Mendelian randomization supports causal effects of inflammatory biomarkers on myopic refractive errors

Abstract

Aims

To determine whether inflammatory biomarkers are causal risk factors for more myopic refractive errors.

Methods

Northern Sweden Population Health Study (NSPHS), providing inflammatory biomarkers data; UK Biobank, providing refractive errors data. 95,619 European men and women aged 40 to 69 years with available information of refractive errors and inflammatory biomakers. Inflammatory biomarkers including ADA, CCL23, CCL25, CD6, CD40, CDCP-1, CST5, CXCL-5, CXCL-6, CXCL-10, IL-10RB, IL-12B, IL-15RA, IL-18R1, MCP-2, MMP-1, TGF-β1, TNF-β, TWEAK and VEGF-A were exposures, and spherical equivalent (SE) using the formula SE = sphere + (cylinder/2) was outcome.

Results

Mendelian randomization analyses showed that each unit increase in VEGF-A, CD6, MCP-2 were causally related to a more myopic refractive errors of 0.040 D/pg.mL-1 (95% confidence interval 0.019 to 0.062; P = 2.031 × 10-4), 0.042 D/pg.mL-1 (0.027 to 0.057; P = 7.361 × 10-8) and 0.016 D/pg.mL-1 (0.004 to 0.028; P = 0.009), and each unit increase in TWEAK was causally related to a less myopic refractive errors of 0.104 D/pg.mL-1 (−0.152 to −0.055; P = 2.878 × 10-5). Tested by the MR-Egger, weighted median, MR-PRESSO, Leave-one-out methods, our results were robust to horizontal pleiotropy and heterogeneity in VEGF-A, MCP-2, CD6, but not in TWEAK.

Conclusions

Our Mendelian Randomization analysis supported the causal effects of VEGF-A, MCP-2, CD6 and TWEAK on myopic refractive errors. These findings are important for providing new indicators for early intervention of myopia to make myopic eyesight threatening consequences less inevitable.

Keywords

Two-sample mendelian randomization refractive errors inflammatory biomarkers myopia early prediction

Introduction

Refractive errors, mainly myopia, are not only the most frequent eye disorders worldwide,^1,2 but also the third global leading cause of blindness in those aged 50 and older in 2020.³ It's estimated that 49.8% of the world population will have myopia and 10% will suffer from high myopia by 2050.⁴ High myopia increases the risk of pathologic ocular changes such as myopic retinopathy, glaucoma, and retinal detachment, all of which can cause irreversible visual impairment.⁵

Over the past decades, many studies suggest that myopic refractive errors have a complex etiology with multiple factors involved, such as environmental exposures and genetics exposures.⁶ Although the exact mechanism of myopic refractive errors is unclear, a growing body of evidence supports that inflammatory biomarkers play crucial roles in ocular growth and refractive development.^7–9 It has been reported that patients with autoimmune diseases such as type 1 diabetes mellitus, systemic lupus erytheomatosus, uveitis,⁸ and allergic conjunctivitis had higher risk of myopia compared with those without.⁹ Interferon γ, IL-6, IFN-γ-induced protein 10 (IP-10), eotaxin, and macrophage inflammatory protein 1α (MIP-1α), were significantly higher in the high myopic group than in the control group.⁹ In high myopes, the inflammatory biomarkers (IL-6, IL-10 and MCP-1),⁷ neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio were upregulated,¹⁰ and the eyes with myopic choroidal neovascularization showed significantly higher VEGF than those without.⁷ long et al. found that the levels of C3 and CH50 were higher in patients with pathologic myopia, suggesting that inflammation induced by complement activation may play an important role in myopia.¹¹ However, it is challenging to identify the role of inflammatory markers in the pathogenesis of myopic refractive errors because of the potential bias due to reverse causation, which refers to the possibility of refractive errors being the cause rather than the consequence of inflammation.

Mendelian randomization (MR) is a method that uses large cross-sectional data combined with genetic information to assess the causal association between exposure and outcome, which can overcome the typical traps of confounders and reverse causation in observational analyses. MR uses SNPs (single nucleotide polymorphism, SNP) strongly associated with exposures as instrumental variables (IVs) to explore the association between exposures and outcomes. Since human genes are rarely disturbed by acquired confounders and the distribution of genes is naturally randomized, MR has strong ability in determing causal associations.

