C-Peptide: A Mediator of the Association Between Serum Uric Acid to Creatinine Ratio and Non-Alcoholic Fatty Liver Disease in a Chinese Population With Normal Serum Uric Acid Levels

Background The data on the relationship between normal-ranged serum uric acid (SUA), β-cell function, and non-alcoholic fatty liver disease (NAFLD) are complicated and insufficient. Moreover, uric acid is excreted by kidney, and SUA levels may be affected by renal function. Thus, we introduced a renal function-normalized index [serum uric acid to creatinine ratio (SUA/Cr)] into the study and explored the association between SUA/Cr, C‐peptide and NAFLD in a Chinese population with normal SUA levels by a cross-sectional analysis. Materials and Methods A total of 282 individuals with normal SUA levels and different glucose tolerance status from a diabetes project were included in the study (mean age = 53.7± 10.5 years; women = 64.50%). NAFLD was diagnosed by abdominal ultrasonography (NAFLD, n=86; without NAFLD, n=196). Trapezoid formula was used to calculate area under the curve of C‐peptide (AUCCP) from 4 points (including 0, 30,60, and 120min) during 2-h oral glucose tolerance test. Spearman correlation analysis was used to explore the correlation between SUA/Cr, AUCCP and NAFLD risk factors. Multiple logistic regression analysis was used to explore the association between SUA/Cr or AUCCP and NAFLD. Mediation analysis was used to explore whether AUCCP mediated the association between SUA/Cr and NAFLD. Results Individuals with NAFLD had significantly higher SUA/Cr and AUCCP than those without NAFLD(P<0.05). Spearman correlation analysis showed that both SUA/Cr and AUCCP were significantly associated with many NAFLD risk factors, and SUA/Cr was positively correlated with AUCCP (P<0.05). Multiple logistic regression analysis indicated that SUA/Cr and AUCCP were positively associated with NAFLD incidence (P<0.05). Medication analysis indicated that SUA/Cr had a significant direct effect on NAFLD (β =0.5854, 95% CI: 0.3232–0.8966), and AUCCP partly mediated the indirect effect of SUA/Cr on NAFLD incidence (β =0.1311, 95% CI: 0.0168–0.4663). Conclusions SUA/Cr was positively associated with NAFLD incidence, and AUCCP partly mediated the association in a Chinese population with normal SUA levels. Thus, we should pay more attention to high-normal SUA and C-peptide levels due to their predictive power in NAFLD incidence.


INTRODUCTION
Non-alcoholic fatty liver disease (NAFLD), characterized by lipid accumulation in liver with no significant alcohol intake, is one of the most common chronic liver diseases in the world. NAFLD can develop into cirrhosis, and even hepatocellular carcinoma and liver failure (1).NAFLD is also closely associated with cardiovascular disease (2), diabetes (3), and obesity (4).Moreover, patients with NAFLD exhibited relative high mortality compared to the general population (5). Hence, finding risk factors and the mechanism of NAFLD is warranted to prevent it.
Serum uric acid (SUA), major product of purine metabolism, is independently associated with NAFLD incidence (6,7). Moreover, the significant association between SUA and NAFLD incidence was also established even in some individuals with normal SUA levels (8,9). SUA is excreted via kidney, and the clearance of SUA is often affected by renal function, while none of the previous studies (6-9) considered the effects from kidney. SUA to creatinine ratio (SUA/Cr) is an index of renal function-normalized SUA, reflecting endogenous UA levels more precisely than SUA. Moreover, SUA/Cr is associated with b-cell function, metabolic syndrome and incident chronic kidney disease (10)(11)(12).However, there has been no study focused on the association between SUA/Cr and NAFLD yet.
In addition to NAFLD, SUA was also associated with islet b-cell function and b-cell secretion in patients with type 2 diabetes (T2DM) and prediabetes (13,14). Moreover, SUA, within normal range, was related to b-cell function in overweight/obesity or male T2DM patients (15). Although the causal relationship between SUA and b-cell function was not conclusive, one study indicated that elevated SUA was the precursor of T2DM (16).There were also some studies reporting the significant association between b-cell secretion and NAFLD. An American study indicated that fasting Cpeptide (FCP) was associated with NAFLD (17). A Chinese study based on obese children also found that FCP is a significant indicator of NAFLD (18). C-peptide and the area under the curve of C-peptide (AUC CP ) could reflect the secretion ability of islet bcell, while the latter is a better indicator of overall and residual islet b-cell secretion compared to FCP. However, the studies involving the association between AUC CP and NAFLD are limited.
Several clinical studies have indicated that increased SUA, within the normal range, is closely associated with many diseases (19,20), and high-normal SUA(elevated SUA within normal range) has already hold our attention. The intrinsic relationship between SUA, b-cell function and NAFLD is complicated, and related studies involving normal-ranged SUA are insufficient. Moreover, renal function may be also a potential confounding factor of the association between SUA and NAFLD. Thus, we performed a cross-sectional study based on a Chinese population with normal SUA levels and introduced an index of renal function-normalized SUA into the study, namely, SUA/Cr, and explored the association between SUA/Cr, AUC CP , and NAFLD.

