Introduction

Bathing in geothermal seawater from the Blue Lagoon (BL) in Iceland has been reported to have a beneficial effect on psoriasis [1, 2]. Additional treatment with narrow-band ultraviolet (NB-UVB) phototherapy further increases the efficacy of the treatment [3-5]. The BL contains geothermal seawater originating from underground reservoirs filled with a mixture of fresh water and seawater. Sampling from the lagoon shows that no pathogenic bacteria thrive in this ecosystem [6]. The fluid in the lagoon has a high level of silica but is moderate in temperature (37 °C) and salinity (2.7%) [7]. Recent data indicate that both the silica mud and two microalgae species growing in the BL promote the integrity of the skin barrier and delay extrinsic skin ageing, thus indicating a biological activity in the lagoon [8].

Psoriasis is mediated by T cells that trigger keratinocytes to hyperproliferate and perpetuate the disease [9]. T helper (h)17 and Th1 cells and the cytokines produced by these cells are found in increased levels within psoriasis plaques [10] as well as in the circulation [11] and are thought to have an important role in psoriatic inflammation. The relationship between Th1 and Th17 cells is still unclear. The tissue-specific localization of T cells is thought to be guided by the skin-homing molecules such as cutaneous lymphocyte-associated antigen (CLA), various chemoattractants and their receptors, including chemokine receptors 4 (CCR4) and 10 (CCR10) [12]. In addition, adhesion molecules are thought to mediate T cell migration and retention in cutaneous tissue, such as the αE (CD103) β7 integrin that is overexpressed in psoriasis skin [13].

The main objective of this study was to evaluate the immunological therapeutic effect of two treatment protocols on psoriasis, focusing on the main inflammatory cytokines and effector T cell phenotypes known to be important for skin homing and tissue retention, thus potentially providing new insight into the immunopathogenesis of psoriasis. Our results confirm the role of Th1 and Th17 effector T cells in psoriasis. It also provides insight into the role of CD8⁺ T cell secreting IFN-γ (Tc1) and IL-17 (Tc17) and CLA⁺/CD103⁺ effector T cells in its immunopathology.

Materials and methods

Statistical analysis

Values are expressed as the mean ± 2 SD. To compare the treatment group with controls, we used the Mann–Whitney U-test. To evaluate the differences between before treatment, during and after treatment, the normality of each type of measurement was evaluated using a KS test based on the residuals from a simple linear model using patient and time as factors. In no case was normality close to being rejected (P > 0.4 in all cases). Hence, one-way repeated-measures anova was used. However, to evaluate the differences between the two treatment groups, two-way repeated-measures anova was used. Three patients who received combined treatment were not evaluated at week 8 because they had started another psoriasis treatment due to exacerbations: two of those patients at week 4 (Fig. 1A; BL3 and BL6) and one patient at week 7 (Fig. 1A; BL1). For these patients, PASI evaluation was made at the time point their study participation was terminated, and they were not included in the analysis at week 8. All measurements were taken using sigmastat 3.1 (Systat Software, San Jose, CA, USA). A P-value ≤ 0.05 was considered statistically significant.

Results

Clinical evaluation

In order to evaluate whether clinical improvement of psoriasis following bathing in geothermal seawater combined with NB-UVB and NB-UVB alone is preceded by changes in systemic inflammatory markers, the clinical efficacy of each treatment regimen was evaluated first. As shown in Fig. 1C, both treatment regimens demonstrated significant clinical improvements. Furthermore, the data suggested that patients receiving combined treatment demonstrated better clinical response, measured by the PASI score, than patients treated only with NB-UVB. This was seen both after one week (% improvement: combined treatment 37.3 ± 10.3 versus NB-UVB treatment 18.3 ± 8.9, P < 0.05) and after three weeks (67.3 ± 11.9 versus 22.0 ± 12.0, P < 0.0001). However, this was not the main aim of the study, and larger cohort and another control group would be needed to fully address this interesting observation.