Here, a two-sample Mendelian randomization analysis was performed to investigate the causal link between inflammatory biomarkers and myopic refractive errors.

Methods

Definitions

Inflammatory biomarkers - We included 21 inflammatory biomarkers from a published GWAS using data of Northern Sweden Population Health Study (NHPHS), which was launched in 2006 in Karesuando county of Norrbotten, Sweden with 1,069 participants, with 892 usable samples (Table 1).^12–15

Table 1.

Detailed information of inflammatory biomarkers.

Biomarkers abbreviation	Biomarkers full name	N	Cohort	No. of genetic instruments
ADA	Adenosine deaminase	871	NSPHS	7
CCL-23	Chemokine ligand 23	871	NSPHS	13
CCL-25	Chemokine ligand 25	871	NSPHS	26
CD6	Cluster of differentiation 6	871	NSPHS	13
CD40	Cluster of differentiation 40	871	NSPHS	22
CDCP-1	CUB-domain containing protein 1	871	NSPHS	8
CST-5	Cystatin 5	549	NSPHS	13
CXCL-5	C-X-C motif ligand 5	871	NSPHS	7
CXCL-6	C-X-C motif ligand 6	871	NSPHS	8
CXCL-10	C-X-C motif ligand 10	871	NSPHS	21
IL-10RB	Interleukin-10 receptor B	871	NSPHS	8
IL-12B	Interleukin-12 subunit β	871	NSPHS	22
IL-15RA	Interleukin-15 receptor α	741	NSPHS	9
IL-18R1	Interleukin-18 receptor	871	NSPHS	93
MCP-2	Monocyte chemoattractant protein 2	871	NSPHS	36
MMP-1	Matrix metallopeptidase 1	720	NSPHS	10
MMP-10	Matrix metallopeptidase 10	871	NSPHS	7
TGF-β1	Transforming growth factor β 1	871	NSPHS	5
TNF-β	Tumor necrosis factor–β	871	NSPHS	6
TWEAK	TNF ligand superfamily member 12	871	NSPHS	4
VEGF-A	Vascular endothelial growth factor A	832	NSPHS	5

Refractive errors (RE) - The genetic data of outcome RE were obtained from the published GWAS,¹⁶ including 95,619 European participants from UK Biobank. The exclusion criteria was participants with conditons that could alter refraction (1) cataract surgery; (2) retinal detachment surgery; (3) laser refractive procedures; (4) keratoconus; (5) ocular or systemic syndromes. Spherical equivalent was assessed as the sphere + (cylinder/2). All participants were measured by non-cycloplegic autorefraction. The value of the spherical equivalent of myopic refractive error is negative, that means in MR results, the interpretation of the β coefficient needs to be interpreted in reverse.

Genotype data

We selected SNPs from the original GWAS article, and the specific selection criteria were based on the reference of this original article.¹⁵ A series of quality control steps were used in the original article to select the instrument SNPs. Although this sample bank has a small sample size due to its expensive cost on self-sequencing, it has the advantages of more accurate measurement and more types of biomarker samples.

The P-value threshold of P = 10⁻⁶ was selected in the original article, and the linkage disequilibrium (LD) threshold was R²= 0.6, which was also followed in this study. The reason for choosing a P-value threshold of P = 10⁻⁶ is that approximately 0.1% of the genome was analyzed for each of the 85 proteins in the original study, so the P-threshold should be at the same level as the multiple test adjustment used in the GWAS. LD threshold of R²= 0.6 was chosen to maintain a meaningful number of genetic instruments and to increase the ability to detect causal effects for as many proteins as possible.¹⁷ After removing data for the above selected exposure SNPs from the outcome trait, we conducted F-test to test whether instrumental SNPs had a weak association with exposure, using formula “F = β²/se²”. If F is far greater than 10, the possibility of weak instrumental variable association is small.

Statistical methods

Two-Sample mr analysis

We performed two-sample MR analyses to evaluate the impact of inflammatory biomarkers on myopia in Europe populations. All analyses were conducted using R software (version 3.4.4) with the TwoSampleMR package.

