Association Between Inflammatory Cytokines and Refractive Errors: A Bidirectional Mendelian Randomization Study

Abstract

Purpose: This investigation aimed to elucidate the causal role of inflammatory cytokines in the risk of developing refractive errors.

Methods: Genetic variants previously associated with inflammatory cytokines served as instrumental variables in genome-wide association studies (GWASs) of European ancestry. Bidirectional two-sample Mendelian randomization (MR) analyses were conducted using summary data from GWAS meta-analyses. Rigorous sensitivity analyses were performed to validate the reliability of the MR results.

Results: We found that, for every unit increase in interleukin 1 receptor antagonist (IL1RA) and interleukin 2 (IL2), there was a corresponding decrease in the prevalence of myopic refractive errors by 0.235 (95% confidence interval [CI], 0.050–0.419 for fixed effects; 95% CI, 0.125–0.345 for random effects) and 0.132 (95% CI, 0.032–0.231 for fixed effects; 95% CI, 0.044–0.220 for random effects), respectively. No substantial causal associations were observed for IL1α, IL1β, IL12p70, or monocyte chemoattractant protein 1 (MCP1) with refractive errors. Conversely, reverse MR analyses failed to indicate a causal influence of refractive errors on IL1RA and IL2.

Conclusions: The present study offers evidence for a causal link between inflammatory cytokines and refractive errors, which could have significant implications for the early detection, surveillance, and management of refractive errors.

Translational Relevance: Our study underscores the importance of IL1RA and IL2 in the prevention and management of refractive errors, suggesting the feasibility of strategies for early identification, continuous surveillance, and the deployment of focused therapeutic approaches.

Introduction

Refractive errors occur when there is an imbalance between the axial length of the eye and its optical power, resulting in blurred vision. They are the second leading cause of blindness worldwide.¹ The incidence of refractive errors, particularly myopia, has seen a stark increase globally over the past six decades.² Countries in Southeast Asia, such as Singapore, China, and South Korea, have seen myopia prevalence rates as high as 70% to 90%.^3–5 It is estimated that, by 2050, 50% of the global population will have myopia (with a prescription greater than −0.50 diopters [D]), and approximately 20% will have high myopia (with a prescription greater than −5.00 D).⁶ Furthermore, high myopia is strongly associated with an increased risk of several ocular conditions, including retinal detachment, glaucoma, cataract, macular retinoschisis, myopia-related maculopathy, and choroidal neovascularization.⁷ 

It is well-established that both genetic and environmental factors contribute to the development of refractive errors.¹ Recent research indicates that inflammatory cytokines play a significant role in ocular growth and the progression of refractive errors.^3,8–10 The complement system, a crucial component of both adaptive and innate immune responses, has been proposed to be involved in the progression of myopia.^8,10 Increased inflammation may affect myopia progression through scleral remodeling, and vice versa, and the progression of myopia might also influence inflammation.¹¹ A study from Korea reported an association between elevated white blood cell counts and an increased prevalence of myopia.³ Additionally, a detailed analysis of the relationships among oxidative stress, inflammation, and metabolic changes in high myopia was conducted using human serum metabolomics.⁹ It was found that interleukin 1 receptor antagonist (IL1RA) levels were significantly lower and monocyte chemoattractant protein 1 (MCP1) levels were significantly higher in high-myopia cataract patients compared to age-related cataract patients.¹² Furthermore, injections of IL1RA were shown to inhibit myofibroblast formation in a rat model of healing by reducing the expression of proinflammatory cytokines, such as IL1α (IL1A), IL1β (IL1B), IL2, and IL12p70.¹² Despite the recognition of refractive errors as an inflammation-related condition, establishing a causal relationship between refractive errors and inflammatory cytokines remains challenging. 

