Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Kidney stones are increasingly common globally, and drug-induced kidney stones are often under-recognized. Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used, and their potential link to kidney stones is an area of concern. The aim of this study is to explore the potential relationship between NSAIDs and kidney stones using Mendelian randomization (MR).
Methods
NSAIDs target genes were identified, and expression quantitative trait locus data were used to find genetic instruments. We conducted a summary data-based MR analysis to verify associations between genetic instruments and inflammatory factors and then assessed the causal relationship between target gene expressions and kidney stones. Sensitivity analyses, including colocalization analysis, horizontal pleiotropy assessment, investigation of weak instrumental variable impact, use of multiple data sources and methods, and testing of associations between inflammatory factors and kidney stone risk, were performed.
Results
Twelve target genes were identified. After comprehensive analysis, only NEU1 showed a significant genetically proxied association with kidney stone risk. A decrease in genetically proxied NEU1 expression was associated with an increased risk of kidney stones (odds ratio = 0.6, 95% CI 0.47–0.77). Sensitivity analyses supported the reliability of this association. No significant association was observed between tumour necrosis factor-alpha (TNF-α) and kidney stones, suggesting that the NEU1-kidney stone association may be independent of TNF-α.
Conclusions
Genetically proxied lower NEU1 expression was associated with higher kidney stone risk. This association appeared independent of TNF-α levels. These findings warrant further mechanistic studies to investigate NEU1-related pathways in nephrolithiasis.
Junfa Liu and Kaiyi Dong have contributed equally to this work.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Introduction
Kidney stones, also known as nephrolithiasis, are increasingly common globally and contribute to a significant economic burden (Huang et al. 2024; Kittanamongkolchai et al. 2018; Zeng et al. 2017). Some cases are associated with specific medications, accounting for 1–2% of all kidney stones (Daudon et al. 2018). However, distinguishing between drug-induced and common kidney stones based on appearance and composition alone is challenging. This under-recognition may obscure the true prevalence of drug-induced kidney stones, leading to ongoing issues until the causative drug is stopped. While drug-induced kidney stones are not common, they warrant attention and further research due to their serious impact and potentially underestimated incidence.
Nonsteroidal anti-inflammatory drugs (NSAIDs) are widely used to alleviate pain and inflammation in conditions such as arthritis, migraine, and postoperative recovery (Bindu et al. 2020; Mazumder et al. 2024; Sousa et al. 2023). As the global population ages and the incidence of NSAID-related diseases increases, the global use of NSAIDs continues to increase. Considering the extensive and long-term use of NSAIDs, attention has been given to their adverse effects. Studies have shown that NSAIDs may lead to various forms of nephrotoxicity, such as functional renal failure (Harirforoosh and Jamali 2009; Prieto-García et al. 2016), interstitial nephritis (Nast 2017), and renal papillary necrosis (Leung et al. 2021). However, the exact mechanisms of NSAID-induced kidney stones remain unclear. Nimesulide, a selective COX-2 inhibitor, has been associated with acute reversible renal failure and the presence of calcium oxalate crystals in renal tubular cells (Van der Niepen et al. 2002), highlighting the need to investigate the link between NSAIDs and kidney stones. Detecting this relationship through observational studies is challenging due to the hidden nature of kidney stones and the common use of NSAIDs. In addition, most studies have focused mainly on renal function or end-stage renal disease. Taking all these factors into consideration, the link between NSAIDs and kidney stones has been severely overlooked. Therefore, thoroughly exploring the potential causal relationship between NSAIDs and kidney stones requires more precise research methodologies.
Anzeige
Mendelian randomization (MR) leverages naturally occurring genetic variants in drug targets to study disease risk. These genetic variants are randomly assigned at conception, making them independent of potential confounding environmental influences. MR estimates lifelong risks with longer follow-up periods than randomized controlled trials (Triozzi et al. 2023; Zheng et al. 2024), allowing for the identification of unintended drug effects and drug repurposing (Swerdlow et al. 2015; Yin et al. 2017). Given that NSAIDs can target multiple pathways, comprehensive analyses incorporating all relevant targets with corresponding instrumental variables are crucial. The aim of our study was to utilize a two-sample MR approach to explore the genetically proxied associations among commonly prescribed NSAIDs, inflammatory factors, and kidney stones.
Method
Study design
The method flow is outlined in Fig. 1. Supplementary Table S1 provides detailed information on the data sources. First, NSAIDs and their target genes were identified. Then, expression quantitative trait locus (eQTL) data from eQTLGen were used to find instrumental variables for target genes. Next, MR analysis was performed via the summary data-based MR (SMR) (Zhu et al. 2016) method to verify the associations between genetic instruments and inflammatory factors (C-reactive protein (Zhou et al. 2024), interleukin-1 alpha, interleukin-1 beta (Galozzi et al. 2021), interleukin-6 (Aliyu et al. 2022), prostaglandin E2 (Murakami and Kudo 2004), and tumour necrosis factor-alpha (TNF-α) (Crorkin et al. 2022)), and target genes that have significant associations with at least one inflammatory factor were screened. SMR analysis was subsequently performed to assess the genetically proxied causal relationship between the expression of these genes and kidney stone genome-wide association studies (GWASs). Finally, multivariate MR (MVMR) was used to assess the genetically proxied association between inflammatory factor levels and kidney stones to determine whether the genetically proxied causal relationship between target gene expression levels and kidney stones is mediated by inflammatory factors. We further conducted a sensitivity analysis to confirm the identified MR associations, such as pleiotropy assessment, colocalization, weak instrument variable bias assessment, and repeated evaluations using multiple analysis methods.