Interestingly, bathing in the Blue Lagoon immediately following skin punch biopsy resulted in no infections and only minor skin irritation resolving in few days. In addition, the above clinical findings were confirmed by the histological Trozak's score where patients in both treatment groups showed a significant histological improvement at week 3 (Trozak's score: BL treatment = 10.3 ± 5.5 versus NB-UVB treatment = 8.0 ± 4.6; Fig. 2).

Immunological evaluation

Skin homing and positive correlation with PASI score

To understand the role of adhesion molecules, chemokines and their receptors in cutaneous lymphocyte homing in patients with psoriasis, we evaluated intercellular adhesion molecule 1 (ICAM-1), E-selectin (CD62E), CD11c, two chemokine receptors (CCR4 and CCR10) and αEβ7 integrin (CD103) on peripheral blood mononuclear cells before, during and after each treatment regimen. In the active stage of the disease (W0) and compared with healthy control, patients with psoriasis had higher percentage of circulating CLA+ T cells expressing CD103 (median 5.7 versus 1.5%; P < 0.05), CCR10 (median 5, 1 versus 1.7%; P < 0.05) and co-expressing CD103/CCR4 (median 11.4 versus 0.8%; P < 0.05) and CCR4/CCR10 (median 3.7 versus 1.2%; P < 0.05) (Fig. 3A). In addition, a positive correlation between PASI and circulating CD103+ T cells (r = 0.6036; P < 0.05) and CLA+ T cells expressing CCR10 (r = 0.7360; P < 0.01) was similarly observed. No therapeutic changes were found regarding the expression of ICAM-1, CD62E, CD11c and other activation markers, such as CD25 and HLA-DR (data not shown).

In addition, patients receiving combined treatment had a significant reduction in CLA+ T cells expressing CCR4 or CD103 (68–74% reduction at W3, P < 0.001), while patients treated with NB-UVB alone did not (Fig. 3A). Furthermore, this reduction in CLA+CCR4+ T cells was predominantly confined to those who also expressed the CD103 integrin. Thus, no CLA+ T cells that co-expressed CD103 and CCR4 were detected in the circulation after 3 weeks (W3) in patients receiving combined treatment (P < 0.05; Fig. 3A). Both treatment groups achieved a significant reduction in CLA+ T cells that expressed CCR10 (71% reduction versus 44% reduction at W3; P < 0.001 versus P < 0.05; Fig. 3A). A marked reduction was also observed of circulating CLA+ T cells that co-expressed CCR4 and CCR10 in the combined treatment group (3.5% before treatment and 0.7% at W3; 80% reduction; P < 0.01; Fig. 3A).

Thus, the increased proportion of skin-homing T cells expressing CD103 and the chemokine receptors CCR4 and CCR10 was significantly reduced following clinical and histological improvements of psoriasis.

Effector T cell phenotype and its clinical correlation in psoriasis

To investigate the expression profile of circulating Th1/Tc1 and Th17/Tc17 cells in patients with psoriasis and its clinical correlation, their phenotypes were investigated amongst both CD4+/CD45RO+ and CD8+/CD45RO+ T cells. As expected in the active stage of the disease, patients with psoriasis had higher percentage of circulating CD4+ T cells expressing IFN-γ, TNF-α, IL-22 and IL-17 as compared with healthy controls (median 5.93 versus 2.06%, 9.08 versus 0.73%, 3.19 versus 0.33% and 4.78 versus 0.42%, respectively, P < 0.05 for all four subsets; Fig. 4A). Furthermore, this was also observed for the CD8+ phenotype expressing IFN-γ, IL-22 and IL-17 (median 6.93 versus 2.37%, 2.39 versus 0.81% and 2.22 versus 0.89%, respectively, P < 0.05 for all three subsets; Fig. 5A).

