Introduction

Stevens-Johnson syndrome (SJS) and its severe variant, toxic epidermal necrolysis (TEN) with spots, are acute inflammatory vesiculobullous reactions of the skin and mucous membranes such as the ocular surface, oral cavity and genitals. They are rare but often associated with inciting drugs and/or infectious agents1,2,3.

The association between human leukocyte antigen (HLA) genotypes and drug-induced severe cutaneous adverse reactions (SCARs) including SJS/TEN has been reported. There was a strong association between HLA-B*58:01 and SCARs, including SJS/TEN and the drug-induced hypersensitivity syndrome (DIHS), induced by the uric acid lowering drug allopurinol. This association was observed in Han Chinese-4, Caucasian-5 and Japanese patients6, suggesting that different ethnic groups share the same risk factor(s) for allopurinol-induced SCARs. HLA-B*15:02 exhibited a very strong association with carbamazepine-induced SJS/TEN in Taiwanese Han Chinese patients7 and HLA-A*31:01 was strongly associated with carbamazepine-induced SCARs including SJS/TEN in Japanese-8 and European patients9. We recently reported that cold medicine-related SJS/TEN with severe mucosal involvement including severe ocular surface complications (SOC) is associated with HLA-A*02:06 and HLA-B*44:03 in Japanese patients10.

The ophthalmologists Mondino et al.11 and the dermatologists Roujeau et al.12,13 reported that HLA-B12 (HLA-Bw44) was significantly increased in Caucasian SJS patients many of whom developed SJS/TEN after taking non-steroidal anti-inflammatory drugs (NSAIDs). HLA-B12 is primarily coded by HLA-B*44:02 or HLA-B*44:03 (http://www.allelefrequencies.net/). The significant association between HLA-B12 and SJS/TEN in Caucasian patients may be attributable to their genetic background.

To determine whether HLA-B*44:03 is a common risk factor for CM-SJS/TEN with SOC in different ethnic groups we used samples from Indian, Brazilian and Korean patients with CM-SJS/TEN with SOC and investigated the association between CM-SJS/TEN with SOC and HLA-B*44:03 and/or HLA-A*02:06.

Methods

Our study was approved by the institutional review boards of the participating institutions. All experimental procedures were conducted in accordance with the principles of the Helsinki Declaration. The purpose of the research and the experimental protocols were explained to all participants and their prior written informed consent was obtained.

Patients and controls

Ophthalmologists diagnosed SJS/TEN 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 surface1,14,15. They defined patients with SOC as those who manifested a pseudomembrane and an epithelial defect on the ocular surface in the acute stage and as patients with ocular sequelae such as dry eye, trichiasis, symblepharon and conjunctival invasion into the cornea in the chronic stage.

As in our previous study, we focused on SJS/TEN with SOC suspected of having been induced by cold medicines such as multi-ingredient cold medications and NSAIDs. As we found earlier that the genetic predisposition might be different between SJS/TEN with and without severe mucosal involvement including SOC10 we focused on patients from different ethnic groups who presented with SJS/TEN with SOC.

Samples from Indian patients with CM-SJS/TEN were collected at the LV Prasad Eye Institute (n = 20; 12 males, 8 females; age range 7 to 63 years; median age 27.1 ± 13.4 (SD) years). Their age at onset ranged from 3 to 42 years (median age at onset, 19.2 ± 12.2 (SD) years; in 8 patients the age at onset was unknown). The drugs administered to these patients and the HLA type (A and B) of patients with CM-SJS/TEN with SOC are shown in Supplemental Table 1. The specific drug(s) were not known in all patients. Healthy volunteers (n = 55; 29 males, 26 females; median age 36.0 ± 11.6 years) served as the Indian controls.