Study Population
A total of 599 individuals from a T2DM project were recruited between 2014 and 2015 (21), and 333 of them completed liver ultrasound. Participants with other liver diseases, estimated glomerular filtration rate <60 ml/min/1.73 m 2 , alcohol consumption >20 g/day, hyperuricemia (SUA ≥ 416 µmol/L for male or SUA ≥ 357 umol/L for female or uric acid-lowering drugs treatment) or missing data were also excluded (n=51). A total of 282 individuals with normal SUA levels and with different glucose tolerance status were included in the study (Figure 1).

Anthropometric and Biochemical Measurements
All participants were asked to complete a questionnaire, including gender, age, and medical history. Blood pressure, waist circumference (WC) and body mass index (BMI) were measured using standard methods. Blood samples for glucose and C-peptide were obtained at 0, 30, 60, and 120 min after a 75-g oral glucose load. Serum C peptide levels were measured by chemiluminescence immunoassay using a Siemens ADIVA Centaur XP analyzer (Siemens Healthcare Diagnostics Inc., Tarrytown, NY, USA), while serum glucose concentrations were assayed using a glucose Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; AUC CP , area under the curve of C-peptide; AUC Glu , area under the curve of glucose; BMI, body mass index; CI, confidence interval; FCP, fasting C-peptide; 2hCP, 2-h postload C-peptide; FGF21, fibroblast growth factor 21; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, homeostasis model assessment of insulin resistance; 2hPG, 2-h post-load glucose; LDL-C, low-density lipoprotein cholesterol, NAFLD, non-alcoholic fatty liver disease; OR, odds ratio; SUA, serum uric acid; SUA/Cr, serum uric acid to creatinine ratio; TC, total cholesterol; T2DM, type 2 diabetes; TG, triglycerides; WC, waist circumference.

Ma et al.
Uric Acid, C-Peptide, and NAFLD Frontiers in Endocrinology | www.frontiersin.org November 2020 | Volume 11 | Article 600472 oxidase assay. Hemoglobin A1c (HbA1c) analysis was analyzed in whole blood using high-performance liquid chromatography. The diagnosis of diabetes and prediabetes were defined based on the1999 World Health Organization criteria after the oral glucose tolerance test (OGTT) (22). Fasting serum alanine transaminase (ALT), aspartate transaminase (AST), SUA, creatinine, and lipids were measured by an automated analyzer. Trapezoid formula was used to calculate AUC for glucose and C-peptide. Insulin resistance and insulin sensitivity was estimated using homeostasis model assessment of insulin resistance (HOMA-IR) (23).

Liver Ultrasonography Evaluation
NAFLD was diagnosed by abdominal ultrasonography. The ultrasound results were assessed by physicians who were blind to the subjects' biochemical results. Except for individuals with significant alcohol consumption, subjects were diagnosed with NAFLD by ultrasonography if at least two of the following three ultrasonic characteristics were positive: bright liver, liver echo greater than kidney, vascular blurring, and deep attenuation of ultrasound signal (24).

