Targeted metabolomic analysis of nitric oxide/L-arginine pathway metabolites in dementia: association with pathology, severity, and structural brain changes

Abstract

L-Arginine/NO pathway is altered in Alzheimer disease (AD). Its clinical relevance and pathway status in vascular dementia (VaD) are unknown. Using targeted metabolomics (a liquid chromatography-mass spectrometry) we assessed L-arginine, L-citrulline, dimethylamine (DMA), asymmetric dimethyl arginine (ADMA) and symmetric dimethylarginine (SDMA) in AD (n = 48), mixed-type dementia (MD; n = 34), VaD (n = 40) and non-demented individuals (n = 140) and determined their clinical relevance (the association with dementia pathology, cognitive impairment, and structural brain damage). L-Arginine, ADMA, L-arginine/ADMA, and L-citrulline levels were decreased in dementia and L-arginine, L-citrulline, age and sex were its independent predictors correctly classifying 91% of cases. L-Arginine and L-arginine/ADMA were differentiating between VaD and AD with moderate accuracy. L-Arginine, L-arginine/ADMA, SDMA, and DMA reflected structural brain changes. DMA and L-citrulline were elevated in patients with strategic infarcts and SDMA, L-arginine/ADMA, and DMA were independent predictors of Hachinski ischemic score. ADMA and SDMA accumulation reflected severity of cognitive impairment. In summary, L-Arginine/NO pathway is altered in neurodegenerative and vascular dementia in association with neurodegenerative and vascular markers of brain damage and severity of cognitive impairment.

Figure 1

Intermediates of NO metabolism in patients with and without dementia: (a) arginine; (b) ADMA; (c) SDMA; (d) citrulline; (e) DMA; (f) Arg/ADMA. Data analyzed using t-test for independent samples. Geometric mean values and 95% confidence interval are given below the dot-plots and presented graphically by inverse triangles and whiskers.

Figure 2

Receiver operating characteristics (ROC) curves: (a) arginine as dementia predictor; (b) citrulline as dementia predictor; (c) arginine and citrulline combined as dementia predictor; (d) model including log(arginine), log(citrulline), age and sex as dementia predictor; (e) arginine as a differential marker for vascular and Alzheimer dementia; (f) model including log(arginine) and BMI as a differential marker for vascular and Alzheimer dementia. AUC, area under ROC curve; sens., sensitivity; spec. specificity.

Figure 3

Intermediates of NO metabolism in study population with respect to dementia pathology: (a) arginine; (b) ADMA; (c) SDMA; (d) citrulline; (e) DMA; (f) Arg/ADMA. Data analyzed using analysis of covariance (ANCOVA) with age as a covariate either in a whole population with significance of age (A) and group (G) effect marked with “1” in superscript. The same analysis was repeated exclusively on dementia patients and the significance of age and group effects was marked as “2” in superscript. Geometric mean values are presented graphically by inverse triangles and 95% confidence interval around mean by whiskers. C-B, controls: blood donors; C-H, controls: non-demented patients; D-A, Alzheimer disease; D-M, mixed-type dementia; D-V, vascular dementia. *p < 0.05; **p < 0.01; ***p < 0.001 (for whole cohort analysis); ^##p < 0.01 (for dementia cohort analysis).

Figure 4

The association of dimethylarginines concentration in serum with the degree of cognitive loss: (a) ADMA and MMSEs in patients with vascular dementia; (b) ADMA and MMSEs in patients with strategic infarcts; (c) SDMA and MMSEs in patients with vascular dementia; (d) SDMA and MMSEs in patients with strategic infarcts; (e) SDMA and CDR in all patients (including non-demented controls); (f) SDMA and CDR in patients with vascular dementia. Data presented as Spearman correlation (panels a–d) or one-way ANOVA with geometric mean reported below the dot-plots; significant between-group differences are marked with connectors with asterisk.

Figure 5

The association between serum concentration of intermediates of NO metabolism and the degree of global cortical atrophy (GCA): (a) arginine in AD + MD patients; (b) Arg/ADMA in AD + MD patients; (c) arginine in MD patients; (d) Arg/ADMA in MD patients; (e) receiver operating characteristic (ROC) curve for Arg/ADMA and age as predictors of severe GCA in AD + MD patients (f) ROC curve for Arg/ADMA as a severe GCA predictor in MD patients. AD, Alzheimer disease; MD, mixed-type dementia; AUC, area under ROC curve; sens., sensitivity; spec. specificity. Data analyzed using t-test for independent samples (panels a–d) with numbers below the dot plots indicating the geometric mean values presented also graphically by inverse triangles with whiskers (95% confidence interval around mean).

Figure 6

The association between serum concentration of intermediates of NO metabolism and the presence of strategic infarcts. (a) DMA in VaD + MD patients; (b) citrulline in VaD patients. Data analyzed using t-test for independent samples. Numbers below the dot plots indicate the mean values. Mean values are additionally presented graphically by inverse triangles and 95% confidence interval around mean by whiskers.

Figure 7

The association between intermediates of NO metabolism and Hachinski ischemic score (HIS). (a) SDMA and HIS in all patients; (b) Arg/ADMA and HIS in all patients; (c) DMA in VaD patients; (d) Arg/ADMA and HID in VaD patients; (e) DMA in patients with strategic infarcts. Data presented as Pearson correlation coefficients. VaD, vascular dementia.