Causal role of circulating inflammatory cytokines in the pathogenesis of IgA nephropathy

To the Editor: Immunoglobulin A nephropathy (IgAN) is the most common form of primary glomerulonephritis worldwide.[1] Observational studies suggest a link between circulating inflammatory cytokines and IgAN.[2] However, the causal roles of these cytokines in IgAN are not yet systematically understood. Mendelian randomization (MR) uses genetic variants as instrumental variables (IVs) in observational epidemiology to infer the causality of modifiable disease risk factors. Since genetic alleles are randomly assigned at conception, this method reduces confounding and reverse causation and has been widely used to explore causal relationships between exposures and diseases. This study explored the causal relationship between circulating inflammatory cytokines and IgAN using MR and identified upstream systemic regulators. The study design is summarized in Figure 1A.[3]Figure 1: The flowchart and inflammatory cytokines in this study. (A) The design and flowchart of MR analysis in this study. (B) The volcano plot illustrates the relationship between 41 and 91 inflammatory cytokines and IgAN. CCL25: C-C motif chemokine ligand 25; IgAN: IgA nephropathy; IL-10: Interleukin-10; IL-10Rα: Interleukin-10 receptor subunit alpha; IL-5: Interleukin-5; IL-7: Interleukin-7; LIF: Leukemia inhibitory factor; MIG: Monokine induced by interferon-γ; MR: Mendelian randomization; S100A12: S100 calcium-binding protein A12; SCGFβ: Stem cell growth factor beta; SNPs: Single-nucleotide polymorphisms; TGFα: Transforming growth factor alpha; TGFβ1: Transforming growth factor beta 1.Genetic instruments for 41 inflammatory cytokines were sourced from a large-scale Genome-wide association studies (GWAS) meta-analysis involving up to 8293 Finnish individuals from three population cohorts.[4] Genetic data for 91 inflammatory cytokines were derived from a recent genome-wide protein quantitative trait locus (pQTL) study, measured using the Olink Target platform across 11 cohorts of 14,824 participants of European ancestry.[5] The comprehensive GWAS summary statistics are available via the GWAS Catalog (https://www.ebi.ac.uk/gwas/home). Summary-level GWAS data for IgAN were extracted from the FinnGen consortium and IEU Open GWAS. The FinnGen consortium (https://finngen.gitbook.io/documentation/v/r9) included 377,277 European individuals (592 cases vs. 376,685 controls). The IEU Open GWAS dataset (https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST90018866/) includes 477,784 participants of European ancestry (15,587 cases vs. 462,197 controls). Summary information for the different databases is provided in Supplementary Table 1, https://links.lww.com/CM9/C561. First, genetic variants were selected for 41 serum cytokines and 91 inflammatory proteins to explore the causality from each inflammatory cytokine to IgAN. Subsequently, estimates from the two IgAN datasets were integrated using a meta-analysis. Based on high-quality MR studies, genetic instrument variables were defined as single nucleotide polymorphisms (SNPs) significantly and independently associated with 41 inflammatory cytokines (5 × 10−6) and 91 inflammatory cytokines (1 × 10−5). Furthermore, to guarantee independence among genetic variants, we implemented a linkage disequilibrium (LD) distance threshold of 10,000 kb and an r2 value threshold of <0.001. After clumping and data harmonization, all 41 cytokines and 91 inflammatory cytokines were found to have more than three SNPs. The strength of the SNPs was assessed using mean F-statistics, with a mean F-statistic >10 indicating robust validity for trait evaluation. In this study, five MR methods were used to investigate the relationship between inflammatory cytokines and IgAN: inverse variance-weighted (IVW), MR–Egger, weighted median, simple mode, and weighted mode. The primary result was determined using the IVW method, with significance set at P <0.05. Cochran’s IVW Q statistic and P-value were used to assess SNP heterogeneity. A statistically significant MR–Egger intercept suggests potential horizontal pleiotropy. A leave-one-out