Causal Associations between Vitamin D Levels and Psoriasis, Atopic Dermatitis, and Vitiligo: A Bidirectional Two-Sample Mendelian Randomization Analysis

Background: Vitamin D level has been reported to be associated with psoriasis, atopic dermatitis, and vitiligo. However, its causal relationship with the risk of these three diseases remains unclear. Methods: We obtained genome-wide association statistics for three measures of circulating vitamin D levels (25(OH)D in 120,618 individuals, and 25(OH)D3 and epimeric form C3-epi-25(OH)D3 in 40,562 individuals) and for the diseases psoriasis (3871 cases and 333,288 controls), atopic dermatitis (21,399 cases and 95,464 controls), and vitiligo (4680 cases and 39,586 controls). We performed Mendelian randomization using inverse-variance weighted, weighted median, MR-Egger, and MR-pleiotropy residual sum and outlier methods. We carried out sensitivity analyses to evaluate the robustness of the results. Results: We showed that elevated vitamin D levels protected individuals from developing psoriasis (OR = 0.995, p = 8.84 × 10−4 for 25(OH)D; OR = 0.997, p = 1.81 × 10−3 for 25(OH)D3; and OR = 0.998, p = 0.044 for C3-epi-25(OH)D3). Genetically predicted risk of atopic dermatitis increased the levels of 25(OH)D (OR = 1.040, p = 7.14 × 10−4) and 25(OH)D3 (OR = 1.208, p = 0.048). A sensitivity analysis suggested the robustness of these causal associations. Conclusions: This study reported causal relationships between circulating vitamin D levels and the risk of psoriasis, atopic dermatitis, and vitiligo. These findings provide potential disease intervention and monitoring targets.


Introduction
Psoriasis, atopic dermatitis, and vitiligo are three chronic inflammatory skin diseases that are commonly characterized by dysregulation of immune function and infiltration of inflammatory cells in skin lesions [1][2][3]. Vitamin D is an important steroid pro-hormone with beneficial effects on regulating immune function and inhibiting inflammation [4]. Epidemiological studies have found associations of blood vitamin D levels with psoriasis, atopic dermatitis, and vitiligo [5]. However, the causal relationship between circulating vitamin D levels and the risk of these three skin diseases remains largely unclear.
Cholecalciferol (D 3 ) and ergocalciferol (D 2 ) are two most important vitamin D compounds in humans. They are synthesized in the epidermis of the skin or ingested from the diet and supplements. The hydroxylation of D 3 and D 2 in the liver leads to 25-hydroxyvitamin D 3 (25(OH)D 3 ) and 25-hydroxyvitamin D 2 (25(OH)D 2 ), respectively. Circulating 25-hydroxyvitamin D (25(OH)D) measures the total amount of 25(OH)D 3 and 25(OH)D 2 and is regularly used to reflect an individual's vitamin D status [6]. In addition, C3-epimer of 25(OH)D 3 (C3-epi-25(OH)D 3 ), a vitamin D metabolite from 25(OH)D 3 through 3β to 3α isomerization transformation, can also affect the amount of bioactive vitamin D conversion [7]. Recent studies have shown that 25(OH)D and its relevant compounds are associated with many diseases, including cancers, kidney, infectious, and autoimmune diseases [8,9]. These findings suggest that vitamin D might play a role in calcium homeostasis, immune regulation, anti-bacterial, and anti-inflammation. Early epidemiological studies have reported associations of blood 25(OH)D or 25(OH)D 3 levels with psoriasis, atopic dermatitis, and vitiligo [10][11][12], but concluded conflicting association results [13][14][15][16]. For example, Ng et al. found a significantly lower blood 25(OH)D level in atopic dermatitis patients than healthy controls and low serum vitamin D levels were associated with disease severity [10]. Vitamin D supplementation might improve disease manifestation [17][18][19] and decrease the risk of psoriasis [20]. However, Lucas et al. suggested no statistically significant difference of blood 25(OH)D 3 levels between atopic dermatitis patients and controls [13]. A cohort study of 4378 children indicated that low 25(OH)D levels at the age of two years old was not a risk factor for the development of atopic dermatitis at age of three years of age [14]. Most of the published studies have focused on the associations between the serum concentrations of 25(OH)D or 25(OH)D 3 and diseases, but the role of C3-epi-25(OH)D 3 in psoriasis, atopic dermatitis, and vitiligo has not been studied yet.
Traditional cross-sectional observations are susceptible to potential confounding factors and reverse causality bias, making causal inference challenging [21]. Mendelian randomization (MR) analysis has recently received increasing attentions. It infers causal effects of exposure on outcomes using genetic instrumental variables (IVs). Compared with cross-sectional observation studies, MR analysis can circumvent confounders and reverse causality [22,23]. Recent genome-wide association studies (GWASs) have identified tens of thousands of genetic variants robustly associated with human complex diseases and traits [24], which as a result provide a wealth of potential IVs and facilitate MR analysis. MR analyses have investigated the causal effects of circulating 25(OH)D levels on the risk of psoriasis, atopic dermatitis, and vitiligo, suggesting that blood 25(OH)D levels confer a significant causal effect on psoriasis but not on atopic dermatitis or vitiligo [25][26][27]. Alternatively, MR suggests that genetically predicted risk of atopic dermatitis increases blood 25(OH)D level [28]. However, the causal relationship between the levels of 25(OH)D 3 and C3-epi-25(OH)D 3 and the risk of psoriasis, atopic dermatitis, and vitiligo remains unclear. In addition, MR analysis is lacking to investigate whether the incidence of psoriasis and vitiligo changes blood vitamin D levels.
Here, we performed a bidirectional two-sample MR to comprehensively evaluate the causal relationships between blood levels of three vitamin D measures including 25(OH)D, 25(OH)D 3 , and C3-epi-25(OH)D 3 and the risk of psoriasis, atopic dermatitis, and vitiligo.