Mendelian randomization (MR) is a genetic epidemiological technique that utilizes single nucleotide polymorphisms (SNPs) as instrumental variables (IVs). It relies on the principle of randomly and independently distributed SNPs during meiosis, which helps to overcome the methodological limitations of traditional studies, such as reverse causation and confounding.¹³ Genome-wide association studies (GWASs) have identified numerous SNPs associated with complex diseases, thereby advancing the field of MR.¹⁴ Building on this knowledge, we conducted a bidirectional MR analysis using recent large-scale GWASs to investigate the causal relationship between refractive errors and inflammatory cytokines. 

Materials and Methods

Study Design

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization (STROBE-MR) guidelines.¹⁵ It was designed as a bidirectional two-sample MR study, and the hypothesis landscape and schematic design are presented in Figure 1. To ensure the validity of causal inferences from MR analyses, the instrumental variables (SNPs) used had to meet three critical assumptions: (1) that the genetic variants have a direct effect on the exposure; (2) that these variants are only associated with the outcome through their influence on the exposure; and (3) that this relationship is consistent across different potential confounders.¹⁶ The study was analyzed in two directions, with inflammatory cytokines serving as both exposure and outcome. We investigated whether individuals with higher inflammatory cytokine levels were more prone to myopia and whether myopic individuals exhibited higher levels of inflammatory cytokines. All datasets used in this investigation were derived from publicly available GWAS summary statistics. Therefore, ethical approvals and informed consents were obtained in the original studies from which these datasets were obtained. Ethics approval was provided by the ethics committee of the Fourth Affiliated Hospital of Soochow University (220152). 

Figure 1.

 

Landscape of this bidirectional Mendelian randomization study design.

Data Sources and Genetic Instruments Selection

Summary statistics for six inflammatory cytokines (IL1A, IL1B, IL1RA, IL2, IL12p70, and MCP1) were obtained from a published GWAS meta-analysis using data from the Northern Finland Birth Cohort 1966 (NFBC1966), Cardiovascular Risk in Young Finns (YFS) study, and the FINRISK1997 and FINRISK2002 studies (Table 1).^17–20 The GWASs were conducted on 47 cytokines in up to 13,365 individuals available in at least one of three Finnish cohorts: NFBC1966, YFS, or FINRISK.¹⁷ 

Table 1.

 

Inflammatory Cytokines

 

The genetic data of the outcome refractive errors (spherical equivalent) were obtained from a large GWAS meta-analysis.²¹ This included GWASs of directly measured spherical equivalent from the UK Biobank, Genetic and Environmental Risk Assessment (GERA), and Consortium for Refractive Error and Myopia (CREAM) studies.^22–24 The necessary data for MR analysis were generated, including beta effect and standard error (SE). Spherical equivalent was calculated as the sphere + (cylinder/2). The negative value of the spherical equivalent for myopia necessitates a reversal in the interpretation of the β coefficient in MR results. 

For the selection of genetic IVs, a genome-wide significance threshold of P < 5 × 10⁻⁶ was employed. Moreover, the IVs were chosen based on the following criteria: (1) To fulfill the assumptions of MR, we conducted a linkage disequilibrium analysis (R² < 0.001; clumping distance = 10,000 kb) utilizing European-based 1000 Genome Projects and excluded the SNPs that did not meet the requirements. (2) When harmonizing exposure and outcome data, palindromic SNPs with intermediate allele frequencies were removed. F-statistics were calculated as (β/SE)2 to evaluate weak instrument strength, with a minimum value of 10 considered sufficient for MR analysis.²⁵ 

