The change of gut microbiota‐derived short‐chain fatty acids in diabetic kidney disease

Abstract Background Previous studies found the dysbiosis of intestinal microbiota in diabetic kidney disease (DKD), especially the decreased SCFA‐producing bacteria. We aimed to investigate the concentration of the stool and serum short‐chain fatty acids (SCFAs), gut microbiota‐derived metabolites, in individuals with DKD and reveal the correlations between SCFAs and renal function. Methods A total of 30 participants with DKD, 30 participants with type 2 diabetes mellitus (DM), and 30 normal controls (NC) in HwaMei Hospital were recruited from 1/1/2018 to 12/31/2019. Participants with DKD were divided into low estimated glomerular filtration rate (eGFR)(eGFR<60ml/min, n=14) and high eGFR (eGFR≥60ml/min, n=16) subgroups. Stool and serum were measured for SCFAs with gas chromatograph‐mass spectrometry. Results The DKD group showed markedly lower levels of fecal acetate, propionate, and butyrate versus NC (p<0.001, p<0.001, p=0.018, respectively) [1027.32(784.21–1357.90)]vs[2064.59(1561.82–2637.44)]μg/g,[929.53(493.65–1344.26)]vs[1684.57(1110.54–2324.69)]μg/g,[851.39(409.57–1611.65)] vs[1440.74(1004.15–2594.73)]μg/g, respectively, and the lowest fecal total SCFAs concentration among the groups. DKD group also had a lower serum caproate concentration than that with diabetes (p=0.020)[0.57(0.47–0.61)]vs[0.65(0.53–0.79)]μmol/L. In the univariate regression analysis, fecal and serum acetate correlated with eGFR (OR=1.013, p=0.072; OR=1.017, p=0.032). The correlation between serum total SCFAs and eGFR showed statistical significance (OR=1.019, p=0.024) unadjusted and a borderline significance (OR=1.024, p=0.063) when adjusted for Hb and LDL. The decrease in serum acetate and total SCFAs were found of borderline significant difference in both subgroups (p=0.055, p=0.050). Conclusion This study provides evidence that in individuals with DKD, serum and fecal SCFAs levels (fecal level in particular) were lowered, and there was a negative correlation between SCFAs and renal function.


| INTRODUC TI ON
Diabetic kidney disease (DKD) is the most serious complication of diabetic mellitus (DM) and the leading cause of chronic kidney disease (CKD) in the world. A recent study indicated that the prevalence of DM in China was 11.2% (95% confidence interval 10.5% to 11.9%), especially in Han ethnicity. 1 About 35% of patients with type 2 DM (T2DM) would eventually develop DKD, with an increased mortality, 2 but the etiology of diabetic kidney disease is yet still unclear.
Recent studies highlighted the involvement of gut-kidney axis in nephropathy. 3,4 Tao et al. demonstrated that gut microbiota composition was associated with the occurrence of DKD, and the individuals with DKD could be accurately distinguished from individuals with diabetes by the variables of two genera (g_Escherichia-Shigella and g_ Prevotella_9). 5 Another study showed that fecal microbiota transplantation could reverse intestinal microbiota dysbiosis and improve renal function in rats with DKD. 6 These suggested that gut microbiota dysbiosis may play an important role in the pathogenesis of DKD.
Besides, studies also indicated that gut microbiota and kidney were interacted via gut-kidney axis, which also participated in kidney injury process. Being one of the major metabolites of microbiota-mediated fiber fermentation process in the gut, short-chain fatty acids (SCFAs) have attracted considerable interest. SCFAs are a subset of fatty acids that contain 6 or less carbon molecules and have shown beneficial effects on kidney. 4,7 SCFAs played a role in biological modulation by attenuating the inflammatory response and reducing mean arterial pressure, via inhibiting histone deacetylases (HDACs) and activating G protein receptor 41(GPR41), GPR43, GPR109a, and Olfr78. 8,9 However, SCFAs presented markedly varied concentrations in different diseases. 10,11 The change of fecal and serum SCFAs levels in DKD remains unclear.
In this study, all 90 participants were included from HwaMei Hospital. Fecal and serum samples were measured for SCFAs with gas chromatograph-mass spectrometry (GC-MS). We reported the substantial variations in the levels of fecal and serum SCFAs among normal controls, participants with diabetes, and participants with DKD. SCFA levels in participants with diabetic kidney disease were further analyzed within subgroups by renal function.  Figure S1. Informed consent for the study and the publication was obtained from each participant.

