Serum uric acid levels and risk of gastrointestinal cancers: a comprehensive Mendelian randomization study

doi:10.21203/rs.3.rs-3407808/v1

Download PDF

Research Article

Serum uric acid levels and risk of gastrointestinal cancers: a comprehensive Mendelian randomization study

https://doi.org/10.21203/rs.3.rs-3407808/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

There has been an ongoing discourse regarding the correlation between serum uric acid (UA) levels and the risk of cancer. Observational studies have presented conflicting findings. This study aims to investigate the potential causal connection between UA levels and five types of gastrointestinal (GI) cancers, including esophageal, gastric, liver, pancreatic, and colorectal cancers.

Methods

We identified 12 single nucleotide polymorphisms (SNPs) significantly linked to UA levels from publicly accessible genome-wide association study (GWAS) data as instrumental variables (IVs). Genetic associations for the five GI cancers were obtained from the FinnGen and UK Biobank studies. Two-sample Mendelian randomization (MR) techniques were employed to evaluate the influence of UA levels on GI cancer. The primary analysis was conducted using the inverse variance-weighted (IVW) approach, and these results were corroborated with four additional methods, namely weighted median, MR-Egger, MR-PRESSO, and MR-Steiger. Multiple sensitivity analyses were carried out to ensure the robustness of our MR findings. Furthermore, a fixed-effects model meta-analysis was employed to amalgamate MR estimates from various data sources.

Results

Our primary analysis disclosed no significant correlation between UA levels and the risk of the five GI cancers (all P values > 0.05). Both sensitivity analyses and supplementary examinations yielded consistent outcomes. We did not observe substantial heterogeneity or horizontal pleiotropy.

Conclusion

This study constitutes the inaugural systematic appraisal of the causal nexus between UA levels and GI cancers utilizing meta-analysis within the MR framework. Our results do not provide substantial evidence for a notable impact of UA levels on the risk of GI cancers. These findings offer valuable insights for the clinical prevention and management of GI cancers.

Uric acid

Gastrointestinal cancers

Mendelian randomization

Genomic association studies

Meta-analysis

GI cancers mainly include esophageal, gastric, liver, pancreatic, and colorectal cancers (Arnold et al. 2020). According to the Global Cancer Statistics report released by the International Agency for Research on Cancer (IARC) in 2020, GI cancers account for more than one quarter of new cancer cases and more than one third of new cancer deaths. They are among the most common and deadly cancers worldwide(Sung et al. 2021). To effectively prevent and treat GI cancers and reduce their threat to human health, it becomes imperative to investigate the molecular mechanisms underlying their occurrence and progression.

UA is the end product of purine degradation, with its concentration in the blood normally ranging between 178 and 360 µmol/L (3 to 6.8 mg/dL). The association between serum UA levels and the risk of GI cancers exhibits complexity. On one hand, as a robust antioxidant accounting for 50% of the total antioxidant capacity in human biological fluids, UA may exert a protective role against cancers(Ames et al. 1981; Ndrepepa 2018). On the other hand, in an inflammatory microenvironment, UA can contribute to the formation and progression of GI tumors due to its proinflammatory effects (Oshima and Oshima 2012). The underlying mechanism may involve the generation of reactive oxygen species (ROS), the activation of the inflammasome, and the production of free radicals mediated by xanthine oxidase (XO) (Chen and Meng 2022; Gasse et al. 2009; Komori et al. 2018; Tsukamoto et al. 2021).

Numerous previous reports have indicated an association between UA and GI cancers. Nevertheless, the findings in these reports exhibit inconsistencies. For example, a cohort study involving a male population revealed that elevated levels of serum UA were linked to a reduced risk of mortality from various cancer types. Notably, this association remained statistically significant even after adjusting for serum total cholesterol and triglyceride levels(Taghizadeh et al. 2014). However, other study has found a positive correlation between high serum UA levels and the majority of tumors(Chen et al. 2014).

MR as a novel epidemiological method has been widely applied in various research fields in recent years(Bourgault et al. 2023; Lin et al. 2023; Xie et al. 2023; Yao et al. 2022). The method employs genetic variations associated with exposure factors, such as SNPs, as IVs to study the causal impact of exposure factors on specific outcomes(Davies et al. 2018). This approach offers two distinct advantages: firstly, the random assignment of SNP alleles at conception provides a potential avenue for mitigating confounding effects that are inherent in traditional observational studies; secondly, the MR design inherently guards against reverse causality because germline genotypes remain impervious to alterations induced by disease (Pingault et al. 2018).

