Findings by an International Collaboration on SJS/TEN With Severe Ocular Complications

Stevens-Johnson Syndrome (SJS) is an acute inflammatory vesiculobullous reaction of the skin and mucosa, e.g., the ocular surface, oral cavity, and genitals. In patients with extensive skin detachment and a poor prognosis, the condition is called toxic epidermal necrolysis (TEN). Not all, but some patients with SJS/TEN manifest severe ocular lesions. Approximately 50% of SJS/TEN patients diagnosed by dermatologists and in burn units suffer from severe ocular complications (SOC) such as severe conjunctivitis with pseudomembrane and ocular surface epithelial defects in the acute stage. In the chronic stage, this results in sequelae such as severe dry eye and visual disturbance. Before 2005, our group of Japanese scientists started focusing on ophthalmic SJS/TEN with SOC. We found that cold medicines were the main causative drugs of SJS/TEN with SOC and that in Japanese patients, HLA-A*02:06 and HLA-B*44:03 were significantly associated with cold medicine-related SJS/TEN with SOC (CM-SJS/TEN with SOC). We expanded our studies and joined scientists from Korea, Brazil, India, Taiwan, Thailand, and the United Kingdom in an international collaboration to detect the genetic predisposition for SJS/TEN with SOC. This collaboration suggested that in Japanese patients, cold medicines, including NSAIDs, were the main causative drugs, and that HLA-A*02:06 was implicated in Japanese and Korean patients and HLA-B*44:03 in Japanese-, Indian-, and European ancestry Brazilian patients. Our joint findings reveal that there are ethnic differences in the HLA types associated with SJS/TEN with SOC.


INTRODUCTION
Stevens-Johnson syndrome (SJS) is an acute inflammatory vesiculobullous reaction of the mucosa of the ocular surface, oral cavity, and genitals, and of the skin. In patients with extensive skin detachment and a poor prognosis, the condition is called toxic epidermal necrolysis (TEN). In the acute stage of SJS/TEN, approximately 50% of patients present with severe ocular lesions such as severe conjunctivitis with pseudomembrane and ocular surface epithelial defects (1).
Ophthalmologists encounter patients not only in the acute-but also the chronic stage. Dermatologists, on the other hand, tend to see SJS/TEN patients only in the acute stage, although in some countries such as France and Germany dermatologists also followed up the patients long time. Our ophthalmologic diagnosis of SJS/TEN was based on a confirmed history of acute-onset high fever, serious mucocutaneous illness with skin eruptions, and involvement of at least two mucosal sites, including the ocular surface (2)(3)(4)(5)(6)(7)(8)(9). SJS/TEN patients with severe ocular complications (SOC) in the acute stage often develop sequelae such as vision loss and very severe dry eye that prevent their having a normal life (10).
We defined acute-stage SOC as a condition with severe conjunctivitis with pseudomembrane and epithelial defects on the ocular surface (cornea and/or conjunctiva) (11). Chronicstage SOC was defined as a condition with sequelae such as severe dry eye, trichiasis, symblepharon, and conjunctival invasion into the cornea (Figure 1A) (10). Ophthalmologists tend to diagnose both SJS and TEN with SOC broadly as "ophthalmic SJS" (Figure 1B) (4).
Dermatologists and others reported anticonvulsants such as carbamazepine and allopurinol (a uric acid-lowering drug) as being the main SJS/TEN-inciting drugs (12), while Japanese dermatologists reported that NSAIDs and multi-ingredient cold medications were main causative drugs for SJS/TEN (13). HLA analyses have shown that a genetic predisposition plays a role in the response to disease-eliciting drugs. Carbamazepine-induced SJS/TEN manifested a very strong association with the HLA-B * 15:02 allele in Taiwanese Han Chinese patients (14), and the HLA-A * 31:01 allele was strongly associated with carbamazepineinduced SJS/TEN in Japanese- (15) and European patients (16), the HLA-B * 57:01 allele was associated with carbamazepineinduced SJS/TEN in European patients (17). Allopurinolinduced SJS/TEN was strongly associated with HLA-B * 58:01 in Han Chinese- (18), European ancestry- (19), and Japanese patients (20). Interestingly, not all patients with carbamazepineinduced SJS/TEN develop SOC (21). Allopurinol has been reported to elicit SJS/TEN without SOC (22).
We reported that about 80% of SJS/TEN with SOC patients seen at the Kyoto Prefectural University of Medicine developed SJS/TEN within several days after taking cold medicines (we recognized the onset of SJS/TEN when the patients had eruptions.) (8).
These included multi-ingredient cold medications and nonsteroidal anti-inflammatory drugs (NSAIDs) (2,4,6,8,23). Our Brazilian collaborators found that 53% of their SJS/TEN with SOC patients had taken cold medicines (24) as had 69% of Thai patients with SJS/TEN with SOC (25), and 50% of Taiwanese patients (26). Our Korean collaborators suspected that NSAIDs and cold medicines were associated with SOC in their SJS/TEN patients (27). These observations suggest that such medicines are major causative drugs in SJS/TEN with SOC patients of different ethnicities.
This mini-review cites the results of our international collaborative efforts to identify the genetic predisposition for SJS/TEN with SOC.

