Metabolic Profile of Patients with Severe Endometriosis: a Prospective Experimental Study

Abstract

Endometriosis is a common disease affecting women in reproductive age. There are several hypotheses on the pathogenesis of this disease. Often, its lesions and symptoms overlap with those of many other medical and surgical conditions, causing a delay in diagnosis. Metabolomics represents a useful diagnostic tool for the study of metabolic changes during a different physiological or pathological status. We used ¹H-NMR to explore metabolic alteration in a cohort of patients with endometriosis in order to contribute to a better understanding of the pathophysiology of the disease and to suggest new useful biomarkers. Thirty-seven patients were recruited for the metabolomic analysis: 22 patients affected by symptomatic endometriosis and 15 not affected by it. Their serum samples were collected and analyzed with ¹H-NMR. Multivariate statistical analysis was conducted, followed by univariate and pathway analyses. Partial Least Square Discriminant Analysis (PLS-DA) was performed to determine the presence of any differences between the non-endometriosis and endometriosis samples (R²X = 0.596, R²Y = 0.713, Q² = 0.635, and p < 0.0001). β-hydroxybutyric acid and glutamine were significantly increased, whereas tryptophan was significantly decreased in the endometriosis patients. ROC curves were built to test the diagnostic power of the metabolites (β-hydroxybutyric acid: AUC = 0.85 CI = 0.71-0.99; glutamine: AUC = 0.83 CI = 0.68-0.98; tryptophan: AUC = 0.75 CI = 0.54-0.95; β-hydroxybutyric acid + glutamine + tryptophan AUC = 0.92 CI = 0.81-1). The metabolomic approach enabled the identification of several metabolic alterations occurring in women with endometriosis. These findings may provide new bases for a better understanding of the pathophysiological mechanisms of the disease and for the discovery of new biomarkers. Trial registration number NCT02337816.

Keywords: 1H-NMR; Biomarkers; Endometriosis; Metabolomics; Multivariate analysis.

Fig. 1

Signal’s assignments of the serum metabolites in an NMR spectrum. *Uncertain attribution; **Extraction contaminant. 1, lipid 1; 2, 2-OH-butyrate; 3, isoleucine; 4, leucine; 5, valine; 6, 3-OH-butyrate; 7, lipid 2; 8, lactate; 9, alanine; 10, lysine; 11, acetate; 12, proline; 13, N-acetyl-groups; 14, methionine; 15, glutamate; 16, pyroglutamate; 17, glutamine; 18, citrate; 19, aspartate; 20, sarcosine; 21, asparagine; 22, creatinine; 23, creatine; 24, ornithine; 25, choline; 26, glucose; 27, betaine; 28, 1,3-dimethylurate*; 29, methanol**; 30, glycine; 31, glycerol; 32, glucitol*; 33, glycylproline*; 34, fructose; 35, threonine; 36, tyrosine; 37, histidine; 38, tryptophan; 39, phenylalanine; and 40, τ-methylhistidine

Fig. 2

¹H-NMR analysis of serum of non-endometriosis (NE) and endometriosis (E) samples. a Supervised model PLS-DA score plot from NE and E samples generated using the ¹H-NMR serum spectra metabolites. b Validation of the models via permutation test (n = 400). c Bar graph of the metabolites significantly different (p < 0.05) between NE and E from the ¹H-NMR analysis and the corresponding ROC curve generated using the single metabolites and then all the metabolites

Fig. 3

Metabolic altered pathways built using MetaboAnalyst 4.0. The analysis showed different altered pathways such as nitrogen metabolism, pyrimidine metabolism, glutamine and glutamate metabolism, and amnoacyl-tRNA biosynthesis