When evaluating the clinical efficacy with its corresponding immunological profile, patients receiving combined treatment showed a marked reduction (81%) in circulating Th17 (IL-23R+CD4+ T cells) after only one week of treatment (Fig. 4A). This was also reflected by a 53% reduction in the amount of IL-23R expressed (MFI) by these cells (P < 0.05, data not shown) and their IL-17/IL-22 cytokine secretion. In contrast, such immunological Th17 inflammatory response improvement was only detected after 8 weeks of NB-UVB treatment (4a). Furthermore, both of the treatment protocols resulted in a significant reduction in Tc17 T cells (producing IL-17 and IL-22; Fig. 5A). Finally, a similar reduction was also noted for the Th1 and Tc1 phenotype (IFN-γ and TNF-α production, Figs. 4A and 5A, P < 0.05).

Discussion

The role of skin-homing, Th1 and Th17 immune response in the immunopathology of psoriasis is demonstrated in this study. In addition, the importance of Tc1 and Tc17 immune response is also suggested. Finally, NB-UVB therapy induced excellent clinical improvement preceded by a reduction in these above systemic inflammatory markers, strongly suggesting that immune modulation mediated the observed clinical effect. Furthermore, an improvement by histological assessment is clearly demonstrated substantially validating the observed clinical improvements by using ‘Trozak's score’ as a measure of treatment efficacy.

There is evidence suggesting that bathing in the geothermal seawater without NB-UVB treatment has a beneficial clinical effect [1, 2]. It has also been noted that the scaling of psoriasis lesions disappears quickly, and the lesions get thinner with less erythema, indicating that bathing in this geothermal seawater has a direct anti-inflammatory effect on psoriatic lesions [2]. Another study demonstrated the beneficial effects of bathing in geothermal seawater where NB-UVB treatment after bathing gave an additional clinical effect compared with NB-UVB treatment alone [5], thus supporting our observation that bathing in the geothermal seawater might provide some additional clinical effect that was further reflected by the reduction in potential pathogenic T cells in the peripheral blood.

Psychological stress has been reported to influence psoriasis severity [17]. Inpatient treatment at the BL clinic in a relaxed environment might reduce stress and thereby indirectly improve the psoriasis lesions in addition to the UVB-induced effects. Immunological studies show that psychological stress increases the numbers of various immunological cells in the peripheral blood of patients with psoriasis, including HLA-DR⁺ T cells, and decreases the numbers of CD25⁺ T cells [18]. However, in our study, the numbers of T cells expressing HLA-DR⁺ and CD25⁺ did not change significantly in the peripheral blood with both treatments, indicating that stress did not influence the outcome of our study.

The therapeutic properties of combined treatment with salt water baths and natural UV radiation (climatotherapy) and bathing in thermal water (spa therapy) have been known since ancient times [21, 22]. Today, it is being practised in many countries in the form of combination treatment of salt or thermal water baths and artificial UV radiation (balneotherapy) [21, 22]. Some studies indicate that the main therapeutic effect of climatotherapy at the Dead Sea can be attributed primarily to the sunshine and secondarily to the hypersaline seawater [21, 23]. Other studies show that balneotherapy with Dead Sea salt solution soaks in combination with NB-UVB therapy is superior to NB-UVB therapy alone [24, 25], which could be attributed to increased photosensitivity of the skin to UV radiation [26, 27]. We do not think that explains the results in our study for two reasons. As mentioned above, there are studies showing that bathing in the geothermal seawater without NB-UVB treatment has a beneficial clinical effect [1, 2]. In addition, the cumulative dose of NB-UVB therapy in this current study was only 10 treatment sessions for patients bathing in geothermal seawater combined with NB-UVB therapy compared with 24 sessions for patients treated with NB-UVB therapy alone.