IVW analysis is the most efficient method in MR with greatest statistic power. It combines Wald estimates for each SNP to get the overall estimates of the correlation of causal effects. P < 0.05/N was considered statistically significant (N = number of exposures * number of outcome), in our study that was P < 0.05/21 = 0.002. In the hypothesis of IVW, we believed that these SNPs are not pleiotropic and considered that the results of GWAS are mostly phenotype standardized, which required us to continue with the pleiotropy and sensitivity analysis.

Pleiotropy and sensitivity analysis

MR-Egger is the method used to assess potential heterogeneity. In the MR-Egger hypothesis, we used intercepts to assess horizontal pleiotropy.^18–20 If the intercept term in MR Egger regression is close to 0, with P > 0.05 then the MR-Egger regression model is very close to the IVW, the estimation accuracy is higher. If the intercept term differs greatly with 0, it indicates that there may be horizontal pleiotropy between these IVs. The weighted median(WM) method can obtain estimates consistent with the final effect when 50% of the IVs are valid instrumental variables, reducing the bias in causality judgments when some IVs are not consistent with the assumptions of the instrumental variables. MR-PRESSO is a method to detect and reduce horizontal pleiotropy in IVW linear regression through removing SNPs related to heterogeneity.²¹ The number of distributions in MR-PRESSO analysis was set to 1000. If the MR-PRESSO analysis detected a significant horizontal pleiotropy, we shall remove the outlier variants and perform MR analysis again. After removing outlier, if heterogeneity was still significant, we would remove all SNPs with P values less than 1 in the MR-PRESSO outlier test and perform MR analysis again.²² MR.RAPS use robust adjusted profile score to correct pleiotropy, with strong robustness in measuring systematic and heterogeneous pleiotropy, and can measure MR outcomes for weak instrumental variables.

At last, we evaluated the heterogeneity for IVs using the Cochran's Q test (P < 0.05 shows heterogeneity) and removed the SNP one by one continuously, while using the remaining SNP continue to do MR analysis, which is “leave one out”. If “leave one out” show existence of SNP with possible potential effects, that means there is reliance of MR on a single SNP.

Patient involvement

No patients were involved in the whole process of data collection and analysis. All our analysis data were derived from publicly published GWAS articles, each included article was approved by the ethical review committees, no registration or protocol was required, and written informed consent was obtained from all participants.

Availability of data and materials

Genome-wide association study data sources of inflammatory biomarkers are publicly available in the original GWAS article, and data on refractive errors can be found in https://www.ebi.ac.uk/gwas/studies/GCST009521. Analysis scripts can be found on https://mrcieu.github.io/TwoSampleMR/.

Results

Two-sample Mendelian randomization analysis for causal link of inflammatory biomarkers with refractive errors

Figure 1 demonstrated that refractive errors was positively correlated with the level of VEGF-A and CD6 at the threthold of P < 0.002. Each unit increase in VEGF-A, CD6 and MCP-2 resulted in a more myopic refractive errors of 0.040 (95% confidence interval 0.019 to 0.062; P = 2.031E-04), 0.042 D/pg.mL⁻¹ (0.027 to 0.057; P = 7.361E-08) and 0.016 D/pg.mL⁻¹ (0.004 to 0.028; P = 0.009). Meanwhile, CCL-25, CST-5, IL-18R1 showed slightly positive correlation with refractive errors at the threshold of P < 0.5.

Refractive errors was negatively associated with the level of TWEAK at the threthold of P < 0.002, and was slightly associated with the level of IL-10RB at the threshold of P < 0.05. Each unit increase in TWEAK was associated with a 0.104 D/pg.mL⁻¹ (−0.152 to −0.055; P = 2.878E-05) reduction in myopic refractive errors, more detailed results are shown in Figure 1.

Sensitivity analyses: results of Mendelian randomisation are robust to potential bias

Table 2 shows the pleiotropy and heterogeneity test results for all inflammatory biomarkers. In MR-Egger, no horizontal pleiotropy was observed (Table 2). Meanwhile, no heterogeneity was found in all of our exposures except for IL-12B (P_IVW= 0.039, P_{MR Egger}= 0.035) and CD40 (P_IVW= 0.038, P_{MR Egger}= 0.034). However, the β of TWEAK in MR Egger was in the opposite direction of the β in the other methods. This means the result of TWEAK was not robust. The estimated effect sizes of the SNPs on both the exposures and refractive errors outcome are displayed in scatter plots (Figure 2).

