ABSTRACT
Obesity is a risk factor for the development of hyperuricemia, both of which were related to gut microbiota. However, whether alterations in the gut microbiota lie in the pathways mediating obesity’s effects on hyperuricemia is less clear. Body mass index (BMI) and serum uric acid (SUA) were separately important indicators of obesity and hyperuricemia. Our study aims to investigate whether BMI-related gut microbiota characteristics would mediate the association between BMI and SUA levels. A total of 6,280 participants from Guangdong Gut Microbiome Project were included in this study. Stool samples were collected for 16S rRNA gene sequencing. The results revealed that BMI was significantly and positively associated with SUA. Meanwhile, BMI was significantly associated with the abundance of 102 gut microbial genera, 16 of which were also significantly associated with SUA. The mediation analysis revealed that the association between BMI and SUA was partially mediated by the abundance of Proteobacteria (proportion mediated: 0.94%, P < 0.05). At the genus level, 25 bacterial genera, including Ralstonia, Oscillospira, Faecalibacterium, etc., could also partially mediate the association of BMI with SUA (the highest proportion is mediated by Ralstonia, proportion mediated: 2.76%, P < 0.05). This study provided evidence for the associations among BMI, gut microbiota, and SUA, and the mediation analysis suggested that the association of BMI with SUA was partially mediated by the gut microbiota.
IMPORTANCE
Using 16S rRNA sequencing analysis, local interpretable machine learning technique analysis and mediation analysis were used to explore the association between BMI with SUA, and the mediating effects of gut microbial dysbiosis in the association were investigated.
KEYWORDS: body mass index, 16S rRNA, microbiota, serum uric acid
INTRODUCTION
Hyperuricemia (HUA) is a metabolic disease characterized by an elevated level of serum uric acid (SUA) (1). It was reported that the prevalence of hyperuricemia in the Chinese adult population increased from 11.1% in 2015–2016 to 14.0% in 2018–2019 (2). Accumulating data have revealed that hyperuricemia is a crucial risk factor for gout, hypertension, cardiovascular, diabetes, atherosclerosis, and chronic kidney disease (3 – 8). Generally, hyperuricemia is attributed to urate overproduction in the liver and insufficient urate elimination in the renal or extra-renal (9). Obesity has also been reported as a risk factor for hyperuricemia (10). As early as 2004, a study of 17,155 children in Japan found that the detection rate of hyperuricemia was 2.5% in normal-weight children, while it was up to 12% in obese children (11). Recently, a cross-sectional study of 144,856 adults showed that overweight people were more likely to develop hyperuricemia (12).
The biological mechanisms of obesity-induced hyperuricemia have been explored extensively in both experimental and epidemiological studies, including the excess ingestion of purine-rich foods, impaired renal clearance of uric acid, and increased production and secretion of uric acid by adipose tissue (10, 13, 14). In addition, lactate accumulation in obesity accelerates renal urate reabsorption via urate transporter 1 (URAT1), and the obesity gene product leptin can affect the renal clearance of uric acid (15, 16). Recent studies have shown that hyperuricema may also be associated with alterations in the gut microbiome (17). Specifically, animal studies have demonstrated that hyperuricemia mice exhibit changes in the diversity and abundance of gut microbiota, which potentially affects amino acid metabolism and promotes purine metabolism disorder and inflammation (18, 19). Similarly, individuals with hyperuricemia have been found to have decreased microbial diversity and an altered gut microbiota composition compared to those with normouricemia (17). In children with hyperuricemia, serum uric acid level was positively correlated with genera Actinomyces, Morganella, and Streptococcus, and negatively associated with the short-chain fatty acids-producing bacteria, such as Alistipes, Faecalibacterium, and Oscillospira, and sulfur-producing bacteria Bilophila (20). Previous studies have shown an association between SUA, gut microbiota, and obesity. For example, Gong et al. demonstrated that the counts of Bifidobacterium significantly decreased in visceral obesity, and serum uric acid may be a mediator between the reduction of Bifidobacterium and the increase of visceral adipose tissue (21). These findings imply that the alterations in the gut microbiota components might partially mediate the association between BMI with SUA.
Whether some gut microbial features lie in the pathways from obesity to hyperuricemia is less clear. In the current study, we investigated the association between BMI and SUA in a community-based population and further investigated the mediating effects of gut microbial dysbiosis in the association. Our findings provided important clues to further elucidate the pathogenesis of elevated SUA levels and aid in the prevention of hyperuricemia resulting from increased BMI or obesity.
MATERIALS AND METHODS
Study design and participants
The Guangdong Gut Microbiome Project (GGMP) was a community-based cross-sectional study cohort conducted in 14 districts in Guangdong Province, China. Multistage cluster sampling method was used to select the population from 14 surveillance sites for the survey. All participants must be at least 18 years of age and signed an informed letter. Detailed information regarding study design could be found in our previous research (22). After removing the samples obtained from those who did not have BMI and UA values in the metadata and those with stool sample sequences of less than 10,000 reads, a total of 6,280 samples were retained in this study. We performed an analysis of the missing rate for the host baseline characteristics (Fig. S1). Those characteristics with more than 5% missing were excluded, and the remaining with less than 5% missing were imputed using random forest algorithm. A comparison of baseline population characteristics before and after imputing missing data showed that the differences were not significant, suggesting that the missing data might be missing randomly (Table S1).