Study Design
To infer the causal relationship between circulating vitamin D levels and the risk of psoriasis, atopic dermatitis, and vitiligo we performed a bidirectional MR analysis. We summarized our study design in Supplementary Table S1. Briefly, we identified genetic IVs from the GWAS for blood vitamin D levels and the risk of psoriasis, atopic dermatitis, and vitiligo. Valid IVs in MR analysis need to fulfill three core assumptions: genetic IVs are robustly associated with the exposure, the association of IVs with the exposure is independent of confounders, and the IVs influence outcome only through the exposure [23]. To ensure IVs fulfill the first assumption, we restricted the study to genetic IVs having genome-wide significant associations with exposure and evaluated their associations by F-statistic [29]. The second and third assumptions of IVs are collectively known as independent from pleiotropy. To ensure IVs fulfill the second and third assumptions, we evaluated them by various sensitivity analyses [30].

Data Sources
We used genome-wide association statistics for circulating vitamin D levels and the risk of psoriasis, atopic dermatitis, and vitiligo from GWAS with the largest sample sizes. All the summary-level statistics are publicly available. Specifically, genome-wide asso-ciation summary statistics for blood 25(OH)D were derived from 120,618 healthy adults of European ancestry while the association statistics for 25(OH)D 3 and C3-epi-25(OH)D 3 came from a GWAS meta-analysis of 40,562 Europeans [31]. We used genome-wide association results for psoriasis reported by the Neale laboratory in the MR base (access ID: ukb-a-100) [32], which contained 3871 psoriasis cases and 333,288 controls in the UK Biobank [33]. We obtained genome-wide association statistics for atopic dermatitis from the Early Genetics and Life course Epidemiology (EAGLE) Consortium GWAS meta-analysis, which was carried out in 21,399 cases and 95,464 controls from 22 cohorts of European ancestry and 4 cohorts of non-European ancestry [34]. We accessed the summary statistics for a GWAS meta-analysis of vitiligo from GWAS Catalog (GCST004785), which were carried out in 4680 cases and 39,586 controls of European ancestry [35]. The details of these GWAS data sources are listed in Supplementary Table S1. The measure methods for vitamin D levels and the disease diagnosis criteria have been described previously in the original studies [31,[33][34][35]].

Selection of Genetic Instruments
To identify IVs for MR analysis, we performed linkage disequilibrium (LD) clumping to identify independent single nucleotide polymorphisms (SNPs; r 2 < 0.001 within 1 Mb window) in the GWAS summary statistics for each exposure, using the 1000 Genomes Project Phase 3 (EUR) as the reference panel [36]. We required instruments robustly associated with the exposure at a level of genome-wide significance (p < 5 × 10 −8 ). We calculated an F-statistic for each IV to assess its association strength with exposure [29], and included IVs with F-statistic > 10 [37]. We excluded palindromic SNPs with intermediate allele frequencies from IVs and harmonized the IVs to ensure their association effects are relative to the same alleles in both exposure and outcome. We used proxy SNPs (LD at r 2 > 0.8 with IV) when the IV was missing in the GWAS summary statistics for outcome.