In this study, we utilize aggregated statistics derived from publicly accessible genome-wide association studies (GWAS) data to investigate the potential causal link between serum UA levels and five GI cancers using a two-sample MR analysis. This research bears substantial clinical and public health implications in its capacity to enhance the efficacy of prevention and treatment of GI cancers, thereby diminishing their impact on human health.

Study design

We conducted a two-sample MR analysis to investigate the causal effect of serum UA levels on GI cancers. The causal relationships between serum UA levels and each specific type of GI cancer were assessed separately. The associations were then integrated using a fixed-effects model meta-analysis. The data used in our analysis comprised publicly available summary-level data from various sources, including GWAS, the FinnGen consortium, and the UK Biobank study. To minimize confounding resulting from population stratification, we focused specifically on non-overlapping populations of European descent. We selected SNPs that were significantly associated with serum UA levels in published GWAS as IVs. Our two-sample MR analysis was based on three key assumptions: (1) the relevance assumption, which posits a strong correlation between the IVs and serum UA levels; (2) the independence assumption, which assumes that the IVs are independent of potential confounding factors; and (3) the exclusion restriction assumption, which posits that the IVs only affect GI cancers through serum UA levels, with no involvement of other pathways(Davies et al. 2018). Figure 1 illustrates the research design flow chart.

Selection of IVs

We obtained GWAS data on serum UA levels from the Global Urate Genetics Consortium (GUGC) in the IEU Open GWAS project database(https://gwas.mrcieu.ac.uk/). The GWAS data were derived from a large-scale meta-analysis of 110,347 European individuals (49,825 women and 60,522 men)(Köttgen et al. 2013). We selected SNPs that exhibited significant genome-wide associations (P < 5×10 − 8) with the exposure variable, serum UA level, as IVs. We calculated the linkage disequilibrium (LD) coefficient r² between SNPs using LDlink in the European 1000 Genomes Project reference panel and applied strict criteria (r²<0.001; Cluster window, 10,000 kb) to screen for independent SNPs. Additionally, we utilized PhenoScanner V2 to search for SNPs and excluded those associated with confounding factors (P < 1 × 10 − 5). These potential confounders included dietary factors such as coffee intake, processed meat intake, bread intake, variation in diet, and alcohol intake frequency, as well as obesity-related traits including obesity, body mass index (BMI), weight, waist circumference, whole body fat mass, trunk fat mass, left and right arm fat mass, and left and right leg fat mass(Ni et al. 2022). To exclude weak IVs, we calculated F-statistics using the formula:\(F=\left(\frac{beta}{se}\right)^2\)(Huang et al. 2021). We retained only SNPs with an F-statistic above 10. We harmonized the selected SNPs across data sets from different studies related to the outcome variable GI and excluded palindromic SNPs with intermediate allele frequencies. Ultimately, we identified 12 SNPs as IVs, as shown in Table 1.

Table 1

Characteristics of SNPs included in the two-sample Mendelian randomization analysis
SNP	Gene	EA	OA	EAF	Beta	SE	P value	F-statistic
rs10761587	A1CF	C	T	0.8845	-0.062	0.0091	9.58E-11	46.42
rs11722228	SLC2A9	T	C	0.3476	0.210	0.0056	1.00E-200	1406.25
rs1394125	UBE2Q2	A	G	0.3425	0.043	0.0063	9.78E-11	46.59
rs17050272	AC073257.2	A	G	0.4499	0.037	0.0061	9.36E-09	36.79
rs17632159	TMEM171	C	G	0.2924	-0.038	0.0061	2.00E-09	38.81
rs1825043	TAPT1-AS1	A	G	0.9540	0.100	0.0130	6.67E-14	59.17
rs2307394	ORC4	C	T	0.3241	0.035	0.0057	7.26E-09	37.70
rs675209	TMEM14C	C	T	0.7301	-0.063	0.0062	1.38E-21	103.25
rs6770152	RP11-894J14.5	T	G	0.5849	-0.048	0.0056	2.66E-16	73.47
rs6830367	CLNK	C	G	0.1943	0.051	0.0072	2.35E-11	50.17
rs7224610	HLF	A	C	0.5686	-0.038	0.0055	4.74E-11	47.74
rs7654258	TAPT1-AS1	T	C	0.9151	-0.099	0.0100	1.09E-19	98.01

SNP, single-nucleotide polymorphism; EA, Effect allele; OA, Other allele; EAF, Effect allele frequency; SE, Standard error.

Data sources for GI cancers

We conducted a genetic susceptibility analysis of five GI cancers, including esophageal cancer, gastric cancer, liver cancer, pancreatic cancer, and colorectal cancer, using GWAS data from the FinnGen consortium and the UK Biobank study.