sensitivity analysis was performed by sequentially removing each SNP to identify potential off-topic IVs. The MR study was performed using the R software, version 4.1.3 (R Foundation, Vienna, Austria) with the TwoSampleMR package (version 0.5.5). A meta-analysis was conducted using Stata (STATA version 18, STATA Corp., Texas, United States). To select strong IVs, all SNPs that strongly and independently predicted cytokines at genome-wide significance (P <5 × 10−8) were first chosen. However, as in other MR analyses based on omics data, an insufficient number of SNPs met this threshold. Consequently, we relaxed the threshold appropriately to obtain a sufficient number of SNPs for analysis. The screening process initially identified 449 IVs for 41 regulators and 2993 IVs for 91 regulators, meeting the locus-wide significance thresholds of 5 × 10−6 and 1 × 10−5, respectively [Supplementary Tables 2 and 3, https://links.lww.com/CM9/C561]. After removing the palindromic SNPs, 364 and 370 IVs were obtained for 41 regulators, and 2198 and 2234 IVs for 91 regulators from the FinnGen consortium and IEU Open GWAS datasets, respectively [Supplementary Tables 4–7, https://links.lww.com/CM9/C561]. In the FinnGen datasets, interleukin-5 (IL-5) (odds ratio [OR] = 0.723, P = 0.045) and Monokine induced by interferon-γ (MIG) (OR = 1.568, P = 0.002) were associated with the risk of IgAN, while in the IEU Open GWAS datasets, stem cell growth factor beta (SCGFβ) (OR = 1.045, P = 0.039) and interleukin-7 (IL-7) (OR = 1.060, P = 0.008) were linked to IgAN risk [Figure 1B, C]. The MR–Egger intercept test revealed no evidence of horizontal pleiotropy in either dataset. The leave-one-out analysis and scatter plot results are provided in Supplementary Figures 1–4, https://links.lww.com/CM9/C560. To increase the statistical power, a meta-analysis of the MR results was conducted from both datasets. Cochran’s Q test revealed heterogeneity for IL-5 (Q = 5.100, QP = 0.024) and MIG (Q = 9.360, QP = 0.002), prompting the use of a random-effects model. A meta-analysis suggested that IL-7 (OR = 1.059, P = 0.009) was associated with an increased risk of IgAN [Figure 1B, Supplementary Figure 5, https://links.lww.com/CM9/C560 and Supplementary Table 8, https://links.lww.com/CM9/C561]. The causal effects of 91 circulating inflammatory cytokines on IgAN were analyzed using data from the IEU Open GWAS and the FinnGen consortium. In the IEU Open GWAS datasets, higher levels of transforming growth factor beta 1 (TGFβ1) (OR = 0.893, P = 0.007) and S100 calcium-binding protein A12 (S100A12) (OR = 0.914, P = 0.020) were associated with a decreased risk of IgAN, while C-C motif chemokine ligand 25 (CCL25) (OR = 1.047, P = 0.030) were linked to an increased risk of IgAN. In the FinnGen datasets, higher levels of transforming growth factor alpha (TGFα) (OR = 1.698, P = 0.018), IL-10 (OR = 1.845, P = 0.006), leukemia inhibitory factor (LIF) (OR = 1.947, P = 0.003), and CD5 (OR = 1.584, P = 0.049) were associated with an increased risk of IgAN, while higher interleukin-10 receptor subunit alpha (IL-10Rα) levels (OR = 0.624, P = 0.041) were linked to lower risk [Figure 1B, Supplementary Figure 6, https://links.lww.com/CM9/C560]. The leave-one-out analysis and scatter plot results are provided in Supplementary Figures 7–14, https://links.lww.com/CM9/C560. The MR–Egger intercept test indicated no horizontal pleiotropy. To enhance the statistical power, a meta-analysis was performed combining the MR results from both data sources. Cochran’s Q test revealed significant heterogeneity for TGFα, IL-10, LIF, and CD5 (all Q P-value <0.05), leading to the use of a random-effects model. This analysis showed that higher CCL25 levels (OR = 1.047, P = 0.030) and lower S100A12 levels (OR = 0.916, P = 0.021) were associated with an increased risk of developing IgAN [Figure 1B, and Supplementary Table 9, https://links.lww.com/CM9/C561]. This study provides a comprehensive assessment of the causal relationship between circulating inflammatory cytokines and IgAN, supported by a series of sensitivity analyses, to ensure the robustness of the findings. The results indicated that