Statistical Analysis

The primary analysis employed both the fixed effects and random effects inverse-variance weighted (IVW) approaches to derive causal estimates, considering the heterogeneity among IVs and improving the reliability and robustness of MR analyses. Moreover, supplementary approaches including MR-Egger, weighted median, and weighted mode were also applied to investigate causal relationships. The MR-Egger regression was used to detect violations of standard IV assumptions and to provide an effect estimate that is robust to such breaches.²⁶ The weighted median estimator remains consistent even when more than 50% of the data are derived from invalid IVs, exhibiting superior finite-sample Type I error rates compared to the IVW approach.²⁷ Although the power of the weighted mode in exploring a causal effect was lower than that of the weighted median and IVW approaches, it surpassed that of MR-Egger regression.²⁸ These approaches serve as sensitivity analyses and confirm the robustness of MR results. Cochran's Q test was performed to evaluate the heterogeneity of different IVs.²⁹ By sequentially excluding each SNP one at a time, leave-one-out analysis was conducted to assess the potential influence of each IV on the overall results. For evaluating horizontal pleiotropy, MR-Egger regression tests were performed.³⁰ If the MR-Egger intercept was close to zero, with P > 0.05, then the MR-Egger regression model was very close to the IVW, indicating high estimation accuracy. The Mendelian Randomization Pleiotropy Residual Sum and Outlier (MR-PRESSO) method was utilized to detect and reduce horizontal pleiotropy in IVW linear regression through excluding SNPs related to heterogeneity.³¹ All statistical analyses were performed with R 4.3.3 (R Foundation for Statistical Computing, Vienna, Austria). A significance level of less than 0.05 (two-sided) was considered statistically significant. 

Results

None of the SNPs that served as IVs for genetically predicted inflammatory cytokines were in linkage disequilibrium with the known refractive errors risk loci (R² < 0.1). The F statistics for the genetic instruments were all above 10 (range, 20.87–561.51), indicating that there was no weak instrument bias in this MR study.^32,33 

Causal Effect of Inflammatory Cytokines on Refractive Errors

The two-sample MR analysis to explore the causal effect of inflammatory cytokines on refractive errors is presented in Figure 2. The analysis suggests that refractive errors were negatively correlated with the level of IL1RA and IL2 based on the IVW method at the threshold of a false discovery rate < 0.05. In detail, each unit increase in IL1RA was linked to a decreased level of myopic refractive errors of 0.235 (fixed effects: 95% confidence interval [CI], 0.050–0.419; random effects: 95% CI, 0.125–0.345). Each unit increase in IL2 was linked to a decreased level of myopic refractive errors of 0.132 (fixed effects: 95% CI, 0.032–0.231; random effects: 95% CI, 0.044–0.220). The absence of horizontal pleiotropy was validated in IL1RA (P = 0.491) and IL2 (P = 0.929) (Fig. 2). In addition, the leave-one-out analysis further confirmed data stability in IL1RA and IL2 (Supplementary Figs. S1, S2). 

Figure 2.

 

Total causal effect of inflammatory cytokines on refractive errors under Mendelian randomization.

Based on the IVW results, no significant causal effect of IL1A, IL1B, IL12p70, or MCP1 on refractive errors was found (Fig. 2). Nevertheless, each unit increase in IL1A was linked to a decreased level of myopic refractive errors of 0.103 (fixed effects: 95% CI, −0.002 to 0.208; random effects: 95% CI, −0.085 to 0.292) (Fig. 2, Supplementary Fig. S3). Each unit increase in IL1B was linked to an increased level of myopic refractive errors of 0.011 (fixed effects: 95% CI, −0.118 to 0.096; random effects: 95% CI, −0.234 to 0.211) (Fig. 2, Supplementary Fig. S4). Each unit increase in IL12p70 was linked to a decreased level of myopic refractive errors of 0.012 (fixed effects: 95% CI, −0.055 to 0.079; random effects: 95% CI, −0.059 to 0.083) (Fig. 2, Supplementary Fig. S5). Each unit increase in MCP1 was linked to a decreased level of myopic refractive errors of 0.123 (fixed effects: 95% CI, 0.024–0.222; random effects: 95% CI, −0.016 to 0.262) (Fig. 2, Supplementary Fig. S6). Although these associations did not reach statistical significance, similar results were also found for the MR-Egger, weighted median, or weighted mode estimates (all P > 0.05). 

