Genetic insights into the risk of metabolic syndrome and its components on psoriasis: A bidirectional Mendelian randomization

The role of metabolic syndrome (MetS) on psoriasis has been explored only in observational studies. We conducted this bidirectional Mendelian randomization (MR) to clarify the causal relationship between MetS and its components and psoriasis. The genetic instruments of MetS and its five components (waist circumference [WC], hypertension, fasting blood glucose [FBG], triglycerides [TG], and high‐density lipoprotein cholesterol [HDL‐C]) were obtained from public genome‐wide association studies (GWAS). Outcome datasets for psoriasis were collected from the FinnGen Biobank Analysis Consortium. The main method was inverse variance weighting, complemented by sensitivity approaches to rectify potential pleiotropy. MetS, WC, and hypertension increase the risk of psoriasis (MetS, odd ratios [OR] = 1.17, 95% confidence interval [CI] 1.08–1.27, p = 1.23e‐04; WC, OR = 1.65, 95% CI  1.42–1.93, p = 1.06e‐10; hypertension, OR = 2.02, 95% CI 1.33–3.07, p = 0.0009). In the reverse analysis, no causal association between psoriasis and MetS and its five components was identified. We provide novel genetic evidence that MetS, WC, and hypertension are risk factors for the development of psoriasis. Early management of MetS and its components may be an effective strategy to decrease the risk of psoriasis.


| INTRODUC TI ON
Psoriasis is characterized as demarcated areas of erythematous plaque with overlying silvery scales on the skin.It is estimated that a total of 60 million patients have psoriasis worldwide, and the incidence rate is 1.72%. 1 Psoriasis is closely associated with genetic elements.3][4] Therefore, early intervention on modifiable risk factors can significantly reduce the incidence of psoriasis.
Generally, a metabolic syndrome (MetS) is a cluster of pathological conditions including glucose abnormalities, hyperlipidemia, central obesity, and hypertension. 5The number of patients with MetS has increased dramatically, and one-quarter of adults in developed countries have MetS. 68][9] However, obvious limitations exist in these studies, such as the study design and limited sample size, which may make causal inference challenging.
Mendelian randomization (MR) is a robust statistical approach using genetic variants (single-nucleotide polymorphisms [SNPs]) to explore the causal effects of MetS and its components on psoriasis, which can overcome the shortcomings in observational studies. 10milar to a randomized clinical trial, genetic variants are allocated randomly during the formation of a zygote during gestation based on the random principle of meiosis. 11,12The results of MR analyses are not susceptible to reverse causality and confusion.At present, no study has been conducted to explore the causal association between psoriasis and MetS and its components.Collectively, we conducted this bidirectional MR analysis to discover their causal links.

| Date sources of exposure and outcome
The summary level exposure datasets for MetS, waist circumference (WC), hypertension, fasting blood glucose (FBG), triglycerides (TG), and high-density lipoprotein cholesterol (HDL-C) were obtained from the largest and most comprehensive GWAS public databases (https:// gwas.mrcieu.ac.uk/).The SNPs associated with MetS (59 677 cases and 231 430 controls), WC (462 166 samples), hypertension (463 010 subjects), TG (441 016), and HDL-C (441 016) were taken from UK Biobank which has more than 400 000 participants. 13As to FBG, the dataset originated from the meta-analyses of glucose and insulinrelated traits consortium that included 281 416 samples. 14 2017, the FinnGen study was launched in Finland and is a population-based study that includes genome information.The genetic information was analyzed with Illumina and Affymetrix chip arrays (Illumina Inc. and Thermo Fisher Scientific), and this study includes 3775 whole genomes and a total of 16 962 023 variants.
We extracted the summary-level outcome datasets regarding psoriasis from the FinnGen Biobank databases.The dataset includes 4510 cases and 212 242 controls.The International Classification of Diseases criteria was used for the diagnosis of psoriasis.GWAS data sources for MetS and its components and psoriasis used in the present study are shown in Table S1.There is no overlap between exposure datasets and outcome datasets.