The FinnGen consortium is a large cohort study that collects health and genetic information using the Finnish health registry system. We utilized the latest released R9 data (https://r9.finngen.fi/), which included a total of 377,277 participants. We performed genome-wide association analyses for each outcome variable, adjusting for gender, age, genotype batch, and the top ten principal components. Outcomes related to GI cancers were determined through codes derived from the International Classification of Diseases (ICD-8, ICD-9, and ICD-10), along with surgical and medicine purchase codes. In our study, we included 566 esophageal cancer cases, 1307 gastric cancer cases, 648 liver cancer cases, 1416 pancreatic cancer cases, and 6509 colorectal cancer cases.

The UK Biobank study is a prospective cohort study that recruited approximately 500,000 adults between 2006 and 2010. Diagnosis of GI cancers relied on codes derived from ICD-9 and ICD-10, as well as surgical procedures and self-reported information. The data utilized in this study were sourced and curated by the Lee Laboratory for Statistical Genetics and Data Science at Seoul National University, Seoul, Republic of Korea (https://www.leelabsg.org/resources). These data have been subjected to adjustments for factors including gender, birth year, genotype batch, and the top four principal components. The study sample consisted of 720 esophageal cancer cases, 554 gastric cancer cases, 4,562 liver cancer cases, 589 pancreatic cancer cases, and 344 colorectal cancer cases. For a more comprehensive understanding of the data sources and definitions of outcome variables, please consult Tables 2 and Supplementary Table 1.

Tables 2 Summary information of the dataset for this study
Exposure/Outcome	Traits	Consortium	Participants	Adjustments	Ethnicity	Web source/Pubmed ID
Exposures	Serum uric acid	GUGC	110,347		European	23263486
Outcomes	Esophageal cancer	FinnGen	410 cases and 308,744 controls	Sex, age, genotyping batch and the first ten genetic principal components	European	FinnGen https://www.finngen.fi/fi
		UK Biobank	720 cases and 393,372 controls	Sex, birth year, genotyping batch and first four principal components	European	UK Biobank https://www.leelabsg.org/resources
	Gastric cancer	FinnGen	1307cases and 287137 controls	Sex, age, genotyping batch and the first ten genetic principal components	European	FinnGen https://www.finngen.fi/fi
		UK Biobank	554 cases and 393,372 controls	Sex, birth year, genotyping batch and first four principal components	European	UK Biobank https://www.leelabsg.org/resources
	Colorectal cancer	FinnGen	6509 cases and 287137 controls	Sex, age, genotyping batch and the first ten genetic principal components	European	FinnGen https://www.finngen.fi/fi
		UK Biobank	4,562 cases and 382,756 controls	Sex, birth year, genotyping batch and first four principal components	European	UK Biobank https://www.leelabsg.org/resources
	Pancreatic cancer	FinnGen	1416 cases and 287137 controls	Sex, age, genotyping batch and the first ten genetic principal components	European	FinnGen https://www.finngen.fi/fi
		UK Biobank	589 cases and 393,372 controls	Sex, birth year, genotyping batch and first four principal components	European	UK Biobank https://www.leelabsg.org/resources
	Liver cancer	FinnGen	648 cases and 259583 controls	Sex, age, genotyping batch and the first ten genetic principal components	European	FinnGen https://www.finngen.fi/fi
		UK Biobank	344 cases and 393,372 controls	Sex, birth year, genotyping batch and first four principal components	European	UK Biobank https://www.leelabsg.org/resources

GUGC, Global Urate Genetics Consortium.

Statistical analysis of MR

To evaluate the causal relationship between serum UA levels and GI cancer, we applied the random-effects multiplicative inverse-variance weighted (IVW) method, a common method in MR studies. The IVW method combines the Wald ratio for each SNP into a single population estimate based on weighted regression of SNP exposure effect and SNP outcome effect(Pierce and Burgess 2013). The IVW method assumes that all genetic variants are valid IVs, and is the most powerful method in MR estimation, but is also susceptible to horizontal pleiotropy. We applied fixed-effects meta-analysis to combine MR estimates from different data sources using the random-effects IVW method. The heterogeneity was assessed with the I² statistic. Low, medium and high heterogeneity were indicated by I² values < 25%, 25–75% and > 75%, respectively.

