Immunological and statistical studies of anti-BP180 antibodies in paraneoplastic pemphigus.

TO THE EDITOR Paraneoplastic pemphigus (PNP) shows clinically intractable stomatitis and conjunctivitis with polymorphous-cutaneous lesions (Anhalt et al., 1990; Hashimoto, 2001). Histopathology shows intraepidermal-acantholytic bullae and keratinocyte apoptosis (Oursler et al., 1992). Most common features revealed by direct immunofluorescence (IF) are deposition of IgG to keratinocyte cell surfaces and C3 to basement membrane zone (BMZ) (Anhalt et al., 1990; Hashimoto, 2001). In addition, in indirect IF, we encounter occasional reactivity with BMZ of normal skin, and more frequently with epidermal side of 1 M NaCl-split skin. BP180 is a transmembranous collagenous protein, whose extracellular NC16a and C-terminal domains were identified as immune-dominant regions in bullous pemphigoid (BP) and mucous membrane pemphigoid, respectively (Giudice et al., 1992; Matsumura et al., 1996; Nie and Hashimoto, 1999; Zillikens et al., 1999; Hashimoto et al., 2012). Lamina lucida-type linear IgA bullous dermatosis reacts with LAD-1, truncated-extracellular domain of BP180 (Ishii et al., 2008). Previous mouse model studies revealed that anti-BP180 antibodies can induce blister formation (Zillikens et al., 1999), whereas the pathogenic role of BP230 is currently unclear. Systemic study for autoantibodies to BP180 in PNP has not been performed, although a few PNP cases showed reactivity with BP180 (Preisz et al., 2004). Although reactivity with BMZ in PNP may be contributed mainly by anti-BP230 antibodies, we suspected a more frequent presence of antibodies to BP180 in PNP sera. In this study, we investigated IgG anti-BP180 antibodies in 59 PNP patients by various methods. Materials and Methods are described in Supplementary Materials online. All results of IF, immunoblotting (IB), and ELISA studies are summarized in Supplementary Table S1 online. Clinical parameters examined are shown in Supplementary Table S2 online. The results of statistical analyses are summarized in Table 1 and Supplementary Table S2 online. BP-like tense blisters were described in six PNP patients (Figure 1a). Histopathology identified apparent subepidermal blister formation in 3 of 30 skin biopsies (Figure 1b). Direct IF for 29 patients detected IgG deposition to keratinocyte cell surfaces in 21 patients (Figure 1c) and C3 deposition to BMZ in 12 patients (Figure 1d). Indirect IF for all 59 PNP sera showed reactivity with BMZ of normal-human skin in 27 sera (Figure 1e), and reactivity with epidermal side of 1 M NaCl-split skin in 26 sera (Figure 1f). One serum reacted with both epidermal and dermal sides (Figure 1g). In IB of normal-human epidermal extract, all 59 PNP sera reacted with both envoplakin and periplakin (Figure 1h). In addition, six sera reacted with BP180-like protein, whereas eight sera reacted with BP230-like protein (Figure 1h). IB of concentrated culture supernatant of HaCaT cells (HaCaT) showed reactivity with LAD-1-like protein in eight sera (Figure 1i). Recombinant proteins (RPs) of BP180 NC16a and C-terminal domains were detected in 7 and 13 sera, respectively (Figure 1j,k). BP sera, but not pemphigus vulgaris, pemphigus foliaceus, and normal sera, reacted with these proteins (Supplementary Figure S1 online). In ELISA, 17 sera reacted with BP180 NC16a domain RP. Index values were 9–50 in 13 sera, 51–100 in 1 serum, and more than 101 in 3 sera. In contrast, only four sera reacted with combined RPs of BP230 N- and C-terminal domains at index values of 9–50. In newly developed ELISA for BP180 C-terminal domain RP, nine sera showed positive reactivity with OD450, with values of 0.2–0.5 in four sera, 0.6–1 in four sera and more than 1 in one serum. Relatively higher reactivity with BP180 in ELISA and/or IB was shown in four of the six PNP patients with BP-like blisters, suggesting that the BP-like blisters may be caused by anti-BP180 antibodies. Statistical analyses for the correlation between BP180 antibodies and clinical