Table 2.

Results for Cochran's Q tests for heterogrneity and MR egger intercept tests for horizontal pleiotropy for the association between inflammatory biomarkers and refractive errors.

Exposure	IVW		MR Egger
	Cochran's Q(df)	p-value	Cochran's Q(df)	p-value	MR EggerIntercept(SE)	Interceptp-value
IL-10RB	3.155 (7)	0.866	2.523 (6)	0.866	−0.012 (0.015)	0.457
IL-12B	32.432 (20)	0 . 039	31.554 (19)	0.035	−0.011 (0.016)	0.476
IL-15RA	10.936 (8)	0.205	10.537 (7)	0.160	0.016 (0.032)	0.623
IL-18R1	64.865 (91)	0.983	63.631 (90)	0.984	−0.004 (0.004)	0.269
VEGF-A	1.806 (4)	0.771	1.241 (3)	0.743	−0.015 (0.019)	0.507
TWEAK	3.838 (3)	0.280	3.409 (2)	0.182	−0.043 (0.085)	0.666
CCL-25	23.596 (23)	0.426	22.675 (22)	0.420	−0.009 (0.009)	0.355
CST-5	7.316 (11)	0.773	6.527 (10)	0.769	−0.015 (0.016)	0.395
MMP-1	6.408 (9)	0.699	4.557 (8)	0.804	−0.027 (0.020)	0.211
MMP-10	7.760 (6)	0.256	6.809 (5)	0.235	−0.018 (0.021)	0.441
CDCP-1	10.517 (7)	0.161	7.296 (6)	0.294	−0.058 (0.035)	0.155
CD6	13.569 (11)	0.258	13.265 (10)	0.209	−0.009 (0.019)	0.643
CD40	32.507 (20)	0.038	31.679 (19)	0.034	0.008 (0.011)	0.489
CXCL-5	2.514 (6)	0.867	2.435 (5)	0.786	0.005 (0.018)	0.790
CXCL-6	1.075 (6)	0.983	1.001 (5)	0.962	−0.004 (0.016)	0.796
CXCL-10	9.372 (18)	0.950	8.403 (17)	0.957	−0.011 (0.011)	0.339
MCP-2	45.336 (35)	0.113	45.099 (34)	0.097	0.003 (0.008)	0.675
ADA	8.372 (6)	0.212	8.372 (5)	0.137	−4.537E-04 (0.022)	0.984
CCL-23	6.774 (11)	0.817	6.732 (10)	0.750	0.003 (0.016)	0.841
TGF-β1	4.244 (3)	0.236	2.952 (2)	0.229	0.034 (0.036)	0.448
TNF-β	8.663 (4)	0.070	7.793 (3)	0.050	−0.011 (0.018)	0.603

The absence of horizontal pleiotropy was confirmed in VEGF-A (P_MR−PRESSO= 0.782), TWEAK (P_MR−PRESSO= 0.405), CD6 (P_MR−PRESSO= 0.343) and MCP-2 (P_MR−PRESSO= 0.121) (Table 3). Moreover, the leave-one-out results further validated data stability. In the absence of heterogeneity and pleiotropy, we may conclude the results of IVW were reliable. Therefore, VEGF-A, MCP-2, CD6 and TWEAK were causally related to refractive errors.

Table 3.

Results of MR PRESSO analysis for associations between inflammatory biomarkers and refractive errors.