Fig. 1
Study design. First, NSAIDs and their target genes were identified. eQTL data were used to find instrumental variables. SMR analysis was performed to verify the associations between genetic factors and inflammatory factors (C-reactive protein, IL-1α, IL-1β, IL-6, PGE2, and TNF-α) and to screen target genes. The causal relationship between gene expression and kidney stones was assessed in the GWAS by SMR. MVMR was used to determine whether the causal relationship is mediated by inflammatory factors. Sensitivity analysis (pleiotropy assessment, colocalization, weak instrument variable bias assessment, and multiple methods) was conducted
We identified NSAIDs, including butylpyrazolidines, acetic acid derivatives and related substances, oxicams, propionic acid derivatives, fenamates, coxibs and others, from the World Health Organization Collaborating Centre for Drug Statistics Methodology (https://www.whocc.no/). We subsequently determined the target genes of these NSAIDs via the DrugBank (Wishart et al. 2006) (https://go.drugbank.com/) database. (Supplementary Table S1a).
Genetic instruments for NSAID target gene expression
To determine the genetic variations related to the expression of NSAID targets, we used a publicly available dataset from the eQTLGen consortium (https://www.eqtlgen.org/) to identify single nucleotide polymorphisms (SNPs) with a minor allele frequency > 0.01. The eQTL SNPs used for this analysis are from cis-regulatory regions (1 MB on each side of the gene). The default p value is 5.0e−8. The eQTLGen consortium is an international research collaboration project aimed at studying the association between gene expression regulatory elements and genetic variations. The consortium integrates data from multiple research institutions and datasets and includes more than 50,000 samples in total.
Anzeige
Outcome GWAS
Kidney stone outcomes were based on kidney stone GWAS summary statistics from the UK Biobank, FinnGen, and a meta-analysis of these two GWAS datasets. The UK Biobank (Sudlow et al. 2015) is a biomedical database and research resource containing genetic, lifestyle, and health information from approximately 500,000 participants aged 40–69 years in the UK (Bycroft et al. 2018). We obtained UK Biobank data from individuals of European ancestry, which involved the phenotype feature "non-cancer” illness code, self-reported: kidney stone/ureter stone/bladder stone" from the Pan-UKBB project, totaling 5530 cases and 415,001 controls. The summary statistics data were downloaded from the Pan-UKBB project portal website (Pan UKBB Team 2024). FinnGen (https://www.finngen.fi/en) is a research project in genomics and personalized medicine. It is a large public–private partnership that has collected and analysed genome and health data from 500,000 Finnish biobank donors (Kurki et al. 2023). FinnGen release R10 has a total sample size of 412,181, consisting of 230,310 females and 181,871 males; it involves the analysis of a total of 21,311,942 variants and provides 2,408 disease endpoints (phenotypes). The GWAS summary statistics for kidney stones were identified using the phenotype N14 calculus of the kidney and ureter, which comprised 10,556 cases and 400,681 controls. The meta-analysis was conducted using PLINK (Purcell et al. 2007) with fixed effects. The resulting summary statistics were employed to carry out MR analyses.
Statistical analyses
SMR analysis
The principle of the SMR method (Zhu et al. 2016) is to explore the causal relationship between gene expression and specific results. Since the causal relationship cannot be determined directly by changing gene expression and observing the results, the SMR method utilizes existing results from GWAS and eQTL studies. Among them, the key genetic tool is usually a SNP. Assuming that genotypes affect phenotypic characteristics, the set result is “y” (such as a certain disease state), the gene expression factor is “x”, and the genetic tool is “g” (SNP). Through eQTL studies, the influence of the genetic tool “g” on gene expression “x” can be estimated. Through GWAS, the influence of the genetic tool “g” on the result “y” can be estimated. In the absence of nongenetic confounding factors, dividing the influence of the genetic tool “g” on the result “y” by its influence on gene expression “x”, the resulting ratio can reflect the causal effect size of gene expression changes on the outcome variable. (Supplementary Fig. S1). In short, the SMR method, with the help of genetic tools as a bridge, infers whether there is a true causal relationship between gene expression and specific results from existing data, avoiding the influence of other interfering factors and making the research results more reliable.
Validation of genetic instruments using inflammatory factor GWAS
To obtain valid instrumental variables for NSAIDs, we used target gene expression in blood (from eQTL data) with the exposure and the levels of the inflammatory factors C-reactive protein, interleukin-1 alpha, interleukin-1 beta, interleukin-6, prostaglandin E2, and TNF-α as the outcomes for SMR analysis. The levels of inflammatory factors were evaluated using GWAS summary data in individuals of European ancestry (Supplementary Table S1b). For SNPs that passed the significance threshold of the SMR test (p < 0.05), the heterogeneity in dependent instruments (HEIDI) test was performed, the correlation of the effect sizes of different genetic instrumental variables was calculated, and statistical tests were performed to determine whether significant heterogeneity was present. The existence of heterogeneity (p value is less than the threshold) may indicate the presence of horizontal pleiotropy. In this study, the p value threshold for the HEIDI test was set at 0.05.
MR analysis of genetically proxied blood target gene expression and kidney stone risk
We performed SMR analysis to assess genetically proxied associations between blood eQTL-derived NSAID target gene expression and kidney stone risk using GWAS data. As the SMR results are based on the association of top SNPs, SMR based on multiple SNPs (SMR-multi) was also performed by incorporating numerous related SNPs at eQTL loci into the SMR analysis to enhance the power of the test (Wu et al. 2018). In the SMR-multi test, the default value of 0.1 was used as the linkage disequilibrium r2 threshold to prune SNPs (eQTLs). To control the genome-wide type I error rate, Bonferroni correction was used to consider multiple tests.
