Causality Between 91 Circulating Inflammatory Proteins and Various Asthma Phenotypes: A Mendelian Randomization Study

Introduction

Asthma is one of the most common chronic respiratory diseases affecting children and adults worldwide. Despite significant reductions in mortality in some countries and regions, the global asthma mortality rate did not change significantly from 2006 to 2012, at about 0.19 deaths per 100,000 people (0.16–0.21).¹ There were approximately 81 million children with asthma worldwide in 2019, with an incidence of 1030.33 cases per 100,000 people (95% CI 683.66–1449.53).² Each phenotype and potential pathogenesis of asthma are different. Markers near the ORMDL3/GSDMB gene have been associated with childhood asthma, and interleukin (IL) 33 and IL1RL1SNP have been associated with atopic asthma. The thymic interstitial lymphoblastopoietin (TSLP) gene has been identified as a protective factor against TH2 asthma risk.³

In recent years, Personalized Medicine has made a multidisciplinary effort to better determine the different phenotypes of allergic diseases by endotyping their mechanisms.⁴ Classifying diseases solely through phenotypes is insufficient to establish adequate or personalized treatments for patients since several phenotypes are superimposable and fluid, varying from one to the other.⁵ Establishing the mechanisms responsible (endotypes) for the phenotypes is much more relevant, as it better demonstrates causality, being more appropriate for the personalized treatment of patients.⁶ The laboratory tools currently available to the practicing clinician fall far short of the scientific knowledge already acquired, and efforts are being made to define endotype markers capable of aiding medical diagnosis. Historically, the classification of asthma phenotypes has evolved a lot. Asthma must be understood not as a single “disease” but as a phenotypic condition, a large “umbrella denomination” for several diseases of very different natures with completely different endotypes.⁷ Even the simplistic classification into “allergic” and “non-allergic”, “extrinsic”, or “intrinsic” is fragile since, as knowledge advances in determining the mechanisms of hypersensitivity, more and more patients previously considered “not-allergic” are now being understood as “allergic”, as tools for the diagnosis of non-IgE-mediated hypersensitivity are being developed.⁸ In this sense, determining inflammatory markers related to the genotypic profile is crucial to assisting the endotype diagnosis, enriching the phenotypic classification, and helping to tailor the patient’s treatment.

Mendelian randomization (MR) is a powerful tool for causal reasoning using single nucleotide polymorphisms (SNPs) as instrumental variables, relying on genome-wide association studies (GWAS) that utilize one or more genetic variants as instrumental variables (IVs) that are strongly associated with the exposure of interest and not influenced by confounding factors. The causal effects of exposure on results can be inferred. Previous studies have shown a significant relationship between inflammatory proteins and asthma phenotypes. Based on this, this study verified the relationship between 91 kinds of circulating inflammatory proteins and six asthma phenotypes through MR Analysis and public GWAS database summary statistics.

Materials and Methods

Study Design

MR Analysis must satisfy three key assumptions to obtain valid results. Specifically, to be used as instrumental variables (IVs) of an exposure factor, the genetic variation must be satisfied: (1) IVs establish a robust association with the studied exposure (correlation hypothesis); (2) IVs are independent of any confounding factors (independence assumption); (3) IVs affects outcomes only by exposure factors (excluding the limiting hypothesis).⁹ Figure 1 depicts the study’s overview in detail.

Figure 1 Overview of Mendelian randomization.

Data Sources

91 circulating inflammatory proteins were derived from a meta-analysis of 11 cohorts.¹⁰ Complete protein GWAS summary statistics can be in https://www.phpc.cam.ac.uk/ceu/proteins and EBI GWAS directory (GCST90274758 - GCST90274848) to download. Various asthma phenotypes are derived from the FinnGen Consortium (R10)¹¹ database, as shown in Table 1. The population selection between the exposure group and the result group did not overlap, and the included populations were all European populations, meeting the requirement that both samples in MR Were from the same genetic background. Supplementary Table 1 summarizes the GWAS details of the 91 circulating inflammatory proteins included. And Supplementary Table 2 describes all SNPs information used as IVs. Our research methods and reports are based on the STROBE-MR guidelines, which are detailed in the Supplementary Table 3. The datasets used in this study were all drawn from publicly published GWAS databases and do not require ethical approval.