We also tested causality using MR-Egger and weighted median methods. The MR-Egger method can obtain pleiotropy-corrected causal estimates from the slope of the regression, and assumes no correlation between horizontal pleiotropy effects and SNP-exposure effects (InSIDE assumption)(Bowden et al. 2015). The MR-Egger method does not impose any restrictions on the average value of the pleiotropic effects, and only requires the InSIDE assumption. Under the InSIDE assumption, the variant-exposure associations are independent of the pleiotropic effects (Bowden et al. 2015). The weighted median method estimates causality if more than 50% of the SNPs are valid instruments (i.e., no horizontal pleiotropy, no association with confounders, and strong association with the exposure)(Burgess et al. 2017). The weighted median method has higher precision than the MR-Egger method(Bowden et al. 2016).

In addition, we detected pleiotropy using MR-Egger intercept and MR-PRESSO methods. The MR-Egger intercept method indicates horizontal pleiotropy between the genetic IVs if the P value is less than 0.05(Bowden et al. 2015). The MR-PRESSO method detects and corrects outliers in IVW linear regression to detect directional pleiotropy(Verbanck et al. 2018). We quantified heterogeneity using Cochran Q statistic, and paid attention to the IVW results if the P value was less than 0.05. The MR analysis assumes that IVs generate effects by first influencing exposure and then influencing the outcome. However, the directionality of this hypothesis needs to be validated. Steiger analysis can be used to calculate the variance explanation of IVs for both exposure and outcome. If the variance explanation of the outcome is smaller than that of the exposure, then the direction of the hypothesis is correct. We conducted repeated analyses using MR-Steiger filtering, which aims to remove SNPs that violate the assumption of no pleiotropy (i.e., SNPs with differences in effect on the outcome greater than the exposure difference). These SNPs indicate the potential presence of reverse causality.

All analyses were performed using R Software 4.3.1 (https://www.R-project.org) with TwoSampleMR (version 0.5.7), MendelianRandomization (version 0.8.0), and MRPRESSO package (1.0).

In order to investigate the potential impact of serum UA levels on GI cancers, we selected 27 SNPs closely associated with serum UA levels from a GWAS meta-analysis of GUGC data, with a P value of less than 5×10^−8. After excluding LD and potential confounding factors, we ultimately chose 12 SNPs as IVs for serum UA levels in the risk analysis of esophageal cancer, gastric cancer, liver cancer, pancreatic cancer, and colorectal cancer. In the GWAS data related to the outcome, all SNPs were retained except rs6770152 for colorectal cancer. These IVs all have F- statistics exceeding 30, indicating their strong instrumentality.

The effect of Uric Acid on GI cancers

Data from each source of GI cancers were analyzed using random-effects multiplicative IVW as MR analysis using IVs of serum UA levels. The results from different sources were pooled by Fixed-Effect Meta-Analysis. The genetic susceptibility of serum UA levels showed no significant association with the increased risk of the five GI cancers (all P value > 0.05). In most of the MR studies on GI cancers, we observed low heterogeneity, including gastric cancer (I² = 0%), liver cancer (I²= 0%), pancreatic cancer (I²= 0%), and colorectal cancer (I²= 0%). However, moderate heterogeneity was observed in esophageal cancer (I² = 69%).

OR, odds ratio; CI, confidence interval; P values are for ORs (95% CIs); I² values < 25%, 25–75%, and > 75% were considered to indicate low, moderate, and high heterogeneity, respectively.

Sensitivity analysis and other analyses

We performed sensitivity analyses using different methods, including weighted median, MR-Egger, and MR-PRESSO. In the sensitivity analysis, the direction of the association between the genetic

susceptibility of serum UA levels and the risk of various GI cancers was consistent in the majority of the results. Cochran's Q statistic showed no evidence of heterogeneity except for the heterogeneity observed in colorectal cancer in the UK Biobank study. MR-PRESSO analysis did not identify any outliers that needed to be excluded. MR-Steiger filtering was conducted for individuals exposed to esophageal cancer, gastric cancer, liver cancer, pancreatic cancer, and colorectal cancer to remove SNPs suggesting reverse causality, and all SNPs passed the filtering. Table 2 presents detailed results of the MR analysis for various types of GI cancers, utilizing data from the UK biobank and FinnGen studies.