Samples from Brazilian patients with CM-SJS/TEN were collected at the Federal University of Sao Paulo (n = 39, 15 males, 24 females; age range 13 to 69 years; median age, 37.1 ± 15.9 years; age range at onset, 3 to 69 years; median age at onset, 24.0 ± 17.2 years). The drugs administered, the ethnicity and the HLA type (A and B) of these CM-SJS/TEN patients with SOC are shown in Supplemental Table 2. Healthy volunteers (n = 134; 55 males, 79 females; median age 41.2 ± 12.8 years) were the Brazilian controls (ethnicity: pardo, n = 66; white, n = 62; black, n = 4, Indian plus white, n = 2).

Samples from Korean patients with CM-SJS/TEN were collected at the Seoul National University College of Medicine, Chonnam National University, Yonsei University and the Catholic University of Korea. There were 31 patients (12 males, 19 females) ranging in age from 4 to 71 years (median age 33.7 ± 19.0 years). Their age at SJS/TEN onset ranged from 3 to 63 years (median age at onset, 23.0 ± 16.1 years). The drugs used and the HLA type (A and B) of these patients with SOC are presented in Supplemental Table 3. The specific drug(s) were not known in all patients. Healthy volunteers (n = 90; 35 males, 55 females; median age 31.7 ± 7.9 years) were the Korean controls.

Samples from Indian subjects were obtained by extracting DNA from whole peripheral blood with the phenol chloroform method. For Brazilian samples, DNA was extracted from whole peripheral blood using the PAX gene blood DNA kit (Qiagen, Hilden, Germany) or from saliva using Oragene DNA (Kyodou International, Kanagawa, Japan). To obtain the samples from Korean subjects, DNA was extracted from whole peripheral blood using the PAXgene Blood DNA kit (Qiagen).

HLA genotyping

For the analysis of HLA-A and HLA-B we performed polymerase chain reaction (PCR) assays followed by hybridization with sequence-specific oligonucleotide probes using commercial bead-based typing kits (Wakunaga, Hiroshima, Japan). Briefly, the target DNA was PCR-amplified with biotinylated primers specifically designed for amplified exons 2 and 3 of HLA-A and -B genes. Then the PCR amplicon was denatured and hybridized to complementary oligonucleotide probes (72 probes for HLA-A, 93 probes for HLA-B) immobilized on fluorescent-coded microsphere beads. At the same time, the biotinylated PCR product was labeled with phycoerythrin-conjugated streptavidin and immediately examined with Luminex 100 (Luminex, Austin, TX, USA). Genotype determination and data analysis were performed automatically using the WAKFLOW typing software (Wakunaga, Hiroshima, Japan) according to the manufacturer's instructions.

Statistical analysis

We compared the carrier frequency and gene frequency of individual HLA alleles in the patients and controls with the χ2-test (Pearson) (JMP version 11 software; SAS Institute Japan Ltd., Tokyo, Japan).

Results

Strong association between HLA-B*44:03 and CM-SJS/TEN with SOC in Indian patients

We genotyped HLA-A and HLA-B in samples from Indian subjects (20 CM-SJS/TEN with SOC patients and 55 controls). Although the number of Indian subjects was small, we found a strong and significant association between their CM-SJS/TEN with SOC and HLA-B*44:03 (carrier frequency: p = 1.07 × 10−5, odds ratio (OR) = 12.25, gene frequency: p = 9.37 × 10−8, OR = 10.88) but not HLA-A*0206 (Table 1).

Table 1 Results of association analyses in patients with CM-SJS/TEN with SOC
Full size table

Significant association between HLA-B*44:03 and CM-SJS/TEN with SOC in Brazilian patients

Next we genotyped HLA-A and HLA-B in samples from Brazilian subjects (39 CM-SJS/TEN with SOC patients and 134 controls). Although the number of Brazilian subjects was small we found a significant association between Brazilian patients with CM-SJS/TEN with SOC and HLA-B*44:03 (carrier frequency: p = 0.0239, OR = 2.74, gene frequency: p = 0.0121, OR = 2.77) but not HLA-A*0206 which is absent in the Brazilian population (Table 1). Interestingly, in Caucasians in the Brazilian samples (Brazilian Caucasian CM-SJS/TEN with SOC patients: n = 15, Brazilian Caucasian controls: n = 62), the association with HLA-B*44:03 was stronger (carrier frequency: p = 0.0037, OR = 6.22, gene frequency: p = 0.0011, OR = 5.99).