| Genetic instrument selection
All instrumental variables (IVs) must satisfy three assumptions: (1)   IVs are significantly associated with exposure; (2) IVs are not related to any confounders of the exposure-outcome association; and (3) IVs affect the outcome only via the exposure.The genetic instruments associated with MetS, WC, hypertension, FBG, TG, and HDL-C should meet the threshold of p < 5 × 10 −8 .Additionally, the linkage disequilibrium (LD) of SNP is regarded as independent SNPs when they reach the criteria (LD r 2 < 0.05 at a window size of 10 000 kb).Furthermore, the MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) approach is used to find the significant outliers that explain the possible pleiotropy. 15The significant outliers should be deleted to reduce possible heterogeneity and pleiotropy.
The F-statistic = (Bets/Se) 2 was used to evaluate the strength of our MR study.A value of more than 10 is an indicator of statistical robustness. 16The qualified SNPs are summarized in Table 1.

| Main statistical analyses
Random effects inverse variance weighting (re-IVW) is used as the main method to test the causal association between MetS WC, hypertension, FBG, TG, HDL-C, and psoriasis since this analysis can provide a robust causal estimate in the absence of directional pleiotropy (no violation of the independence assumption). 17A Bonferroni-corrected p < 0.0041 (0.05/12) in this study was deemed as the significant threshold.All statistical analyses were performed using "TwoSampleMR," "mr.raps," "cause," and "forest plot" packages in R software (version 4.2.0;R Foundation for Statistical Computing, Vienna, Austria).

| Ethical approval
All exposure and outcome datasets in our study were retracted from public de-identified public data/studies.All studies were approved by the ethics committee previously and informed consent was obtained from all participants in the original studies.

| Sensitivity analyses
To assess potential heterogeneity and pleiotropy, we chose another five methods to perform sensitivity analyses, including MR-Egger, weighted median, maximum likelihood, MR.RAPS, and MR-PRESSO methods.On the assumption of instrument strength independent of direct effect (InSIDE), the egger intercept term in the regression model was introduced to represent the directional pleiotropy in MR-Egger analysis.MR-Egger can obtain a robust causal effect, even if all the IVs are invalid. 18When the intercept term is close to zero, horizontal pleiotropy does not exist and the results of both IVW and MR-Egger are similar. 18Where half of the SNPs are invalid, consistent findings can be obtained by aggregating numerous estimates into a single total effect using the weighted median approach. 19The maximum likelihood method with a low standard error may be influenced by a small sample. 20Robust causal estimates can be yielded in the condition of the normal distribution of pleiotropy in the MR.RAPS analysis when systematic and idiosyncratic pleiotropy exits. 21When weak SNPs exist, MR-RAPS analysis can provide higher statistical power. 22MR-PRESSO was used to find outliers and remove them, and then obtain a robust estimate. 15The value of the global test in MR-PRESSO analysis was used to account for horizontal pleiotropy.
Cochran's Q test was performed to explain the possible heterogeneity among variant-specific estimates.In addition, the leave-one-out analysis was used to illustrate the robustness of the results when each IV was removed in turn.

| Casual effect of genetically predicted MetS and its components on psoriasis in the MR analysis
The overview of our MR analyses is shown in Figure 1.All IVs satisfied the three MR assumptions.The characteristics of IVs in MR analysis are presented in Table S2.
The results of IVW analysis show that genetically predicted MetS contribute to a 1.17-fold risk of psoriasis (95% confidence interval [CI] 1.08-1.27,p = 1.23e-04,Table 2).The scatter plot in Figure 2a reveals that with the increasing effect of SNPs on psoriasis, the SNP effect on MetS is intensified, indicating the risk of psoriasis is increased by MetS.In the sensitivity analysis, the results of Cochran's Q test (Table 2) and funnel plots in Figure 3a demonstrate no sign of heterogeneity (p > 0.05).In MR-Egger and MR-PRESSO analysis, no evidence of pleiotropy was observed (all p > 0.05).In addition, no influential SNP was detected using leave-one-out analysis after removing each SNP in turn (Figure 4a).The forest plot in Figure 5a displays the effect estimate of each SNP on psoriasis.
For MetS components, the increasing risk of psoriasis was increased by WC (OR = 1.65, 95% CI 1.42-1.93,p = 1.06e-10,  (all p > 0.05).The leave-one-out analysis does not detect influence while removing any SNP in turn (Figure 4b-f).The forest plot in Figure 5b-f shows the casual estimates for each IV on psoriasis.

| Casual effect of genetically predicted psoriasis on MetS and its components in the reverse MR analysis
All IVs satisfied the three MR assumptions.The characteristics of (IVs in MR analysis are presented in Table S3. Although genetically predicted psoriasis results in the increasing level of HDL-C (OR = 1.005, 95% CI 1.000-1.010,p = 0.0023, Table S4, Figure S1), the causal association is not supported by the CAUSE method in the sensitivity analysis (p > 0.0041).No causal relationship was observed between psoriasis and the risk of MetS, WC, hypertension, FBG, and FBG (all p > 0.05).In Cochran's Q test, although the p < 0.05 is detected in the results between psoriasis and HDL-C, no evidence of heterogeneity was observed due to the symmetry of the MR results (Figure S2).No significant outliers in MR-Egger and MR-PRESSO analysis (all p > 0.05) were found.

TA B L E 2
The causal effect of MetS and its components on psoriasis in the MR analysis.Moreover, no influential outliers were observed in the leave-one-out analysis (Figure S3).The causal estimate of each SNP is presented in Figure S4.

| DISCUSS ION
We provide novel genetic evidence that MetS, WC, and hypertension are risk factors for psoriasis in this MR study.However, no causal association between psoriasis and MetS and its five components was identified in our reverse MR study.
Many epidemiological studies have demonstrated that MetS is prevalent among patients with psoriasis.In a population study that included 398 701 individuals conducted in Spain, researchers found that MetS increased by 13.2% in patients with psoriasis compared with controls (p < 0.001, OR = 2.21). 23A similar conclusion was also obtained in a South African population 24 and in a Chinese psoriatic population. 25Moreover, the incidence of MetS is positively related to psoriasis severity.A study involving 44 715 individuals found that increasing odds of MetS was observed in patients with mild (22%), moderate (56%), and severe psoriasis (98%). 26However, several studies did not find a relationship between MetS and psoriasis. 27,28The inconsistent conclusions may be a result of obvious confounders in observational studies, such as study design and the retrospective nature of the study.In our bidirectional MR study, we found that MetS is positively related to the risk of psoriasis.
Among the individual components of MetS, obesity had a higher prevalence in patients with psoriasis.In a prospective study that Health Study (HUNT), was associated with a 1.26-fold risk of psoriasis (95% CI 1.15-1.39). 29The current evidence supports the association between hypertension and psoriasis. 30,31In a metaanalysis consisting of 24 observational studies, a dose-dependent trend was found with ORs of 1.49 for severe psoriasis and 1.30 for mild psoriasis. 32Moreover, difficult-to-control hypertension and demand for cardiovascular interventions and surgery necessory in psoriasis. 15In a 10 year follow-up study of 256 356 adults (42 726   in the hypertension group and 213 630 in the control group), the incidence of psoriasis in the hypertension group was higher compared to the control group (hazard ratio [HR] = 1.54, 95% CI 1.47-1.61,p < 0.001), the positive association remained stable after adjusting for comorbidities such as diabetes and dyslipidemia. 33In our MR study, we found that WC and hypertension increased the risk of psoriasis, and the inverse relationship was not supported in our reverse MR analysis.
At present, no study has been conducted on the relationship between FBG and psoriasis.In contrast, many studies have focused on the association between diabetes and psoriasis.It is well-known that diabetes is prevalent in patients with psoriasis. 34,35A metaanalysis of 44 studies found that diabetes in patients with psoriasis had a higher risk than in those without psoriasis (OR = 1.76, 95% CI 1.59-1.96),and with even higher ORs for patients with severe psoriasis (OR = 2.10, 95% CI 1.73-2.55). 35The relationship still exists after adjusting age, sex, and smoking. 36Not all studies identify the association between plasma lipid profiles and psoriasis. 34,37This inconsistency may be the result the heterogeneity of criteria for MetS.In our MR and reverse analyses, we did not identify an obvious causal association between FBG, TG, HDL-C, and psoriasis.Some common potential pathological processes are observed among MetS, WC, hypertension, and psoriasis.It is well-known that the hallmark of psoriasis is sustained inflammation. 38Multiple inflammatory cytokines and immune cells are involved in the pathological process of psoriasis. 39Th17 cells play an important role in the development of psoriasis because they can secrete several crucial pro-inflammatory cytokines such as IL-17, TNFα, and IL-22.1][42] IL-17 is involved in the inflammation process and insulin resistance, suggesting that it is a potential mediator linking MetS and psoriasis. 43An anti-IL-17A monoclonal antibody like secukinumab is an effective drug for the treatment of plaque psoriasis, and weight and WC are also decreased in patients using secukinumab. 44Similarly, the anti-TNFα agents approved by the Food and Drug Administration to treat moderate and severe psoriasis, such as infliximab, adalimumab, and etanercept, have been shown to improve MetS and blood pressure. 45 our bidirectional MR analyses, we found that MetS, WC, and hypertension are risk factors for psoriasis, but not FBG, TG, and Some strengths of our study should be highlighted.First, it is the first study to assess the bidirectional, causal association between MetS and psoriasis.MR analysis can overcome confounders such as retrospective study type and reverse causation.Second, the large sample in our MR study supports the reliability of our conclusions.
Some limitations should also be mentioned.The current study was conducted on populations of European ancestry, which leads to the difficulty in applying the findings to other races.Moreover, there is a huge quantity variation between cases and controls in outcome datasets.In future studies, replication in other ethnicities and large studies with more samples should be performed to verify the conclusions.