Table 2 Mendelian randomization results for different types of gastrointestinal cancer from UK biobank and FinnGen studies

Gastrointestinal

cancers

Inverse variance weighted

Weighted median

MR-Egger

estimate

MR-Egger

intercept

MR-PRESSO

Cochran's Q test,

P value

OR (95% CI)

P value

OR (95% CI)

P value

OR (95% CI)

P value

Outliers

Global test, P value

UK Biobank

Esophageal cancer

0.848

0.261

1.28(0.83,1.96)

0.236

1.35(0.82,2.24)

0.327

1.43(0.72,2.84)

0.682

0.885

Gastric cancer

0.894

0.422

1.22(0.75,1.99)

0.554

1.19(0.67,2.12)

0.814

1.10(0.50,2.41)

0.745

0.931

Liver cancer

0.571

0.851

1.06(0.58,1.96)

0.647

1.19(0.57,2.46)

0.388

1.57(0.59,4.18)

0.339

0.662

Pancreatic cancer

0.148

0.619

1.15(0.66,2.02)

0.199

1.44(0.83,2.49)

0.249

1.73(0.72,4.18)

0.272

0.252

Colorectal cancer

0.002

0.992

1.00(0.75,1.33)

0.841

0.98(0.81,1.19)

0.886

1.04(0.65,1.66)

0.847

0.052

FinnGen

Esophageal cancer

0.212

0.157

0.67(0.39,1.16)

0.020

0.50(0.28,0.90)

0.016

0.32(0.15,0.69)

0.037

0.275

Gastric cancer

0.787

0.561

0.91(0.66,1.25)

0.210

0.79(0.55,1.14)

0.080

0.60(0.36,1.00)

0.072

0.931

Liver cancer

0.861

0.853

1.04(0.67,1.63)

0.811

0.94(0.56,1.57)

0.843

0.93(0.45,1.92)

0.696

0.871

Pancreatic cancer

0.216

0.738

1.06(0.75,1.50)

0.319

1.19(0.84,1.68)

0.239

1.41(0.82,2.43)

0.216

0.318

Colorectal cancer

0.090

0.229

0.89(0.74,1.07)

0.172

0.89(0.76,1.05)

0.787

1.04(0.78,1.39)

0.216

0.205

P value < 0.05 indicates statistical significance; CI, confidence interval; OR, odds ratio; MR-PRESSO, Mendelian randomization-pleiotropy residual sum and outlier.

This study employed a two-sample MR approach to systematically investigate the causal relationship between serum UA levels and digestive system cancers (including esophageal, gastric, hepatic, pancreatic, and colorectal cancers). None of the different MR analysis methods and the meta-analysis using the IVW method supported the susceptibility of serum UA levels to GI cancers risks.

Our findings are consistent with some previous studies and can be explained from the perspective of the pathogenic mechanism. Ames et al. proposed that UA can clear singlet oxygen and hydroxyl radicals, protecting red cell membranes from oxidative damage, reducing the oxidants produced by hemoglobin reacting with peroxides, thus avoiding aging and cancer caused by oxidants and free radicals(Ames et al. 1981). This viewpoint has been supported by other studies(Becker 1993; Song et al. 2019). Additionally, Itahana et al. proposed that the UA transporter SLC2A9 is a direct target gene of the tumor suppressor p53. By transporting UA as a source of antioxidants, it participates in antioxidant defense, reduces the generation of ROS, prevents DNA damage and cell death, and may thus contribute to suppressing cancer development(Itahana et al. 2015). Recent experimental results in hepatocellular carcinoma further confirmed this conclusion, demonstrating that SLC2A9 could be a novel tumor suppressor gene and a potential target for treating liver cancer(Han et al. 2019).

As of now, a unanimous conclusion has not been reached regarding whether UA can increase or decrease the risk of digestive system cancers. Observational studies frequently yield inconsistent results. For instance, in a recent cohort study with an average follow-up of 6.6 years, involving 444,000 participants without a history of cancer and 920 cancer patients. This study found that in men, serum UA levels exhibited a U-shaped relationship with the risk of liver cancer, while in women, serum UA levels were positively correlated with the risk of pancreatic cancer(Huang et al. 2020). Another cohort study, which included 8,408 newly diagnosed gout patients and 25,010 controls, found a significant positive correlation between gout and colorectal, liver, and stomach cancers but no significant association with esophageal and pancreatic cancers(Chen et al. 2014). Boffetta et al. reported an increased risk of colon, liver, and pancreatic cancer in patients with gout(Boffetta et al. 2009). However, another cohort study, which included 256 patients with colorectal cancer, failed to find a significant association between serum UA levels and colorectal cancer(Kühn et al. 2017). Observational studies are susceptible to reverse causality and confounding factors, making them prone to bias.

