Dietary protein restriction is commonly recommended in moderate to advanced chronic kidney disease (CKD) patients; however, the effectiveness of consumption of low protein diet (LPD, <0.8 g/kg body weight/day) to preserve renal function remains an area of continuous debate [1
]. Low protein intake can ameliorate proteinuria, through the regulation of intraglomerular pressure and angiotensin pathway, decrease sodium loading and reduce urea and nitrogenous wastes; consequently, limits uremia [4
]. However, nutritional imbalance and protein energy wasting represent key concerns, in regards to dietary adherence and surveillance, particularly in special populations (such as children in growth age or elderly patients). While the implementation of very low protein diet (VLPD, 0.4–0.6 g/kg body weight/day) supplemented with ketoanalogues amino acids can retard renal progression without development of caloric or nutritional detriments [1
]; however, experimental model has demonstrated that VLPD enhance inflammation, malnutrition and aortic calcification [6
]. The precise benefit and pathophysiology of LPD on gut-renal axis remain partially elucidated in CKD patients.
The diet constitutes the substrate for intestinal fermentation affecting the gut microbiota and leading to production of diverse metabolites causing metabolic disarrangements [7
]. Little is known about the nutrient-associated microbiome changes in CKD patients, considering all the intestinal dysbiosis and dietary restrictions presented in renal patients. Significant reduction in serum levels of p-cresyl sulfate (pCS) and changes of gut microbiota were found in moderate CKD patients receiving 6-month of LPD. However, no clustering of pattern was observed in the gut microbiota of patients consuming LPD or free diet [9
]. The implementation of VLPD was associated with a reduced abundance of Proteobacteria and increased level of Blautia, Faecalibacterium and Coprococcus and Roseburia species and the concomitant reduction in serum levels of both indoxyl sulfate (IS) and pCS, compared to those patients receiving free diet [10
]. Despite significant changes in the abundance of selected intestinal microbes associated with dietary protein deprivation, however, modifications of bacterial functional capability and other related metabolites secondary to this adaptive change of gut microbiota remain unclear. Hence, in the present study, we aim to explore the change in intestinal microbiota, related metabolomic profiling and bacterial functional capability associated with LPD in CKD patients.
Knowledge on diet–microbiome–metabolite interaction of CKD patients remains mandatory to support long-term dietary interventions, which allows modulation of an individual’s enterotype to preserve renal function. Comprehensive discernments of relationships between change of gut microbiota and serum metabolomic profiling associated with different dietary instruction remain incompletely understood in CKD patients. Here, we reported significant change of composition and diversity of gut microbiota and its associated functional shift in strong association with circulating metabolites in CKD patients receiving LPD. CKD-LPD patients had significant reduction in the relative abundance of many butyrate-producing bacteria (family Lachnospiraceae and Bacteroidaceae) associated with enrichment of functional module of metabolism of carbohydrate, specifically, the butanoate metabolism. The abundances of these microbes were highly correlated with daily protein intake. Consequently, CKD-LPD patients had lower serum levels of acetic acid, heptanoic acid and nonanoic acid. In addition, the serum levels of glyco λ-muricholic acid was significantly increased in patients under dietary protein restriction compared to CKD-NPD. Correspondently, the abundance of microorganism responsible for fermentation of secondary bile acid (family Veillonellaceae, genera Megasphaera) was increased with differential enrichment of microbial gene associated with secondary bile acid biosynthesis. Overall, our analyses reveal signatures and functions of gut microbiota related to dietary protein restriction in CKD patients.
Consistent with our [11
] and many other reports [22
], in which a descending trend in α-diversity was associated with the CKD severity, the decrease in gut microbial diversity between CKD and normal control subjects was also noted in the present study. However, the difference of ⍺-diversity between CKD-LPD and CKD-NPD patients was not as prominent in our investigation as in others [10
]. In contrast with these reports, our study has demonstrated a strong bacterial community dissimilarity (β-diversity) between patients receiving LPD vs. NPD. The discrepancy observed between studies may in part be explained in the limited sample size between studies and also in the supplementation given to patients having LPD. The addition of inulin [24
] and ketoanalogues amino acids [10
] may have an influence on the gut microbiota by providing extra substrates for microbial nutrient metabolism compared to our mere LPD intervention.
The SCFAs are gut-derived metabolites produced from fermentation of dietary fiber by anaerobic microbes. Acetate, propionate, and butyrate are the three most common SCFAs and exert many renoprotective properties, such as anti-inflammation, anti-atherosclerosis, anti-oxidative functions [25
]. Our and other previous investigations have indicated a reduction in levels of SCFA, especially the butyrate, associated with decreased butyrate-producing bacteria (family Lactobacillaceae and Prevotellaceae) in CKD patients [11
]. With the restriction of dietary protein intake, the abundance of many of these bacteria (Pseudobutyrivibrio, Lachnospira, Eubacterium_hallii_group, Roseburia, Coprococcus, Fusicatenibacter, Anaerostipes, Lachnoclostridium and Prevotellaceae_NK3B31) and serum levels of specific SCFA/MCFA (acetic acid, heptanoic acid and nonanoic acid) were lower compared to CKD-NPD or non-CKD controls. In contrast to the decrease in abundance of most of the butyrate-producing bacteria observed in our study, the abundance of Faecalibacterium prausnitzii (a main butyrate-producer) was increased in our and other patients consuming LPD [10
] than in CKD-NPD patients; in spite of this increase, however, the abundance of this microbe in CKD-LPD patients remained low compared to the abundance of non-CKD control. The negligible effect of LPD in increasing levels of SCFA in CKD patients may in part be attributed by the severe gut dysbiosis caused by uremic milieu and also by the reduced fiber intake in CKD patients. The relationships between levels of SCFA/MCFA and the outcome of patients receiving dietary protein restriction deserve further study.