| Fecal and serum sample collection
Fresh fecal samples were collected and a portion of 200mg was utilized for each test. Blood samples were collected in the fasting status and serum was obtained by centrifugation at 3,500rpm for 5min at 4℃. These samples were then stored at −80°C until usage. One fecal sample and one serum sample in DKD group were later found not usable and were excluded in the study. 30 serum samples in NC group were not collected from the physical examination center.
Hence, 30 fecal samples in NC group, 30 fecal and serum samples in the diabetes group, and 29 fecal and serum samples in the group with DKD were used for data determination.

| Determination of SCFAs using gas chromatograph-mass spectrometry (GC-MS)
The analysis was performed using the GC-MS 7890A-5975C (Agilent Technology, USA). A FFAP capillary column (30m×0.25mm×0.25μm) was used for chromatographic separation, and helium (1 mL/min) was used as the carrier gas. The stepwise chromatographic thermal conditions were as follows: 100°C for 1 min, 5°C/min to 160°C, 40°C/ min to 240°C, maintaining for 10 min. The mass spectrometer was set to scan mode at m/z 100-300 and selected ion monitoring mode at m/z 60 for acetate, butyrate, iso-valerate, valerate, and caproate, maintaining for 4.72min、 7.34min、8.90min、8.03min, and 11.26min respectively, as well as m/z 73 for propionate and iso-butyrate for 5.90min and 6.31min separately.

| The correlations between SCFAs and the biochemical indicators
Correlations between the fecal SCFAs and clinical indicators were estimated by Spearman's correlation analysis ( Figure 3). As ex-

| The subgroup analysis of fecal and serum SCFAs in DKD
To study the fecal and serum SCFAs in patients with various renal function, we categorized the DKD patients into two subgroups according to the eGFR level, the low GFR subgroup (eGFR<60ml/min, n=14), and the high GFR subgroup (eGFR≥60ml/min, n=16). The baseline data of the two subgroups were shown in Table S1. Age, gender, and BMI between the two groups were matched with no statistical difference (p>0.05). UACR, serum creatinine, and blood urea nitrogen were higher (p<0.05) in the low GFR subgroup compared with the high GFR subgroup with statistically significant difference.
There were no differences in fecal SCFAs between the two subgroups (p>0.05). As shown in Table S2, serum acetate and total SCFAs were lower and with borderline significant in the low GFR subgroup versus the high GFR subgroup (p=0.055, p=0.050, respectively). However, other SCFAs had no difference between these two subgroups (p>0.05).