Sensitivity analyses
Colocalization analysis
For statistically significant MR associations, whether genetically proxied drug target gene expression and the occurrence of kidney stones shared a common causal variant in a given region was examined by applying a Bayesian localization approach (Giambartolomei et al. 2014) to test the exclusion restriction assumption. Typically, a posterior probability greater than 0.80 is considered supportive of the existence of a common causal variant. In the coloc (v3.1) R package, genetic variants with statistically significant MR associations for gene expression and outcomes are evaluated using default parameters (Wang et al. 2020).
Horizontal pleiotropy assessment
In MR, genetic variants are used as instrumental variables to infer the causal effect of an exposure factor on an outcome variable. However, if the genetic instrumental variable has other independent pathways (i.e., horizontal pleiotropy) in addition to affecting the outcome variable through the exposure factor, then the results of MR may be biased. We used the HEIDl (Zhu et al. 2016) (heterogeneity in dependent instruments) test to compare whether the effect sizes of different genetic instrumental variables on the outcome variable were heterogeneous. If heterogeneity was present, it could possibly indicate the presence of horizontal pleiotropy. Specifically, the HEIDI test judges whether heterogeneity is significant by calculating the correlation of the effect sizes of different genetic instrumental variables and performing statistical tests. If the p value of the HEIDI test is less than the set threshold (set as 0.05 in this study), evidence of horizontal pleiotropy is usually present. In addition to applying the HEIDl test, we also applied the generalized MR-Egger regression method and used the pleiotropy test of the MR-Egger intercept term to quantify the level of pleiotropy in the MR analysis. The intercept of MR-Egger regression can be interpreted as the average horizontal pleiotropy of all instrumental variables (Burgess and Thompson 2017). An intercept that is not significantly different from zero may evidence the lack of horizontal pleiotropy (Bowden and Holmes 2019).
Investigating the impact of a weak instrumental variable
We examined the impact of weak instruments on our exposure. An F statistic that is less than 10 is usually considered a potentially weak instrument problem; that is, the genetic variations as instrumental variables cannot effectively represent the exposure factor (Burgess and Thompson 2011).
Reducing bias with multiple data sources
The summary GWAS statistics and meta-analysis results from the UK Biobank and FinnGen genome-wide association studies were used as the outcome data for the MR analysis of kidney stones, minimizing potential bias and enhancing the reliability and generalizability of the analysis results.
Estimating relationships and effects with multiple methods
To verify the genetically proxied causal relationship between drug target genes and kidney stones estimated using SMR, we repeated the estimation using five methods of two-sample MR, namely, MR-Egger (Bowden et al. 2015), weighted median, inverse variance weighting (Bowden et al. 2015), simple mode, and weighted mode, to verify the reliability of the results.
Test genetically proxied association between inflammatory factors and kidney stone risk
To determine whether the genetically proxied association between drug target gene expression and the risk of kidney stones is mediated by changes in inflammatory factor levels or whether it may be driven independently, MVMR (Burgess and Thompson 2015) was used to dissect the causal pathway. The GWAS data of inflammatory factors were used as exposures, and the results of a meta-analysis of GWAS from both the UK Biobank and FinnGen were used as the outcome for kidney stones.
Results
Selection and validation of the genetic instrument
A total of 31 commonly used NSAIDs were identified. As expected, some overlap was observed between drug categories and related targets. Therefore, a total of 54 target genes in the DrugBank database (Supplementary Table S2a) were identified. Among them, 39 genes were annotated as expressed in blood in eQTLGen (Supplementary Table S2b). Among these genes, 37 were strongly associated with eQTL SNPs in eQTLGen (F statistic > 10, P_eQTL < 5e–8), and the expression of 13 genes was associated with at least one inflammatory factor when P_SMR < 0.05. A total of 12 associations were confirmed by the HEIDI outlier test (p_HEIDI > 0.01). These 12 target genes were further considered in the MR analysis, with kidney stones as the outcome (Supplementary Table S3).
MR analysis of the association of genetically proxied target gene expression in the blood and the risk of kidney stones
In the MR genetically proxied association analysis of the expression of 12 target genes in blood obtained through screening and the risk of kidney stones, we identified a significant association for only NEU1. The significance threshold was p < 4.2 × 10⁻3 (0.05/12) after Bonferroni correction. A decrease in genetically proxied NEU1 expression by 1 standard deviation was associated with a decrease in TNF-α by 0.277 pg/mL. The p value of the HEIDI outlier test was 0.85, indicating no evidence of heterogeneity (Supplementary Table S3). A decrease in genetically proxied NEU1 expression was associated with an increased risk of kidney stones, with an odds ratio of 0.6 (95% CI 0.47–0.77) (Fig. 2). The p value of the HEIDI outlier test was 0.054, indicating no evidence of heterogeneity. In addition, the p value of the SMR test for kidney stone risk (3.72 × 10⁻5) was similar to that of the SMR multivariate test (9.03 × 10⁻5). These findings suggest that the results of the univariate SMR test and the multivariate SMR test for kidney stone risk were relatively consistent, implying that the genetically proxied relationship between the expression of NEU1 and kidney stone risk was stable and not strongly affected by the introduction of additional variables in the multivariate analysis. Notably, kidney stone outcomes were based on meta-analysis of kidney stone GWAS summary statistics from two GWAS datasets, the UK Biobank and FinnGen.
Fig. 2
Genetically proxied association of drug target gene expression in blood with kidney stone risk. In the MR analysis of the genetically proxied expression of 12 target genes related to blood and kidney stone risk, only NEU1 showed a significant genetically proxied association (Bonferroni-corrected p < 4.2 × 10⁻3). A 1 standard deviation decrease in NEU1 expression was associated with a 0.277 pg/mL decrease in TNF-α. Decreased genetically proxied NEU1 expression was associated with increased kidney stone risk (odds ratio = 0.6, 95% CI 0.47–0.77)
Since only NEU1 expression was significantly associated with the risk of kidney stones, we performed a colocalization analysis. We found that the probability (PP.H4) that SNPs associated with NEU1 and kidney stones are shared by the same causal variant was 98.5%, indicating that kidney stones and the genetically proxied drug target NEU1 were strongly colocalized.
Reducing bias with multiple data sources
We used the meta-analyzed GWAS data from the UK Biobank and FinnGen as the outcome dataset for MR analysis. We also conducted MR analysis using each of these two data sources separately. The results show that when the UK Biobank and FinnGen were used as outcomes for kidney stones, the analysis results were consistent. Supplementary Tables S4 and S5 present the relationships between 12 genetically proxied drug target genes and TNF-α and the GWASs of the two data sources and their meta-analyses for kidney stones, respectively.
Horizontal pleiotropy assessment
The p values of the HEIDI test for the relationship between NEU1 and TNF-α and the relationship between NEU1 and kidney stones were 0.054 (Table 1) and 0.85 (Supplementary Table S3), respectively, both of which are greater than the set threshold of 0.05. This finding indicates that in these two relationships, not enough evidence was available to suggest the existence of horizontal pleiotropy. That is, the genetic variation representing the drug target NEU1 mainly acts through the expected pathway and was unlikely to involve other independent pathways that affect the outcome variable. The significant genetically proxied relationship between the drug target NEU1 and the inflammatory factor TNF-α and kidney stones obtained by the SMR method is relatively reliable in the current analysis.
Table 1
Genetically proxied association drug target gene expression in blood with kidney stone risk
Gene
Probe-Chr
Probe_bp
topSNP
top-SNP_chr
topSNP_bp
Effect_allele
Other_allele
Freq_Effect_allele
eQTL association
Beta
Se
p value
SLC7A11
4
139124377
rs28364685
4
139164388
G
A
0.0457256
0.216242
0.0196773
4.30E−28
NEU1
6
31828059
rs3117576
6
31726794
C
T
0.0745527
−0.191577
0.0223862
1.15E−17
AKR1B1
7
134135569
rs73164856
7
134233851
T
C
0.392644
0.28106
0.00836027
8.99E−248
CA2
8
86384901
rs7814461
8
86958004
C
G
0.482107
− 0.350069
0.00774806
0.00E+00
MYC
8
128750677
rs10956401
8
129002419
A
G
0.385686
− 0.0808447
0.00909205
6.01E−19
RXRA
9
137270687
rs1045570
9
137332311
T
G
0.163022
0.0967898
0.0125388
1.17E−14
PTGDR2
11
60620928
rs530963
11
60617834
C
A
0.44831
0.113773
0.00893557
3.90E−37
ACAT1
11
108005373
rs4550189
11
107986344
G
A
0.415507
0.377586
0.0116896
6.77E−229
PTGES3
12
57069643
rs11835120
12
57049774
A
G
0.33996
0.0963568
0.00936885
8.25E−25
NFKBIA
14
35872336
rs72664840
14
35596323
T
C
0.151093
0.413327
0.0150032
4.52E−167
GP1BA
17
4836958
rs67059207
17
4869093
C
T
0.207753
0.0971946
0.0100672
4.70E−22
PCNA
20
5101435
rs16990612
20
5089494
G
A
0.0904573
− 0.175067
0.0135367
2.94E−38
Gene
GWAS associatione association
MR association
Multi_p
HEIDI Test
Beta
Se
p value
Beta
Se
p value
p value
Number of SNPs
SLC7A11
0.0016
0.0252126
9.49E−01
0.00739912
0.116596
9.49E−01
1.12E−01
0.542
10
NEU1
0.0971
0.02063
2.52E−06
−0.506846
0.122897
3.72E−05
9.03E−05
0.054
19
AKR1B1
0.0153
0.0125935
2.24E−01
0.0544368
0.0448365
2.25E−01
2.31E−04
0.091
20
CA2
0.0054
0.0124705
6.65E−01
− 0.0154255
0.0356247
6.65E−01
2.35E−01
0.151
20
MYC
− 0.0189
0.0129488
1.44E−01
0.233782
0.162312
1.50E−01
5.48E−01
0.154
17
RXRA
− 0.0001
0.0117335
9.93E−01
− 0.00103317
0.121226
9.93E−01
8.68E−01
0.494
8
PTGDR2
− 0.0233
0.0125949
6.43E−02
− 0.204794
0.111864
6.71E−02
1.46E−01
0.110
20
ACAT1
− 0.0033
0.0125879
7.93E−01
− 0.00873973
0.0333388
7.93E−01
2.92E−01
0.399
20
PTGES3
− 0.0036
0.0134655
7.89E−01
− 0.0373611
0.139794
7.89E−01
7.89E−01
0.451
20
NFKBIA
0.0192
0.0177172
2.79E−01
0.0464523
0.0428979
2.79E−01
7.73E−01
0.110
20
GP1BA
0.0256
0.0143625
7.47E−02
0.263389
0.150267
7.96E−02
1.97E−01
0.768
20
PCNA
0.031
0.0225385
1.69E−01
− 0.177075
0.129468
1.71E−01
1.71E−01
0.946
10
Bold in italics indicates significant association
eQTL expression quantitative trait locus, GWAS genome-wide association study, HEIDI heterogeneity in dependent instruments, MR Mendelian randomization, SNP single nucleotide polymorphism
We also applied the generalized MR-Egger regression method. Using the pleiotropy test of the MR Egger intercept term, we quantified the pleiotropy level of the genetically proxied relationship between NEU1 and kidney stones. The results revealed that the Egger intercept was 0.007 and the p value of the intercept term was 0.773 (Supplementary Table S6), indicating that no evidence of horizontal pleiotropy was observed; this means that the genetically proxied relationship between NEU1 and kidney stones obtained from MR analysis is relatively reliable and is unlikely to be significantly affected by horizontal pleiotropy.