In the past, there have been inconsistent results in MR studies regarding the relationship between serum UA levels and cancers. Kobylecki et al. used two epidemiological designs, namely observational studies and MR studies based on the SLC2A9 rs7442295 genotype, to assess the potential causal effect of high serum UA levels on cancer incidence and overall mortality. They found that high serum UA levels were associated with high cancer incidence and overall mortality in both observational and genetic aspects, which contradicts our conclusions(Kobylecki et al. 2017). However, this difference may be due to the fact that they only used one genetic variant (SLC2A9 rs7442295) as an IV for serum UA levels and did not fully consider other genetic factors influencing serum UA levels. Furthermore, SLC2A9 has varying expression levels in different organs, but the study did not perform MR analysis for specific organ cancers(Kobylecki et al. 2017). Another study conducted by Jiang et al. investigated the relationship between serum UA levels and eight specific cancers, including bladder cancer, breast cancer, colorectal cancer, lung cancer, prostate cancer, renal cancer, skin cancer, and thyroid cancer. Consistent with our research findings, this study did not find a causal relationship between serum UA levels and colorectal cancer(Jiang et al. 2021). However, it is worth noting that this study only used data from the UK Biobank, and individuals in this database are generally healthier than the general population, so selection bias cannot be ruled out. In addition, the study did not cover other GI cancers(Jiang et al. 2021).

Compared to previous studies, our research is the first to utilize a Meta-analysis approach in MR to investigate the causal relationship between serum UA levels and specific GI cancers. Our study has several advantages. Firstly, we extensively leveraged GWAS data from multiple large genetic consortia, including thousands of patients and hundreds of thousands of control individuals, making our MR study one of the most comprehensive and largest-scale investigations into the association between serum UA levels and GI cancer risk. Secondly, we positioned the MR method at the core of our study to mitigate potential confounding factors and reverse causality that could affect outcomes in observational studies. Finally, we employed various MR analytical strategies, including sensitivity analyses and supplementary analyses, to combine results from GI cancers data derived from multiple databases using meta-analysis, ensuring the robustness of our research findings.

However, our study also has some limitations to consider. Firstly, serum UA levels vary by gender and age, with males and postmenopausal females typically having higher serum UA levels (Fini et al. 2012). Since our dataset does not provide sufficiently detailed individual information, we were unable to perform gender and age-stratified analyses. Additionally, to avoid selective bias resulting from population stratification, our sample primarily consisted of individuals of European descent, which may impact the external generalizability of our study results. Therefore, further research is needed to validate the applicability of these findings in other racial populations.

In summary, this study is the first to use a meta-analysis approach in MR to investigate the causal relationship between serum UA levels and specific GI cancers. Our results do not provide substantial evidence for a notable impact of UA levels on the risk of GI cancers. These findings offer valuable insights for the clinical prevention and management of GI cancers. However, our study also has some limitations, such as the lack of gender and age-stratified analyses and the restricted generalizability to other racial populations. Therefore, further research is needed to validate the applicability of these findings in different settings and populations.

UA: Uric acid

GI: Gastrointestinal

SNP: Single nucleotide polymorphism

IV: Instrumental variable

MR: Mendelian randomization

GWAS: Genome-wide association study

IVW: Inverse variance-weighted

LD: Linkage disequilibrium

ICD: International Classification of Diseases

XO: Xanthine oxidase

ROS: Reactive oxygen species

InSIDE: Instrument strength independent of direct effect

OR: Odds ratio

CI: Confidence interval

GUGC: Global Urate Genetics Consortium

Supplementary Information

Supplementary Table 1.

Acknowledgements

We are grateful to the GUGC, UK Biobank, and FinnGen consortia for providing the data used in this study.

Author contributions

ZJX and YGL conceived and designed the study. ZJX conducted the statistical analysis, created the figures, and drafted the manuscript. WFN and YGL supervised the data interpretation and revised the manuscript critically. All authors read and approved the final version of the manuscript.

Funding

Funding This research received no external funding.

Availability of data and materials

Availability of data and materials The data supporting the findings of this study are publicly available from the sources indicated in the manuscript.

Conflict of interest The authors declare that they have no competing interests.

Ethical approval Ethical approval was obtained from the original studies. This article does not involve any studies with human participants or animals conducted by any of the authors.

Consent for publication Not applicable.