features (Supplementary Table S2a online) revealed that skin lesions on the extremities were seen significantly more frequently in PNP patients with positive reactivity with BP180 NC16a or C-terminal domain RP by ELISA or IB (P<0.05) (Supplementary Table S2b, c online). These results suggested that anti-BP180 autoantibodies may induce BP-like skin lesions on the trauma-prone extremities. Other clinical features, including age, gender, neoplasm, respiratory diseases, and mucocutaneous lesions, showed no significant correlation with anti-BP180 antibodies. Statistical analyses of sensitivity among indirect IF, IB, and ELISA revealed that indirect IF of both normal-human skin and 1 M NaCl-split skin showed higher sensitivity than IB of epidermal extracts for native BP180, IB of RP of either BP180 NC16a or C-terminal domain, and IB of HaCaT for LAD-1 (Table 1a, b). This higher sensitivity of IF compared with IB may be caused by anti-BP230 antibodies, which are detected in most PNP sera by immunoprecipitation. In contrast, sensitivity of indirect IF either of normal-human skin (P=0.578) or of 1 M NaCl-split skin (P=0.709) was not significantly different from ELISA of BP180 NC16a or C-terminal domain RPs (Table 1a, b). Most sera that are positive in indirect IF of normal-human skin or 1 M NaCl-split skin were also positive in BP180 NC16a domain RP ELISA (P=0.578), whereas normal-human skin they were not positive in IB of BP180 NC16a or C-terminal domain RPs (P=0.057). In addition, sensitivity of ELISA of the NC16a domain only (P=0.022), but not of NC16a and C-terminal domains (P=0.176) and of the C-terminal domain only (P=0.344), was significantly higher than that of the corresponding IB. These results showed that IF and ELISA, particularly IF, were generally more sensitive than IB. Curiously, significant differences were found between the results of IB and ELISA using the same RP such as NC16a domain (Supplementary Table S1 online). We also noticed a similar difference in tests for BP sera (unpublished data). Different procedures for IB and ELISA may cause the discrepant results. Finally, our studies of IB of epidermal extracts and ELISA of BP230 RPs detected anti-BP230 antibodies in only 10 patients, indicating that PNP sera may react with PNP-specific conformational epitopes on BP230, which are detected only by immunoprecipitation. In this study, IF studies showed reactivity with BMZ in PNP patients. Combined results of IB of BP180 RPs and HaCaT, and of ELISAs of BP180 NC16a and C-terminal domain RPs showed that 42 and 40.67%, respectively, of PNP sera reacted with BP180. These results indicated that anti-BP180 autoantibodies are relatively frequently detected and have a role in BP-like blister formation in PNP. The authors state no conflict of interest. We gratefully acknowledge Ms Michiru Kubo and Ms Kyoko Hiromatsu for technical assistance, and Ms Mihoko Ikeda, Ms Tomoko Tashima, Ms Shoko Nakamura, and Ms Mami Nishida for secretarial work. We thank the patients for their participation. This study was supported by Grants-in-Aid for Scientific Research (no. 20390308, 20591331, 21659271, 23591634, 23791298, 23791299, 23791300, 23791301, 24659534, 24591672, 24591640, 24791185), and Supported Program for the Strategic Research Foundation at Private Universities 2011–2015 from the Ministry of Education, Culture, Sports, Science and Technology; and by “Research on Measures for Intractable Diseases” Project: matching fund subsidy (H23-028 to K. Iwatsuki, and H24-038 to T. Hashimoto) from the Ministry of Health, Labour and Welfare. The study was also supported by grants from the Kaibara Morikazu Medical Science Promotion Foundation, Ishibashi Foundation, Kanae Foundation for the Promotion of Medical Science, Takeda Science Foundation, Chuo Mitsui Trust and Banking Company, Limited, and Nakatomi Foundation. Author contributions Tamihiro Kawakami contributed to this study mainly for statistical analyses as a part time instructor at Kurume University. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper