Immune cells and the risk of acne vulgaris: A Mendelian randomization study

Abstract

Acne vulgaris is a chronic inflammatory disease affecting the pilosebaceous glands. It primarily manifests on the face, chest, and back, and is prevalent among adolescents of both genders. The mechanism behind it is significantly associated with the inflammatory response triggered by microbial infections and the infiltration of immune cells. However, the precise causal role of immune cells in acne vulgaris is unclear. To this end, we conducted a 2-sample Mendelian randomization analysis to evaluate the relationship between immune cells and the likelihood of developing acne vulgaris. Instrumental variables were chosen from single nucleotide polymorphisms linked to immune cells, as determined by an extensive genome-wide association study. To explore the potential causal relationship between the 731 immune cell traits and the risk of acne vulgaris, we performed a Mendelian randomization analysis using an inverse variance weighting approach. Moreover, sensitivity, heterogeneity and pleiotropy analyses were performed to ensure the reliability of the findings. In addition, a Finnish dataset was used to validate the results of the forward Mendelian randomization analysis and further multivariate Mendelian randomization analyses were performed. Our study identified 31 immune phenotypes causally associated with acne vulgaris. After validation using the FinnGen database, 3 types of immune cells were identified as being associated with the development of acne vulgaris, including secreting regulatory T (Treg) AC (odds ratio [OR] = 0.967, 95% confidence interval [CI] = 0.946–0.989, P = .004), TD DN(CD4‐ CD8‐) %T cells (OR = 1.079, 95% CI = 1.012–1.149, P = .018), CD25 on CD39+ secreting Treg (OR = 0.951, 95% CI = 0.912–0.992, P = .022). Furthermore, multivariate Mendelian randomization analysis revealed that only 1 immune cell, secreting Treg AC, was causally associated with acne vulgaris. This research established a causal link between Treg cells and acne vulgaris, potentially serving as a predictive marker for diagnosing acne vulgaris and advancing new immunotherapy approaches.

1. Introduction

Global statistics indicate that acne vulgaris is the eighth most prevalent condition, affecting around 9.4% of the global population.^[1] The main characteristics of acne vulgaris involve noninflammatory lesions, inflammatory lesions, and persistent erythema, hyperpigmentation, and scarring.^[2] Multiple factors are involved in the pathogenesis of acne vulgaris, bacterial infection, pilosebaceous duct obstruction, sebaceous hypersecretion and inflammation are closely related to its occurrence.^[3] These factors interact to cause hair follicle plugging and promote bacterial multiplication, ultimately triggering an inflammatory response and the formation of skin lesions.^[4] Admittedly, the occurrence of acne vulgaris is influenced by a combination of factors, including individual genetic predisposition, lifestyle choices, and environmental influences.

In recent times, there has been an increasing focus on the role of immune cells in the onset and progression of acne vulgaris.^[5] Studies have shown that acne vulgaris is not only caused by excessive sebum secretion but may also be a manifestation of immune dysfunction.^[6] The infiltration of immune cells can trigger a local inflammatory response, which can promote the development of acne vulgaris.^[7] In this process, Cutibacterium acnes can activate the helper T cells in the skin, especially Th1 and Th17 cells, which further aggravates the local inflammatory response and aggravates the pathological process of acne vulgaris.^[8] In addition, although regulatory T (Treg) cells also appear to play a role in the progression of acne vulgaris, the specific mechanisms are not fully understood. Although a large number of studies have confirmed the close association between acne vulgaris and the immune system, whether this association is causal still needs to be further explored.

Mendelian randomization (MR) is a research approach used to identify the causal link between instrumental variables, exposure, and outcomes by randomly choosing genetic variables to serve as instrumental variables within a population.^[9,10] Conventional case-control research has contributed to uncovering the connection between immune cells and acne vulgaris. Nonetheless, it is crucial to recognize that these studies possess specific drawbacks, including limited sample sizes and the presence of uncontrolled confounding variables. In contrast, MR Studies have unique advantages: larger sample sizes and fewer factors that can confound causality.^[11] We used a Mendelian randomization 2-sample analysis to assess the genetic causality of 731 immune cell traits associated with acne vulgaris. The aim is to provide a new perspective for the prevention, treatment, and diagnosis of acne vulgaris.

2. Materials and methods

2.1. Overall study design

Two-way MR studies follow the framework shown in Figure 1. Causal associations between 731 immunophenotypes and acne vulgaris were evaluated based on 2-sample MR analysis. To ensure the reliability of the results, MR analysis needs to fulfill 3 key assumptions: Association assumption: the instrumental variables should have a strong correlation with the exposure; Independence assumption: there must be independence between the instrumental variables and any confounding factors; Exclusivity assumption: the influence of instrumental variables on outcomes occurs solely through the exposure.^[12,13]

Figure 1.

Study design to explore the association between immune cells and the risk of acne vulgaris.

Firstly, we obtained data on 731 immune cell traits and conducted a univariate MR analysis to investigate their association with acne vulgaris. The findings from the univariate MR analysis were subsequently validated using data from the FinnGen. Based on these validation results, we further employed reverse Mendelian randomization analysis to evaluate potential reverse causality between the exposures and outcomes. To minimize the influence of intercellular interactions, we also performed multivariate Mendelian randomization (MVMR) analysis.

2.2. Data sources

2.2.1. Summary statistics for immune cell

The immunophenotypic data of this study were obtained from the genome-wide association study (GWAS) dataset numbered GCST90001391–GCST90002121, which contains 731 immune cells traits.^[14] This GWAS analysis was based on 3757 European adults, testing about 22 million single nucleotide polymorphisms (SNPs) and adjusting for sex and age. The dataset comprised 118 absolute cell counts, 389 median fluorescence intensity measurements, 32 morphological characteristics, and 192 relative cell counts.^[14]

2.2.2. Summary statistics for acne vulgaris

Aggregate acne vulgaris statistics were from GWAS and meta-analyses of 3 independent European pedigrees.^[15] The ultimate sample included 34,422 cases paired with 364,991 controls. Variations in the reporting of acne vulgaris cases were observed among cohorts, encompassing clinical assessment records, the extraction of ICD-10 codes from electronic health records, and diagnoses reported by individuals themselves.^[15]

The validation dataset for the FinnGen L12_ACNE phenotype is based on a large population cohort from Finnish biobanks and national health registries. It includes 3245 acne vulgaris patients and 394,105 controls, providing a robust resource for genetic analysis related to acne vulgaris.

2.3. Instrumental variables selection

In order to improve the dependability of our results concerning the link between immune cells and the likelihood of acne vulgaris, we established a range of quality control procedures aimed at identifying the most efficient genetic instruments. We carefully selected suitable genetic variants as instrumental variables (IVs) from publicly available databases of GWAS. Initially, we established a threshold of P < 1 × 10^‐5 for screening SNPs associated with immune cells. At this threshold, there were too few SNP loci with 3 immune cells (SNP number < 3), so these 3 immune cells were not selected for subsequent MR analysis. Second, we used genome-wide data from the European 1000 Genomes Project as a reference and eliminated SNPs associated with linkage disequilibrium. The linkage disequilibrium criterion was set as r² < 0.001 and the window size was 10,000 kb. Highly correlated SNPs were removed to ensure independence between SNPs.^[16,17] At last, in order to evaluate if the included SNPs were impacted by the weak instrumental variables, we also computed the F-statistic. When the F-statistic associated with the SNPs is found to be <10, it indicates that these SNPs are being affected by weak instrumental variables.

2.4. Statistical analysis

2.4.1. Forward Mendelian analysis

The primary aim of this research was to explore the possible causal link between immune cells and acne vulgaris through various research techniques. We employed the inverse variance weighting (IVW) method as our main analytical approach. This method integrates β values with the standard error associated with their causal estimates.^[18] The rationale for the IVW approach is to determine the inverse variance of each instrumental variable as a weight on the assumption that all instrumental variables are valid. This approach represents the most traditional technique used in MR analysis.^[19,20]

In order to evaluate the reliability of the preliminary analysis, multiple sensitivity assessments were performed to confirm the validity of the findings.^[21] The MR-Egger intercept test was utilized to evaluate whether horizontal pleiotropy was present. A P-value for the intercept that fell below .05 was deemed statistically significant, indicating a potential presence of horizontal pleiotropy. Additionally, the MR-PRESSO test was utilized to identify and remove outliers, as well as mitigate the impact of moderate pleiotropy. This method detects potential pleiotropic outliers and reevaluates the causal effect estimates after their exclusion.^[22,23] In the present research, the Cochran Q test was employed to assess the variability among SNPs. When heterogeneity was detected (P-value < .05), the MR effect could be estimated using a direct random effects model.^[24]

2.4.2. Verification of the robustness of Mendelian randomization results

In this study’s Mendelian randomization analysis, we utilized a separate sample from the Finnish dataset for validation purposes, aiming to reduce the likelihood of chance and improve the reliability of the initial screening outcomes. This dataset was derived from the FinnGen project and contains large-scale genetic data from European populations. The robustness and generalization of our main results were further verified by repeating the analyses in independent populations. This validation step ensured the reliability of the results and was able to reduce the impact of potential population structure and bias on the analysis. If the results of Mendelian randomization obtained from the primary screening were inconsistent with those derived during the validation phase, they were excluded from consideration.

2.4.3. Reverse Mendelian analysis

To assess a potential causal relationship between acne vulgaris and immune cells, we performed a reverse Mendelian analysis in which acne vulgaris was used as the exposure factor and immune cells as the outcome.^[25] We selected SNPs as potential IVs by applying a stringent significance threshold of P < 5 × 10^‐8 across sites. The anti-Mendelian analysis method is consistent with the forward MR method.

2.4.4. Multivariate Mendelian analysis

Given the interaction between the various immune cells, they have the potential to be confounding factors. Therefore, we further performed multivariate Mendelian analysis. Multivariate Mendelian analysis^[26] functions as a method that improves the accuracy of determining the impact of exposure on outcomes by considering several potential confounding variables. It reduces error and improves the accuracy of the results by adjusting for the effect of these confounders on the outcome variables.^[27] In studies with multivariate Mendelian analysis, MVMR–IVW was used as the main method.

2.5. Data analysis

The odds ratio (OR) along with the 95% confidence interval (CI) was employed to determine the relative risk associated with the presence of exposure. All MR analyses were performed with the use of the “TwoSampleMR (0.5.6),” “MendelianRandomization (0.9.0)”, and “MR-PRESSO (1.0)” packages in R 4.3.1 software.

2.6. Ethical statement

All GWAS databases utilized in this research are available to the public, and every original study included has undergone ethical review and obtained approval.

3. Results

3.1. Characteristics of the selected SNPs

A total of 731 immunophenotypes were included in this study. After preliminary screening and data processing by Plink, a total of 11,747 SNPs were screened as effective instrumental variables. The F-statistic for all instrumental variables was >10, effectively eliminating the effect of weak instrumental variables in this study. Reverse MR analysis identified 17 SNPs strongly associated with acne vulgaris, all of which had F-statistics > 10, suggesting the robustness of IVs (Fig. 1).

3.2. Single variable MR analysis and verification results

In this study, a 2-sample Mendelian Randomization analysis was conducted to investigate the causal link between immune cells and acne vulgaris, predominantly utilizing the inverse variance weighting method. We identified 31 immunophenotypes of immune cells that are potentially causally related to acne vulgaris (Table S1, Supplemental Digital Content, https://links.lww.com/MD/Q533). Subsequently, we validated these 31 immune cells using the FinnGen dataset and showed that only 3 immune cells passed the validation, which is consistent with our findings in the preliminary MR analysis (Fig. 2). Immune cells that have not been validated may be due to population differences between different data sets, may be affected by pleiotropy, or covariate confounding, thus affecting the consistency of the results.

Figure 2.

The primary screening and validation forest plots illustrate the causal involvement of immune cells in acne vulgaris. TD DN = double negative TD.