Estimating genetically proxied relationships and effects with multiple methods
We employed five methods of 2-Sample MR (MR-Egger, weighted median, inverse variance weighting, simple mode, and weighted mode) to repeatedly estimate the genetically proxied relationship between NEU1 and kidney stones. The results showed that all the β values were negative and the p values were statistically significant (Supplementary Table S6). Moreover, the β value of the SMR method was also negative and statistically significant (Table 1). The SMR method integrates gene expression data and GWAS results, whereas the 2-sample MR method utilizes GWAS data from two independent samples. Both methods analysed the genetically proxied causal relationships between drug target genes and kidney stones from different data sources and perspectives. The consistent negative β values and significant p values from multiple methods indicate that the genetically proxied negative causal relationship between the drug target gene NEU1 and kidney stones was unlikely to be accidental and demonstrate that this causal relationship was stable under different data and analysis methods.
Test genetically proxied association between inflammatory factors and kidney stone risk
Sensitivity analysis was carried out to determine whether the genetically proxied association between NEU1 expression and kidney stone risk is mediated by changes in TNF-α levels or whether this association is independently driven. MVMR was used to dissect the causal pathway. The results indicate that no significant genetically proxied association was observed between TNF-α and kidney stones (Supplementary Table S5). Based on the results of the above MR analysis, genetically proxied NEU1 expression was significantly associated with both TNF-α levels and kidney stone risk; this implies that the genetically proxied association between NEU1 expression and kidney stone risk may not be mediated by TNF-α levels and may be independently driven.
Based on the results of MR analysis and sensitivity analysis, the overall results of this study are presented in Fig. 3. We verified the effectiveness of the genetic instrument of NEU1 by demonstrating a significant correlation between NEU1 and TNF-α levels. Additionally, through SMR analysis, we found a significant genetically proxied negative correlation between the expression of NEU1 and the risk of kidney stones. Moreover, this relationship is unlikely to be mediated by changes in the level of TNF-α.
Fig. 3
Schematic representation of the research results. The effectiveness of the NEU1 genetic instrument was verified by its significant correlation with TNF-α. Summary data-based Mendelian randomization analysis revealed a significant negative genetically proxied correlation between NEU1 expression and kidney stone risk, which was unlikely to be mediated by changes in TNF-α levels
Exploratory ethnic-specific analyses of NSAID target genes
We performed SMR analysis on the kidney stone GWAS datasets of South Asian, Hispanic or Latin American, and African American or Afro-Caribbean ethnic groups. The results revealed that an increase in the genetically proxied PTGDR2 expression was significantly associated with an increased risk of kidney stones in the South Asian population, with an odds ratio of 23.01 and a p value of 1.28E−03 (Supplementary Table S7), suggesting that associations may vary by ethnic group. However, the extreme magnitude of this odds ratio should be interpreted with caution.
The role of drug metabolites in the genetically proxied relationship between NEU1 and kidney stones
Previous studies have shown that drugs can induce stone formation through their metabolic effects on calcium, oxalate, phosphate, uric acid, or purine metabolism, or by altering urine (Daudon et al. 2018; Khan et al. 2016). To investigate whether such mechanisms might explain the observed genetically proxied association between NEU1 and kidney stones, we analyzed metabolites of NEU1-targeting NSAIDs (aspirin and celecoxib). Among all known metabolites (aspirin: salicylic acid, salicyluric acid, the ether or phenolic glucuronide and the ester or acyl glucuronide; celecoxib: hydroxycelecoxib, carboxycelecoxib and celecoxib glucuronide), only salicylic acid had available GWAS data. We found that genetically proxied NEU1 expression was positively associated with salicylic acid levels (β = 0.18, p = 4.2 × 10⁻3), but no significant association was observed between salicylic acid and kidney stone risk (Supplementary Fig. S2).
These results indicate that although NEU1 is involved in aspirin metabolism, the aspirin metabolite salicylic acid does not appear to mediate the genetically proxied association between NEU1 and kidney stones.
Discussion
In this study, our results indicate that lower expression of genetically proxied NEU1 was associated with a greater risk of kidney stones. We also explored whether the association between genetically proxied NEU1 expression and kidney stone risk could be influenced by the level of inflammatory factors, specifically TNF-α, which is linked to NEU1. However, we found no significant relationship between TNF-α and the use of NSAIDs. This MR study provides further evidence supporting the idea that NSAIDs may contribute to the development of kidney stones.
Approximately 10% of the global population experiences chronic pain, leading to a high demand for NSAIDs (Sinniah et al. 2021). Aspirin is the most widely used NSAID, with increasing global usage (Ittaman et al. 2014). In addition to traditional uses, NSAIDs have been repurposed for various medical purposes (Bindu et al. 2020; Sousa et al. 2023). Given the widespread use of NSAIDs, even minor adverse reactions can have significant social impacts. Evaluating the risk–benefit balance of NSAIDs is essential for strategic pain and inflammation management (Mazumder et al. 2024).
Previous studies have linked nimesulide to acute reversible renal failure and the presence of calcium oxalate crystals in renal tubular cells. While this finding points to a potential connection between NSAIDs and kidney stones, limited research in this area is attributed to challenges in diagnosing NSAID-induced kidney stones. These stones typically form when drugs or their byproducts become part of the calculi or disrupt metabolic processes essential for stone formation. Identifying the latter mechanism is difficult, as these stones are similar in composition to common calculi. More importantly, the hidden nature of kidney function and the routine use of NSAIDs make it difficult for most studies to prove that exposure to NSAIDs occurs before the onset of kidney disease. Consequently, any associations drawn from epidemiological data are more likely coincidental than causal. In addition, most studies have focused mainly on renal function and end-stage renal disease, which has led to the neglect of kidney stones. The situation is further complicated when patients receive multiple potentially stone-causing drugs concurrently or when kidney stones are discovered years after treatment initiation or drug cessation. The limitations of conventional clinical studies highlight the need for innovative research approaches to better understand the relationship between NSAIDs and kidney stones.