Outcome	Exposure (Inflammatory biomarkers)	MR-presso
Outcome	Exposure (Inflammatory biomarkers)	β (Se)	p- value
Refractive Errors	IL-10RB	−0.023 (0.007)	0.907
Refractive Errors	IL-12B	0.010 (0.011)	0.034
Refractive Errors	IL-15RA	0.025 (0.013)	0.243
Refractive Errors	IL-18R1	0.008 (0.003)	0.980
Refractive Errors	VEGF-A	0.040 (0.007)	0.782
Refractive Errors	TWEAK	−0.104 (0.025)	0.405
Refractive Errors	CCL-25	0.016 (0.006)	0.503
Refractive Errors	CST-5	0.019 (0.007)	0.777
Refractive Errors	MMP-1	−0.020 (0.010)	0.753
Refractive Errors	MMP-10	−0.024 (0.019)	0.316
Refractive Errors	CDCP-1	−0.021 (0.018)	0.191
Refractive Errors	CD6	0.042 (0.008)	0.343
Refractive Errors	CD40	−0.015 (0.008)	0.060
Refractive Errors	CXCL-5	−0.002 (0.007)	0.880
Refractive Errors	CXCL-6	−0.001 (0.005)	0.989
Refractive Errors	CXCL-10	−0.007 (0.006)	0.957
Refractive Errors	MCP-2	0.016 (0.006)	0.121
Refractive Errors	ADA	−0.020 (0.012)	0.265
Refractive Errors	CCL-23	0.015 (0.009)	0.836
Refractive Errors	TGF-β1	0.041 (0.028)	0.310
Refractive Errors	TNFβ	−0.020 (0.016)	0.253

Discussion

In this study, we used summary statistics from GWASs to identify the causal relationships between 21 inflammatory biomarkers and myopic refractive errors. We observed some evidence for a detrimental effect of greater levels of VEGF-A, CD6, MCP-2 on the risk of myopic refractive errors. Additionally, greater levels of TWEAK was associated with lower risk of myopic refractive errors. To our knowledge, no published Mendelian randomization study had reported the causal association between inflammation and refractive errors.

Comparison with other studies

Inflammation in the etiology of myopic refractive errors has not yet been fully clarified. Previous studies have shown possible links between inflammation and myopic refractive errors. McBrien et al. reported increased scleral MMP2 activity in tree shrew eyes during myopia development.²³ Yuan et al. used undiluted aqueous humor samples from a senile cataract population and found a strong positive association between eye globe axial length and IL-6, MMP-2.²⁴ Wei et al. found that VEGF, MCP-1, IL-5, IL-6, IP-10, IFN-γ, eotaxin, MIP-1α, IL-4, and G-CSF were significantly higher in the high myopic patients group than control group.²⁵ Rada et al. reported decreased expression of TIMP-2 in the sclera of form-deprived chick eyes.²⁶

Therefore, both clinical and experimental results support that inflammation may contribute to myopia development. Despite the associations between inflammation and refractive errors reported by many of these previous studies, however, they have not shown causality. Also, because in clinical settings, there is often no reasonable reason to extract eye fluid from refractive errors patients for research, the sample size of the current studies on inflammation and myopic refractive errors is small, or most of them are animal models. In our study, we used mendelian randomization analyses to find that VEGF-A, MCP-2, CD6 and TWEAK were causally associated with myopic refractive errors, although the result of TWEAK need to draw with caution.

Among them, the effect of VEGF-A on myopic refractive errors was consistent with published experimental studies. VEGF has long been noted to be lower in myopic eyes.^24,27–30 However, these findings have no consensus on specific mechanism about the connection between low VEGF and high myopia. VEGF is a key regulator of angiogenesis during fetal life. In the normally developing retina, VEGF is released in response to higher oxygen demand in retinal tissue, which leads to the development of blood vessels from the optic nerve to the periphery. Sawada et al.³¹ suggested the low VEGF in myopic eyes is due to less VEGF production by the retina, which may be supported by the finding of Grunwald et al. that repeated intravitreal injection of anti-VEGF drugs promote geographic atrophy of the membrane retina.³² However, Hu et al. suggested that low VEGF is caused by the larger size of the eyeball in myopic eyes and greater effect of dilution.²⁷

Besides, there have also been conflicting experimental results on the association between VEGF and myopic refractive errors. Wei et al. found that the vitreous VEGF levels were higher in the high myopic group than in the control group.²⁵ Although the sample size of this study was small (36 high myopic patients and 42 controls) and there may be confounding factors, it still needs our attention because it may imply that VEGF regulation is different in different parts of the eye. The mechanism for lower VEGF levels in myopic eyes may be more complex than the proposed diluting effect and low retina production, so further research is needed to verify the hypothesis.