HLA TYPES ASSOCIATED WITH SJS/TEN WITH SOC
The extreme rarity of cutaneous and ocular surface reactions to drugs led us to suspect individual susceptibility. Therefore, we entered an international collaboration to analyze the association between HLA genotypes and SJS/TEN with SOC.
We reported that the main causative drugs for SJS/TEN with SOC in Japanese patients were cold medicines, including multiingredient cold medications and NSAIDs taken to combat the common cold. As we also found that acetaminophen, present in various cold medicines, was the most frequently implicated causative drug (2,30), we focused on acetaminophen-related SJS/TEN with SOC. Analysis of the involved HLA types revealed that HLA-A * 02:06 was strongly associated with acetaminophenrelated SJS/TEN with SOC [80 patients, 113 controls; OR = 5.4, p = 8.0 × 10 −7 ] (30).

Korea
Together with our Korean collaborators we investigated the HLA types (HLA-A * 02:06 and HLA-B * 44:03) that were associated with CM-SJS/TEN with SOC in Japanese patients. We compared ours with samples from Korean patients and found that in Koreans, CM-SJS/TEN with SOC was also significantly associated with HLA-A * 02:06 (31 patients, 90 controls; OR = 3.0, p = 0.018), but not with HLA-B * 44:03 (3).
Our Korean collaborators suspected that NSAIDs and cold medicines were associated with SOC in Korean patients with SJS/TEN (27). They reported that allopurinol-induced SJS/TEN might not elicit serious acute or chronic complications of the ocular surface (22).

Brazil
Together with our Brazilian collaborators we investigated the HLA types (HLA-A * 02:06 and HLA-B * 44:03) that were associated with Japanese CM-SJS/TEN with SOC. Comparison of our and Brazilian samples revealed that in Brazilian CM-SJS/TEN with SOC, there was a significant association with HLA-B * 44:03 (39 patients, 134 controls; OR = 2.7, p = 0.024), but not with HLA-A * 02:06, a genotype not found in all Brazilian population (3). Interestingly, focused on European ancestry of Brazilians, the association with HLA-B * 44:03 was stronger (15 patients, 62 controls; OR = 6.2, p = 0.0037) than in all Brazilians (3).

India
Together with our Indian collaborators we investigated the HLA types (HLA-A * 02:06 and HLA-B * 44:03) associated with Japanese CM-SJS/TEN with SOC.
In samples from Indian patients with CM-SJS/TEN with SOC there was a significant association with HLA-B * 44:03 (20 patients, 55 controls; OR = 12.3, p = 1.1 × 10 −5 ), but not with HLA-A * 02:06 (3). Although the number of Indian patients was small, the association between Indian CM-SJS/TEN with SOC and HLA-B * 44:03 was strong and significant (3).
According to Kannabiran et al. (33), Indian ophthalmologists found it difficult to obtain a detailed history of disease onset from their SJS/TEN with SOC patients and in many patients they could not identify causative drugs. HLA analysis showed that HLA-A
In Thailand, as in the USA and UK, cold medicines, especially acetaminophen (paracetamol), are widely-used over-the-counter drugs. Elsewhere we reported that in Japan, acetaminophen is the most frequently included drug in various cold medicines (2,30). Therefore, we focused on Japanese acetaminophen-related SJS/TEN with SOC and analyzed the HLA types (30). Together with our Thai collaborators we also investigated Thai patients with acetaminophen-related SJS/TEN with SOC and analyzed the HLA types.

Taiwan
Our Taiwanese collaborators found that the main causative drugs in 26 Han Chinese with SJS/TEN with SOC were cold medicines; in 13 of 26 patients with SOC, cold medicines were the causative drugs, in none of 7 patients without SOC they identified cold medications as causative (26). Their findings echoed earlier studies that implicated cold medicines in the development of SOC in 80% of Japanese SJS/TEN patients (2,35), 53% of Brazilian patients (24), and 69% of Thai patients (25).
Together with our Taiwanese collaborators we performed HLA analysis of SJS/TEN with SOC in the Han Chinese and found that HLA-A * 02:07 (26 patients, 98 controls; OR =3.2, p = 0.049) was associated with their development of the disease. Our focus on CM-SJS/TEN with SOC revealed that HLA-A * 02:07 (13 patients, 98 controls; OR = 5.6, p = 0.016) was strongly associated with the development of SOC among Han Chinese CM-SJS/TEN patients (26). Single amino acid substitutions in major histocompatibility complex (MHC) class I molecules were found to play a role in distinct peptide repertoires. For example, three HLA-A2 subtypes, i.e., HLA-A * 02:04, HLA-A * 02:06, and HLA-A * 02:07, differed by only a single amino acid residue substitution; each harbored the HLA-A * 02:01 molecule at the floor of their binding grooves. Allele-specific peptide motifs for each HLA-A2 subtype differed substantially from the HLA-A * 02:01 motif in the dominant anchor residues (36). Although the carrier-and gene frequency of HLA-A * 02:06 in Japanese patients with CM-SJS/TEN with SOCs was significantly higher than in the control group, the frequency of HLA-A * 02:07 was similar in both groups (2). We found that the expression of HLA-A * 02:07 but not of HLA-A * 02:06 was associated with CM-SJS/TEN with SOC in the Han Chinese patients (26).
No HLA-B * 44:03 expression was detected in Han Chinese SJS/TEN patients or the controls (26), a finding compatible with earlier studies that showed that only 0.41-0.63% of the Taiwanese Han Chinese population harbored HLA-B * 44:03 (37,38). This observation suggests a genetic diversity in the pathogenesis of SJS among different ethnic groups although, because the number of samples was small, these studies must be expanded to include more samples.

