-9 [23]. Moreover, in a model of cultured human keratinocytes, stimulation with antibodies to human BP180 led to high levels of tPA expression and release [22]. Our previous data [4, 9] confirm the involvement of fibrinolysis activation and coagulation activation in human BP blister fluid, as shown by high levels of d-dimer (a marker of fibrin degradation) and prothrombin fragment F1+2 (a marker of thrombin generation).

At systemic level, the increase in PAI-1 levels indicates that fibrinolysis is inhibited in BP. The positive correlation between plasma PAI-1 and CRP levels may be due to the fact that both are acute phase reactants. Interestingly, PAI-1 levels correlated significantly with both disease severity and blood eosinophilia, which is found frequently in the blood stream of patients with active BP [4]. Considering that the evaluation of disease severity in BP has only recently been standardized [29], and that in the patients of the present study there was no mucosal involvement, for evaluating the disease extent we adopted an easy system based on the percentage of involved body surface area, also used by other groups [30, 31].

Anti-BP180 autoantibody levels correlated with coagulation activation markers but not with PAI-1, probably because PAI-1 expression is more affected by inflammation than by autoantibody production. Although some studies indicated a correlation between disease severity and anti-BP180 autoantibody serum levels [32], other studies failed to find such a correlation [33], in accordance with our present data. A clear explanation for the discrepancy between autoantibody titres and BP severity is still lacking; however, some hypotheses have been proposed, including the phenomenon of ‘epitope spreading’, the switch between IgG subclasses and the production of non-pathogenic antibodies by long-lived plasma cells [33].

We provide evidence that the beneficial clinical effects induced by systemic corticosteroid treatment are associated with a significant decrease in PAI-1 levels. This finding supports the view that the normalization of fibrinolysis is probably related to the reduction in skin inflammation and blister formation observed in BP patients. We also found that the markers of coagulation activation decreased significantly during the clinical remission induced by immunosuppressive treatment, thus confirming our previous data [4].

The limitation of the present study is the relatively small number of patients, which is due to the low incidence of cases of BP (one in 100 000 per year in Italy [34]), but it may be counterbalanced by the clear-cut differences observed.

Overall, the reduction in fibrinolysis inhibition and coagulation observed after treatment may not only contribute to the healing of the cutaneous manifestations, but also reduce thrombotic risk as a whole.

Acknowledgements

The study was supported by ‘Fondo Interno per la Ricerca Scientifica e Tecnologica’, University of Milan.

Disclosure

None.

Ancillary