While there are no studies, to our knowledge, looking at the role of MCP-2, CD6 and TWEAK on myopic refractive errors, our findings may provide new insight into possible inflammatory mechanism on myopic refractive errors. TWEAK is a cytokine of the TNF superfamily that activates the Fn14 receptor and are abundantly expressed in injured kidneys and heart. TWEAK may regulate cell proliferation, apoptosis, differentiation, fibrosis and inflammation.³³ TWEAK can also inhibit the expression of tissue protective factors such as anti-aging factor Klotho and mitochondrial biogenesis peroxisome proliferator-activated receptor γ coactivator-1α (PGC-1α).³⁴ High levels of TWEAK induce cardiac remodeling, and promote inflammation, fibroblast proliferation. Although it's not clear whether TWEAK works in the eye, these effects of TWEAK may give it a role in scleral remodeling, which may in turn leads to myopic refractive errors.

MCP-2, also known as CCL-8, is a monocyte chemotactic factor and higher levels of this cytokine suggest an early inflammatory response. It can activate many different immune cells including monocytes, T cells, natural killer cells, eosinophils, and basophils.³⁵ Previous studies have shown that retinal degeneration involves neuroinflammation, and the proinflammatory chemokines MCP-1 and MCP-2 were significantly increased in the eyes of patients with retinitis pigmentosa,³⁶ MCP-1 was significantly elevated in the vitreous of high myopia group.²⁵ In our study, although the P value for MCP-2 was greater than 0.05/21 = 0.002, it is also very close, at 0.009. This suggests that MCP-2 is also likely to be a potential protective factor in myopic refractive errors.

CD6 is a highly specific marker of T cells and a risk factor for multiple sclerosis.³⁷ CD6 possesses possibly the longest cytoplasmic tail of the known receptors of leukocytes, containing amongst other signaling motifs nine tyrosine residues. CD6 is being recognized as an important target for treatment of several autoimmune diseases.³⁸ The activation of the complement system, which is strongly regulated in the human body to avoid overstimulation and damage caused by inflammation, may also be involved in the pathogenesis of myopic refractive errors. The complement system is dysregulated in a variety of eye diseases, including glaucoma, autoimmune uveitis, diabetic retinopathy, and age-related macular degeneration.³⁹

Different from previous studies, TNF-a was not significantly associated with myopic refractive errors in our study, but the β effect values showed a same direction of action as previous studies, that is, the expression level of TNF-α was up-regulated in myopic eyes.^8,40 This may be related to inducing retinal inflammation and activating the lens growth pathway.⁴⁰ More studies on the causal relationship between inflammatory biomarkers as IL6, MMP2, IFN-γ, MIP-1α, NFκB and myopic refractive errors with larger sample sizes and more powerful instrumental variables remain to be explored.

Strengths and limitations of this study

The main strength of the present study is the use of two-sample MR which has strong power for testing causal hypotheses in epidemiology, with the ability to overcome the drawbacks of traditional observational epidemiology such as reverse causation and residual confounding. Mendelian randomization uses genotype to group patients—randomly allocated at conception and so it forms a natural randomized controlled trial in which genetic variants are used as instrumental variables for an environmental exposure to make causal inferences about the impact of the exposure on the outcome. Also, our study used GWAS that contains the most types of inflammatory biomarkers known so far for analysis, making the results more comprehensive. Our study is the first known study to explore the causal relationship between ocular inflammatory biomarkers and myopic refractive errors, which may fill a gap in the relevant field.

There are a few limitations of this study. First, the exclusion criteria were not found in the original GWAS of inflammatory biomarkers, and there may be other diseases affecting myopic refractive errors. Second, although steps were taken to minimize the risk of bias from the potential use of invalid instrumental variables, such bias cannot be completely ruled out. Without knowing the function of SNPs used in mendelian randomization, it's possible that some SNPs may influence the outcome through other pathway besides exposure.^21,41 Third, in order to reduce the potential energy loss in the two-sample MR, a larger LD threshold (R²= 0.6) was used in the original paper to increase the genetic instrumental variables for each biomarker, which may raise the possibility that selected SNPs are related to other genetic variables, so that SNPs influence outcomes not just through the SNP-exposure-outcome pathway, but through the other SNP-exposure-outcome pathway, violating the third hypothesis and thus cause bias to results.

