In this study, we performed a UPLC-MS/MS-based metabolomics analysis to reveal the metabolic profile of 20 AAs in plasma, urine, and saliva samples of patients with DKD. After that, to predict the presence of DKD, we established a distinct diagnostic model based on six differential AAs including plasma histidine and valine; urine histidine, valine, and proline; and, saliva arginine with high accuracy, and we confirmed its performance using the validation set (92.2% by diagnostic feature).
In our study, we found decreased plasma levels of histidine and valine in patients with DKD, these results are consistent with our previous findings(Zhou et al., 2021). Further exploration of the key metabolites and enzymes of histidine metabolism indicated enhanced carnosine hydrolysis, decreased degradation of homocarnosine and anserine, and increased histidine methylation in the circulating blood of patients with DKD. Histidine exerts anti-inflammatory and antioxidant effects by a combination of free radical elimination and metal chelation(Babizhayev, 2006). Insufficient plasma histidine levels are associated with persistent inflammation and oxidative stress in renal diseases, which can be relieved by dietary histidine supplementation(Lee et al., 2005; Watanabe et al., 2008), and the abnormal levels are probably caused by histidine metabolism abnormalities(Lee et al., 2005). Carnosine is an important HDP with renoprotective effects in patients with T2DM and DKD. In rodent models of T2DM, oral carnosine supplementation alleviates structural and functional renal damage via anti-inflammatory, antioxidant, anti-glycation, and reactive carbonyl-quenching mechanisms(Holeček, 2020). Results of a clinical trial showed that carnosine supplementation over 12 weeks improved glucose tolerance(Albrecht et al., 2017; de Courten et al., 2016). Our results are consistent with those of another study indicating that overexpression of CNDP1, the metabolic enzyme of carnosine, is associated with a the risk for poor diabetes control (and a high risk for T2DM and DKD). CNDP2 is another carnosinase that mainly hydrolyzes other HDP substrates (homocarnosine and anserine) under nonphysiological conditions(Boldyrev et al., 2013). In our study, homocarnosine was elevated in patients with DKD due to attenuated activation of the downstream enzyme CNDP2; homocarnosine is endogenously synthesized in skeletal muscle from histidine and γ-aminobutyric acid and its concentration is barely influenced by a low dietary protein intake.
Little is known about the function of the methylated form of carnosine. Anserine is the exogenous HDP synthesized through carnosine methylation in most mammals, fish, and amphibians but not in humans(Derave et al., 2019). By exerting the same effect as carnosine, anserine may improve glucose metabolism, proteinuria, and vascular permeability under diabetic conditions due to its resulting glycation inhibition, oxidative damage reduction, and antioxidant activity enhancement(Derave et al., 2019). Our data show that the plasma levels of anserine were much higher in patients with DKD than in other participants; this may be attributable to a decrease in anserine degradation due to the low expression of CNDP2(Peters et al., 2011). 1-MH is derived from the degradation of exogenous anserine (which is present in the meat of most fish, but not in humans) and is a biomarker of protein intake(Stifel & Herman, 1971). In contrast, 3-MH is endogenously synthesized only in muscle by the methylation histidine residues and it gets excreted in the urine without being metabolized(Holeček, 2020). We found that 3-MH was increased in participants of the DKD group, whereas 1-MH was unchanged in all participants. The 3-MH increases are likely due to higher muscle protein turnover or sarcopenia rather than to restricted dietary protein intake(Kim et al., 2010) .
Valine metabolism was also altered in patients with DKD. Valine is transaminated by BCAT in mitochondria to generate 3-methyl-2-oxobutyrate, and then it gets dehydrolyzed by the BCKDH complex, which produces isobutyryl-coenzyme A, a raw material for odd chain fatty acid (OCFA) synthesis. Valine contributes to the increased levels of liver and circulating OCFA in patients with T2DM via 2 distinct mechanisms–increased α-oxidation and de novo OCFA lipogenesis(Bishop et al., 2020). We observed an increase in 3-methyl-2-oxobutyrate in patients with T2DM, but the circulating levels of BCAT and BCKDH, the enzymes responsible for its metabolism, were not significantly altered. This may be because valine is mainly metabolized in the mitochondria of skeletal muscle and the levels are lower in other peripheral tissues and in the circulation. In our study, we found that patients with DKD had lower plasma levels of 3- methyl-2-oxobutyrate and abnormal upregulation of circulating BCAT1 and BCKDHB, which can promote valine metabolism(Nie et al., 2018). Additional studies are needed to determine how the association between valine metabolism and a restricted dietary intake contributes to DKD development.