Table 1 Detailed Information for the Various Asthma Phenotypes Data

Instrumental Variables Selection

Firstly, SNPS of outcomes and 91 circulating inflammatory proteins were identified using a significance threshold of p< 5×10⁻⁸. However, for certain inflammatory proteins, the determination of the number of SNPS is limited under these conditions. In order to obtain more positive SNPS, we adjusted the threshold to 5×10⁻⁶ with reference to previously published papers.¹² Second, the SNPs were grouped together to eliminate linkage imbalances (kb = 10,000, r²= 0.001).¹³ During coordination, SNPS are excluded if they do not agree with the intermediate allele frequency or palindrome. Finally, we calculate the strength of each SNP through the F statistic. Using the following formulas:

K is the number of instrumental variables included. R² represents the proportion of variance in phenotype explained by a single SNP. N represents the sample size, Beta represents the estimated genetic effect of the SNPs on exposure, EAF represents the effect of allele frequency, and represents the standard error of Beta.¹⁴ SNPS with F statistic >10 are considered to have strong correlation.¹⁵ So if F statistic <10, they are excluded. The outlier SNPS that affect the level pluripotency are eliminated by the outlier test of the MR-PRESSO method.¹⁶

MR Analysis

IVW was used as the main research method. MR-Egger, Weighted median method, Simple mode method, and Weighted mode method are used as supplements. Results were presented as OR and 95% CI and p<0.05 was considered statistically consistent. First, in the absence of either heterogeneity or pleiotropy, the IVW estimate was chosen as the preferred option. Secondly, in the case of heterogeneity but no pleiotropy, the weighted median method is the preferred analysis method.

Sensitivity Analysis

Heterogeneity among IVs was evaluated by Cochran’s Q statistic. MR Pleiotropic test was used for MR-Egger and intercept value was returned to evaluate horizontal pleiotropic. Since 91 inflammatory proteins were examined, we applied FDR correction to reduce the incidence of Class I errors. A positive result means that the difference is significant before correction (p<0.05), and the difference is still significant after correction (p<0.05). It was significant before correction (p<0.05), but not significant after FDR correction, indicating a suggestive association result with potential causality.¹⁷ A sensitivity analysis called “leave-one-out” was conducted by leaving out one SNP at a time to assess if a single SNP had a disproportionate effect on the overall estimation. All analyses were performed using the TwoSampleMR package in R4.3.3

Result

Instrumental Variables result

After the clumped function was used to remove the linkage disequilibrium, a total of 1817 SNPs remained for 91 inflammatory proteins. After calculation, it is concluded that the F-value corresponding to each single SNP in this study is greater than 10, that is, there is no weak instrumental variable bias.

MR Analysis

Forward MR Analysis results

The specific results of forward MR Analysis are shown in Table 2. The results showed that CD40, IL-8, CD6, and TNFRSF9 had a potential causal relationship with asthma. Higher genetically determined CD40 levels (increased 1-SD) were found to be associated with a reduced chance of asthma by the IVW approach, with a 6.6% reduction in asthma risk for each additional 1 SD (OR=0.934; 95% CI: 0.886–0.984; P = 0.010); According to the weighted median method (OR=0.932; 95% CI: 0.877–0.991; P=0.043). Elevated levels of IL-8, CD6, and TNFRSF9 were associated with increased asthma risk, with each additional SD increasing by 12.2% (OR=1.122; 95% CI: 1.023–1.230; P=0.014), 7.9% (OR=1.079; 95% CI: 1.017–1.145; P=0.011) and 7.1% (OR=1.071; 95% CI: 1.000–1.146; P=0.050). Scatter plots, funnel plots, and leave plots of MR Analyses of potentially causal inflammatory factors and asthma are shown in Supplemental Figure S1. There was no statistical significance between the intercept term of the above inflammatory factor MR-Egger and 0 (P>0.05), indicating that there was no horizontal pleiotropy. Cochran’s Q test did not observe the statistical significance of heterogeneity among SNPS (P>0.05).

Table 2 MR Analysis Results of Causal Association Between Inflammatory Proteins and Asthma

Reverse MR Analysis Results

The reverse MR Analysis results are shown in Table 3. The results showed that asthma was positively correlated with CD40 and CD6 levels (P < 0.05). There was no significant causal relationship between asthma and IL-8 and TNFRSF9 (P > 0.05).

Table 3 Reverse MR Analysis Results of Causal Association Between Inflammatory Proteins and Asthma

Subgroup MR Analysis

91 Circulating Inflammatory Proteins and Allergic Asthma

As shown in Table 4, a total of 4 possible causal relationships were identified between the circulating inflammatory proteins and the risk of developing allergic asthma. The reverse MR Analysis results are shown in Table 5.