Association between HLA-A*02:06 and Korean patients with CM-SJS/TEN with SOC

We also genotyped HLA-A and HLA-B in samples from Koreans (31 patients with CM-SJS/TEN with SOC and 90 controls). Although the number of Korean patients was small we found a significant association between patients with CM-SJS/TEN with SOC and HLA-A*0206 (carrier frequency: p = 0.0362, OR = 3.00, gene frequency: p = 0.0263, OR = 2.46) but not HLA-B*44:03 (Table 1).

Discussion

We previously reported that in the Japanese, CM-SJS/TEN with severe mucosal involvement including SOC was associated with HLA-A*02:06 and HLA-B*44:0310. In the present study we investigated whether the association with these alleles is shared by other ethnic groups. We found that HLA-B*44:03 was strongly associated with CM-SJS/TEN with SOC in the Indian population which is genetically close to European populations16 and significantly associated in the Brazilian population which is comprised of individuals with different ethnic backgrounds. There was no association between HLA-B*44:03 and CM-SJS/TEN with SOC in the Korean population. HLA-A*02:06 was weakly associated in the Korean population which is genetically close to the Japanese, but not in the Indian and Brazilian population.

HLA-B12 (HLA-Bw44) was significantly increased in Caucasian SJS patients many of whom developed SJS/TEN after taking NSAIDs11,12,13. Because HLA-B12 is primarily coded by HLA-B*44:02 or HLA-B*44:03 (http://www.allelefrequencies.net/), the significant association of HLA-B12 with SJS/TEN in Caucasian patients may be attributable to the association with the HLA-B*44:03 genotype.

We also found that in Brazilian Caucasian patients with CM-SJS/TEN with SOC, the significant association with HLA-B*44:03 was stronger than in the entire study population of Brazilians with CM-SJS/TEN with SOC. To determine whether HLA-B*44:03 is a common marker for CM-SJS/TEN with SOC in Caucasian, HLA analysis of European patients with CM-SJS/TEN with SOC is needed.

Although HLA-A*02:06 was strongly associated with the Japanese CM-SJS/TEN with SOC and the Korean and Japanese population is genetically close16, in Korean patients CM-SJS/TEN with SOC was not strongly associated with HLA-A*02:06. To determine whether HLA-A*02:06 is a common marker for CM-SJS/TEN with SOC in East Asian populations further investigations using a larger number of samples are needed.

We also performed a meta-analysis by adding our previously-reported samples10. We used Cochran-Mantel-Haenszel statistics and found that both HLA-A*02:06 and HLA-B*44:03 are significantly associated with CM-SJS/TEN with SOC (Supplemental Table 4).

SCARs including SJS/TEN and DIHS induced by allopurinol were commonly and strongly associated with HLA-B* 58:01 in patients of different ethnic backgrounds including Han Chinese-4, Caucasian-5 and Japanese patients6. This observation suggests that different ethnic groups share the same risk factor(s) for allopurinol-induced SCARs.

With respect to carbamazepine-induced SJS/TEN, different HLA alleles are associated. HLA-B*15:02 is associated in Taiwanese Han Chinese patients7 and HLA-A*31:01 in Japanese-8 and European patients9.

In CM-SJS/TEN with SOC, the associated alleles we identified are HLA-A*02:06 in Japanese and Korean patients and HLA-B*44:03 in Indian-, Brazilian- and Japanese patients. Studies are underway to determine whether other HLA alleles are associated with CM-SJS/TEN with SOC in other populations.