Statistical Analysis
Normally distributed continuous variables were recorded as mean ± standard deviation. Non-normal distribution parameters were transformed or presented as the median (25th-75th percentile). Categorical data were presented as number and percentages (n, %). Differences between groups were compared by Student's t test or Chi-squared test or Mann-Whitney's U-test. Spearman correlation analysis was used to explore the association between SUA/Cr or AUC CP and potential NAFLD risk factors. Multiple logistic regression analysis was used to evaluate the association between SUA/Cr or AUC CP and NAFLD. Mediation models were established to explore whether AUC CP mediated the association between SUA/Cr and NAFLD. Analyses were conducted using the SPSS (version 22.0). P < 0.05 (2-tailed) was considered statistically significant.

Clinical Characteristics of Individuals in NAFLD and Non-NAFLD Groups
According to the liver ultrasonography, individuals were divided into a NAFLD group (n=86) and a non-NAFLD group (n=196). As recorded in Table 1, individuals with NAFLD had higher BMI, WC, total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), HbA1c, 2-h post-load glucose (2hPG), area under the curve of glucose (AUC Glu ), FCP, 2-h post-load C-peptide (2hCP), AUC CP , HOMA-IR, ALT, AST, and SUA/Cr but lower high-density lipoprotein cholesterol (HDL-C) compared to those without NAFLD(P<0.05). In addition, individuals with NAFLD were more likely to have higher prevalence of diabetes than those without NAFLD ( Table 1).

Mediated Effect of AUC CP on the Association Between SUA/Cr and NAFLD
Both SUA/Cr and AUC CP were positively associated with NAFLD incidence, while SUA/Cr was positively correlated with AUC CP , suggesting a mechanistic link between SUA/Cr and NAFLD, possibly explained by AUC CP . To explore the internal relationships between AUC CP , SUA/Cr and NAFLD, we conducted mediation analysis to explore whether AUC CP mediated the association between SUA/Cr and NAFLD incidence.

