Toxic epidermal necrolysis and Stevens-Johnson syndrome are induced by soluble Fas ligand

Abstract

The pathogeneses of toxic epidermal necrolysis (TEN) and Stevens-Johnson syndrome (SJS), both severe blistering diseases usually associated with drug intake, are not fully elucidated. Histologically, both TEN and SJS are characterized by extensive keratinocyte apoptosis. Previous studies have shown that keratinocyte apoptosis in TEN and SJS was induced by a suicidal interaction between Fas and Fas ligand (FasL), which are both expressed by keratinocytes. However, our preliminary examinations demonstrated that FasL is hardly detected on keratinocytes. We hypothesized that soluble FasL (sFasL) is secreted by peripheral blood mononuclear cells (PBMCs), and this interacts with the Fas expressed on keratinocytes in TEN and SJS. To justify this hypothesis, we investigated whether sFasL secreted by PBMCs could induce the keratinocyte apoptosis in TEN and SJS. Enzyme-linked immunosorbent assay analysis demonstrated that there was no significant sFasL increase in any samples of healthy controls (<40 pg/ml, n = 14) and patients with an ordinary erythema multiforme-type drug eruption (41.5 +/- 3.1 pg/ml, n = 14), whereas high concentrations are detected in all samples of TEN and SJS patients (TEN: 131.5 +/- 57.4 pg/ml, n = 8; SJS: 119.1 +/- 41.0 pg/ml, n = 14) (P < 0.0001). In vitro analysis using cultured keratinocytes revealed that the sera of TEN and SJS patients induced abundant keratinocyte apoptosis compared to erythema multiforme-type drug eruption sera. Furthermore, on stimulation with the causal drug, PBMCs obtained from TEN and SJS patients secreted high levels of sFasL. Taken together, these results indicate that sFasL secreted by PBMCs, not keratinocytes, plays a crucial role in the apoptosis and pathomechanism of TEN and SJS, and that the serum sFasL level may be a good indicator for the early diagnosis of TEN and SJS.

Figure 1.

High levels of sFasL were detected in serum samples from patients only with TEN and SJS, but not with EMDE or healthy controls. Serum sFasL was examined by sFasL ELISA described in the Materials and Methods. Serum sFasL concentrations increased in all samples of TEN (n = 8) and SJS (n = 14) patients, whereas there was no significant increase in sFasL in any samples of patients with EMDE (n = 14) or healthy controls (n = 14) (the limit of detection was 40 pg/ml). The bar indicates average values and an asterisk indicates statistical significance as compared with EMDE or controls (*, P < 0.0001).

Figure 2.

Serum sFasL in patients with TEN and SJS decreased significantly between 3 to 6 days after the start of the disease course. Sera from six patients with TEN and SJS were obtained at the beginning of the disease and after 3 to 6 days. Levels of sFasL significantly declined from 105.5 ± 25.4 pg/ml to 46.7 ± 10.4 pg/ml during the 3 to 6 days after the disease onset (n = 6; *, P < 0.01).

Figure 3.

Fas is expressed on keratinocytes in TEN skin lesions, however FasL expression is not detected. Skin sections obtained from TEN patients (a–d) or healthy controls (e, f, h, i) were subjected to H&E staining (a, e), immunohistochemical staining with anti-Fas (b, f) or Fas L mAb (c, h), or TUNEL analysis (d, i). Histological examinations showed numerous apoptotic keratinocytes (H&E) (a). Fas and FasL expression in skin were analyzed as described in the Materials and Methods. Fas expression was observed in keratinocytes of TEN skin lesions as well as normal controls (b, f), whereas no apparent FasL expression was observed in keratinocytes of both samples (c, h). FasL expression was detected in lesion of contact dermatitis (g). In the TUNEL assay, apoptotic cells were identified in situ by histochemical techniques with staining of double-stranded DNA breaks as described in the Materials and Methods. Skin sections from individuals with TEN showed numerous keratinocyte apoptotic figures (d), whereas no apoptoses were detected in normal skin (i).

Figure 4.

Keratinocyte apoptosis is induced by sera of TEN and SJS patients. Human keratinocytes were incubated at 37°C for 24 hours with medium containing 1% serum healthy control sera or 0.1% and 1% TEN patients’ sera. Apoptotic cells were identified by the TUNEL assay (a, b) or propidium iodide staining using FACS analysis (c, d). Sera of TEN patients induced a greater number of apoptotic keratinocytes (a, c), whereas keratinocytes cultured in medium containing normal serum showed significantly less apoptotic cell death (b, d). Apoptotic cells identified by the TUNEL assay were measured (e). *, P < 0.001.

Figure 5.

Keratinocyte apoptosis induced by sera of TEN and SJS patients was mediated through Fas-FasL interaction. Human keratinocytes were incubated at 37°C for 24 hours with medium containing 1% TEN patients’ sera plus an inhibitory anti-FasL mAb (1 and 10 μg/ml) or sFasL (100 ng/ml). Apoptotic cells were identified by the TUNEL assay. Culture with an inhibitory anti-FasL mAb (10 μg/ml) significantly reduced the number of apoptotic cells in a dose-dependent manner. Furthermore, sFasL mediated keratinocyte apoptosis. *, P < 0.001.

Figure 6.

PBMCs of a TEN patient produced sFasL and expressed FasL mRNA after stimulation with a causal drug. PBMCs freshly isolated from a TEN patient (1 × 10⁶/ml) were cultured with the causal drug, carbamazepine (10 and 100 μg/ml) for 24 hours. The amounts of sFasL in supernatants were measured by the sFasL ELISA kit. FasL mRNA of PBMCs was examined by reverse transcriptase-polymerase chain reaction as described in the Materials and Methods. Large amounts of sFasL were detected after the addition of carbamazepine (a). In contrast, no sFasL were found without carbamazepine (*, P < 0.001, compared with 10 or 100 μg/ml of carbamazepine). Similarly, FasL mRNA was detected only with stimulation by 10 and 100 μg/ml of carbamazepine (b).