The immune cell double negative TD (CD4‐ CD8‐) %T cells (primary screening: OR = 1.038, 95% CI = 1.008–1.069, P = .012; validation: OR = 1.079, 95% CI = 1.012–1.149, P = .018) had a positive causal relationship with the occurrence of acne vulgaris. We also noted that the MR analysis based on the IVW method showed an inverse causal relationship between immune cells secreting Treg AC (primary screening: OR = 0.984, 95% CI = 0.971–0.998, P = .032; validation: OR = 0.967, 95% CI = 0.946–0.989, P = .004) and acne vulgaris. There was a significant causal relationship between CD25 on CD39+ secreting Treg (primary screening: OR = 0.973, 95% CI = 0.954–0.993, P = .008; Validation: OR = 0.951, 95% CI = 0.912–0.992, P = .022) and the decreased risk of acne vulgaris (Fig. 2).

In sensitivity analysis, the MR-Egger findings indicated the absence of horizontal pleiotropy among the instrumental variables (P > .05). Cochran Q test showed that there was no heterogeneity in this study (P > .05). Additionally, the results of the MR-PRESSO global test analysis showed that the included SNPs did not have any outliers, which was consistent with the IVW results (Table 1).

Table 1.

Mendelian randomization analyses were used to estimate the association between the 3 immune cells and the risk of acne vulgaris.

Exposure: immune cells	nSNPs	Outcome	Heterogeneity test			Horizontal pleiotropy test
						MR-Egger intercept test		MR-PRESSO global test
			Method	Cochran Q	P-value	Intercept	P	P
Secreting Treg AC	15	Acne	IVW	14.267	.429	‐0.001	.733	.944
TDDN (CD4‐ CD8‐)% T cell	16		IVW	14.282	.504	‐0.001	.835	.241
CD25 on CD39+ secreting Treg	10		IVW	9.296	.41	‐0.002	.632	.444

The Cochran Q test was used to assess heterogeneity between SNP-specific estimates, and the MR-Egger intercept test and the MR-PRESSO global test were used to test for horizontal pleiotropy.

IVW = inverse variance weighting, SNPs = single nucleotide polymorphisms, TD DN = double negative TD.

3.3. Reverse MR analysis results

To investigate the causal connection between acne vulgaris and the 3 identified positive immune cell characteristics, reverse Mendelian randomization analysis was conducted through a 2-sample univariate approach. After performing MR reverse analysis using IVW method, we found no evidence that acne vulgaris was associated with secreting Treg AC (OR = 0.902, 95% CI = 0.745–1.093, P = .296), double negative TD (CD4‐ CD8‐) %T cells (OR = 1.003,95%CI = 0.801–1.256, P = .976), and CD25 on CD39+ secreting Treg (OR = 1.119, 95% CI = 0.929–1.348, P = .233) (Fig. 3).

Figure 3.

Forest plots show the reverse causality of acne vulgaris on immune cells. TD DN = double negative TD.

3.4. Multivariable MR

To evaluate the results of this study and to control for the mutual association between the different exposure variables, a multivariate Mendelian randomization analysis was used. We assessed 3 specific immune-cell signatures linked to acne vulgaris through multivariable MR analysis, aiming to account for the influence of various positive immune-cell signatures on the results (Fig. 4).

Figure 4.

Causal effect of genetically predicted secreting Treg AC and acne vulgaris using MVMR. MVMR = multivariate Mendelian randomization, TD DN = double negative TD.

The results showed that only 1 immune cell had an independent causal role in acne vulgaris: secreting Treg AC (OR = 0.977, 95% CI = 0.956–0.998, P = .037) reduced the risk of acne vulgaris (Fig. 4). To identify and eliminate horizontal pleiotropy, we used the MR-PRESSO method for analysis. The results of MR-PRESSO global test showed a P-value of .573 (P > .05), which indicated that there was no significant horizontal pleiotropy between immunophenotype and acne vulgaris.

4. Discussions

This research is the initial endeavor to explore the causal link between immune cells and acne vulgaris through MR analysis. Findings indicated that 3 types of immune cells displayed a causal association with acne vulgaris, as confirmed by the Finnish dataset, whereas reverse MR analysis revealed no influence of acne vulgaris on immune cells. Further MVMR showed that only 1 types of immune cells, secreting Treg AC, affected the pathogenesis of acne vulgaris. This provides a new perspective for further understanding the pathogenesis, targets and prevention strategies of acne vulgaris.