United Kingdom
Our UK collaborators found that 9 of their 28 patients with SJS/TEN with SOC (32%) had taken cold medicines (39). Together with our UK collaborators we analyzed the association of HLA-A, HLA-B, and HLA-C alleles with SJS/TEN in 33 patients residing in the UK (28 patients with-and 5 without SOC) and in age-matched controls. There was a statistically significant and novel negative allele association with HLA-B * 07:02 (25 patients, 15 controls; OR = 0.16, p = 0.012) and with HLA-C * 07:02 (25 patients, 15 controls; OR = 0.09, p = 0.030) in a sub-group of European ancestry SJS/TEN patients (both with and without SOC) but not in their controls. This finding identified these alleles as being protective (39). Interestingly, a focus on European ancestry patients with SJS/TEN with SOC revealed only the association with HLA-B * 07:02 (23 patients, 15 controls; OR = 0.17, p = 0.027), but not with HLA-C * 07:02. When the focus was directed on European ancestry patients with CM-SJS/TEN with SOC, both associations with HLA-B * 07:02 and HLA-C * 07:02 disappeared (39). Thus, although HLA-B * 07:02 was associated with SJS/TEN with SOC in European ancestry, it may not be a biomarker for CM-SJS/TEN with SOC in that population. Because the number of samples was small, these studies must be expanded to include more samples.

DISCUSSION
A summary of our collaborative HLA analyses is shown in Table 1. It shows that HLA-B * 44:03 was significantly associated with CM-SJS/TEN with SOC in the Japanese (2), in Brazilians, especially European ancestry Brazilians (3,24,32), in Indian patients (3,33), and in Thais (25,34). HLA-A * 02:06 was significantly associated with CM-SJS/TEN with SOC in the Japanese (2) and in Koreans (3,31). Ma et al. (26) suggested that HLA-A * 02:07, differing by only a single amino acid residue substitution from HLA-A * 02:06, might be significantly associated with CM-SJS/TEN with SOC in Taiwanese patients.
We think that a common function of cold medicines such as acetaminophen, dipyrone, and NSAIDs is highly implicated in the onset of SJS/TEN with SOC (4,23,30,40).
The common function of cold medicines is the suppression of prostaglandin E 2 (PGE 2 ) production which suppress mucocutaneous inflammation. PGE 2 acts on EP3 (PGE 2 receptor 3) in the epidermis (41) and the mucosal epithelium (42,43) and negatively regulates mucocutaneous inflammation. We suspect that cold medicines that include acetaminophen and dipyrone could upregulate inflammatory responses by suppressing the production of PGE 2 which suppress mucocutaneous inflammation, that they augment abnormal immune responses, and that they elicit the induction of SJS/TEN with SOC (4,23,30,40).
Besides HLA types, we investigated other SJS/TEN with SOC susceptibility genes. Our genome-wide association study revealed IKZF1 to be a susceptibility gene for CM-SJS/TEN with SOC in Japanese-, Korean-, and Indian populations (6). It was also significantly associated with CM-SJS/TEN with SOC in Thai patients (44). Consequently, IKZF1 may be a universal marker for CM-SJS/TEN with SOC (6,44). Elsewhere (45) we documented that IKZF1 regulates mucocutaneous inflammation. We reported that IKZF1 transgenic mice developed spontaneous mucocutaneous inflammations such as ocular surface-and oral inflammation and dermatitis (45).
In the Japanese population we identified PTGER3 as a susceptibility gene for CM-SJS/TEN with SOC (8), and we reported that HLA-A * 02:06 and PTGER3 polymorphisms exerted additive effects in Japanese and Korean patients with CM-SJS/TEN with SOC (OR = 10.8 and 14.2, respectively) (46).
We also suggest that in addition to microbial infections and cold medicines, the combination of multiple gene polymorphisms and their interactions contributes strongly to the onset of CM-SJS/TEN with SOC. Abnormal Innate Immunity might strongly contribute the pathology of SJS/TEN with SOC (4,23).
Despite the genetic diversity in SJS/TEN with SOC among different ethnic groups, to prevent its onset and to reduce the incidence of blindness due to SJS/TEN, efforts must continue to identify the genetic predisposition for SJS/TEN with SOC.

AUTHOR CONTRIBUTIONS
MU wrote this mini review.

FUNDING
This work was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of the Japanese government, by the JSPS Core-to-Core Program, A. Advanced Research Networks, and partly supported by grants-in-aid for scientific research from the Japanese Ministry of Health, Labor, and Welfare.