Causal Effect of Refractive Errors on Inflammatory Cytokines

With genetic liability for refractive errors as the exposure, the results of reverse MR analyses showed no evidence of the causal effect of refractive errors on IL1RA by the IVW (β, −0.007; fixed effects: 95% CI, −0.022 to 0.008, P = 0.379; random effects: 95% CI, −0.023 to 0.010, P = 0.409) and IL2 by the IVW (β, −0.007; fixed effects: 95% CI, −0.031 to 0.017, P = 0.547; random effects: 95% CI, −0.031 to 0.017, P = 0.548). 

Sensitivity Analyses

The pleiotropy and heterogeneity examinations for the inflammatory cytokines are shown in Table 2. There was no evidence of horizontal pleiotropy in MR-Egger regression (all P > 0.05 for intercept). In addition, no heterogeneity was found in any of our exposures except IL1A (IVW Cochran's Q = 29.143, P = 0.001; MR-Egger Cochran's Q = 27.915, P < 0.001), IL1B (IVW Cochran's Q = 21.557, P = 0.001; MR-Egger Cochran's Q = 21.080, P < 0.001), and MCP1 (IVW Cochran's Q = 29.626, P = 0.013; MR-Egger Cochran's Q = 29.625, P = 0.009). Considering the absence of heterogeneity and pleiotropy in IL1RA and IL2, the IVW results were reliable. As a result, our MR analysis suggests that IL1RA and IL2 may be causally related to refractive errors, although further research is necessary to confirm these findings. 

Table 2.

 

Results for Cochran's Tests for Heterogeneity and MR-Egger Intercept Tests for Horizontal Pleiotropy for the Association Between Inflammatory Cytokines and Refractive Errors

 

Discussion

In this bidirectional two-sample MR study, we established that both IL1RA and IL2 exert a negative causal influence on the onset of myopia, suggesting a protective role for higher levels of these cytokines against the risk of myopic refractive errors. Converse MR analyses failed to detect any significant causal effect of refractive errors on IL1RA and IL2, as evidenced by the IVW approach and combined analyses. To the best of our knowledge, this is the first MR study to uncover a causal relationship among IL1RA, IL2, and refractive errors. 

Previous research has hinted at a link between inflammatory cytokines and refractive errors, yet the exact mechanisms remain elusive.³⁴ Children suffering from inflammatory conditions, such as allergic conjunctivitis, uveitis, and systemic lupus erythematosus, have a greater prevalence of myopia than those without such conditions.^35,36 Inflammation-related factors such as tumor necrosis factor-α (TNF-α), IL6, IL8, MCP1, and nuclear factor kappa B (NK-κB) are often elevated in patients with allergic conjunctivitis, whereas IL10 and the inhibitor of kappa B are reduced.³⁵ Additionally, the levels of inflammatory cytokines in the aqueous humor have been associated with myopia development.³⁷ Patients with high myopia exhibit a distinct inflammatory and fibrogenic profile compared to those with normal vision.³⁸ The complement decay-accelerating factor (CD55) has been shown to inhibit myopia progression by suppressing complement activation and reducing inflammation.⁸ Inflammatory cytokines, such as NK-κB, IL6, and IL8, may contribute to the chronic inflammatory state in myopia, and anti-inflammatory agents may contribute to managing inflammation and slowing myopia progression.³⁹ Despite these findings, the causal relationship between inflammatory cytokines and refractive errors has yet to be definitively established, in part due to the small sample sizes typical of such studies. To address this gap, we conducted bidirectional two-sample MR analyses and found that IL1RA and IL2 are causally associated with myopia. 