Biological sample collection and testing
All participants had a face-to-face interview followed by a medical examination and sample collection by medical professionals. Information on anthropometrics, physiological and biochemical indicators, dietary habits together with the use of drugs and antibiotics was recorded. Fresh peripheral blood samples and stool samples were collected from all participants and sent to the laboratory authorized by Guangdong Center for Disease Control and Prevention through cold chain transport. In our previous study, we described the method for processing the samples (22). In short, we extracted total bacterial DNA from stool samples. The barcode primers (5′ to 3′) used to amplify the V4 region of the 16S rRNA gene were V4F, GTGYCAGCMGCCGCGGTAA and V4R, GGACTACNVGGGTWTCTAAT. PCR amplification conditions were as follows: initial denaturation at 94°C for 5 minutes, followed by 30 cycles of denaturation at 94°C for 30 seconds, annealing at 52°C for 30 seconds, extension at 72°C for 45 seconds, and final extension at 72°C for 5 minutes. Next-generation sequencing of the PCR products was performed on an Illumina HiSeq 2500 platform at the Beijing Genome Institute (BGI; Beijing, China).
Definition of key measures
Hyperuricemia was defined as SUA >420 µmol/L for men, >360 µmol/L for women based on diagnostic criteria (23). Here, we employed four operational taxonomic unit (OTU)-based α-diversity indices: Chao1, Shannon, observed OTUs, and phylogenic diversity (PD) whole tree. And assessed four bacteria (Bacteroidetes, Proteobacteria, Firmicutes, and Verrucomicrobia) at the phylum level due to the high abundance (> 92%) observed in our previous study (22).
Bioinformatics and biostatistics
The procedure for pre-processing the raw sequences was described in our earlier paper (22). In brief, QIIME V.1.9.1 was used to merge the data, perform quality control assessments, and analyze the data at the sequence level with the non-reference statistical denoising deblur method to avoid misidentification of different operational taxonomic unit-based data derived from misclustered sequences (24). Representative sequences were aligned using the PyNAST algorithm (25). The Ribosomal Database Project classifier against the Greengenes database (version 13.8) was used to perform taxonomic profiling (26, 27).
The missing values of the host characteristic were analyzed and visualized in this study with the R package “naniar” (version 0.6.1) and imputed with the R package “missForest” (version 1.5) (28). The completed metadata data set was then divided into a 70% training set, 20% validation set, and 10% unseen test to establish the random forest classification model, which could distinguish between healthy participants and those with high serum uric acid. In addition, the categorical variables used in this model were coded by One-Hot encoding method and the target class-imbalance fixed by Synthetic Minority Oversampling TEchnique (SMOTE). SHapley Additive exPlanations (SHAP) analysis was adapted to better assess the model’s feature importance. Briefly, SHAP is a game theoretic approach that combines optimal credit allocation with local explanations by using classical Shapley values to explain the output of the machine learning model (29). The Python library “Pycaret” (version 2.3.10) was used to perform the above process and to optimize the hyperparameters of the random forest model.
In the microbiome mediation analysis section, we first excluded the data at genus level with a prevalence under 5% to reduce statistical noise from low-prevalence microbiota. Using the cleaned data, Spearman correlations were calculated between BMI and genus-level microbes, and significantly associated microbes were examined using a threshold of False Discovery Rate (FDR) <0.1. The R package “bruceR” (version 0.8.9) was then used to calculate whether microbiota played a mediating role between BMI and UA and its value. In brief, a mediator formula (microbiota = covariates + BMI) and an outcome formula (UA = covariates + BMI +microbiota) were created to calculate the direct effect, indirect effect, and the total effect. The mediated proportion was calculated by dividing indirect effect by total effect. All analyses were performed on the 6,280 samples and the top 15 baseline characteristics associated with risks of HUA were chosen as covariates in all correlation and mediation analyses: age, gender, triglyceride (TG), blood urea nitrogen (BUN), high density lipoprotein (HDL), hemoglobin (Hb), low density lipoprotein (LDL), fasting blood glucose (FBG), systolic blood pressure (SBP), diastolic blood pressure (DBP), alaninetransaminase (ALT), total cholesterol (TCHO), grains, livestock_meat, and fruits.
RESULTS
Characteristics of study participants
The study population was previously described in the GGMP, which contains 7,009 individuals from 14 districts within Guangdong Province, China. Participants with missing information on BMI and uric acid were excluded. Finally, 6,280 participants remained for subsequent analysis.
The participants were divided into two groups: the normal (Nor) group and the hyperuricemia (HUA) group (Table 1). The number of participants with HUA were 1,494 (23.78%) with a median SUA level of 445 µmol/L (IQR: 411–490) and a median BMI level of 24.6 kg/m2 (IQR: 22.3–26.9). Most of the clinical indicators are statistically different between the two groups. Compared with Nor group, those with HUA had higher levels of SBP, DBP, FBG, TCHO, TG, LDL, Hb, ALT, and BUN, but lower levels of HDL (Table 1).
TABLE 1.