The impact of genetic variants used as instruments in MR design is established at conception, enabling MR studies to assess the prolonged effects of the exposure in question on outcome risk (Chauquet et al. 2021; Ference et al. 2021; Gill et al. 2020). The features of MR make it uniquely advantageous for studying long-term adverse drug reactions. In this study, we used a two-sample MR method to find a genetically proxied association between decreased expression of NEU1 and increased risk of kidney stone (OR = 0.6, 95% CI 0.47–0.77), indicating that NSAIDs targeting NEU1 gene expression may increase the risk of kidney stone. The OR value of 0.6 is a noteworthy result. Although MR cannot replace clinical trial evidence, it provides additional evidence for studying the relationship between NSAIDs and kidney stones. This direction deserves further in-depth research.
The NEU1 gene is a target of aspirin and celecoxib and reduces the expression of NEU1. NEU1 is one of the four mammalian neuraminidase isoenzymes involved in cellular signalling during the immune response process (Bonten et al. 2014; Seyrantepe et al. 2010; Sieve et al. 2018a, b; Sieve et al. 2018a, b). In mammals, NEU1 has been shown to be involved in diverse physiological and pathological processes (Lillehoj et al. 2022; Lorenz et al. 2021). Previous studies have shown that NEU1 is significantly upregulated in the fibrotic kidneys of patients and mice (Chen et al. 2023). In addition, NEU1 is upregulated in various types of chronic kidney disease, including IgA nephropathy, diabetic nephropathy, lupus nephritis, focal segmental glomerulosclerosis, and membranous nephropathy (Chen et al. 2023). NEU1 can regulate the activation of toll-like receptors (TLRs) on macrophages and lead to the production of nitric oxide and pro-inflammatory cytokines (Abdulkhalek et al. 2011). The formation of kidney stones is related to oxidative stress and immune inflammation (Joshi et al. 2015a, b; Joshi et al. 2015a, b; Khan et al. 2016; Taguchi et al. 2016). Macrophages, which can induce inflammation, engulf crystals, and exert anti-adhesion effects, are crucial for the continuous crystal formation process (Dominguez-Gutierrez et al. 2018; Taguchi et al. 2014). The infiltration and polarization of macrophages in the development of kidney stones have been confirmed in some studies (Taguchi et al. 2021; Xi et al. 2019). Cell surface NEU1 activates phagocytosis in macrophages and dendritic cells through desialylation of surface receptors, thus contributing to their functional integrity (Seyrantepe et al. 2010). Moreover, NEU1 promotes TGF-β-induced epithelial-mesenchymal transition, which affects the composition of the extracellular matrix. Alterations in the extracellular matrix can promote the adhesion and aggregation of crystals, leading to their retention in renal tissue (Chen et al. 2023). Here, our results revealed that decreased genetically proxied expression of NEU1 could increase the risk of kidney stone. We speculated that aspirin may increase the risk of kidney stones by targeting NEU1. Although our study focused on NSAIDs, other non-NSAIDs, such as oseltamivir (an antiviral drug), have been shown to regulate the expression of NEU1. Oseltamivir specifically inhibits the activity of NEU1, which is involved in the desialylation of glycoproteins and plays a key role in the immune response and inflammation (Abdulkhalek et al. 2011). Given the role of NEU1 in inflammation and its association with the risk of kidney stones, it is reasonable to speculate that oseltamivir may have a potential impact on the risk of kidney stone formation by regulating the expression of NEU1.
Anzeige
Previous studies have shown that cytokines and inflammatory substances trigger a cascade reaction promoting crystal formation in the kidney (Wang et al. 2024). Excessive ROS production triggers the NF-κB signalling cascade, inducing inflammation and crystal aggregation, ultimately leading to the formation of kidney stones (Huang et al. 2008; Khan and Thamilselvan 2000). Inhibiting NLRP3 upstream can reduce stone formation and recurrence by decreasing inflammation and crystal damage in renal tubular epithelial cells (So et al. 2018). In our study, we analysed 12 target genes related to inflammatory factors. The goal of this study was to analyse whether NSAIDs cause kidney stones through inflammatory factors. The NEU1 gene is significantly associated with TNF-α through a genetically proxied association, but there is no evidence of a genetically proxied causal relationship between genetically estimated TNF-α and the risk of kidney stones. NSAIDs may cause kidney stones through other pathways.