IL1RA, a potent inhibitor of IL1 receptors, plays a crucial role in modulating inflammation.⁴⁰ Animal models have indicated that IL1RA can suppress inflammation in corneal grafts, with higher levels of IL1RA correlating with a reduced inflammatory response.⁴¹ Our research extends these findings, revealing a significant negative association between IL1RA levels and the severity of myopic refractive errors; an increase in IL1RA was associated with a 0.235 reduction in the level of myopia. Furthermore, higher levels of IL1RA were associated with less pronounced myopia. These findings align with prior studies and provide additional evidence of a link between IL1RA and refractive errors. For example, Wang and colleagues⁴² observed significantly lower levels of IL1RA in the aqueous humor of patients with high myopia compared to controls. The expression of IL1RA in aqueous humor was also lower in high-myopia cataract patients than in a control group.¹² When IL1RA levels are reduced in these high-myopia patients, the balance between pro-inflammatory and anti-inflammatory signals is disrupted, leading to an exaggerated inflammatory response. This pro-inflammatory environment may contribute to the pathological changes associated with myopia, including scleral remodeling, extracellular matrix degradation, and excessive axial elongation.¹¹ Therefore, the downregulation of IL1RA might indicate a pro-inflammatory state that could play a significant role in promoting the onset and progression of myopia. 

IL2 is a powerful immune-modulating cytokine that is essential for supporting the survival and growth of regulatory T (Treg) cells. These cells are pivotal in maintaining peripheral tolerance and in regulating ongoing inflammation and preventing autoimmunity.⁴³ A relative deficiency in IL2 can lead to imbalances in Treg cell homeostasis, potentially worsening the cycle of tolerance breakdown and chronic inflammation seen in certain autoimmune conditions.⁴⁴ Low-dose IL2 therapy has emerged as a potential treatment approach and an alternative to anti-IL2 agents in managing autoimmune diseases.⁴⁵ A significant negative correlation between IL2 concentration and intraocular pressure has been reported in primary open-angle glaucoma patients.⁴⁶ This finding suggests that IL2 may play a role in regulating intraocular pressure, potentially through its immunomodulatory effects. Interestingly, a strong bidirectional genetic causal link has been identified between myopia and primary open-angle glaucoma, with intraocular pressure serving as a key mediating factor.⁴⁷ These findings suggest that IL2 may influence myopia development, at least in part, by modulating intraocular pressure. Our research suggests that an increase in IL2 levels is associated with a reduction in the severity of myopic refractive errors, indicating a protective role for IL2 in modulating the immune response in myopia. 

IL1RA and IL2 play a significant role in immune regulation, exerting their effects through distinct mechanisms and pathways; however, in certain circumstances, there may be potential interactions and regulatory relationships between them. These interactions are crucial for maintaining the balance and function of the immune system and have the potential to influence the onset and progression of myopia.⁴⁸ Consequently, they have potential applications in the prevention and treatment of myopia. IL1RA has been attributed to the immunosuppressive effects of mesenchymal stem cells, which promote expansion of immunosuppressive Treg cells and attenuate the generation of inflammatory Th1 and Th17 cells.⁴⁹ IL2 is also an immune regulatory cytokine that modulates the proliferation and differentiation of T cells. A negative correlation between the level of several inflammatory cytokines in the aqueous humor and axial length was reported by Zhang et al.⁴⁸ This result somewhat helps to explain our findings. Given these findings, downregulation of IL1RA and IL2 may serve as a biomarker for identifying individuals at risk of developing myopia or progressing to high myopia. Moreover, targeting IL1RA and IL2 pathways through therapeutic interventions could offer novel strategies for preventing or slowing myopia progression. We proposed the hypothesis that an increase in IL1RA and IL2 could postpone the development of axial length through reducing ocular inflammation, decreasing extracellular matrix degradation and slowing down the excessive growth of the eye. Further research should focus on elucidating the precise mechanisms by which IL1RA and IL2 modulate ocular inflammation and growth and exploring the potential of IL1RA-based and IL2-based therapies in clinical settings. 