MR Analysis
To test the causal effect of exposure on outcome, we performed MR analysis using four approaches, namely inverse-variance weighted (IVW) [38], weighted median [39], MR-Egger [40], and MR pleiotropy residual sum and outlier (MR-PRESSO) [41]. These four approaches make different assumptions and use different strategies to handle IVs with horizontal pleiotropy effects. The IVW method relies on the assumption that no pleiotropy existed and assumes all SNPs are valid genetic instruments [38]. The weighted median method assumes at least 50% of the IVs are valid [39]. The MR-Egger method provides a causal estimate even all IVs are invalid [40]. The MR-PRESSO method detects possible IV outliers through a global test and provides an unbiased causal estimation by removing the identified outliers [41].
To evaluate the robustness of the causal estimates in MR analyses, we used Cochran's Q tests [42] to detect heterogeneity of causal effects across IVs, and the MR-Egger intercept [43] and MR-PRESSO global test to detect horizontal pleiotropy [41]. We additionally performed a leave-one-out analysis [44] to evaluate whether the MR analysis results were driven by any single IV.
All analyses were performed in R software version 4.1.1 using TwoSampleMR (version 0.5.6; https://github.com/MRCIEU/TwoSampleMR, accessed on 18 July 2022) and MR-PRESSO (version 1.0; https://github.com/rondolab/MR-PRESSO, accessed on 18 July 2022) packages. We considered significant causal estimations at p < 0.05 in any of the four MR approaches. We presented the IVW results in this manuscript unless stated otherwise.

Causal Effects of Vitamin D Levels on the Risk of Psoriasis, Atopic Dermatitis, and Vitiligo
To evaluate the causal effects of circulating vitamin D levels on the risk of psoriasis, atopic dermatitis, and vitiligo, we performed MR analyses. We identified 14, 9, and 3 IVs for 25(OH)D, 25(OH)D 3 , and C3-epi-25(OH)D 3 levels, respectively. The characteristics of these IVs are shown in Supplementary Table S2.
The IVW analysis identified a significant causal effect of blood 25(OH)D level on developing psoriasis. Each one standard deviation (SD) decrease in circulating 25(OH)D level was associated with approximately 5% increased risk of psoriasis (odds ratio ( Table S3). Leave-one-out analyses suggested that none of the single IV could drive these causal effects (Supplementary Figure S2).

Causal Effects of the Risk of Psoriasis, Atopic Dermatitis, and Vitiligo on Vitamin D Levels
To check if there is a bidirectional relationship between diseases and vitamin D levels, we further performed reverse two-sample MR analysis. We identified 29, 13, and 48 IVs for the risk of psoriasis, atopic dermatitis, and vitiligo, respectively. The characteristics of these IVs are shown in Supplementary Table S2.
The IVW analysis suggested that the genetically predicted risk of atopic dermatitis was significantly associated with increasing levels of blood 25(OH)D (OR: 1.040, 95%; CI: 1.017-1.064, p = 7.14 × 10 −4 ; Figures 3 and 4A). The causal estimate in MR-PRESSO showed a stable association and the weighted median and MR-Egger methods also provided similar causal association (Figure 3). In addition, MR-Egger methods detected a significant We did not observe any significant evidence for causal effects of the three types of circulating vitamin D levels on atopic dermatitis or vitiligo (p > 0.05) (Figure 1 and Supplementary Table S3).

Causal Effects of the Risk of Psoriasis, Atopic Dermatitis, and Vitiligo on Vitamin D Levels
To check if there is a bidirectional relationship between diseases and vitamin D levels, we further performed reverse two-sample MR analysis. We identified 29, 13, and 48 IVs for the risk of psoriasis, atopic dermatitis, and vitiligo, respectively. The characteristics of these IVs are shown in Supplementary Table S2.
The IVW analysis suggested that the genetically predicted risk of atopic dermatitis was significantly associated with increasing levels of blood 25(OH)D (OR: 1.040, 95%; CI: 1.017-1.064, p = 7.14 × 10 −4 ; Figures 3 and 4A). The causal estimate in MR-PRESSO showed a stable association and the weighted median and MR-Egger methods also provided similar causal association (Figure 3). In addition, MR-Egger methods detected a significant causal effect of atopic dermatitis on the levels of 25(OH)D 3 (OR: 1.208, 95%; CI: 1.031-1.415; p = 0.048; Figures 3 and 4B). Sensitivity analyses detected a weak heterogeneity effect and horizontal pleiotropy in the analyses for 25(OH)D (Cochran's Q test: P Q >0.05; MR-Egger intercept test: P intercept < 0.05; MR-PRESSO global test: P global > 0.05) but not in the analyses for 25(OH)D 3 (Supplementary Table S3). Leave-one-out analyses indicated none single IV affected these causal estimations (Supplementary Figure S3).  Table S3). Leave-one-out analyses indicated none single IV affected these causal estimations (Supplementary Figure S3). Our analyses found the genetically predicted risk of neither psoriasis nor vitiligo influenced the circulating vitamin D levels (Figure 3 and Supplementary Table S3).