Table 4 Effects of the Relationship Between Meaningful Circulating Inflammatory Proteins and Allergic Asthma in MR Analysis

Table 5 Effects of the Relationship Between Meaningful Allergic Asthma and Circulating Inflammatory Proteins in Reverse MR Analysis

In the MR analysis, the level of IL-2 (OR=1.193; 95% CI: 1.018–1.398; P=0.029, P-FDR=2.640) was significantly associated with an increased risk of allergic asthma. And levels of SCF (OR=0.910; 95% CI: 0.833–0.995; P=0.039, P-FDR=1.183), TRAIL (OR=0.909; 95% CI: 0.831–0.994; P=0.036, P-FDR=1.640), TWEAK (OR=0.838; 95% CI: 0.704–0.997; P=0.046, P-FDR=1.050) was associated with a reduced risk of allergic asthma. According to Cochran’s Q-test, there was no evidence of heterogeneity in the IVW model (Table 4). As anticipated, no evidence of horizontal pleiotropy was found according to the MR-Egger intercept (Table 4). Heterogeneity and sensitivity analysis results corroborated the accuracy of the results. Scatter plots and funnel plots also supported these findings (Supplementary Figure S2). The leave-one-out method further validated the robustness of the data (Supplementary Figure S2). It is also regrettable that there were no significant positive results after the FDR correction.

91 Circulating Inflammatory Proteins and Non-Allergic Asthma

IVW analysis revealed 6 possible causal relationships between the circulating inflammatory proteins and the risk of non-allergic asthma (Table 6). The reverse MR Analysis results are shown in Table 7.The level of CCL19 (OR=1.140; 95% CI: 1.011–1.286; P=0.033; P-FDR=0.592), CST5 (OR=1.184; 95% CI: 1.075–1.305; P=0.0001; P-FDR=0.028), IL-7 (OR=1.272; 95% CI: 1.017–1.591; P=0.035; P-FDR=0.534) was significantly associated with an increased risk of allergic asthma. The level of CX3CL1 (OR=0.731; 95% CI: 0.611–0.874; P=0.0001; P-FDR=0.055), LAP TGF-beta-1 (OR=0.797; 95% CI: 0.669–0.950; P=0.011; P-FDR=0.342), ST1A1 (OR=0.870; 95% CI: 0.766–0.988; P=0.032; P-FDR=0.731) were associated with reduced risk of non-allergic asthma. Among them, there were significant positive results of CST5 after FDR correction, suggesting that CST5 levels were associated with an increase in non-allergic asthma. No significant heterogeneity or pleiotropy was found in Cochran’s Q-test, MR-Egger intercept, heterogeneity, and sensitivity analysis (Table 6). Scatter plots, funnel plots, and leave-one-out plots are shown results in Supplementary Figure S3. Subsequently, different colors in Table 6 represent the results for each meaningful circulating inflammatory protein.

Table 6 Effects of the Relationship Between Meaningful Circulating Inflammatory Proteins and Non-Allergic Asthma in MR Analysis

Table 7 Effects of the Relationship Between Meaningful Non-Allergic Asthma and Circulating Inflammatory Proteins in Reverse MR Analysis

91 Circulating Inflammatory Proteins and Suggestive for Childhood Asthma

ARTN, CCL25, IL-18R1, LIF-R, and TNF were all potentially causally associated with childhood asthma (<16 years), in which LIF-R was significantly positive (P-FDR=0.003). Higher genetically determined ARTN and IL-18R1 levels were associated with an increased risk of asthma in children (<16 years of age) by IVW, with each additional SD increasing the risk of asthma by 25.2% (OR=1.252; 95% CI: 1.017–1.542; P=0.034) and 11.1% (OR=1.111; 95% CI: 1.007–1.227; P=0.036). CCL25, LIF-R, and TNF levels were associated with reduced asthma risk in children, with each increase of 1 SD associated with a 1.4% decrease in asthma risk (OR=0.869; 95% CI: 0.776 to 0.972; P=0.014), 2.7% (OR=0.723; 95% CI: 0.620–0.842; P=0.000) and 1.5% (OR=0.845; 95% CI: 0.725–0.985; P=0.031). MR Analysis scatterplot, funnel plot, and left plot are shown in supplementary Figure S4. No evidence of horizontal pleiotropy was found at the MR-Egger intercept (Table 8). Heterogeneity and sensitivity analysis confirmed the accuracy of the results. The reverse MR Analysis results are shown in Table 9.

