Stevens–Johnson Syndrome and Toxic Epidermal Necrolysis: An Update

Roni P. Dodiuk-Gad; Wen-Hung Chung; Laurence Valeyrie-Allanore; Neil H. Shear


Am J Clin Dermatol. 2015;16(6):475-493. 

In This Article

2 Pathogenesis

SJS/TEN mostly results from a cumulative effect of aligned risks related to the structure of a drug and to the patient's genetic predisposition (human leukocyte antigen [HLA] alleles, drug metabolism characteristics, and T cell clonotypes) (Fig. 1).

Figure 1.

''Swiss cheese'' risk model of Stevens–Johnson syndrome and toxic epidermal necrolysis (SJS/TEN). CYP cytochrome P450, HLA human leukocyte antigen

2.1 Genetic Associations of SJS/TEN

Over the past few decades, important progress has been made in understanding the pathogenic mechanisms of SJS/TEN—specifically, the important role of HLA alleles. The pathogenesis of inducing cytotoxic responses in SJS/TEN is generated by recognition of culprit drugs by specific HLA molecules. In 1987, the genetic susceptibility of HLA alleles was first proposed by Roujeau et al.,[4] who identified weak associations of HLA-B*12 and oxicam- and sulfonamide-related TEN in Europeans. More notable evidence of genetic susceptibility to SJS/TEN was reported in 2004 by Chung et al.,[5] who found that HLA-B*15:02 is strongly associated with carbamazepine (CBZ)-induced SJS/TEN. The early associations were made in an era of serological typing of HLA. Currently, in an era of sequence-based typing and deep sequencing, which allows resolution to four digits, identification of more specific and stronger associations are possible. Hence, many other HLA associations in SJS/TEN have been discovered,[4,6–21] as are detailed in Table 1.

HLA associations with specific drug-induced SJS/TEN can be restricted to certain ethnicities and phenotypes. The strength of HLA associations with specific drug-induced SJS/TEN has been found to be related to the prevalence of the susceptibility allele in the ethnic population. The association of HLA-B*15:02 and carbamazepine-induced SJS/TEN has been reported in Han Chinese, Thai, Indian,[22] and Malaysian populations[23] but not in Europeans, who carry the HLA-B*15:02 allele in low frequency (<1 %).[24] The association is also present in Han Chinese ancestry of Europeans. This ethnic difference can be tracked by historical evolution. The US Food & Drug Administration and Asian health administrations have recommended HLA-B*15:02 screening for new carbamazepine users of Asian ancestry since 2007.[25] In contrast, the HLA-B*58:01 allele with allopurinol-induced SJS/TEN is common to all populations, being found in Han Chinese, Thai, Japanese, Korean, and European populations.[26]

There are also phenotype-specific characteristics involved in carbamazepine hypersensitivity. HLA-B*15:02 and other B75 serotype HLA alleles are strongly associated with carbamazepine-induced SJS/TEN but are not associated with carbamazepine-induced drug reaction with eosinophilia and systemic symptoms (DRESS); however, HLAA* 31:01 is strongly associated with carbamazepine-induced DRESS but less so with carbamazepine-induced SJS/TEN.[27]

Association with a specific HLA risk allele appears to be necessary but not sufficient for development of SJS/TEN. Additional factors such as individual differences in drug metabolism or clearance may also play an important role in SJS/TEN development, recovery, or prognosis. Drug clearance is known to be important for preventing exceptional damage, which can occur as a result of retention of the drug in the body. Shear and Spielberg first recognized potential pharmacogenetic associations and risks of altered drug metabolism for SJS/TEN and DRESS with anticonvulsants and sulfonamides.[28,29]

A recent genome-wide association study by Chung et al.[30] reported that genetic variants of cytochrome P450 2C (CYP2C) are strongly associated with phenytoin-related SJS/TEN. They identified 16 significant single-nucleotide polymorphisms (SNP) in CYP2C genes at 10q23.33. Further studies showed that CYP2C9*3 variants, which reduce CYP2C9 enzymatic activity, were significantly related to phenytoin-induced SJS/TEN. CYP2C9*3 has been known to be related to the drug's metabolism and can attenuate the clearance of phenytoin.[31,32] Patients with phenytoininduced SJS/TEN who carried CYP2C9*3 showed delayed clearance of plasma phenytoin, resulting in increasing phenytoin toxicity in the body. Another study showing that genetic variability in a metabolizing enzyme can contribute to SJS/TEN examined nevirapine-induced SJS/TEN. CYP2B6 G516T and T983C SNPs were found to be associated with SJS/TEN susceptibility.[33] The issue of drug metabolism or clearance is another key factor in consideration of the risk of developing SJS/TEN.