Discussion
In this study we carried out a bidirectional MR analysis and tested causal relationships between circulating vitamin D levels and the risk of three common skin inflammatory diseases: psoriasis, atopic dermatitis, and vitiligo. We identified a significant causal effect of elevated blood vitamin D levels on decreasing the risk of psoriasis and suggested that the incidence of atopic dermatitis raise blood vitamin D levels. These findings highlight a role of blood vitamin D levels as potential disease intervention and monitoring targets.
Early epidemiological studies have shown lower serum vitamin D levels in patients with psoriasis than healthy controls [12]. Administrating vitamin D supplementation led to reduced mortality in patients with psoriasis [17] and topical vitamin D analogues demonstrated effectiveness in treating psoriasis [9]. A recent prospective study suggested that low 25(OH)D level was associated with high risk of psoriasis using both cox proportional hazard model and MR analysis [27]. It used 69 genetic variants that were identified in a GWAS meta-analysis of 443,734 European individuals as IVs in the MR analysis [45]. However, the causal relationship of blood 25(OH)D3 and C3-epi-25(OH)D3 with psoriasis is yet to be studied. Our study validated the causal effect of 25(OH)D on the risk of psoriasis using a set of IVs which we identified in a different GWAS data of 120,618 Europeans, and performed MR analysis with different psoriasis GWAS summary statistics from the MR base [31,32]. In addition, we identified low levels of circulating 25(OH)D3 and C3-epi-25(OH)D3 increased the risk of psoriasis. To our best knowledge, we provided the causal evidence of blood 25(OH)D3 and C3-epi-25(OH)D3 with psoriasis for the first time. Their causal effects are in line with the reported one of 25(OH)D levels. Vitamin D receptors are widely present in keratinocytes, dendritic cells, macrophages, and T cells [4]. The active form of vitamin D interacts with its receptors within these immune cells, which as a result promotes antibacterial or antiviral innate responses and attenuates adaptive immunity [4]. In psoriasis, abnormal vitamin D metabolism might play a role in activating T cells and regulating cytokine secretion [12]. The serum levels of pro-inflammatory cytokines such as IFN-7, TNF-α, IL-1β, IL-6, IL-8, and IL-17 significantly decreased in patients with psoriasis after a three-month administration of vitamin D supplement [46]. These results together suggest the potential of vitamin D in reducing systemic inflammation in psoriatic patients. Our analyses found the genetically predicted risk of neither psoriasis nor vitiligo influenced the circulating vitamin D levels (Figure 3 and Supplementary Table S3).