Table 8 Results of MR Analysis Between Causally Associated Inflammatory Proteins and Asthma in Children (<16 y)

Table 9 Effects of the Relationship Between Meaningful Asthma in Children and Circulating Inflammatory Proteins in Reverse MR Analysis

91 Circulating Inflammatory Proteins and Suggestive for Eosinophilic Asthma

Four possible causal relationships were found between the circulating inflammatory proteins and suggestive of eosinophilic asthma. The level of CXCL6 (OR=1.246; 95% CI: 1.019–1.524; P=0.032; P-FDR=0.970), CST5 (OR=1.209; 95% CI: 1.005–1.453; P=0.044; P-FDR=1.002), SLAMF1 (OR=1.318; 95% CI: 1.059–1.641; P=0.013; P-FDR=1.223) suggest a potential causal relationship with increased the risk of suggestive for eosinophilic asthma. The level of TNFB (OR=0.834; 95% CI: 0.709–0.982; P=0.030; P-FDR=1.350) was associated with a reduced risk of suggestive eosinophilic asthma. Unfortunately, there were no significant positive circulating inflammatory proteins after FDR correction. The results of MR-Egger and MR-PRESSO tests confirmed that there is no horizontal pleiotropy and the outcomes from Cochrane’s Q test demonstrated that there is no obvious heterogeneity among the selected SNPs (Table 10). Scatter plots, funnel plots, and leave-one-out plots are shown results in Supplementary Figure S5. The reverse MR Analysis results are shown in Table 11.

Table 10 Effects of the Relationship Between Meaningful Circulating Inflammatory Proteins and Suggestive for Eosinophilic Asthma in MR Analysis

Table 11 Effects of the Relationship Between Meaningful Suggestive for Eosinophilic Asthma and Circulating Inflammatory Proteins in Reverse MR Analysis

91 Circulating Inflammatory Proteins and Suggestive for Obesity Related Asthma

IVW analysis revealed 3 possible causal relationships between the 91 circulating inflammatory proteins and the risk of obesity-related asthma. After the FDR correction, there was no significant strong correlation. There is only a potential causal relationship between the three inflammatory proteins associated with obesity-related asthma. The level of FGF-5 (OR=1.104; 95% CI: 1.018–1.198; P=0.017; P-FDR=0.790), IL-17A (OR=1.235; 95% CI: 1.021–1.493; P=0.030; P-FDR=0.900) suggests a potential causal relationship with increased the risk of obesity-related asthma. The level of PD-L1 (OR=0.781; 95% CI: 0.663–0.922; P=0.003; P-FDR=0.310) was associated with a reduced risk of obesity related asthma. No significant heterogeneity or pleiotropy was found in Cochran’s Q-test, MR-Egger intercept, heterogeneity, and sensitivity analysis (Table 12). Scatter plots, funnel plots, and leave-one-out plots are shown results in Supplementary Figure S6. The reverse MR Analysis results are shown in Table 13.

Table 12 Effects of the Relationship Between Meaningful Circulating Inflammatory Proteins and Obesity Related Asthma in MR Analysis

Table 13 Effects of the Relationship Between Meaningful Suggestive for Obesity Related Asthma and Circulating Inflammatory Proteins in Reverse MR Analysis

Discussion

This study is the first to examine the causal relationship between 91 circulating inflammatory proteins and asthma and their phenotypes. A total of 30 potentially causal results were obtained, and after FDR correction, the occurrence of errors was reduced, and finally, two inflammatory protein result with a significant causal relationship was identified. Due to the limitations of clinical trial selection, there are no studies to explore the causal relationship between CST5 and non-allergic asthma, LIF-R and the onset of asthma in children. The results with potential causality, it is still groundbreaking and has potential value for future research.

The findings suggest that for asthma, elevated levels of IL-8, CD6, and TNFRSF9 may lead to an increased risk, and elevated levels of CD40 may lead to a decreased risk. Consistent with the results of Marc-Malovrh M, the elevation of IL-5 and IL-8 in phlegm eosinophils may stimulate airway remodeling and may be a useful non-invasive biomarker and therapeutic target for accelerating FEV1 decline in asthmatic patients.¹⁸ Similarly, in the Matsuda S experiment on patients with first wheezing and confirmed asthma, serum IL-8, and IL-12 values at first wheezing were significantly higher than those in acute asthma or control groups.¹⁹ Previous studies have shown that CD6 is related to the pathogenesis of autoimmune diseases.²⁰ Although no clinical or animal studies have been conducted to date on its association with asthma. However, it is presumed that CD6 is involved in the development and activation of T cells.²¹ Therefore, CD6 may influence asthma development by affecting T cells. We have observed that TNFRSF9 levels are associated with an increased risk of asthma, but there are currently no studies to support this result, and further research is necessary to fully understand the role of this protein in maintaining health or promoting disease development.