General characteristics of study participants a
| Characteristics | HUA (N = 1,494) | Nor (N = 4,786) | P-value |
|---|---|---|---|
| BMI, median (IQR) | 24.6 (22.3–26.9) | 22.6 (20.5–24.9) | <0.001 |
| UA (µmol/L), median (IQR) | 445 (411–490) | 298 (256–342) | <0.001 |
| Age, median (IQR) | 56 (45–65) | 53 (43–63) | <0.001 |
| Gender (female), n (%) | 654 (44) | 2,834 (59) | <0.001 |
| Education, n (%) | 0.001 | ||
| Middle school or lower | 1,082 (72) | 3,684 (76.9) | |
| High school or professional college | 238 (15.9) | 679 (14.1) | |
| University | 174 (11.6) | 423 (8.8) | |
| Smoke now, n (%) | 435 (29) | 1,173 (25) | <0.001 |
| Sleep_time (min), median (IQR) | 480 (420–480) | 480 (420–510) | 0.03 |
| SBP (mmHg), median (IQR) | 133 (121–148) | 126 (115–142) | <0.001 |
| DBP (mmHg), median (IQR) | 79 (72–87) | 76 (69–83) | <0.001 |
| b HR (/min), median (IQR) | 76 (70–84) | 76 (69–83) | 0.603 |
| FBG (mmol/L), median (IQR) | 5.43 (5.03–5.98) | 5.27 (4.88–5.74) | <0.001 |
| TCHO (mmol/L), median (IQR) | 5.33 (4.82–5.86) | 5.18 (4.68–5.72) | <0.001 |
| TG (mmol/L), median (IQR) | 1.42 (1.01–2.14) | 0.98 (0.72–1.44) | <0.001 |
| HDL (mmol/L), median (IQR) | 1.12 (0.92–1.37) | 1.26 (1.04–1.51) | <0.001 |
| LDL (mmol/L), median (IQR) | 3.40 (2.81–4.07) | 3.12 (2.56–3.74) | <0.001 |
| Hb (g/L), median (IQR) | 148 (134–159) | 142 (130–153) | <0.001 |
| ALT (U/L), median (IQR) | 17 (12–25) | 15 (11–21) | <0.001 |
| BUN (mmol/L), median (IQR) | 5.39 (4.55–6.49) | 4.99 (4.17–5.95) | <0.001 |
| Gout, n (%) | 97 (6.5) | 109 (2.3) | <0.001 |
| c Bristol stool type, n (%) | 0.025 | ||
| 1&2 | 96 (6.4) | 440 (9.1) | |
| 3&4 | 1,154 (77.2) | 3,624 (75.7) | |
| 5&6 | 236 (15.7) | 704 (14.7) | |
| 7 | 8 (0.5) | 18 (0.3) | |
| Synbiotics use within 1 month, n (%) | 47 (3.2) | 121 (2.6) | 0.198 |
| Antibiotics use within 1 month, n (%) | 86 (5.9) | 289 (6.2) | 0.707 |
| Drug use within 3 days, n (%) | 295 (20) | 764 (16) | <0.001 |
| Grains (g), median (IQR) | 106,412 (65,700–160,052) | 109,500 (73,000–164,250) | 0.003 |
| Vegetables (g), median (IQR) | 109,500 (73,000–164,250) | 109,500 (73,000–164,250) | 0.725 |
| Fruits (g), median (IQR) | 15,600 (5,200–39,000) | 15,600 (4,320–36,500) | 0.005 |
| Livestock meat (g), median (IQR) | 36,500 (17,238–73,000) | 36,500 (16,900–70,905) | 0.063 |
Kruskal-Wallis rank sum test for continuous variables and Pearson's chi-squared test for categorical variables.
HR, heart rate.
Types 1 and 2 describe stool that is hard to pass and may point to constipation. Types 3 and 4 describe stool that is well-formed and easy to pass. Types 5 and 6 may point to loose stools. Type 7 describes very loose stools or fully liquid diarrhea.
Associations between BMI and hyperuricemia using local interpretable machine learning technique
We build a random forest classifier model with the clinical characteristics shown in Table 1 to evaluate their contribution on distinguishing subjects with hyperuricemia from healthy counterparts, which helped us to discover the associations between characteristics and hyperuricemia. By using the built-in random forest’s Gini impurity method, we found that BMI was the second most contributed feature among all characteristics (Fig. S2). Though the traditional global explanation approach, such as mean decrease in Gini impurity used in the random forest model, is widely applied and easy to use, it could only list out the feature contribution level in the model and could not offer the direction of the feature’s impact on the model. However, the local interpretable machine learning technique such as SHAP could allow us to supplement the deficiencies mentioned above. In Fig. 1A, the characteristics were ranked based on their mean absolute SHAP values, which were in line with the global importance of the feature, and each point indicated a value for a subject. Meanwhile, those points locate on the positive side of the horizontal axis, which means they have a higher probability of hyperuricemia, and the other points that lay on the negative side of the axis represent a lower probability of hyperuricemia. Combined with the color of the dot, which shows the original value of the feature, we could simply find that the one with a higher BMI seems to have a higher probability of hyperuricemia. It is noteworthy that the SHAP technique helps us to take an insight into the relationship between BMI and hyperuricemia. Individuals with a BMI above 24 are more likely to develop hyperuricemia than those who are thinner (Fig. 1B).
Fig 1.