Discussion
In this study we carried out a bidirectional MR analysis and tested causal relationships between circulating vitamin D levels and the risk of three common skin inflammatory diseases: psoriasis, atopic dermatitis, and vitiligo. We identified a significant causal effect of elevated blood vitamin D levels on decreasing the risk of psoriasis and suggested that the incidence of atopic dermatitis raise blood vitamin D levels. These findings highlight a role of blood vitamin D levels as potential disease intervention and monitoring targets.
Early epidemiological studies have shown lower serum vitamin D levels in patients with psoriasis than healthy controls [12]. Administrating vitamin D supplementation led to reduced mortality in patients with psoriasis [17] and topical vitamin D analogues demonstrated effectiveness in treating psoriasis [9]. A recent prospective study suggested that low 25(OH)D level was associated with high risk of psoriasis using both cox proportional hazard model and MR analysis [27]. It used 69 genetic variants that were identified in a GWAS meta-analysis of 443,734 European individuals as IVs in the MR analysis [45]. However, the causal relationship of blood 25(OH)D 3 and C3-epi-25(OH)D 3 with psoriasis is yet to be studied. Our study validated the causal effect of 25(OH)D on the risk of psoriasis using a set of IVs which we identified in a different GWAS data of 120,618 Europeans, and performed MR analysis with different psoriasis GWAS summary statistics from the MR base [31,32]. In addition, we identified low levels of circulating 25(OH)D 3 and C3-epi-25(OH)D 3 increased the risk of psoriasis. To our best knowledge, we provided the causal evidence of blood 25(OH)D 3 and C3-epi-25(OH)D 3 with psoriasis for the first time. Their causal effects are in line with the reported one of 25(OH)D levels. Vitamin D receptors are widely present in keratinocytes, dendritic cells, macrophages, and T cells [4]. The active form of vitamin D interacts with its receptors within these immune cells, which as a result promotes antibacterial or antiviral innate responses and attenuates adaptive immunity [4]. In psoriasis, abnormal vitamin D metabolism might play a role in activating T cells and regulating cytokine secretion [12]. The serum levels of pro-inflammatory cytokines such as IFN-7, TNF-α, IL-1β, IL-6, IL-8, and IL-17 significantly decreased in patients with psoriasis after a three-month administration of vitamin D supplement [46]. These results together suggest the potential of vitamin D in reducing systemic inflammation in psoriatic patients.
Previous MR analyses using GWAS data of vitamin D in 443,734 individuals suggested no causal effect of blood 25(OH)D on developing atopic dermatitis but a significant causal effect of atopic dermatitis on increasing blood 25(OH)D levels [25,28,45]. Our study achieved consistent results between 25(OH)D and atopic dermatitis by using GWAS summary statistics of different datasets. Furthermore, we extended the investigation further to the causal relationships between the levels of two extra vitamin D measures 25(OH)D 3 and C3-epi-25(OH)D 3 and the risk of atopic dermatitis. We found no significant causal effect of blood 25(OH)D 3 or C3-epi-25(OH)D 3 on the risk of atopic dermatitis but suggest that developing atopic dermatitis raised blood 25(OH)D 3 levels. Previous MR identified no causal effects of circulating 25(OH)D or 25(OH)D 3 on the risk of vitiligo [26]. We achieved similar results, suggesting no causal effects of 25(OH)D, 25(OH)D 3 , or C3-epi-25(OH)D 3 on vitiligo. In addition, the reverse associations were not statistically significant either. To our best knowledge, this is the first MR analysis to evaluate whether the genetically predicted risk of vitiligo consequently affects blood vitamin D levels. Despite of the common characteristics of immune dysregulation and inflammation among psoriasis, atopic dermatitis, and vitiligo our study identified their heterogeneous causal relationships with blood vitamin D levels.
Our study did not support a causal role of vitamin D levels on atopic dermatitis and vitiligo. This finding was in line with previous cross-sectional and perspective studies, which detected no supporting evidence for vitamin D being a risk factor [14,15,47]. However, our findings of atopic dermatitis raising blood vitamin D levels contradicted previous epidemiological reports that atopic dermatitis or vitiligo patients were associated with vitamin D deficiency [10,11]. This result suggested that disease biological pathways might influence blood vitamin D levels in atopic dermatitis. As a result, our finding has the potential to disentangle disease etiology and help understand disease complications for atopic dermatitis.
We acknowledge that these causal inferences might reflect the mixed effects of insufficient statistical power and unmeasured confounding factors. The interpretation of current MR results needs careful consideration. A recent paper highlighted the potential issue of non-linear genetic effects of vitamin D SNPs on vitamin D [48]. The impact of the potential non-linear effects on MR is worthy of further investigation. In addition, further studies are warranted to validate the causal relationship and to determine the molecular mechanisms underlying the causal effects.

Conclusions
In summary, we confirmed the previous reported causal relationship between 25(OH)D level and psoriasis and identified causal associations of lower 25(OH)D 3 and C3-epi-25(OH)D 3 levels with the increased risk of psoriasis and consequences of increased 25(OH)D 3 levels on atopic dermatitis for the first time. These findings provide intervention targets for psoriasis and disease monitoring biomarkers for atopic dermatitis.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nu14245284/s1, Table S1: GWAS data sources used in the present study. Table S2: The characteristics of instrumental variables (IVs) (LD clumping r 2 < 0.001). Table  S3: Mendelian randomization analysis results for IVs with LD r 2 < 0.001. Figure S1: Study design. Figure S2: Leave-one-out sensitivity analysis for MR analysis. Figure S3: Leave-one-out sensitivity analysis for reverse MR analysis.  Institutional Review Board Statement: The data were sourced from a publicly available database and ethical approval was not applicable.
Informed Consent Statement: Ethical approval and consent to participate were not needed for this current study because it is a secondary analysis of previously published data. In all original studies related to this study, ethical approval, and consent to participate were obtained.