The interaction between CD40 and its ligand CD40L controls humoral and cell-mediated immune responses, and CD40 expressed on connected airway epithelial cells can up-regulate the expression of inflammatory mediators.^22,23 In a test on the association of CD40 polymorphism with asthma risk and serum IgE level in a Korean population, the results showed that CD40 gene polymorphism regulates the expression of CD40 in B cells through translation, and has a genetic influence on the production of IgE in asthmatic patients.²⁴

Genetically determined ARTN levels have a potential causal relationship with childhood asthma. In a case-control study that included 423 children with asthma and 414 non-asthmatic controls, 17 SNPS were significantly associated with asthma, of which 6 ARTN-related SNPS were still significantly associated with asthma after adjustment.²⁵ This is the same as our result. The IL18R1 gene is a strong candidate for asthma. CCL25, LIF-R, and TNF were all associated with reduced asthma in children. The research results of Stenstad H and Wurbel M. A fully prove that CCL25 may be involved in T cell development.^26,28 Different from the results of MR, the results of Sen Y. showed that the CCR9/CCL25 signaling pathway could interact with CD226 signaling to activate asthma NKT cells, leading to airway hyperreactivity and inflammation, which aggravated asthma.²⁹ IL-18 promotes airway inflammation and promotes auxiliary T-2 response.^30,31 Induces IgE, IL-4 and IL-13, and histamine release from basophils.^32,33 In earlier reports, IL-18 levels in blood and sputum were found to increase significantly during asthma attacks.³⁴ The results of the Zheng X study showed that the serum LIF levels in subjects with atopic asthma were higher than those in non-atopic subjects, and LIF links nerve and allergic inflammatory processes.³⁵ There is also evidence that LIF may also play an important role in regulating neuro-immune system interactions during acute inflammatory injury and subsequent healing and recovery processes. Therefore, LIF can be used as a mediator of bidirectional crosstalk between nervous tissue and the immune system.³⁶

TNF is a potent pro-inflammatory cytokine that has been linked to asthma and atopic activity. Among them, TNF-ɑ is involved in the development of allergic diseases, especially asthma and atopic dermatitis. It plays a dual role in the regulation of immune response, not only as a pro-inflammatory mediator to initiate strong inflammatory response, but also as an immunosuppressive mediator to inhibit the development of autoimmune diseases and tumorigenesis, and plays a crucial role in maintaining immune homeostasis by limiting the degree and duration of inflammatory processes.³⁷ Castro J showed elevated levels of TNF-α in airway biopsies and bronchoalveolar lavage fluid in asthmatic patients. However, no significant association was found between TNFB alleles and atopic properties.³⁸ There have been no clear studies on the relationship between CCL25, LIF-R, and TNF and childhood asthma in terms of age.

Observational and longitudinal studies have identified biomarkers that can distinguish patients with allergic asthma from healthy controls. Genetically predicted IL-2 levels are associated with an increased risk of allergic asthma. Kannan AK’s³⁹ experimental results suggest that IL-2-induced T cell kinases regulate TH2-mediated allergic airway inflammation by inhibiting IFN-γin naive CD4+ T cells. Genetically determined levels of SCF, TRAIL, and TWEAK are potentially associated with reduced risk of allergic asthma. Interestingly, the study⁴⁰ results show that allergic asthma is characterized by elevated cytokine SCF, and measuring its expression in serum protein levels or mRNA levels in PBMC will be an important parameter in diagnosing allergic asthma. This is contrary to our results. Since there were no significant positive results for this inflammatory protein after FDR correction, more clinical studies are needed to confirm the causal relationship between this inflammatory protein and allergic asthma. TRAIL may promote or conversely address inflammation in asthma by inducing apoptosis in a variety of cells.⁴⁰ This is consistent with our results. McGrath EE’s⁴¹ findings showed that animals lacking TRAIL showed delayed neutrophil apoptosis and increased neutrophil inflammation. The administration of exogenous TRAIL restored the WT phenotype of TRAIL-deficient mice and, importantly, accelerated neutrophil apoptosis and neutrophil count reduction in WT mice. TWEAK binds to receptor factor-induced type 14 (Fn14) and is involved in a variety of pathological processes, including angiogenesis, cell proliferation and death, inflammation, and carcinogenesis.⁴² Airway remodeling in asthma can accelerate HASMC cell proliferation and migration by activating the NF-κB pathway from the TWEAK/Fn14 axis.⁴³ TWEAK and TGF-β1 have a synergistic effect in epithelial-mesenchymal transition and may contribute to chronic airway change and remodeling.⁴⁴