Association of BMI and HUA, revealed by the interpretable machine learning approach. (A) The summary plot shows the SHAP values of all individuals. The position of the point on the horizontal axis quantifies the effect of a single characteristic on the random forest classifier’s prediction for a particular individual. Colors represent the actual value of the features, with blue and red representing low and high. (B) The dependence plot illustrates the trend of the SHAP value for BMI changes with BMI.
Association of BMI with the gut microbiota
There was no significant correlation between BMI and α-diversity of gut microbiota. However, among the four bacteria at the phylum level (Bacteroidetes, Proteobacteria, Firmicutes, and Verrucomicrobia), the abundance of Proteobacteria and Firmicutes increased and that of Verrucomicrobia and Bacteroidetes decreased with increasing BMI (Table 2). We also found 102 genera that were significantly related to BMI (FDR 0.1). Among them, the abundance of 50 genera showed a decreasing trend with increasing BMI, while the abundance of other species showed an increasing trend (Table S2). For example, Blautia and Ralstonia were positively associated with BMI. In contrast, Parabacteroides species, which may help maintain intestinal homeostasis, were inversely associated with BMI. In addition, some butyric acid-producing bacteria, such as Butyricimonas and Oscillospira, were also negatively associated with BMI (Table S2).
TABLE 2.
The associations of BMI with the α-diversity and phylum level of the gut mircobiota a
| β | P | FDR | 95% CI | |
|---|---|---|---|---|
| PD whole tree | −0.012 | 0.408 | 0.536 | (−0.042, 0.017) |
| Observed species | −0.132 | 0.499 | 0.653 | (−0.513, 0.25) |
| Chao1 | −0.284 | 0.213 | 0.301 | (−0.732, 0.163) |
| Shannon | 0 | 0.98 | 0.98 | (−0.008, 0.007) |
| Verrucomicrobia | −0.002 | 0 | 0 | (−0.003, –0.001) |
| Proteobacteria | 0.002 | 0.006 | 0.054 | (0.001, 0.003) |
| Firmicutes | 0.004 | 0 | 0 | (0.003, 0.006) |
| Bacteroidetes | −0.004 | 0 | 0 | (−0.006, –0.003) |
Model was adjusted for age, gender, TG, BUN, HDL, Hb, LDL, FBG, SBP, DBP, ALT, TCHO, grains, livestock_meat, and fruits.
Association of gut microbiota and SUA
The results of the association between gut microbiota and SUA revealed that Chao1, one of α-diversity indices, was negatively associated with SUA (β = −0.158, 95% CI: −0.270, –0.045, P = 0.006) (Fig. 2A). In the analyses of the four phylum levels, increased abundance of Proteobacteria was associated with higher SUA (β = 35.152, 95% CI: 11.447, 58.858, P = 0.004), whereas Bacteroidetes, Firmicutes, and Verrucomicrobia were not significantly associated with SUA (Fig. 2B). Next, we only chose BMI-associated genera in the analysis of the association between SUA and gut microbes. We found 16 genera were also significantly associated with SUA (P < 0.05). Especially, Odoribacter and Pyramidobacter were negatively associated with BMI and SUA, and Megamonas, Clostridium, Acinetobacter, Chryseobacterium, Novosphingobium, etc., were positively associated with BMI and SUA (Table S3).
Fig 2.
The association of gut microbiota with SUA. (A) The associations of the α-diversity of the gut microbiota with SUA. (B) The associations of phylum of gut microbiota with SUA. Model was adjusted for age, gender, TG, BUN, HDL, Hb, LDL, FBG, SBP, DBP, ALT, TCHO, grains, livestock_meat, and fruits.
The mediation effects of the gut microbes on the association between BMI and SUA
Mediation analysis indicated that the associations of BMI with SUA were partially mediated by gut microbiota. For example, at the phylum level, the relative abundance of Proteobacteria was positively associated with higher BMI and SUA (proportion mediated: 0.94%, P < 0.05) (Fig. S3). However, Bacteroidetes, Firmicutes, and Verrucomicrobia did not mediate the association of BMI with SUA (Table S4).
Furthermore, we evaluated the mediating effects of gut microbiota diversity at the genus level. The results showed that 25 bacteria could partially mediate the association of BMI with SUA (Fig. 3; Fig. S4 and Table S5). Particularly, 15 bacteria (Ralstonia, f_Comamonadaceae, f_Caulobacteraceae, Bradyrhizobium, Halomonas, Ochrobactrum, Pseudidiomarina, Odoribacter, Delftia, Oscillospira, Faecalibacterium, f_Rikenellaceae, f_Methylobacteriaceae, f_[Barnesiellaceae], Sediminibacterium) had mediation effect values greater than 1%. For instance, 2.76% of the total effects of BMI were mediated by Ralstonia, and 1.23% of the total effects of BMI were mediated by Oscillospira (Fig. 3). To make our results more reliable, we used partial sampling and randomly selected 5,280, 4,280, and 3,280 samples for mediated effects analysis. We found 10 species were better reproduced, such as Ralstonia, Bradyrhizobium, Halomonas, Ochrobactrum, Odoribacter, Oscillospira, etc. (Fig. S5 and Table S6).
Fig 3.
