Intradermal injection of human sweat enhances production of specific IgG against antigens solved in the sweat

To the editor Sweat plays a crucial role in maintaining skin homeostasis by regulating body temperature, moisturizing the skin surface, and protecting against infectious microorganisms through its antimicrobial peptides and proteases. However, sweat is also a major factor that exacerbates atopic dermatitis (AD) and triggers the development of symptoms of cholinergic urticaria (CholU). Both conditions involve type I hypersensitivity to sweat antigens [1], with the primary antigen identified as MGL_1304, a protein secreted by Malassezia globosa, a skin-resident fungus [2,3]. In these diseases, leakage of sweat from the sweat glands into the skin is observed [4]. Recent studies have revealed that insufficient production of Claudin-3, an essential component of tight junctions in the sweat duct/gland epithelium, may cause sweat to leak into the dermis in AD [5]. Human sweat contains various cytokines [6], such as IL-1α, IL-1β, and IL-31, in addition to Malassezia-derived antigens. Thus, it is plausible that sweat leakage could trigger diverse biological reactions beyond type I hypersensitivity. IL-1 enhances adaptive immunity by promoting T cell-antigen-presenting cell interactions and strengthening T cell priming [7]. Dai et al. [6] reported that IL-1α, IL-1β, and IL-31 in human sweat can activate epidermal keratinocytes [5]. However, to our knowledge, no studies have explored the biological effects of human sweat on dermal skin. Given the elevated levels of IgE against various antigens in AD, we investigated whether human sweat could contribute to percutaneous sensitization to external antigens through the epidermis or by direct delivery to the dermis by the leakage of sweat with antigen from the sweat glands and ducts. Sweat samples were collected from 7 healthy volunteers, sterilized using a 0.2 mm filter (Pall Corp., WA, USA), pooled, and used for the experiments. As a sweat antigen sensitization model, ovalbumin (OVA) was used as an immunogen and administered with or without crude sweat sample to study the effects of components in the sweat other than antigens on sensitization. OVA was dissolved in saline or sweat at a concentration of 1 μg/μl and served. The indicated amounts of the immunogen were administered to the backs of HR-1 hairless mice, which possess adequate administration area on the body without the influence of hair follicles, by either a patch for epidermal administration or a microneedle adjusted with an adaptor to reach the dermal layer for dermal administration (Fig. 1A). FITC-labeled OVA was administered to the back of mice to confirm the location of the administered antigen, by each administration method, and inspected in frozen sections of the skin stained with DAPI under a fluorescent microscope. The antigen was properly inserted into the epidermis and dermis by epidermal patch and dermal injection, respectively (Fig. 1B). Antigen samples were administered 6 times over a 2-week period, and blood samples were collected before and after administrations. OVA-specific IgG and IgE in the serum were measured using ELISA (Fig. 1C).Figure 1.: Protocol of OVA immunization and induction of OVA-specific serum IgG in Hos: HR-1 hairless mice (Hoshino Lab. animal, Inc., Ibaraki). (A) For epidermal sensitization, 200 μl of sample was placed on a sterile gauze CARELEAVES patch (NICHIBAN, Tokyo), and applied to 2 places on the back skin. For dermal sensitization, 2 μl of the sample was injected to 10 places on the back skin by the injection system using 34-G micro needle (TERUMO, Tokyo). (B) The location of samples delivered by each immunization method was confirmed with FITC-labeled OVA (Thermo Fisher, Waltham, MA, USA). The arrow indicates FITC-OVA delivered into the tissue. Green: FITC-OVA, Blue: DAPI. (C) Saline, sweat, OVA (1 μg/μl)/saline, or OVA (1 μg/μl)/sweat were administered to HR-1 hairless mouse 3 times a week over 2 weeks. (D) OVA-specific serum IgG concentration measure by ELISA on day 18 after the start of immunization. The protein levels are expressed as means ± SD (n = 3). Data were evaluated using the two-way ANOVA analysis (two-way ANOVA; interaction, P = 0.027, η² = 0.43; 95%CI, 0.04–1.00) based on Brown-Forsythe test (F = 1.8998, P = 0.1363) followed by Bonferroni’s multiple comparisons test. Statistical significance was set at ***P < 0.005, *P <0.05.In the dermal administration group, OVA-specific IgG production was higher when OVA was dissolved in sweat (582.8 ± 317.0 ng/ml) compared to when it was dissolved in saline (159.3 ± 137.9 ng/ml) (Fig. 1D). In the epidermal administration groups, there was no difference in OVA-specific IgG production between the OVA/saline and OVA/sweat groups (137.2 ± 118.9 ng/ml vs 129.3 ± 132.8 ng/ml), and no difference was observed when compared to dermal administration with OVA/saline (159.3 ± 137.9 ng/ml) (Fig. 1D). OVA-specific IgE levels did not reach the detection limit of ELISA (data not shown). As these data analyses were performed with N = 3, further validation of the data is needed through larger-scale experiments. We did not measure the amounts of OVA that entered and retained in the epidermis and dermis, respectively. However, the anti-OVA IgG levels induced by epidermal and dermal applications of OVA in saline without sweat were comparable. Thus, anti-OVA IgG was induced by either trans-epidermal or direct-dermal administrations; however, an adjuvant effect was observed only when sweat was administered directly to the dermis along with the antigen. Given the presence of cytokines as well as antigens in human sweat (eg, IL-1α and IL-1β), the delivery of human sweat into the dermis may activate antigen-presenting cells, such as dendritic cells and macrophages, and promote their interactions with T cells to induce antigen-specific antibody production [7]. The inability to detect OVA-specific IgE may be attributed to the non-Th2 skewed genetic background of the HR-1 mice and/or short immunization protocol used in this study. Reduced barrier function has been suggested as a risk factor for antigen sensitization in AD [8]. However, data on the skin barrier function in CholU are limited. On the other hand, sweat leakage into the dermis has been reported in both AD and CholU [4,5]. Thus, sensitization to sweat antigens in AD and CholU may result from the impairment of barrier function in the sweat gland and/or duct, rather than epidermal keratinocytes. A measure to enhance and/or recover barrier function of sweat gland and duct may be a target of new treatments to prevent the development of hypersensitivity and/or aggravation in AD and CholU. Further analysis is needed to understand how specific cytokines and proteins in sweat influence dermal immune responses, including IgE production, using many mice including AD models and an extended immunization period. Acknowledgments We thank the volunteers who donated their sweat, and Tomoko Kawaguchi for her excellent technical assistance. We especially thank Minoru Hattori for his efforts in providing detailed statistical analysis. This study was supported by JSPS KAKENHI Grant Number JP20K17319. Ethics Statement Clinical investigations of sweat samples were performed after obtaining written informed consent. The study protocol was approved by the Ethics Committee of the Hiroshima University Graduate School of Biomedical and Health Sciences (Committee chairman: Dr. Hitoshi Okamura) (institutional approval no. E-2018-1420). The animal experiments were approved by the Animal Experiment Committee of the same university (institutional approval no. A22-196). Conflicts of interest The authors have no conflicts of interest.