| DISCUSS ION
It is the first study to investigate fecal and serum SCFAs simultaneously in individuals with DKD. In this study, fecal acetate, propionate, butyrate, and total SCFAs were markedly lower in the DKD group. Serum acetate and total SCFAs were also found lower in the low GFR subgroup. Furthermore, fecal and serum acetate seem to be respectively correlated with eGFR in DKD patients. Besides, serum total SCFAs seem to be an independent factor for renal function.
SCFAs are end products of bacterial carbohydrate fermentation, and function as an important energy source and signaling molecules. 14 The concentration of SCFAs varies among different diseases.
In DKD mice, there was a significant decrease in propionic acid and butyric acid contents in DKD progression. 15  The gut microbiota, yielding SCFAs as the major products, was also believed to involve with DKD. Studies have clearly outlined the changes in microbiota in DKD patients, 5,17 that the richness of gut microbiota and the variation of bacteria population were found different in DKD compared to DM 5 and SCFAs-producing bacteria Prevotella declining in DKD patients. 5 We speculated that this reduction of SCFAs-producing bacteria was accompanied by the decrease of yielding SCFAs. Maybe this was the result of the lowest fecal SCFAs levels in DKD.
Despite the finding of fecal SCFAs changes, there has not been a defined study on the subsequent serum SCFAs in DKD patients.
Our study revealed that the serum acetate was lower in the low GFR subgroup than in the high GFR subgroup with a borderline significant difference. This change is postulated to be caused by changes in medication, gastrointestinal microecology, and host physiology and pathology. However, we noticed that the main types of SCFAs, including acetate, propionate, butyrate, and valerate did not change significantly in DKD group versus DM group, which was unexpected given recent literature identifying a significant decline in SCFAs-producing bacteria with advancing kidney disease. 17 Wang et al demonstrated that serum acetate and butyrate level was significantly lower in CKD 5 patients than in CKD 1-4 patients. 16 Jadoon et al found a significant graded decrease in the concentration of acetate, but the plasma valerate concentration increased in patients with advancing kidney disease than in mild CKD patients. 18 Paradoxically, in streptozotocin (STZ)-induced DKD rats, serum acetate levels were markedly elevated compared with controls. 6 The conclusions indicated by our study vary from the above studies, assuming that being associated with the small sample size and the few participants with CKD 5, as well as the low peripheral concentration of SCFAs, which may mitigate the changes. 16 Furthermore, the discrepancies of SCFAs change were possibly due to different etiology of CKD, various severities of the disease, and different animal models. 6 Meanwhile, intestinal microecology is known to be complex and each type of bacteria plays a role when the ecology changes. Therefore, it is significant to investigate the types and concentrations of SCFAs in a larger group of DKD patients. Notably, we identified a significant decline of the level of serum caproate in DKD patients than in DM patients in our study. It in line with the study that serum caproate concentration decreased in CKD 3 patients compared to non-CKD participants conducted by Wu et al. 11 SCFAs diffuse through the intestinal mucosa and enter the bloodstream via the portal vein. 19,20 Samuel et al found that the intestinal absorption of SCFA seems to be influenced by the Gprotein-coupled receptor (GPCR), which is broadly distributed in mammalian organisms. 21 However, serum SCFAs were not in parallel with fecal SCFAs changes in DM and DKD patients in our study. It is assumed that SCFAs measured in circulation may not be utilized in fecal SCFAs excretion, therefore fecal SCFAs may be more accurate in revealing SCFAs absorption or production. 22  ileal, jejunal, and gastric epithelial tight junction in different models of CKD in rats and in cultured human colonocytes exposed to uremic human plasma. 23 28 In the recent studies, SCFAs played an important effect on multiple aspects of renal physiology, inhibiting inflammation, immunity, and fibrosis, decreasing blood pressure, and adjusting energy metabolism. 29 Protective effects of SCFAs on DKD have also been reported, via activation of GPCRs and the inhibition of HDAC activity.
Administration of sodium butyrate (NaBu), the major members of SCFAs, ameliorates mesangial matrix expansion, fibrosis, and inflammation in the kidneys of STZ-induced diabetic rats. 30 and details of the physiological effects are sparse. Previous work has identified these as ligands for GPCR, 35 which influence a variety of metabolic, immune, and vascular processes. 36 In this study, we did not use nutrition diaries, but all participants were interviewed for dietary habits and were explicitly asked for special dietary habits. Since all participants reported a Chinese omnivorous diet without any special dietary habits, dietary habits were unlikely to be a major confounder in the investigated subjects. However, there are some limitations in our cross-section study, consequently we could not demonstrate the causal relationship between fecal, serum SCFAs, and the presence of DKD. This monocentric study included a small number of patients in China, prudence needs to be taken when trying to extrapolate our data to other populations. Besides, the composition and construction of gut microbiota in participants were not analyzed, therefore the relationship between fecal and serum SCFAs and gut microbiota was not identified.
In conclusion, this study provides evidence for quantitative reduction of gut microbial products-SCFAs (fecal acetate, propionate, and butyrate in particular) in DKD patients, demonstrating the association of SCFAs with renal function in DKD.

ACK N OWLED G M ENTS
We sincerely appreciate the participation of each patient in the study. The GC/MS was performed by Metabo-Profile Biotechnology (Shanghai) Co., Ltd.

CO N FLI C T O F I NTE R E S T S
The authors declare that they have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated and/or analyzed during this study are available from the corresponding authors upon reasonable request.