The mediating effects (%, proportion mediated) of bacterial genus on the associations of BMI with SUA. The mediation analyses were adjusted for age, gender, TG, BUN, HDL, Hb, LDL, FBG, SBP, DBP, ALT, TCHO, grains, livestock_meat, and fruits.
DISCUSSION
Obesity proceeded to the incidence of hyperuricemia, and BMI was positively correlated with serum uric acid levels in previous studies, which was confirmed by our findings (30, 31). Studies clarified that hyperuricemia in obesity is mainly attributed to the increased intake of purines and an impaired renal clearance of UA (13, 32). Also, xanthine oxidoreductase activity increased in adipose tissue, which degrades hypoxanthine and xanthine to uric acid (33). Increasing evidence shows that the gut microbiota plays an important role in the development of obesity and obesity-related complications, such as hyperlipidemia and diabetes (34). However, little research has been done on the roles of gut microbiota in the development of obesity-related hyperuricemia. In the present study, we first provided evidence supporting the mediator role of gut microbiota characteristics alterations in the association between BMI and SUA.
The association between obesity and human gut microbiota was explored in several studies over the past few years (35 – 37). For example, studies described that obese individuals exhibited a higher Firmicutes/Bacteroidetes ratio, compared to the normal-weight individuals (38). In our assessment of the relationship between BMI and gut microbiota, we found Proteobacteria was positively associated with BMI level. Additionally, several gut microbes were found to be associated with BMI in recent studies. Deng et al. found that higher BMI was associated with lower abundances of four species (Clostridium hathewayi, Parabacteroides unclassified, Lachnospiraceae bacterium 3 1 57FAA CT1, and Lachnospiraceae bacterium 7 1 58FAA) (39). A cross-sectional study among Chinese college students observed that Anaerotruncus, Parabacteroides, and Alistipes were negatively related to BMI, which was consistent with our findings. We also identified that 52 genera were positively associated with BMI, and 50 genera were negatively associated with BMI. For example, Blautia and Ralstonia were positively associated with BMI, but some butyric acid-producing bacteria, such as Butyricimonas and Oscillospira, were negatively associated with BMI.
The diversity and abundance of gut microbiota were altered in hyperuricemia. However, due to the high complexity of gut microbiota, affected by many confounding factors such as geography, environment, and diet, how and what gut microbes may be associated with SUA level or hyperuricemia were not clear. Wei et al. demonstrated low relative abundances of genus Coprococcus were associated with high levels of serum urate (17), which was confirmed by our findings. In this study, we found α-diversity (Chao1) was negatively associated with SUA level. At the phylum level, the increase in the abundance of Proteobacteria was associated with higher SUA. Furthermore, we found 16 genera that related to BMI were also significantly associated with SUA.
The mediation analysis suggested that the association of BMI with SUA was partially mediated by the gut microbiota. The majority of Proteobacteria was positively associated with BMI and SUA and mediated 0.94% of the total effects of BMI on SUA. A recent study reported that compared to the controls, Proteobacteria was more abundant in tophaceous gout, caused by hyperuricemia (40). Therefore, we hypothesize that Proteobacteria may drive the hyperuricemia and play an important role in the development of hyperuricemia. Furthermore, we only used BMI-associated microbes in the next mediation analysis associated between BMI and SUA. We found that 25 genera may partially mediate the association of BMI with SUA. Besides, we can get better reproducibility through partial random sampling. Among them, Ralstonia exerts the largest mediating effect in the association of BMI and SUA (2.79%). It has been observed that the fecal abundance of Ralstonia pickettii was increased in obese subjects with pre-diabetes and Diabetes mellitus type 2 (T2DM) (41). Ralstonia pickettii was also found to facilitate the transfer of lipopolysaccharide to the blood, and it explains the increased low-grade inflammation observed in obese/diabetic patients (42). These further demonstrate that Ralstonia may be important for the development of obesity. We also found that Rikenellaceae was negatively associated with SUA and mediated the effects of BMI. Previous studies have also revealed that the Rikenellaceae family can produce short-chain fatty acid butyrate and helps to maintain good gut condition (43). This suggests that Rikenellaceae is beneficial for host metabolism. However, all the results should be identified using the shotgun metagenomic sequencing technology or prospective, large-sample, multicenter studies in the future.
Intestinal microecology may be a new treatment for hyperuricemia by regulating intestinal flora and restoring intestinal microecological homeostasis through microbial agents or fecal bacteria transplantation. But for obese people, although our study suggests that 25 genera exert mediating effects, it is worth noting that these effect sizes are small. Therefore, approaches to prevent or ameliorate hyperuricemia by intervening gut microbiota in obese people remain to be further verified.
Strengths and limitations
Our study has some strengths. First, this is one of the largest Eastern-population-based gut microbiome data sets, ensuring sufficient statistical power to test the associations among BMI, the gut microbiome, and SUA (22). Second, the large sample size allowed us to adequately adjust for many covariates, including clinical indicators and dietary intake. These covariates are also important factors related to SUA and the gut microbiome.
We admit that our study has some limitations. First, our study is a cross-sectional study; therefore, we cannot describe the causality between BMI, the gut microbiota, and SUA. Second, our study explored the association between BMI and gut microbiota by using data from 16S rRNA sequencing; thus, we cannot detect gut microbes at the species level. It should be further validated whether the identified associations in our cohort are specific in the future larger studies. Finally, the mediation effect size was relatively low. Whether microbiota can be used as a target for prevention/amelioration of hyperuricemia in obese people needs to be further confirmed.