We observed lower levels of histidine and valine, as well as higher levels of proline in urine of patients with DKD than those in patients with T2DM or healthy controls. Hyperglycemia is considered the principal cause of the significantly altered urinary AA excretion pattern of patients with T2DM(Verrey et al., 2009). Bingham reported the prevalence of glycosuria-related AAs in all types of diabetes (Bingham et al., 2001). Bidi S found that urine levels of aromatic amino acids, such as phenylalanine and tryptophan, sulfur-containing amino acids including cysteine, and basic amino acids, such as arginine, were significantly higher in patients with T2DM than in healthy controls(Bidi et al., 2020). However, the published data on urine amino acid patterns in DKD is scarce. Kim NH reported that the levels of methionine, valine, and leucine in urine increased in 8-week-old DKD mice, but then decreased to normal healthy control levels at 20 weeks (Kim et al., 2018). Bingham found that the levels of urinary AAs were high in T2DM regardless of the presence of chronic renal failure. Urinary AA levels reflect alterations in blood AAs levels(Babizhayev et al., 1994). Our findings revealed that the urine levels of histidine and valine were significantly lower in patients with DKD than those in patients with T2DM and healthy controls, a fact that may be due to the reduction of histidine and valine in plasma. Indeed, we found the levels of histidine and valine in urine to be moderately to strongly correlated with the levels in plasma.
The proline plasma levels have been shown to be increased in patients with T2DM, obesity, and insulin resistance; such persistent alterations lead ultimately to impaired insulin secretion, systemic glucose homeostasis disruption, and other dysfunctions(Liu et al., 2016). The localized levels of proline get increased due to excessive redox reactions under chronic inflammation and they lead to fibrotic tissue remodeling(Distler et al., 2019). Proline is indispensable in local tissues for collagen synthesis, but large accumulations result in maladapted tissue architectures(Tarbit et al., 2019; Vettore et al., 2021). The increased urinary excretion and decreased plasma levels of proline in patients with DKD in our study are probably due to the local inflammation and fibrosis in the kidney. However, the mechanisms leading to increased AA excretion are unclear. Glucose may also cause decrease AA reabsorption in the kidneys by depolarizing the electrical gradient of sodium-dependent amino acid transporters in the proximal renal tubules.
We found the saliva levels of most AAs in patients with T2DM and DKD to be higher than those in healthy controls. However, most studies on the levels of AAs in saliva have been conducted on animals(Distler et al., 2019). The AA contents in saliva may inadequately reflect those in serum, because the components of saliva are passively diffused from serum and also actively transported through the oral mucosa or gums(Gross et al., 2005; Rheinberger & Böger, 2014). We found the saliva levels of arginine in the DKD group were higher than in T2DM and healthy control groups. However, after analyzing possible correlations of the arginine levels between saliva and plasma, we found none. The mechanism causing the differing arginine levels in the DKD group remain to be elucidated.
To validate the diagnostic performance for AAs differing in the DKD group when compared to those in the other groups in plasma, urine, and saliva, we conducted a logistic regression analysis to generate an optimal diagnostic model. The AUC of our diagnostic model was 0.957 in the corresponding ROC curve, with high specificity and sensitivity. The glomerular filtration rate (eGFR) is the gold standard for evaluating renal function(Gross et al., 2005; Rheinberger & Böger, 2014), but abnormalities are usually undetectable before the occurrence of substantial kidney injury in patients with DKD. DKD is usually diagnosed on the basis of histological features after an invasive puncture biopsy that poses risks of hematuria or perirenal hematoma(Stiles et al., 2000). The lack of suitable early detection methods is a major obstacle to the prevention and treatment of DKD. For our study, we applied targeted metabolomics to plasma, urine, and saliva AA levels, and we found the altered metabolic profiles of 20 AAs in patients with DKD. After screening the AAs with significant difference levels in different biofluids, we were able to establish a simple and accurate DKD diagnostic model.
In conclusion, we report here the metabolomic profiles of 20 AAs of plasma, urine, and saliva from patients with DKD. Logistic regression revealed a distinct diagnostic equation based on six differential AAs. ROC analysis demonstrated that the diagnostic performance of this equation was higher than the diagnostic performance of individual AAs, and the AUC of the diagnostic model was 0.957. The accuracy of the diagnostic model was 92.2% in the validation set, and these results support the use of the model as a viable approach for identifying at-risk populations through the detection of metabolites. Therefore, our signature model based on AA levels may help identify patients at risk of DKD.