2.2 Immunological Mechanisms of SJS/TEN

SJS/TEN is mostly recognized as an immune disorder elicited by drugs. SJS/TEN is a delayed-type drug hypersensitivity reaction, with a typical latency of 4–28 days and with cases rarely occurring as long as 8 weeks following initiation of the implicated drug.[34]

The drugs are of low molecular weight and often serve as foreign antigens that are recognized by T cell receptors (TCRs) to activate adaptive immune responses. Several proposed concepts have been found to explain how small-molecule compounds are recognized by TCRs. In some cases, drugs interact directly with the TCRs involved in presenting to HLA molecules of antigen-presenting cells (APCs). This model is known as the p-i [pharmacological interaction of drugs with immune receptors] concept.[35] An example is that carbamazepine cannot bind covalently to peptides or proteins but is able to bind with low affinity to TCRs and provoke T cell activation.[36,37] Drugs can also interact with TCRs by a drug–peptide complex presented to HLA molecules of APCs in what is known as the hapten concept. The b-lactams that form covalent binding to lysine residues are an example.[38] Upon noncovalent binding of specific drugs presented to HLA molecules and TCRs, the HLA–drug–TCRs may initiate a series of immune reactions, which result in activation of CD8+ cytotoxic T cell–mediated and natural killer (NK) cell–mediated cytotoxicity. Recently, the shared and restricted TCR usage subtype in carbamazepine-induced SJS/TEN was identified,[39] and it was demonstrated that the endogenous peptide-loaded HLA-B*15:02 molecule presented carbamazepine to cytotoxic T cells without the involvement of intracellular drug metabolism or antigen processing (Fig. 2).[37]

Figure 2.

Direct interaction between the HLA-B*1502 molecule and carbamazepine activates the T cell with a restricted T cell receptor (TCR). HLA human leukocyte antigen

Cytokines are involved in the pathogenesis of SJS/TEN. Several studies have shown that tumor necrosis factor (TNF)-α is strongly expressed in SJS/TEN lesions and correlates proportionally with disease development.[40–42] TNF-α can induce cell apoptosis, activation, and differentiation, and an inflammatory response.[43,44] In addition, interferon (IFN)-γ is a common cytokine involved in delayed-type drug hypersensitivity, including SJS/TEN. IFN-γ is often expressed in the superficial dermis and epidermis of SJS/TEN lesions.[41,42] IFN-γ is known to induce antigen presentation and thus stimulate cell-mediated immunity by upregulation of HLA molecules.[45–47] TNF-α, IFN-γ, several cytokines, and chemokine receptors that are responsible for proliferation, trafficking, and activation of T cells have been found in the skin lesions, blister fluids, blister cells, peripheral blood mononuclear cells, or plasma of SJS/TEN patients. These cytokines/chemokines include interleukin (IL)-2, IL-5, IL-6, IL-10, IL-12, IL-13, IL-15, IL-18, chemokine (C–C motif) receptor (CCR) 3, chemokine (C-X-C motif) receptor (CXCR) 3, CXCR4, and CCR10.[40–42,48–50]

The major theory to explain the severe epidermal detachment of SJS/TEN is CD8+ cytotoxic T cell–mediated and NK cell–mediated cytotoxicity. It has been well established that epidermal detachment in SJS/TEN is due to keratinocyte apoptosis. Recently, studies have shown that keratinocyte apoptosis in lesional skin and blister fluid in SJS/TEN patients is associated with greatly increased numbers of CD8+ cytotoxic T cells and NK cells.[51,52] While CD8+ cytotoxic T cells and NK cells are activated, they subsequently carry out the cellular-mediated immune reactions directed at keratinocytes in an HLA class I–restricted manner. Upon activation of these responses, various cytotoxic signaling molecules, including granulysin, perforin/granzyme B, and Fas/Fas ligand, are relayed to the skin lesions to induce keratinocyte apoptosis.[52–54] More importantly, Chung et al. found that granulysin—a cytotoxic protein produced by cytotoxic T cells or NK cells—acts as a key mediator responsible for disseminated keratinocyte death.[52,55] Furthermore, granulysin is not only a cytotoxic protein but also a chemoattractant and proinflammatory activator, which can promote monocyte expression of chemokine (C–C motif) ligand (CCL) 20[56] and is capable of promoting antigen-presenting (dendritic) cell and leukocyte recruitment (specifically, a granulysin 15 kD subunit, which is largely produced by CD8+ T cells and NK T cells).[57] Granulysin-positive cells in fixed drug eruptions have been found to be similar to those observed in SJS/TEN.[58]

2.3 Environmental Factors

SJS/TEN can be secondary to infection with Mycoplasma pneumoniae or herpes simplex virus, but the full pathogenesis remains unclear.[59–61] Human enterovirus has not previously been recognized to be associated with SJS.[59,60] By comparison, erythema multiforme major (EMM) is mainly caused by viruses, usually involving the palms and soles, and the patient experiences rapid healing without sequelae.

A recent study found that a new variant of coxsackie virus (CV) A6, which belongs to the human Enterovirus genus and causes severe mucocutaneous blistering reactions, mainly mediated by cytotoxic T cells and NK cells expressing granulysin, mimics the histopathological features of SJS or EMM in children.[62] In fact, there are still about 20 % of SJS/TEN cases without an identified causality.[63,64] Potential risk factors for this unusual presentation of virus infection as a cause of SJS/TEN have not been elucidated.