Conclusion
Our study demonstrated that there was a mediating effect of gut microbiota between BMI and SUA, which suggested the microbiota has an effect on increasing uric acid levels in obese people.
Supplementary Material
ACKNOWLEDGMENTS
We thank all participants and staff from the Guangdong Gut Microbiome Project.
This work was supported by the National Natural Science Foundation of China (81800746).
The authors declare that they have no competing interests.
Contributor Information
Yan He, Email: yanhe@i.smu.edu.cn.
Wei-Hua Chen, Huazhong University of Science and Technology, Wuhan, China .
Xiaoyu Gao, Jinlin University, Jinlin, China .
Khemlal Nirmalkar, Arizona State University, Tempe, Arizona, USA .
DATA AVAILABILITY
The raw data for 16S rRNA gene sequences are available in NCBI under accession number PRJEB18535.
SUPPLEMENTAL MATERIAL
The following material is available online at https://doi.org/10.1128/spectrum.01140-23.
Supplemental figures and tables.
An accounting of the reviewer comments and feedback.
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.
REFERENCES
- 1. Lu Y-H, Chang Y-P, Li T, Han F, Li C-J, Li X-Y, Xue M, Cheng Y, Meng Z-Y, Han Z, Sun B, Chen L-M. 2020. Empagliflozin attenuates hyperuricemia by upregulation of ABCG2 via AMPK/AKT/CREB signaling pathway in type 2 diabetic mice. Int J Biol Sci 16:529–542. doi: 10.7150/ijbs.33007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Zhang M, Zhu X, Wu J, Huang Z, Zhao Z, Zhang X, Xue Y, Wan W, Li C, Zhang W, Wang L, Zhou M, Zou H, Wang L. 2021. Prevalence of hyperuricemia among Chinese adults: findings from two nationally representative cross-sectional surveys in 2015-16 And 2018-19. Front Immunol 12:791983. doi: 10.3389/fimmu.2021.791983 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Bjornstad P, Laffel L, Lynch J, El Ghormli L, Weinstock RS, Tollefsen SE, Nadeau KJ, TODAY Study Group . 2019. Elevated serum uric acid is associated with greater risk for hypertension and diabetic kidney diseases in obese adolescents with type 2 diabetes: an observational analysis from the treatment options for type 2 diabetes in adolescents and youth (TODAY) study. Diabetes Care 42:1120–1128. doi: 10.2337/dc18-2147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Filippatos GS, Ahmed MI, Gladden JD, Mujib M, Aban IB, Love TE, Sanders PW, Pitt B, Anker SD, Ahmed A. 2011. Hyperuricaemia, chronic kidney disease, and outcomes in heart failure: potential mechanistic insights from epidemiological data. Eur Heart J 32:712–720. doi: 10.1093/eurheartj/ehq473 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Li B, Chen L, Hu X, Tan T, Yang J, Bao W, Rong S. 2023. Association of serum uric acid with all-cause and cardiovascular mortality in diabetes. Diabetes Care 46:425–433. doi: 10.2337/dc22-1339 [DOI] [PubMed] [Google Scholar]
- 6. Lippi G, Targher G, Montagnana M, Salvagno GL, Guidi GC. 2008. High serum uric acid as a novel risk factor for type 2 diabetes: response to Dehghan et al. Diabetes Care 31:e68. doi: 10.2337/dc08-0468 [DOI] [PubMed] [Google Scholar]
- 7. Rock KL, Kataoka H, Lai J-J. 2013. Uric acid as a danger signal in gout and its comorbidities. Nat Rev Rheumatol 9:13–23. doi: 10.1038/nrrheum.2012.143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rodrigues TC, Maahs DM, Johnson RJ, Jalal DI, Kinney GL, Rivard C, Rewers M, Snell-Bergeon JK. 2010. Serum uric acid predicts progression of subclinical coronary atherosclerosis in individuals without renal disease. Diabetes Care 33:2471–2473. doi: 10.2337/dc10-1007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Ichida K. 2009. What lies behind serum urate concentration? Insights from genetic and genomic studies. Genome Med 1:118. doi: 10.1186/gm118 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Lee Y-S, Kim S-H, Yuk HJ, Lee G-J, Kim D-S. 2018. Tetragonia tetragonoides (Pall.) kuntze (New Zealand spinach) prevents obesity and hyperuricemia in high-fat diet-induced obese mice. Nutrients 10:1087. doi: 10.3390/nu10081087 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Ogura T, Matsuura K, Matsumoto Y, Mimura Y, Kishida M, Otsuka F, Tobe K. 2004. Recent trends of hyperuricemia and obesity in Japanese male adolescents, 1991 through 2002. Metabolism 53:448–453. doi: 10.1016/j.metabol.2003.11.017 [DOI] [PubMed] [Google Scholar]
- 12. Liu D-M, Jiang L, Gan L, Su Y, Li F. 2019. Association between serum uric acid level and body mass index in sex- and age-specific groups in Southwestern China. Endocr Pract 25:438–445. doi: 10.4158/EP-2018-0426 [DOI] [PubMed] [Google Scholar]