Our study has several limitations. First, we did not estimate any genetically proxied associations between nonsteroidal anti-inflammatory drugs and other inflammatory factors or subtypes of kidney stones. Further research is essential to explore these potential relationships. Second, by utilizing data from participants of British white ethnicity in the UK Biobank and data of European ancestry from the FinnGen study, the generalizability of our results may be restricted. The usage rates of NSAIDs vary across different countries. In some countries, physiotherapists are authorized to prescribe such drugs to patients, whereas in other countries, only doctors have the authority to directly prescribe these medications to patients (Bissell et al. 2008; Shaikh et al. 2021). To some extent, the differences in prescription regulations among countries have affected the prescription rates of NSAIDs. The usage frequency, dosage, and duration of NSAID use vary, and these factors may have different impacts on kidney stones. In addition, there are differences in genetic composition and epigenetic modifications among different ethnic groups. These differences are likely to result in varying expression levels of NEU1. Third, genetic variation reflects the impact of changes in drug target expression on the risk of kidney stones. However, it cannot be directly compared with pharmacological inhibition. In MR, lifelong exposure differs from the shorter duration of pharmacological intervention. Fourth, although our MVMR analysis did not reveal a significant genetically proxied association between TNF-α and the risk of kidney stones, we acknowledge that this conclusion is based on the negative results of an underpowered analysis (the calculated statistical power of this model was 0.2834), indicating that our study may lack sufficient power to detect a true association. Therefore, we urge caution in interpreting this result. Fifth, due to the lack of sex-stratified data in the kidney stone GWAS dataset used in this study, we were unable to evaluate the sex-specific effects of the genetically proxied association between NEU1 and kidney stones, restricting our ability to determine whether sex differences have a role in this association. Sixth, although various sensitivity analyses were conducted to test the assumption of MR analysis, the instrumental variable assumption cannot be empirically verified. Pleiotropy or confounding factors may potentially bias the current estimates. Finally, although our MR analysis supports the genetically proxied association between NEU1 inhibition and kidney stones, the specific involvement of crystal-pathway mechanisms remains hypothetical. This is because GWAS data lack intermediate phenotypes of crystal deposition, and MR cannot resolve the spatiotemporal dynamics of pathological processes (Burgess and Thompson 2017).
However, our study has several unique strengths. Human genetics continuously provides evidence for new drug discovery and drug safety. We used eQTL genetic variants as MR instrumental variables. Multiple datasets were exploited to investigate causality, and observational study and RCT limitations were avoided by using a two-sample MR design. By integrating multiple datasets, our results can more comprehensively reflect the relationship between NSAIDs and kidney stone risk in different populations. To some extent, the MR method compensates for the inability to use traditional RCTs for research; this provides new insights from a genetic perspective. Our findings suggest the need for the cautious use of NSAIDs in individuals at high risk of kidney stones. When prescribing NSAIDs, clinicians must carefully weigh the therapeutic benefits against the potential risk of kidney stones. For patients using NSAIDs, especially those with known risk factors for kidney stones, regular monitoring of renal function, including serum creatinine levels and urinalysis, is essential. Additionally, for individuals at high risk of kidney stones, clinicians should consider alternative analgesics that do not target NEU1 expression. When prescribing NSAIDs, it is important to carefully assess the dosage and frequency of use based on the individual patient's circumstances to minimize the risk of kidney stones.
Conclusion
In this study, we found that the risk of kidney stones was negatively genetically proxied associated with the expression of NEU1. Moreover, the genetically proxied association between NEU1 and kidney stone was likely independent of the inflammatory factor TNF-α. Our research provides some insights and possible directions for further exploration of the relationship between NSAIDs and kidney stones.
Anzeige
Acknowledgements
None.
Declarations
Conflict of interest
The authors announce that there are no conflict of interests.
Consent for publication
No content in this manuscript requires consent for publication.
Ethics approval and consent to participate
Publicly available GWAS data were used for this study. Relevant informed consent and ethical approval were obtained for the original study. Therefore, additional ethical approval was not required in the present study.
Anzeige
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Abdulkhalek S, Amith SR, Franchuk SL et al (2011) Neu1 sialidase and matrix metalloproteinase-9 cross-talk is essential for Toll-like receptor activation and cellular signaling. J Biol Chem 286(42):36532–36549. https://doi.org/10.1074/jbc.M111.237578CrossRefPubMedPubMedCentral
Aliyu M, Zohora FT, Anka AU, Ali K, Maleknia S, Saffarioun M, Azizi G (2022) Interleukin-6 cytokine: an overview of the immune regulation, immune dysregulation, and therapeutic approach. Int Immunopharmacol 111:109130. https://doi.org/10.1016/j.intimp.2022.109130CrossRefPubMed
Bowden J, Davey Smith G, Burgess S (2015) Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol 44(2):512–525. https://doi.org/10.1093/ije/dyv080CrossRefPubMedPubMedCentral