Ames B, Cathcart R, Schwiers E and Hochstein P (1981) Uric acid provides an antioxidant defense in humans against oxidant- and radical-caused aging and cancer: a hypothesis. Proceedings of the National Academy of Sciences of the United States of America 78(11), 6858-6862. https://doi.org/10.1073/pnas.78.11.6858.
Arnold M, Abnet C, Neale R, Vignat J, Giovannucci E, McGlynn K and Bray F (2020) Global Burden of 5 Major Types of Gastrointestinal Cancer. Gastroenterology 159(1), 335-349.e315. https://doi.org/10.1053/j.gastro.2020.02.068.
Becker B (1993) Towards the physiological function of uric acid. Free radical biology & medicine 14(6), 615-631. https://doi.org/10.1016/0891-5849(93)90143-i.
Boffetta P, Nordenvall C, Nyrén O and Ye W (2009) A prospective study of gout and cancer. European journal of cancer prevention : the official journal of the European Cancer Prevention Organisation (ECP) 18(2), 127-132. https://doi.org/10.1097/CEJ.0b013e328313631a.
Bourgault J, Abner E, Manikpurage H, Pujol-Gualdo N, Laisk T, Gobeil É, Gagnon E, Girard A, Mitchell P, Thériault S, Esko T, Mathieu P and Arsenault B (2023) Proteome-Wide Mendelian Randomization Identifies Causal Links Between Blood Proteins and Acute Pancreatitis. Gastroenterology 164(6), 953-965.e953. https://doi.org/10.1053/j.gastro.2023.01.028.
Bowden J, Davey Smith G and Burgess S (2015) Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. International journal of epidemiology 44(2), 512-525. https://doi.org/10.1093/ije/dyv080.
Bowden J, Davey Smith G, Haycock P and Burgess S (2016) Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genetic epidemiology 40(4), 304-314. https://doi.org/10.1002/gepi.21965.
Burgess S, Bowden J, Fall T, Ingelsson E and Thompson S (2017) Sensitivity Analyses for Robust Causal Inference from Mendelian Randomization Analyses with Multiple Genetic Variants. Epidemiology (Cambridge, Mass.) 28(1), 30-42. https://doi.org/10.1097/ede.0000000000000559.
Chen C, Yen J and Chang S (2014) Gout patients have an increased risk of developing most cancers, especially urological cancers. Scandinavian journal of rheumatology 43(5), 385-390. https://doi.org/10.3109/03009742.2013.878387.
Chen M and Meng L (2022) The double faced role of xanthine oxidoreductase in cancer. Acta pharmacologica Sinica 43(7), 1623-1632. https://doi.org/10.1038/s41401-021-00800-7.
Davies N, Holmes M and Davey Smith G (2018) Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ (Clinical research ed.) 362, k601. https://doi.org/10.1136/bmj.k601.
Gasse P, Riteau N, Charron S, Girre S, Fick L, Pétrilli V, Tschopp J, Lagente V, Quesniaux V, Ryffel B and Couillin I (2009) Uric acid is a danger signal activating NALP3 inflammasome in lung injury inflammation and fibrosis. American journal of respiratory and critical care medicine 179(10), 903-913. https://doi.org/10.1164/rccm.200808-1274OC.
Han X, Yang J, Li D and Guo Z (2019) Overexpression of Uric Acid Transporter SLC2A9 Inhibits Proliferation of Hepatocellular Carcinoma Cells. Oncology research 27(5), 533-540. https://doi.org/10.3727/096504018x15199489058224.
Huang C, Huang J, Mi N, Lin Y, He Q, Lu Y, Yue P, Bai B, Zhang J, Zhang C, Cai T, Fu W, Gao L, Li X, Yuan J and Meng W (2020) Associations between serum uric acid and hepatobiliary-pancreatic cancer: A cohort study. World Journal of Gastroenterology 26(44), 7061-7075. https://doi.org/10.3748/wjg.v26.i44.7061.
Huang W, Xiao J, Ji J and Chen L (2021) Association of lipid-lowering drugs with COVID-19 outcomes from a Mendelian randomization study. eLife 10. https://doi.org/10.7554/eLife.73873.
Itahana Y, Han R, Barbier S, Lei Z, Rozen S and Itahana K (2015) The uric acid transporter SLC2A9 is a direct target gene of the tumor suppressor p53 contributing to antioxidant defense. Oncogene 34(14), 1799-1810. https://doi.org/10.1038/onc.2014.119.
Jiang M, Ren L, Chen S and Li G (2021) Serum Uric Acid Levels and Risk of Eight Site-Specific Cancers: A Mendelian Randomization Study. Frontiers in genetics 12, 608311. https://doi.org/10.3389/fgene.2021.608311.
Kobylecki C, Afzal S and Nordestgaard B (2017) Plasma Urate, Cancer Incidence, and All-Cause Mortality: A Mendelian Randomization Study. Clinical chemistry 63(6), 1151-1160. https://doi.org/10.1373/clinchem.2016.268185.