DISCUSSION
This study, based on a Chinese population with normal SUA levels, indicated that individuals with NAFLD had higher SUA/   Cr and AUC CP than those without NAFLD. Both SUA/Cr and AUC CP were significantly correlated with many conventional risk factors of NAFLD, and the correlation between SUA/Cr and AUC CP was positive. In addition, SUA/Cr was positively associated with NAFLD incidence, and AUC CP partly mediated the indirect effect of SUA/Cr on NAFLD incidence in a population with normal SUA levels.
The association between SUA and NAFLD has been explored for a long time. A Chinese study contained 21,798 subjects revealed that SUA was significantly associated with NAFLD incidence (25). A prospective observational study demonstrated that high SUA independently predicted 3-year's incidence of NAFLD (26). Moreover, even in individuals with normal SUA levels, increased SUA was independently associated with NAFLD (8,9). UA is the end product of human purine metabolism and is excreted by the kidney. SUA level will increase due to its impaired clearance in individuals with impaired renal function (27).However, many studies ignored the effect of kidney on SUA, while SUA/Cr is a renal function-normalized index and may be a more precise indicator than SUA. A Chinese study based on 713 diabetics revealed that SUA/Cr was significantly associated with b-cell function (11). Another study suggested that SUA/Cr in T2DM patients was closely linked to metabolic syndrome and its components (10). Furthermore, a longitudinal study indicated that SUA/Cr had stronger associations with chronic kidney disease than SUA alone (12). Similarly, our present study firstly demonstrated that SUA/Cr was an independent risk factor of NAFLD in individuals with normal SUA levels, and mediation analysis indicated SUA/Cr had direct effect on NAFLD. Although the detailed mechanism of NAFLD remains uncertain, many studies have indicated that there is a close association between SUA and NAFLD. SUA may function as a pro-oxidant and react with oxidants, inducing the production of free radicals and oxidative stress (28), which are critical factors in the development of NAFLD (29).Thus, SUA may have direct effect on NAFLD as a pro-oxidant.
We also found the effect of SUA/Cr on NAFLD partly via AUC CP . Several studies have indicated that FCP was independently associated with NAFLD (17,18),while we used a more accurate index (AUC CP ), which could reflect overall b-cell secretion. Equimolar amount of C-peptide is produced when insulin is secreted. However, insulin, not C-peptide, is partly cleared in liver with first-pass hepatic extraction (30). Thus, serum C-peptide was a well-established marker of the endogenous insulin secretion. We found that individuals with NAFLD had higher AUC CP and HOMA-IR, and AUC CP was significantly correlated with HOMA-IR. We also found AUC CP partly mediated the association between SUA/Cr and NAFLD. Although the causal relationship between UA and insulin resistance was not conclusive, elevated SUA may aggravate insulin resistance to some extent. A clinical study indicated that elevated SUA was the precursor of T2DM (16). SUA could induce endothelial dysfunction and inhibit nitric oxide bioavailability, which is involved in insulin resistance (31). Thus, higher AUC CP may represent higher endogenous insulin secretion and may be a compensatory of insulin resistance due to higher SUA, namely, higher AUC CP is a sign of insulin resistance, which plays important role in the progress of NAFLD (32). In other hand, many metabolic regulators such as follistatin and fibroblast growth factor (FGF21) were regulated by islet hormone (33,34). Moreover, fold changes of C-peptide during an OGTT were inversely associated with those of FGF21 in individuals with normal glucose tolerance (35), and FGF21 related signal pathways played important roles in the progression of NAFLD (36). In addition, previous studies have indicated that C-peptide may be also a predictive marker of the severity of the cardiovascular disease (37) and mortality (38), which suggested C-peptide was a bioactive peptide with other potential physiological functions. Thus, AUC CP may partly mediate the association between SUA/Cr and NAFLD via insulin resistance and other potential physiological function.
Besides AUC CP , other mechanism may be also involved in the association between SUA and NAFLD. A previous study indicated that SUA, within normal range, was positively associated with inflammation markers (39), which may be the important mediator in the development of NAFLD (40). Basic studies showed that SUA also caused hepatic steatosis and liver fat accumulation via endoplasmic reticulum stress (41) and mitochondrial oxidative stress (42).In addition, SUA may generate from fructose metabolism, which could induce hepatic steatosis (43). Overall, high-normal SUA was positively associated with NAFLD incidence via its direct pro-oxidant effect, C-peptide, and other signal pathways. Thus, high-normal SUA and C-peptide levels are important factors in the pathological process of NAFLD, and we should pay enough attention to these indicators. Our present study had some advantages. First, present study introduced SUA/Cr as a newer index into the study and revealed that there was a significant association between SUA/Cr and NAFLD incidence, which brought a more accurate predictor of NAFLD. Second, we explored the internal relationship between SUA/Cr, AUC CP , and NAFLD by different statistical methods, which strengthens our understanding of their internal relationship. Third, our present study based on a population with normal SUA levels and found its strong predictive ability for NAFLD, which suggested that high-normal SUA should cause our attention. Our resent study also had some limitations. First, liver biopsy has been established as the gold diagnosis standard of NAFLD, while NAFLD was determined by ultrasonography with no histologic confirmation in our present study. Nevertheless, ultrasonography is the widely-used methodology to detect NAFLD because of safety, availability, and economy. Second, the nature of the cross-sectional study and the relatively small sample size were also the limitations of the present study, and therefore larger scale and longitude studies are warranted in the future.

CONCLUSIONS
SUA/Cr was positively associated with NAFLD incidence, and AUC CP partly mediated the association between SUA/Cr and NAFLD incidence in a Chinese population with normal SUA levels. This finding indicates that we should pay more attention to high-normal SUA and C-peptide levels due to their predictive power in NAFLD incidence.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/ supplementary material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of Peking Union Medical College Hospital. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
CFM conducted the research, performed the statistical analysis, and wrote the first draft of the manuscript. YiL, SH, JZ, PL, CXM, FP, HZ, LX, and WL contributed to the discussion, conducted the research, and collected the data. YuL designed the study and revised the manuscript. All authors contributed to the article and approved the submitted version.