Chauquet S, Zhu Z, O’Donovan MC, Walters JTR, Wray NR, Shah S (2021) Association of antihypertensive drug target genes with psychiatric disorders: a mendelian randomization study. JAMA Psychiatr 78(6):623–631. https://doi.org/10.1001/jamapsychiatry.2021.0005CrossRef
Crorkin P, Hao S, Ferreri NR (2022) Responses to Ang II (angiotensin II), salt intake, and lipopolysaccharide reveal the diverse actions of TNF-α (tumor necrosis factor-α) on blood pressure and renal function. Hypertension 79(12):2656–2670. https://doi.org/10.1161/hypertensionaha.122.19464CrossRefPubMed
Daudon M, Frochot V, Bazin D, Jungers P (2018) Drug-induced kidney stones and crystalline nephropathy: pathophysiology, prevention and treatment. Drugs 78(2):163–201. https://doi.org/10.1007/s40265-017-0853-7CrossRefPubMed
Galozzi P, Bindoli S, Doria A, Sfriso P (2021) The revisited role of interleukin-1 alpha and beta in autoimmune and inflammatory disorders and in comorbidities. Autoimmun Rev 20(4):102785. https://doi.org/10.1016/j.autrev.2021.102785CrossRefPubMed
Giambartolomei C, Vukcevic D, Schadt EE, Franke L, Hingorani AD, Wallace C, Plagnol V (2014) Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet 10(5):e1004383. https://doi.org/10.1371/journal.pgen.1004383CrossRefPubMedPubMedCentral
Gill D, Walker VM, Martin RM, Davies NM, Tzoulaki I (2020) Comparison with randomized controlled trials as a strategy for evaluating instruments in Mendelian randomization. Int J Epidemiol 49(4):1404–1406. https://doi.org/10.1093/ije/dyz236CrossRefPubMed
Huang HS, Ma MC, Chen J (2008) Chronic L-arginine administration increases oxidative and nitrosative stress in rat hyperoxaluric kidneys and excessive crystal deposition. Am J Physiol Renal Physiol 295(2):F388-396. https://doi.org/10.1152/ajprenal.00405.2007CrossRefPubMed
Huang K, Peng Z, Zha C, Li W, Deng G, Chen X, Luo Y, Ji Z, Wang Q, Jiang K (2024) Inflammatory factors and the risk of urolithiasis: a bidirectional Mendelian randomization study. Front Med (Lausanne) 11:1432275. https://doi.org/10.3389/fmed.2024.1432275CrossRefPubMed
Joshi S, Wang W, Peck AB, Khan SR (2015b) Activation of the NLRP3 inflammasome in association with calcium oxalate crystal induced reactive oxygen species in kidneys. J Urol 193(5):1684–1691. https://doi.org/10.1016/j.juro.2014.11.093CrossRefPubMed
Khan SR, Thamilselvan S (2000) Nephrolithiasis: a consequence of renal epithelial cell exposure to oxalate and calcium oxalate crystals. Mol Urol 4(4):305–312PubMed
Lorenz L, Amann B, Hirmer S, Degroote RL, Hauck SM, Deeg CA (2021) NEU1 is more abundant in uveitic retina with concomitant desialylation of retinal cells. Glycobiology 31(7):873–883. https://doi.org/10.1093/glycob/cwab014CrossRefPubMed
Mazumder S, Bindu S, Debsharma S, Bandyopadhyay U (2024) Induction of mitochondrial toxicity by non-steroidal anti-inflammatory drugs (NSAIDs): the ultimate trade-off governing the therapeutic merits and demerits of these wonder drugs. Biochem Pharmacol 228:116283. https://doi.org/10.1016/j.bcp.2024.116283CrossRefPubMed
Seyrantepe V, Iannello A, Liang F, Kanshin E, Jayanth P, Samarani S, Szewczuk MR, Ahmad A, Pshezhetsky AV (2010) Regulation of phagocytosis in macrophages by neuraminidase 1. J Biol Chem 285(1):206–215. https://doi.org/10.1074/jbc.M109.055475CrossRefPubMed
Shaikh S, Tharani R, Saad Khan M, Chughtai MRB, Alam B (2021) Physiotherapists’ knowledge, usage and attitude towards non-steroidal anti-inflammatory drugs (NSAIDs) in Karachi. Pakistan Int J Risk Saf Med 32(2):109–121. https://doi.org/10.3233/jrs-202007CrossRefPubMed
Sieve I, Ricke-Hoch M, Kasten M, Battmer K, Stapel B, Falk CS, Leisegang MS, Haverich A, Scherr M, Hilfiker-Kleiner D (2018b) A positive feedback loop between IL-1β, LPS and NEU1 may promote atherosclerosis by enhancing a pro-inflammatory state in monocytes and macrophages. Vascul Pharmacol 103–105:16–28. https://doi.org/10.1016/j.vph.2018.01.005CrossRefPubMed
Sousa SM, Xavier CPR, Vasconcelos MH, Palmeira A (2023) Repurposing some of the well-known non-steroid anti-inflammatory drugs (NSAIDs) for cancer treatment. Curr Top Med Chem 23(13):1171–1195. https://doi.org/10.2174/1568026623666230130150029CrossRefPubMed
Triozzi JL, Hsi RS, Wang G, Akwo EA, Wheless L, Chen HC, Tao R, Ikizler TA, Robinson-Cohen C, Hung AM (2023) Mendelian randomization analysis of genetic proxies of thiazide diuretics and the reduction of kidney stone risk. JAMA Netw Open 6(11):e2343290. https://doi.org/10.1001/jamanetworkopen.2023.43290CrossRefPubMedPubMedCentral
Van der Niepen P, Janssen van Doorn K, Van den Houte K, Verbeelen D (2002) Nimesulide and acute renal failure caused by oxalate precipitation. Nephrol Dial Transplant 17(2):315–316. https://doi.org/10.1093/ndt/17.2.315CrossRefPubMed
Wang G, Sarkar A, Carbonetto P, Stephens M (2020) A simple new approach to variable selection in regression, with application to genetic fine mapping. J R Stat Soc Series B Stat Methodol 82(5):1273–1300. https://doi.org/10.1111/rssb.12388CrossRefPubMedPubMedCentral
Wishart DS, Knox C, Guo AC, Shrivastava S, Hassanali M, Stothard P, Chang Z, Woolsey J (2006) DrugBank: a comprehensive resource for in silico drug discovery and exploration. Nucleic Acids Res 34(Database issue):D668-672. https://doi.org/10.1093/nar/gkj067CrossRefPubMed
Xi J, Chen Y, Jing J, Zhang Y, Liang C, Hao Z, Zhang L (2019) Sirtuin 3 suppresses the formation of renal calcium oxalate crystals through promoting M2 polarization of macrophages. J Cell Physiol 234(7):11463–11473. https://doi.org/10.1002/jcp.27803CrossRefPubMed
Yin P, Anttila V, Siewert KM, Palotie A, Davey Smith G, Voight BF (2017) Serum calcium and risk of migraine: a Mendelian randomization study. Hum Mol Genet 26(4):820–828. https://doi.org/10.1093/hmg/ddw416CrossRefPubMed
Zeng G, Mai Z, Xia S et al (2017) Prevalence of kidney stones in China: an ultrasonography based cross-sectional study. BJU Int 120(1):109–116. https://doi.org/10.1111/bju.13828CrossRefPubMed
Zhu Z, Zhang F, Hu H et al (2016) Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nat Genet 48(5):481–487. https://doi.org/10.1038/ng.3538CrossRefPubMed