Conclusions

In conclusion, this is a study based on genes related to inflammatory factors, providing strong evidence for subsequent screening of inflammatory factors and their related genes to achieve early diagnosis of myopia. If the persistently rises of inflammation can have a protective or harmful effect on myopic refractive errors, then the early observation of specific inflammatory biomarkers can help with early prediction of myopic refractive errors and the use of drugs that inhibit inflammation may alleviate myopic refractive errors.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Beijing Science Foundation for Distinguished Young Scholars, National Natural Science Foundation of China, Capital Health Research and Development of Special Fund, (grant number JQ20029, 82071000, 2024-2G-1081).

ORCID iDs

Ning-Li Wang

Shi-Ming Li

References

GBD 2019 Blindness and Vision Impairment Collaborators, Vision Loss Expert Group of the Global Burden of Disease Study. Causes of blindness and vision impairment in 2020 and trends over 30 years, and prevalence of avoidable blindness in relation to VISION 2020: the right to sight: an analysis for the global burden of disease study. Lancet Glob Health 2021; 9: e144–e160.

Bullimore

Ritchey

Shah

, et al. The risks and benefits of myopia control. Ophthalmology 2021; 128: 1561–1579.

Modjtahedi

Abbott

Fong

, et al. Reducing the global burden of myopia by delaying the onset of myopia and reducing myopic progression in children: the academy’s task force on myopia. Ophthalmology 2021; 128: 816–826. https://www.aaojournal.org/article/S0161-6420(20)31043-5/fulltext (accessed 14 January 2023).

Holden

Fricke

Wilson

, et al. Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050. Ophthalmology 2016; 123: 1036–1042.

Morgan

French

Ashby

, et al. The epidemics of myopia: aetiology and prevention. Prog Retin Eye Res 2018; 62: 134–149.

Medina

. The cause of myopia development and progression: theory, evidence, and treatment. Surv Ophthalmol 2022; 67: 488–509.

Zhang

Mao

Chen

, et al. Difference in aqueous concentration and vitreous mass of cytokines in high myopias with and without choroidal neovascularization. Front Med (Lausanne) 2022; 9: 1029425.

, et al. Role of chronic inflammation in myopia progression: clinical evidence and experimental validation. EBioMedicine 2016; 10. https://pubmed.ncbi.nlm.nih.gov/27470424/ (accessed 15 October 2022).

Wei

C-C

Kung

Y-J

Chen

, et al. Allergic conjunctivitis-induced retinal inflammation promotes myopia progression. EBioMedicine 2018; 28: 274–286.

10.

Wang

Zhao

, et al. Assessment of neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio in patients with high myopia. BMC Ophthalmol 2022; 22: 464.

11.

Long

, et al. C-reactive protein and complement components in patients with pathological myopia. Optom Vis Sci 2013; 90: 501–506.

12.

Enroth

Maturi

Berggrund

, et al. Systemic and specific effects of antihypertensive and lipid-lowering medication on plasma protein biomarkers for cardiovascular diseases. Sci Rep 2018; 8: 5531.

13.

Ahsan

Rask-Andersen

, et al. The relative contribution of DNA methylation and genetic variants on protein biomarkers for human diseases. PLoS Genet 2017; 13: e1007005.

14.

Enroth

Johansson

Enroth

, et al. Strong effects of genetic and lifestyle factors on biomarker variation and use of personalized cutoffs. Nat Commun 2014; 5: 4684.

15.

Karlsson

Höglund

, et al. Causal effects of inflammatory protein biomarkers on inflammatory diseases. Sci Adv 2021; 7: eabl4359.

16.

Ghorbani Mojarrad

Plotnikov

Williams

, et al. Association between polygenic risk score and risk of myopia. JAMA Ophthalmol 2020; 138: 7–13. https://doi.org/10.1001/jamaophthalmol.2019.4421 (accessed 13 October 2022).

17.

Burgess

Dudbridge

Thompson

. Combining information on multiple instrumental variables in Mendelian randomization: comparison of allele score and summarized data methods. Stat Med 2016; 35: 1880–1906.

18.