genetically predicted levels of IL-7 and CCL25 were positively associated with an increased risk of IgAN, whereas higher levels of S100A12 were associated with a reduced risk. The identification of IL-7, CCL25, and S100A12 as upstream regulators has significant clinical implications. As noninvasive biomarkers, these cytokines could be utilized in developing blood-based diagnostic tools for the early detection and monitoring of IgAN risk, thereby reducing the need for invasive procedures, such as kidney biopsies. Specifically, elevated IL-7 or CCL25 levels, along with lower S100A12 levels in blood samples, may serve as indicators of increased disease risk, facilitating earlier intervention. These insights enhance our understanding of complex immune system interactions in IgAN and may guide future research toward more effective and targeted therapeutic approaches. This study has several strengths. First, this study systematically assessed the causal relationships between circulating inflammatory cytokines and IgAN using the most recent summary data available. Traditional epidemiological studies are susceptible to bias owing to confounding factors and reverse causation. In this two-sample MR study, conclusions were based on genetic IVs, ensuring that the direction of the exposure-to-outcome relationship was clear and not influenced by external factors. Multiple MR analysis methods were employed for causal inference to enhance the robustness of the results against horizontal pleiotropy. Additionally, a meta-analysis was conducted to assess the causal association between inflammatory factors and IgAN, utilizing data from two large GWAS to enhance the statistical power. However, this study had some limitations. First, regarding the selection of IVs, the traditional threshold is typically set at P <5 × 10–8. However, as in other MR analyses based on omics data, an insufficient number of SNPs met this threshold. Consequently, we relaxed the threshold appropriately to obtain a sufficient number of SNPs for analysis. Although this approach may potentially increase the number of false positives, it allows for a more comprehensive evaluation of the association between inflammatory regulators and IgAN. Second, owing to the lack of GWAS data for IgAN in East Asian populations, this study was primarily derived from European populations, limiting the generalizability of our findings to other ethnic groups. Finally, the inherent limitations associated with the MR analysis must be acknowledged. MR relies on the assumption that genetic instruments accurately predict the exposure, which may not always be true. Although minimal horizontal pleiotropy was detected in our study, its potential impact cannot be entirely excluded. Because SNPs explain only a small proportion of phenotypic variance, MR analyses often require large sample sizes to achieve adequate statistical power. In this study, two IgAN datasets with sufficiently large sample sizes were selected to improve the statistical power. Despite their narrow confidence intervals, the findings related to SCGFβ and CCL25 provide valuable insights into the pathogenesis of IgAN. These cytokines had a modest but statistically significant association with IgAN, suggesting their potential involvement in disease mechanisms. The narrow confidence intervals around the ORs indicated precision in the estimates, which strengthened the robustness of the associations. However, further research is needed to confirm these findings and to fully elucidate their roles in IgAN pathogenesis. In conclusion, this MR study identified three upstream inflammatory cytokines in IgAN, providing insights into noninvasive risk prediction and potential therapeutic targets. However, the precise association between these inflammatory cytokines and IgAN requires further validation in larger cohorts. Funding This work was supported by the Multidisciplinary Clinical Research Innovation Project (No. CYDXK202213), the Natural Science Foundation of China (No. 62476181), the Jinzhongzi project of Beijing Chao-Yang Hospital (No. CYJZ202203), the Natural Science Foundation of China (No. 82170686). Conflicts of interest None.