However, the agents responsible for these beneficial effects of bathing in saline or thermal water have not been fully elucidated but most likely involve chemical [26, 28, 29], thermal [30], mechanical [2] and immunomodulatory effects [28, 31]. Furthermore, studies have shown that bathing in salt solutions has been associated with increased photosensitivity of the skin to UV radiation [26, 27]. Even though balneotherapy and spa therapy are widely used, the immune modulatory mechanisms are only partly understood. Few studies have shown immunomodulatory effects on epidermal Langerhans cells, inhibition of Th1 differentiation and cytokine production from keratinocytes [28, 31]. One recent study from Korea [32] showed that thermal spring water suppressed the expression of pro-inflammatory cytokines in human keratinocytes ‘in vitro’ as well as the differentiation of mouse CD4⁺ T cells into Th1, Th2 and Th17 cells.

CCR4 has been found to be abundantly expressed on circulating T cells with a skin-homing CLA⁺ phenotype [33] in normal subjects as well as in patients with psoriasis [34], which is consistent with our results. In contrast, CCR10 and CD103 are weakly expressed in the peripheral blood of normal subjects and nearly undetected in normal skin [35, 36]. In addition, CCR10 is expressed by a minority (approximately 30%) of circulating CLA⁺ T cells [37]. However, both CCR10 and CD103 have been found in the inflamed psoriatic lesions [35, 36]. Their involvement in the immunopathogenesis of psoriasis is further suggested by our findings demonstrating the increased proportion of circulating skin-homing CLA⁺ T cells co-expressing the tissue retention integrin CD103 and/or the chemokine receptors CCR4 and CCR10. More importantly, they had a positive correlation with the clinical improvements observed in the study, thus implicating the role of directing CCR4⁺/CCR10⁺ and CD103⁺ subset of skin-homing T cells (CLA⁺) into psoriasis plaques during the active stage of the disease. CLA+, CD103+ T cells, various adhesion molecules as well as activation markers did not change significantly during or after both treatment protocols. These findings demonstrate that the changes observed in our study are not only explained by the significant overall decrease in the inflammatory state of the patients following the treatment protocols. Today, it is known that CCR6 is a common chemokine receptor on Th17 T cells [38], but it is not included in our study. It is unfortunate, but at the time that our study was conducted, the role of CCR6 as a Th17 marker was being debated and unclear.

The immunopathogenesis of psoriasis has been connected to both Th1 and Th17 effector cells, and our observation that IL-17, IL-22 and IFNγ levels in the blood of patients with psoriasis returned to baseline with effective therapy supports this notion [9-11, 39]. Furthermore, the increased proportion of IL-17-/IL-22-producing CD8⁺ T cells in the peripheral blood compared to healthy controls suggests their involvement in the immunopathogenesis of psoriasis, which has also been implicated by others [40]. In addition, the involvement of Tc17 cells in the immunopathogenesis was also evident by the positive correlation with individual clinical improvement measures. Similar to our findings, the therapeutic effectiveness of NB-UVB therapy has been associated with the corresponding Th1/Th17 pathway in psoriasis. In addition, in that study the role of innate immunity in psoriasis was suggested [41]. This has particularly been evaluated by the role of various Toll-like receptors in psoriasis. Thus, the expression of TLR2 has been found to be overexpressed in keratinocytes in psoriatic lesions [42], a finding also observed in our study with an increased expression of TLR2 on circulating monocytes (CD14⁺) and dendritic cells (CD11c⁺) in the peripheral blood of patients with psoriasis (data not shown). This study reflects the complexity behind the immunopathogenesis of psoriasis. It also reflects the following major confounding immunological elements. First, it confirms the importance of IFN-γ-, TNF-α-, IL-17- and IL-22-driven inflammatory response. Secondly, it suggests that these inflammatory cytokines are originating from both CD4⁺ and CD8⁺ T cells. Finally, this suggests that the inflammatory response is most likely predominantly driven by skin-homing tissue retaining T cells expressing the chemokine receptors CCR4 and CCR10.

Funding

This work was supported by the Landspitali University Hospital Research Fund, the Icelandic Technology Development Fund and the Blue Lagoon Ltd.

Conflict of interest

This study was conducted in collaboration with Blue Lagoon Ltd. and Landspitali University Hospital of Iceland.