Hemani

Bowden

Davey Smith

. Evaluating the potential role of pleiotropy in Mendelian randomization studies. Hum Mol Genet 2018; 27: R195–R208.

19.

Burgess

Thompson

. Interpreting findings from Mendelian randomization using the MR-egger method. Eur J Epidemiol 2017; 32: 377–389.

20.

Rees

JMB

Wood

Burgess

. Extending the MR-egger method for multivariable Mendelian randomization to correct for both measured and unmeasured pleiotropy. Stat Med 2017; 36: 4705–4718.

21.

Verbanck

Chen

C-Y

Neale

, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet 2018; 50: 693–698.

22.

Huang

, et al. Mendelian randomization study of inflammatory bowel disease and bone mineral density. BMC Med 2020; 18: 312.

23.

McBrien

. Regulation of scleral metabolism in myopia and the role of transforming growth factor-beta. Exp Eye Res 2013; 114: 128–140.

24.

Yuan

Wang

, et al. Inflammatory cytokines in highly myopic eyes. Sci Rep 2019; 9: 3517.

25.

Wei

Zhuang

Fan

, et al. Proinflammatory and angiogenesis-related cytokines in vitreous samples of highly myopic patients. Cytokine 2021; 137: 155308.

26.

Rada

Perry

Slover

, et al. Gelatinase A and TIMP-2 expression in the fibrous sclera of myopic and recovering chick eyes. Invest Ophthalmol Vis Sci 1999; 40: 3091–3099.

27.

Liu

Deng

, et al. Intravitreal vascular endothelial growth factor concentration and axial length. Retina 2015; 35: 435–439.

28.

Zhang

Jeong

Mori

, et al. Vascular endothelial growth factor from retinal pigment epithelium is essential in choriocapillaris and axial length maintenance. PNAS Nexus 2022; 1: pgac166.

29.

Hong

Yang

, et al. Relationship between aqueous humor levels of cytokines and axial length in patients with diabetic retinopathy. Asia Pac J Ophthalmol (Phila) 2020; 9: 149–155.

30.

Shin

Nam

Park

, et al. Aqueous humor concentrations of vascular endothelial growth factor and pigment epithelium-derived factor in high myopic patients. Mol Vis 2012; 18: 2265–2270.

31.

Sawada

Miyake

Kakinoki

, et al. Negative correlation between aqueous vascular endothelial growth factor levels and axial length. Jpn J Ophthalmol 2011; 55: 401–404.

32.

Grunwald

Pistilli

Daniel

, et al. Incidence and growth of geographic atrophy during 5 years of comparison of age-related macular degeneration treatments trials. Ophthalmology 2017; 124: 97–104.

33.

Sanz

Sanchez-Niño

Ortiz

. TWEAK, a multifunctional cytokine in kidney injury. Kidney Int 2011; 80: 708–718.

34.

Poveda

Vázquez-Sánchez

Sanz

, et al. TWEAK-Fn14 as a common pathway in the heart and the kidneys in cardiorenal syndrome. J Pathol 2021; 254: 5–19.

35.

Irak

Bayram

Cifci

, et al. Serum levels of NLRC4 and MCP-2/CCL8 in patients with active Crohn’s disease. PLoS One 2021; 16: e0260034.

36.

Dikalova

Kirilyuk

Dikalov

. Antihypertensive effect of mitochondria-targeted proxyl nitroxides. Redox Biol 2015; 4: 355–362.

37.

Fehervari

. CD6 Targeting. Nat Immunol 2017; 18: 373.

38.

Gonçalves

Henriques

Santos

, et al. CD6, A rheostat-type signalosome that tunes T cell activation. Front Immunol 2018; 9: 2994.

39.

Clark

Bishop

. The eye as a complement dysregulation hotspot. Semin Immunopathol 2018; 40: 65–74.

40.

Cheng

Zhang

, et al. TRPV4-induced Müller cell gliosis and TNF-α elevation-mediated retinal ganglion cell apoptosis in glaucomatous rats via JAK2/STAT3/NF-κB pathway. J Neuroinflammation 2021; 18: 271.

41.

Torday

. Pleiotropy, the physiologic basis for biologic fields. Prog Biophys Mol Biol 2018; 136: 37–39.