Komori H, Yamada K and Tamai I (2018) Hyperuricemia enhances intracellular urate accumulation via down-regulation of cell-surface BCRP/ABCG2 expression in vascular endothelial cells. Biochimica et biophysica acta. Biomembranes 1860(5), 973-980. https://doi.org/10.1016/j.bbamem.2018.01.006.
Köttgen A, Albrecht E, Teumer A et al (2013) Genome-wide association analyses identify 18 new loci associated with serum urate concentrations. Nature Genetics 45(2), 145-154. https://doi.org/10.1038/ng.2500.
Kühn T, Sookthai D, Graf M, Schübel R, Freisling H, Johnson T, Katzke V and Kaaks R (2017) Albumin, bilirubin, uric acid and cancer risk: results from a prospective population-based study. British Journal of Cancer 117(10), 1572-1579. https://doi.org/10.1038/bjc.2017.313.
Lin J, Zhou J and Xu Y (2023) Potential drug targets for multiple sclerosis identified through Mendelian randomization analysis. Brain : a journal of neurology 146(8), 3364-3372. https://doi.org/10.1093/brain/awad070.
Ndrepepa G (2018) Uric acid and cardiovascular disease. Clinica chimica acta; international journal of clinical chemistry 484, 150-163. https://doi.org/10.1016/j.cca.2018.05.046.
Ni J, Li X, Zhang H, Xu Q, Wei X, Feng G, Zhao M, Zhang Z, Zhang L, Shen G and Li B (2022) Mendelian randomization study of causal link from gut microbiota to colorectal cancer. BMC cancer 22(1), 1371. https://doi.org/10.1186/s12885-022-10483-w.
Oshima H and Oshima M (2012) The inflammatory network in the gastrointestinal tumor microenvironment: lessons from mouse models. Journal of gastroenterology 47(2), 97-106. https://doi.org/10.1007/s00535-011-0523-6.
Pierce B and Burgess S (2013) Efficient design for Mendelian randomization studies: subsample and 2-sample instrumental variable estimators. American journal of epidemiology 178(7), 1177-1184. https://doi.org/10.1093/aje/kwt084.
Pingault J, O'Reilly P, Schoeler T, Ploubidis G, Rijsdijk F and Dudbridge F (2018) Using genetic data to strengthen causal inference in observational research. Nature reviews. Genetics 19(9), 566-580. https://doi.org/10.1038/s41576-018-0020-3.
Song Y, Tang L, Han J, Gao Y, Tang B, Shao M, Yuan W, Ge W, Huang X, Yao T, Bian X, Li S, Cao W and Zhang H (2019) Uric Acid Provides Protective Role in Red Blood Cells by Antioxidant Defense: A Hypothetical Analysis. Oxidative medicine and cellular longevity 2019, 3435174. https://doi.org/10.1155/2019/3435174.
Sung H, Ferlay J, Siegel R, Laversanne M, Soerjomataram I, Jemal A and Bray F (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA: a Cancer Journal For Clinicians 71(3), 209-249. https://doi.org/10.3322/caac.21660.
Taghizadeh N, Vonk J and Boezen H (2014) Serum uric acid levels and cancer mortality risk among males in a large general population-based cohort study. Cancer causes & control : CCC 25(8), 1075-1080. https://doi.org/10.1007/s10552-014-0408-0.
Tsukamoto T, Tsujii M, Odake K, Iino T, Nakamura T, Matsumine A and Sudo A (2021) Febuxostat reduces muscle wasting in tumor-bearing mice with LM8 osteosarcoma cells via inhibition of reactive oxygen species generation. Free radical research 55(7), 810-820. https://doi.org/10.1080/10715762.2021.1947502.
Verbanck M, Chen C, Neale B and Do R (2018) Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nature Genetics 50(5), 693-698. https://doi.org/10.1038/s41588-018-0099-7.
Xie J, Huang H, Liu Z, Li Y, Yu C, Xu L and Xu C (2023) The associations between modifiable risk factors and nonalcoholic fatty liver disease: A comprehensive Mendelian randomization study. Hepatology (Baltimore, Md.) 77(3), 949-964. https://doi.org/10.1002/hep.32728.
Yao S, Zhang M, Dong S, Wang J, Zhang K, Guo J, Guo Y and Yang T (2022) Bidirectional two-sample Mendelian randomization analysis identifies causal associations between relative carbohydrate intake and depression. Nature human behaviour 6(11), 1569-1576. https://doi.org/10.1038/s41562-022-01412-9.

No competing interests reported.

SupplementaryTable1.xlsx

Download PDF

Version 1

posted

You are reading this latest preprint version

Serum uric acid levels and risk of gastrointestinal cancers: a comprehensive Mendelian randomization study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Introduction

Materials and methods

Study design

Selection of IVs

Data sources for GI cancers

Statistical analysis of MR

Result

The effect of Uric Acid on GI cancers

Sensitivity analysis and other analyses

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1