Our results are similar to those of previous studies. Levels of CCL19, CST5, and IL-7 may reflect the risk of non-allergic asthma. Genetically, CX3CL1, LAP TGF-beta-1, and ST1A1 levels were associated with a reduced risk of non-allergic asthma. Previous findings related to asthma or allergic asthma, but there was uncertainty about whether it was related to non-allergic asthma. Nakano K’s⁴⁵ results show recombinant CCL19 increased the phosphorylation of STAT5 in mice with allergic asthma and induced the expression of TH2 cells and genes associated with the IL-2 signaling pathway. There is a clear causal relationship between IL-7 and asthma. Defects in the human and mouse IL-7 pathway lead to severe combined immune deficiency caused by lymphocytopenia.⁴⁶ However, due to the lack of specific studies on IL-7 and non-allergic asthma, a clear positive and negative causal relationship needs to be supported by other results. Similarly, studies have demonstrated that loss of CX3CR1 signaling unexpectedly leads to severe impairment of lung function. The CX3CL1/CX3CR1 axis preserves lung function during fungus-associated allergic airway inflammation through non-classical immunomodulatory mechanisms.⁴⁷ Unfortunately, there are no relevant studies on LAP TGF-beta-1 and ST1A1 levels and asthma, and it is impossible to know whether there is a causal relationship in clinical practice.

For eosinophilic asthma, only TNFB has been studied, but most of them are at the genetic level. Tumor necrosis factor (TNF; TNFA and TNFB) are major pro-inflammatory cytokines that are important in the pathogenesis of asthma.⁴⁸ A gene-level study from Japan showed that TNFA and TNFB genes encoding TNF-α and TNF-β, respectively, were associated with atopic asthma.⁴⁹ However, results from Egypt showed that TNFA-1031C >T and TNFA-308G >A polymorphisms were strongly associated with asthma risk (p = 0.007 and p = 0.000), but TNFB +252A>G polymorphisms were not (p = 0.6).⁵⁰

For obesity-related asthma, FGF-5, IL-17A, and PD-L1 levels are causally related to it at the genetic level. Among them, the relationship between PD-L1 and obesity asthma was positive. The IL-17 axis is involved in the pathogenesis of ovalbumin-sensitized rat disease and is one of the therapeutic targets for obesity-related asthma subjects.⁵¹ Unlike MR Analysis, PD-L1 has a bidirectional effect on the mechanism of asthma. PD-L1 can interact with PD-1 to activate Th2 and make it produce more IL-4, resulting in airway hyperreactivity. However, PD-L1 can also bind to PD-1 on the surface of immune cells (such as T cells) to promote the immunosuppressive effect of PD-1, such as reducing T cell proliferation, cytokine secretion, and cytotoxicity, and contributing to tissue homeostasis in the inflammatory response.⁵² Unfortunately, the clinical correlation between FGF-5 and asthma is not high, and there is no specific study.

This study also has certain limitations: (1) Although some MR Results showed a potential causal relationship, it was never demonstrated in clinical or animal experiments, such as TNFRSF9, CXCL6, CST5, FGF-5 and SLAMF1. Therefore, whether this inflammatory factor has practical significance with asthma or its phenotype needs more research to clarify; (2) The experimental verification of the age level of disease is insufficient, and more studies are needed to confirm it; (3) Due to the fact that the population used is European, there may also be some limitations in the generalization of the results to other ethnic groups. (4) Since this study is a statistical analysis, it is not possible to explore the mechanism between various inflammatory factors and diseases, and more follow-up studies may be needed to clarify it.

In summary, this MR Analysis explored the causal relationship between 91 inflammatory factors and asthma and 5 phenotypes, providing new evidence for the demonstration of various inflammatory factors and asthma and their phenotypes, and verified previous studies to provide directions for future research.