Phenylacetylglutamine as a novel biomarker of type 2 diabetes with distal symmetric polyneuropathy by metabolomics

Abstract

Purpose: Type 2 diabetes mellitus (T2DM) with distal symmetric polyneuropathy (DSPN) is a disease involving the nervous system caused by metabolic disorder, while the metabolic spectrum and key metabolites remain poorly defined.

Methods: Plasma samples of 30 healthy controls, 30 T2DM patients, and 60 DSPN patients were subjected to nontargeted metabolomics. Potential biomarkers of DSPN were screened based on univariate and multivariate statistical analyses, ROC curve analysis, and logistic regression. Finally, another 22 patients with T2DM who developed DSPN after follow-up were selected for validation of the new biomarker based on target metabolomics.

Results: Compared with the control group and the T2DM group, 6 metabolites showed differences in the DSPN group (P < 0.05; FDR < 0.1; VIP > 1) and a rising step trend was observed. Among them, phenylacetylglutamine (PAG) and sorbitol displayed an excellent discriminatory ability and associated with disease severity. The verification results demonstrated that when T2DM progressed to DSPN, the phenylacetylglutamine content increased significantly (P = 0.004).

Conclusion: The discovered and verified endogenous metabolite PAG may be a novel potential biomarker of DSPN and involved in the disease pathogenesis.

Keywords: Biomarker; Distal symmetric polyneuropathy; Metabolomics; Phenylacetylglutamine; Type 2 diabetes.

Fig. 1

A brief flowchart of the study

Fig. 2

Identification of differential metabolites of DSPN. a A PCA plot between three groups. b Venn diagram of significant metabolites from the two pairwise comparisons. c Box plots diagram of six differential metabolites showing the same trend. The “+” in the figure represents means. The abscissa shows groups, and the ordinate represents relative intensity normalized by the peak value. *P < 0.05, **P < 0.01, and ***P < 0.001 when compared with control group, respectively; ^#P < 0.05, ^##P < 0.01, and ^###P < 0.001 when compared with DSPN group

Fig. 3

Identification of the potential biomarkers in DSPN. a ROC curve of each metabolite. b ROC curve for the final model. c Box plots for the comparison of six metabolites between the mild DSPN group and the moderate–severe DSPN group. ***P < 0.001

Fig. 4

Details for PAG metabolism pathway. a The relative content difference of mapped 6 metabolites in the pathway by box plots. b The pathway flowchart of PAG metabolism. Red arrows indicate changes in metabolite content in the DSPN group compared with the Con/T2DM groups. Gray fonts within dotted lines indicate metabolism sites

Fig. 5

Details for Sorbitol metabolism pathway. a The relative content difference of mapped 4 metabolites in the pathway by box plots. b The pathway flowchart of sorbitol metabolism. Red arrows indicate changes in metabolite content in the DSPN group compared with the Con/T2DM groups. AR aldose reductase, SDH sorbitol dehydrogenase

Fig. 6

Box plot of PAG content change in the validation cohort. **P < 0.01

Fig. 7

Diagram of the possible potential mechanism of PAG in DSPN state. Solid lines represent confirmed mechanisms, and dotted lines represent putative mechanisms. CNS, central nervous system. PNS, peripheral nervous system. α1, α2A, α2B, β, β1, β2, adrenergic receptor subtype