The concentrations of these inflammatory cytokines may have predictive value for changes in refractive errors and possibly provide a useful prognostic modality. In addition to the six inflammatory cytokines explored in this study, some other inflammatory cytokines, including NF-κB, TGF-β, IL6, IL8, and TNF-α, have also been reported to be associated with the chronic inflammatory state observed in myopia, indicating a profound connection between the onset of myopia, inflammatory processes, and fibrosis.³⁹ Building on these findings, clinical intervention strategies such as anti-inflammatory treatments and immunomodulatory approaches can be considered. The efficacy and safety of anti-inflammatory drugs or immunomodulatory agents in delaying or preventing the progression of myopia should be investigated based on IL1RA and IL2, which could potentially lead to the development of new therapeutic options for refractive errors. Similar to many complex diseases, myopia is attributable to the interaction of genetic and environmental factors that results in excessive axial eye growth.⁵⁰ Although our study primarily focused on the association between inflammatory cytokines and myopia, it is important to acknowledge the significant role of environmental factors in myopia development. Environmental influences, such as near-work activities, limited outdoor exposure, and prolonged screen time, have been consistently linked to an increased risk of myopia onset and progression.⁵¹ These factors are thought to contribute to myopia, potentially promoting a pro-inflammatory microenvironment and exacerbating inflammatory responses within ocular tissues. 

The primary strengths of this study lie in its application of a bidirectional two-sample MR design and the utilization of extensive genotype and phenotype data from large-scale GWAS meta-analyses. These methodological choices were instrumental in enhancing the validity and credibility of our findings. However, several significant limitations must be acknowledged. First, the datasets used in this study are limited to individuals of European ancestry, a limitation arising from the lack of GWAS summary statistics for other ancestral backgrounds. Therefore, it is crucial to acknowledge that the generalizability of the findings from this study to individuals of non-European ancestry or other racial and ethnic groups may be relatively poor. Further research involving more diverse datasets and incorporating detailed genetic and demographic information is necessary to enhance the applicability and relevance of the conclusions of this study across a broader range of racial and ethnic backgrounds. Although this selection may reduce the potential for population stratification bias, it also compromises the generalizability of our results to diverse populations. Second, it is crucial to acknowledge that the genetic instruments used in MR analyses represent average effects across an individual's lifespan. As such, the true complexity and diversity of the biological interactions between inflammatory cytokines and refractive errors may not be fully captured by our research. Future randomized controlled trials should be designed to explore the intricacies of the relationship between inflammatory cytokines and refractive errors in a more detailed and nuanced manner. Furthermore, future studies should delve into other potential influencing factors that may contribute to the development of myopia, such as environmental factors and lifestyle choices. Exploring the role of environmental elements, such as lighting conditions, outdoor activity levels, and exposure to digital screens, can provide valuable insights into the complex interplay between the environment and the genetic predispositions that lead to myopia. Converse MR analyses failed to detect any significant causal effect of refractive errors on IL1RA and IL2 in this study, but the possible influence of myopia on inflammation cytokines should also be further analyzed in future studies. 

Conclusions

Our study provides strong evidence for a causal relationship between inflammatory cytokines and the progression of refractive errors, as IL1RA and IL2 were shown to offer protection against the development of myopia. These findings underscore the importance of IL1RA and IL2 in the prevention and management of refractive errors, suggesting the feasibility of strategies for early identification, continuous surveillance, and deployment of focused therapeutic approaches. To further clarify the complexities of this connection, there is an urgent need for continued investigation into the detailed mechanisms. 

Acknowledgments

Supported by grants from the National Key R&D Program of China (2024YFC2510800, 2024YFC2510801) and the Suzhou Medical Innovation Application Research Project (SZM2023027). 

Author Contributions: The work presented here was carried out in collaboration among all authors. PCW, and ZXF defined the research theme, discussed analyses, interpretation, and presentation. XY and DXX drafted the manuscript, analyzed the data, developed the algorithm and interpreted the results. WY, ZXY, and CYJ helped to perform the statistical analysis, figure interpretation and reference collection. All authors read and approved the final manuscript. 

Disclosure: Y. Xu, None; X.-X. Dong, None; Y. Wang, None; X.-Y. Zhuang, None; Y.-J. Chen, None; X.-F. Zhang, None; C.-W. Pan, None