- 13. Choi HK, Atkinson K, Karlson EW, Willett W, Curhan G. 2004. Purine-rich foods, dairy and protein intake, and the risk of gout in men. N Engl J Med 350:1093–1103. doi: 10.1056/NEJMoa035700 [DOI] [PubMed] [Google Scholar]
- 14. Yamashita S, Matsuzawa Y, Tokunaga K, Fujioka S, Tarui S. 1986. Studies on the impaired metabolism of uric acid in obese subjects: marked reduction of renal urate excretion and its improvement by a low-calorie diet. Int J Obes 10:255–264. [PubMed] [Google Scholar]
- 15. de Oliveira EP, Burini RC. 2012. High plasma uric acid concentration: causes and consequences. Diabetol Metab Syndr 4:12. doi: 10.1186/1758-5996-4-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Ichida K, Matsuo H, Takada T, Nakayama A, Murakami K, Shimizu T, Yamanashi Y, Kasuga H, Nakashima H, Nakamura T, Takada Y, Kawamura Y, Inoue H, Okada C, Utsumi Y, Ikebuchi Y, Ito K, Nakamura M, Shinohara Y, Hosoyamada M, Sakurai Y, Shinomiya N, Hosoya T, Suzuki H. 2012. Decreased extra-renal urate excretion is a common cause of hyperuricemia. Nat Commun 3:764. doi: 10.1038/ncomms1756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wei J, Zhang Y, Dalbeth N, Terkeltaub R, Yang T, Wang Y, Yang Z, Li J, Wu Z, Zeng C, Lei G. 2022. Association between gut microbiota and elevated serum urate in two independent cohorts. Arthritis Rheumatol 74:682–691. doi: 10.1002/art.42009 [DOI] [PubMed] [Google Scholar]
- 18. Song S, Lou Y, Mao Y, Wen X, Fan M, He Z, Shen Y, Wen C, Shao T. 2022. Alteration of gut microbiome and correlated amino acid metabolism contribute to hyperuricemia and Th17-driven inflammation in Uox-KO mice. Front Immunol 13:804306. doi: 10.3389/fimmu.2022.804306 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Liu X, Lv Q, Ren H, Gao L, Zhao P, Yang X, Yang G, Xu D, Wang G, Yang W, Wang P, Wang Z, Xing S. 2020. The altered gut microbiota of high-purine-induced hyperuricemia rats and its correlation with hyperuricemia. PeerJ 8:e8664. doi: 10.7717/peerj.8664 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Yuan X, Chen R, Zhang Y, Lin X, Yang X. 2022. Altered gut microbiota in children with hyperuricemia. Front Endocrinol (Lausanne) 13:848715. doi: 10.3389/fendo.2022.848715 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Gong H, Gao H, Ren Q, He J. 2022. The abundance of bifidobacterium in relation to visceral obesity and serum uric acid. Sci Rep 12:13073. doi: 10.1038/s41598-022-17417-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. He Y, Wu W, Zheng H-M, Li P, McDonald D, Sheng H-F, Chen M-X, Chen Z-H, Ji G-Y, Zheng Z-D-X, Mujagond P, Chen X-J, Rong Z-H, Chen P, Lyu L-Y, Wang X, Wu C-B, Yu N, Xu Y-J, Yin J, Raes J, Knight R, Ma W-J, Zhou H-W. 2018. Regional variation limits applications of healthy gut microbiome reference ranges and disease models. Nat Med 24:1940. doi: 10.1038/s41591-018-0219-z [DOI] [PubMed] [Google Scholar]
- 23. Zhang Q, Lou S, Meng Z, Ren X. 2011. Gender and age impacts on the correlations between hyperuricemia and metabolic syndrome in Chinese. Clin Rheumatol 30:777–787. doi: 10.1007/s10067-010-1660-7 [DOI] [PubMed] [Google Scholar]
- 24. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, Peña AG, Goodrich JK, Gordon JI, Huttley GA, Kelley ST, Knights D, Koenig JE, Ley RE, Lozupone CA, McDonald D, Muegge BD, Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ, Walters WA, Widmann J, Yatsunenko T, Zaneveld J, Knight R. 2010. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7:335–336. doi: 10.1038/nmeth.f.303 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Caporaso JG, Bittinger K, Bushman FD, DeSantis TZ, Andersen GL, Knight R. 2010. PyNAST: a flflexible tool for aligning sequences to a template alignment. Bioinformatics 26:266–267. doi: 10.1093/bioinformatics/btp636 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Wang Q, Garrity GM, Tiedje JM, Cole JR. 2007. Naive Bayesian classififier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol 73:5261–5267. doi: 10.1128/AEM.00062-07 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Huber T, Dalevi D, Hu P, Andersen GL. 2006. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol 72:5069–5072. doi: 10.1128/AEM.03006-05 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Stekhoven DJ, Bühlmann P. 2012. MissForest--non-parametric missing value imputation for mixed-type data. Bioinformatics 28:112–118. doi: 10.1093/bioinformatics/btr597 [DOI] [PubMed] [Google Scholar]
- 29. Lundberg SM, Su-In L. 2017. A unified approach to interpreting model predictions, . In Advances in neural information processing systems. Vol. 30. https://proceedings.neurips.cc/paper_files/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf. [Google Scholar]