| CON CLUS ION
Our bidirectional MR study supports MetS, WC, and hypertension as risk factors for psoriasis.In reverse MR analysis, no association was detected.The early management of those factors may be helpful in decreasing the risk and severity of psoriasis.
Figure 2d-f).The scatter plot in Figure 2bf revealed the effects of WC, hypertension, FBG, TG, FBG, and HDL-C on psoriasis.In sensitivity analysis, the Cochran's Q value demonstrated an obvious heterogeneity between hypertension, TG, and psoriasis (p < 0.05), while the funnel plot in Figure 3e shows a symmetry of the MR results, indicating no evidence of heterogeneity.The MR-Egger and MR-PRESSO analysis did not find any significant outliers and no significant outliers were removed F I G U R E 1 The flow chart of our Mendelian randomization (MR) analysis.HDL-C, high-density lipoprotein cholesterol; IVs, instrumental variables; MetS, metabolic syndrome; SNP, single nucleotide polymorphism.| 1395 LIU et al.

F I G U R E 2
The scatter plots of the association between genetically predicted (a) metabolic syndrome (MetS), (b) waist circumference (WC), (c) hypertension, (d) fasting blood glucose (FBG), (e) triglycerides (TG), (f) high-density lipoprotein cholesterol (HDL-C) on psoriasis in the Mendelian randomization (MR) analysis.IVW, inverse variance weighting; RAPS, robust adjusted profile score; SNP, single-nucleotide polymorphism.F I G U R E 3 The funnel plots of the association genetically between (a) metabolic syndrome, (b) waist circumference, (c) hypertension, (d) fasting blood glucose (e) triglycerides, and (f) high-density lipoprotein cholesterol on psoriasis in the Mendelian randomization (MR) analysis.IVW, inverse-variance weighted.

included 33 734
people in the population-based Nord-Trøndelag F I G U R E 4 The leave-one-out analysis of the association between genetically (a) metabolic syndrome, (b) waist circumference (c) hypertension, (d) fasting blood sugar, (e) triglycerides, and (f) high-density lipoprotein cholesterol on psoriasis in the Mendelian randomization analysis.OR, odds ratios.

F I G U R E 5
Forest plots of the association between (a) metabolic syndrome, (b) waist circumference, (c) hypertension, (d) fasting blood glucose, (e) triglycerides, and (f) high-density lipoprotein cholesterol on stroke in the Mendelian randomization analysis.OR, odds ratios.| 1399 LIU et al.
The R 2 and F-statistics for the genetic instruments in the MR analyses.
13,14Ethical approval (North West Multi-centre Research Ethics Committee [MREC] and The Coordinating Ethics Committee of the Helsinki and Uusimaa Hospital District) was thus exempted from our study TA B L E 1 Abbreviations: FBG, fasting blood glucose; HDL-C, high-density lipoprotein cholesterol; MetS, metabolic syndrome; MR, Mendelian randomization; NAFLD, non-alcoholic fatty liver disease; Re-MR, reverse mendelian randomization; SNP, single nucleotide polymorphism; TG, triglycerides; WC, waist circumference.because we used only the summary level data.The Strengthening the Reporting of Observational Studies in Epidemiology using Mendelian Randomization (STROBE-MR) check list is provided in Supporting Information.