Acne vulgaris refers to the clinical and histological manifestations of localized or diffuse pathological immune processes in the hair follicles and sebaceous glands of the skin.^[28] Alterations in the number and function of immune cells and bacterial-mediated inflammatory responses are commonly seen in patients with acne vulgaris.^[29] Our findings regarding the causal relationship between immune-cell profiles, especially for Treg cells, and acne vulgaris confirm previous epidemiologic and laboratory studies. T cells are divided into 3 main subsets: helper T cells, cytotoxic T cells, and Treg cells. Our findings revealed a significant inverse correlation between secreting Treg AC expression and the risk of acne vulgaris. Treg cells suppress the function of other immune cells through the secretion of anti-inflammatory cytokines, including interleukin (IL)-10, IL-2, and transforming growth factor-β. This suppression leads to a decrease in the activity of antigen-presenting cells and T cells, resulting in a lower production of inflammatory cytokines and antibodies.^[30,31] Tregs can be induced from conventional T cells and demonstrate strong immunosuppressive capabilities.^[32] Treg cells play an important role in the regulation of immune system, homeostasis and prevention of autoimmunity. In autoimmune diseases, tumors, and various other pathological conditions, the quantity and role of Treg cells could potentially be altered. As in certain diseases, an imbalance in the number or function of Tregs may lead to immunosuppression or may cause the immune system to respond to stress. Secreting Treg AC may reflect a subset of Treg cells with a strong secretory function. These Tregs regulate the activity of antigen-presenting cells by secreting inhibitory cytokines (e.g., IL-10, transforming growth factor-β, and IL-35).^[33,34] The enhancement of Treg may inhibit the over-activation of antigen-presenting cells, and then reduce the production of pro-inflammatory factors (such as IL-23), and finally inhibit the differentiation of Th17 cells and IL-17-mediated inflammatory response.^[35,36] This mechanism helps to control local immune hyperactivation triggered by Propionibacterium acnes, thereby reducing the risk of acne vulgaris. In inflammatory skin disorders such as acne vulgaris, secreting Treg AC can effectively inhibit excessive local immune responses and alleviate inflammatory damage, making them significant protective factors in the development and progression of acne vulgaris.^[37]

However, several limitations of this study must be acknowledged. First, the results of our study are exclusively from a population of European ancestry. This may limit generalizability to other ethnic groups because of potential genetic and environmental heterogeneity. Second, the lack of key individual-level data (e.g., age, and sex) in the original data set precluded stratified analyses and hindered exploration of subgroup specific effects. Third, while Mendelian randomization provides robust genetic evidence of causality, further validation by multidimensional approaches, such as prospective cohort studies and mechanistic experiments, is still necessary. Finally, clinical validation was only feasible for secreting Treg AC through retrospective clinical records, as routine measurement data for other immune phenotypes were unavailable; consequently, their causal roles require substantiation via targeted experimental assays.

Future research should focus on the following directions: Firstly, deeply analyze the molecular mechanism of secreting Treg AC and acne vulgaris, focusing on its specific functions, key cytokines and regulation of signaling pathways. Secondly, in vitro model (cell co-culture) and in vivo model (animal experiment) were used to verify the reliability of MR results and explore its therapeutic potential in other immune-related diseases. Finally, the validation was extended to multi-ethnic populations to ensure the universality of the conclusions. These directions will clarify the immunomodulatory role of secreting Treg AC, provide scientific basis for the precise treatment strategy of acne vulgaris and related diseases, and promote the progress of the field.

5. Conclusion

According to our Mendelian randomization results, there appears to be a potential causal relationship between the cell level of secreting Treg AC and a reduced risk of developing acne vulgaris. Further investigations are necessary to validate these discoveries, explore the underlying pathological mechanisms, and formulate innovative approaches for assessing risk, predicting prognosis, and treating acne vulgaris.

Author contributions

Conceptualization: Wenjuan Wu, Jiankang Yang.

Data curation: Xiaoyi Yang.

Formal analysis: Xiaoyi Yang.

Methodology: Wenjuan Wu, Jiankang Yang.

Supervision: Wenjuan Wu.

Validation: Jiankang Yang.

Writing – original draft: Xiaoyi Yang.

Writing – review & editing: Wenjuan Wu, Jiankang Yang.