- 30. Yu J, Sun H, Zhu J, Wei X, Shi H, Shen B, Ren L, He Y, Zhang R, Zhang M, Peng H. 2021. Asymptomatic hyperuricemia and metabolically unhealthy obesity: a cross-sectional analysis in the tianning cohort. Diabetes Metab Syndr Obes 14:1367–1374. doi: 10.2147/DMSO.S301363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Hong C, Zhang Q, Chen Y, Lu Y, Chen L, He Y, Li J, Ma S, Jiang J, Zhang X, Hu J, Ding Y, Zhang M, Peng H. 2022. Elevated uric acid mediates the effect of obesity on hypertension development: a causal mediation analysis in a prospective longitudinal study. Clin Epidemiol 14:463–473. doi: 10.2147/CLEP.S363429 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Snaith ML. 2001. Gout: diet and uric acid revisited. Lancet 358:525. doi: 10.1016/S0140-6736(01)05745-2 [DOI] [PubMed] [Google Scholar]
- 33. Tsushima Y, Nishizawa H, Tochino Y, Nakatsuji H, Sekimoto R, Nagao H, Shirakura T, Kato K, Imaizumi K, Takahashi H, Tamura M, Maeda N, Funahashi T, Shimomura I. 2013. Uric acid secretion from adipose tissue and its increase in obesity. J Biol Chem 288:27138–27149. doi: 10.1074/jbc.M113.485094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Gao X, Zhang M, Xue J, Huang J, Zhuang R, Zhou X, Zhang H, Fu Q, Hao Y. 2018. Body mass index differences in the gut microbiota are gender specific. Front Microbiol 9:1250. doi: 10.3389/fmicb.2018.01250 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Festi D, Schiumerini R, Eusebi LH, Marasco G, Taddia M, Colecchia A. 2014. Gut microbiota and metabolic syndrome. World J Gastroenterol 20:16079–16094. doi: 10.3748/wjg.v20.i43.16079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Borgo F, Garbossa S, Riva A, Severgnini M, Luigiano C, Benetti A, Pontiroli AE, Morace G, Borghi E. 2018. Body mass index and sex affect diverse microbial niches within the gut. Front Microbiol 9:213. doi: 10.3389/fmicb.2018.00213 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Ley RE, Turnbaugh PJ, Klein S, Gordon JI. 2006. Microbial ecology: human gut microbes associated with obesity. Nature 444:1022–1023. doi: 10.1038/4441022a [DOI] [PubMed] [Google Scholar]
- 38. Magne F, Gotteland M, Gauthier L, Zazueta A, Pesoa S, Navarrete P, Balamurugan R. 2020. The firmicutes/bacteroidetes ratio: a relevant marker of gut dysbiosis in obese patients? Nutrients 12:1474. doi: 10.3390/nu12051474 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Deng K, Shuai M, Zhang Z, Jiang Z, Fu Y, Shen L, Zheng J-S, Chen Y-M. 2022. Temporal relationship among adiposity, gut microbiota, and insulin resistance in a longitudinal human cohort. BMC Med 20:171. doi: 10.1186/s12916-022-02376-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Méndez-Salazar EO, Vázquez-Mellado J, Casimiro-Soriguer CS, Dopazo J, Çubuk C, Zamudio-Cuevas Y, Francisco-Balderas A, Martínez-Flores K, Fernández-Torres J, Lozada-Pérez C, Pineda C, Sánchez-González A, Silveira LH, Burguete-García AI, Orbe-Orihuela C, Lagunas-Martínez A, Vazquez-Gomez A, López-Reyes A, Palacios-González B, Martínez-Nava GA. 2021. Taxonomic variations in the gut microbiome of gout patients with and without tophi might have a functional impact on urate metabolism. Mol Med 27:50. doi: 10.1186/s10020-021-00311-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Udayappan SD, Kovatcheva-Datchary P, Bakker GJ, Havik SR, Herrema H, Cani PD, Bouter KE, Belzer C, Witjes JJ, Vrieze A, de Sonnaville ESV, Chaplin A, van Raalte DH, Aalvink S, Dallinga-Thie GM, Heilig HGHJ, Bergström G, van der Meij S, van Wagensveld BA, Hoekstra JBL, Holleman F, Stroes ESG, Groen AK, Bäckhed F, de Vos WM, Nieuwdorp M. 2017. Intestinal Ralstonia pickettii augments glucose intolerance in obesity. PLoS One 12:e0181693. doi: 10.1371/journal.pone.0181693 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Zhang-Sun W, Tercé F, Burcelin R, Novikov A, Serino M, Caroff M. 2019. Structure function relationships in three lipids A from the ralstonia genus rising in obese patients. Biochimie 159:72–80. doi: 10.1016/j.biochi.2019.01.015 [DOI] [PubMed] [Google Scholar]
- 43. Lee Y, Kim N, Werlinger P, Suh D-A, Lee H, Cho J-H, Cheng J. 2022. Probiotic characterization of Lactobacillus brevis MJM60390 and in vivo assessment of its antihyperuricemic activity. J Med Food 25:367–380. doi: 10.1089/jmf.2021.K.0171 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental figures and tables.
An accounting of the reviewer comments and feedback.
Data Availability Statement
The raw data for 16S rRNA gene sequences are available in NCBI under accession number PRJEB18535.