The serum concentrations of total bile acids are increased in renal patients because of the reduction in glomerular filtration and derangement of bile acid metabolism secondary to the gut dysbiosis of CKD [29
]. Increased serum levels of taurocholic acid, taurochenodeoxycholic acid, taurohyocholic acid and tauro α-muricholic acid were associated with death in ESRD patients [30
]. Secondary bile acids are derived by gut microbes via the biotransformation of primary bile acids produced in liver. Reciprocally, the bile acid receptor, also known as farnesoid X receptor (FXR), expressed at high concentration in both ileum and liver, can exert negative feedback on the liver production of primary bile acids, in the situation of elevated levels of secondary bile acids [31
]. The biological functions of secondary bile acids are pleotropic and remain elusive. They are proposed to have roles on host energy production, intestinal immunity, oxidative damage, colonic carcinogenesis and dysmetabolism, such as diabetes or obesity [31
]. In a germ-free mice model, tauro-conjugated muricholic acids act on FXR of ileum to result on the suppression of bile acid synthesis in the liver [34
]. The expansion of glyco λ-muricholic acid observed in CKD-LPD patients may affect the regulation of bile acids burden in adapting change of dysbiosis or of specific dietary pattern in CKD patients; however, the exact roles need further investigation.
Nutritional or iatrogenic therapy alters intestinal microbiota resulting in alleviation of serum levels of IS and pCS in patients receiving LPD, oral vancomycin, prebiotics or probiotics [10
]. The change of tryptophanase-producing bacteria was remarkable in our CKD-LPD patients as well as several studies (family: clostridiacea, ruminococcaceae, lachnospiraceae and genera: roseburia, faecalibacterium) [10
]. Unexpectedly, the serum levels of IS and pCS did not vary in our patients receiving LPD, in spite of significant change of microbiota and gene marker associated with phenylalanine, tyrosine and tryptophan biosynthesis. The difference on the source of dietary protein (red meat or soy bean) can shape gut microbial composition. A Western diet characterized by animal protein and fat was associated with predominance of Bacteroides enterotypes versus the Prevotella enterotype observed in carbohydrates-based diet [8
]. The small sample size and the low residual renal function of the CKD-LPD group may also contribute to this divergence. Recently, simultaneous measurements of levels of p-cresol, indol and indol-3-acetic acid in feces, plasma, and urine of different stages of CKD patients found that intestinal generation of these toxins did not contribute to the difference of concentration detected in their serum. The renal tubular clearance represented the key determinant of serum concentration of these solutes [37
]. However, the differences on the abundance of selected microbes and other gut-producing metabolites between patients with distinct protein intake remained significant in the resampling subset of subjects having comparable renal function (Table 4
). It is likely that LPD might eventually lead to greater uremic symptoms even if rate of GFR progression is slower given the uremic toxin differences. The possibility of renoprotection associated with lowering of gut-producing uremic toxins, induced by the manipulation of dietary protein, remains to be proven in large trials.
In addition to the changes of microbial gene abundances related to metabolism modules (Figure 3
), enrichments of D-alanine metabolism, synthesis/degradation of ketone bodies and glutathione metabolism were noted in CKD-LPD patients. D-alanine metabolism intervenes in the glucose-alanine cycle of gluconeogenesis and participates actively in process of protein synthesis. The ketone bodies are substrates contributing to lipogenesis and sterol biosynthesis in anabolic condition and can also reduce oxidative stress by inhibiting reactive oxygen species production and increasing antioxidant proteins to prevent lipid peroxidation and protein oxidation during periods of starvation [38
]. Likewise, gut microbiota interacts, through the modification of substrate availability secondary to protein restriction, with the diet leading to metabolic pathway reprogramming and impacting on host functioning to adapt this nutrient manipulation. Together with the synergistic changes of metabolomic profile observed in patients receiving LPD, the findings of this study illustrated remodeling and adaption of gut microbiota and their genetic potential to compensate possible energy wasting in face of dietary protein restriction in renal patients.
The causality of the association should be interpreted with caution because of the cross-sectional design, small number of patients and limited dietary intervention of only three-months in this study. Several shortcomings should be also addressed, including unique ethnic group, inference of functional capacities of bacterial communities based on 16S rRNA gene sequencing and unavailability of fecal concentration of metabolites to reflect their intestinal generation. However, detailed dietary recall, accurate recording of daily protein intake from 24h urine nitrogen estimates and matching of common confounding characteristics from baseline may all minimize bias of the study and strengthen the conjecture of our supposition. Further prospective longitudinal or randomized studies with breakthrough methodologies, such as shotgun metagenomic sequencing, may help to elucidate the function of LPD intervention on mysterious intestinal microbiome–host metabolite synergies in order to preserve renal function of CKD patients.