Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 18.
Published in final edited form as: Metabolomics. 2021 Jan 18;17(2):13. doi: 10.1007/s11306-020-01757-0

Sex and race differences of cerebrospinal fluid metabolites in healthy individuals.

Zackery W Reavis 1,6, Nikhil Mirjankar 2, Srikant Sarangi 2, Stephen H Boyle 3, Cynthia M Kuhn 1, Wayne R Matson 2, Michael A Babyak 3, Samantha A Matson 2, Ilene C Siegler 3, Rima Kaddurah-Daouk 3, Edward C Suarez 3, Redford B Williams 3, Katherine Grichnik 4, Mark Stafford-Smith 5, Anastasia Georgiades 3
PMCID: PMC8041469  NIHMSID: NIHMS1679875  PMID: 33462762

Abstract

Introduction:

Analyses of cerebrospinal fluid (CSF) metabolites in large, healthy populations have been limited and potential demographic moderators of brain metabolism are largely unknown.

Objective:

Our objective in this study was to examine sex and race differences in 33 CSF metabolites within a sample of 129 healthy individuals (37 African American women, 29 white women, 38 African American men, and 25 white men).

Methods:

CSF metabolites were measured with a targeted electrochemistry-based metabolomics platform. Sex and race differences were quantified with both univariate and multivariate analyses. Type I error was controlled for by using a Bonferroni adjustment (.05/33 = .0015).

Results:

Multivariate Canonical Variate Analysis (CVA) of the 33 metabolites showed correct classification of sex at an average rate of 80.6% and correct classification of race at an average rate of 88.4%. Univariate analyses revealed that men had significantly higher concentrations of cysteine (p<0.001), uric acid (p<0.001), and N-acetylserotonin (p = 0.049), while women had significantly higher concentrations of 5-hydroxyindoleacetic acid (5-HIAA) (p=0.001). African American participants had significantly higher concentrations of 3-hydroxykynurenine (p=0.018), while white participants had significantly higher concentrations of kynurenine (p<0.001), indoleacetic acid (p<0.001), xanthine (p=0.001), tocopherol-alpha (p=0.007), ), cysteine (p=0.029), melatonin (p=0.036), 7-methylxanthine (p=0.037), and xanthosine (p = 0.001). After the Bonferroni adjustment, the effects for cysteine, uric acid, and 5-HIAA were still significant from the analysis of sex differences and kynurenine and indoleacetic acid were still significant from the analysis of race differences.

Conclusion:

Several of the metabolites assayed in this study have been associated with mental health disorders and neurological diseases. Our data provide some novel information regarding normal variations by sex and race in CSF metabolite levels within the tryptophan, tyrosine and purine pathways, which may help to enhance our understanding of mechanisms underlying sex and race differences and potentially prove useful in the future treatment of disease.

2. INTRODUCTION

Cerebrospinal fluid (CSF) contains products of brain metabolism, some of which have been extensively studied to understand disorders that affect the central nervous system (CNS), such as major depressive disorder, Alzheimer’s disease, and Parkinson’s disease (Kaddurah-Daouk, Yuan, Boyle, Matson, Wang, Zeng, Zhu et al. 2012; Nagata, Hirayama, Ikeda, Shirahata, Shoji, Maruyama, Kayano et al. 2018; Olivola, Pierantozzi, Imbriani, Liguori, Stampanoni Bassi, Conti, D'Angelo et al. 2014). CSF metabolite analyses in large, healthy populations have been limited and the potential sex and race differences in brain metabolism that might be useful in the era of precision medicine are largely unknown. Examining such differences in healthy populations could be of potential importance as they might highlight vulnerabilities that could help explain sex and race related disease disparities. Two previous studies have found higher levels of the dopamine and serotonin metabolites, homovanillic acid (HVA) and 5-hydroxyindoleacetic acid (5-HIAA) respectively, in women than men (Blennow, Wallin, Gottfries, Karlsson, Mansson, Skoog, Wikkelso et al. 1993; Brewerton, Putnam, Lewine, Risch 2018) Another study of 43 healthy controls (but only 13 women) found that kynurenic acid was higher in women than men but found no sex differences in CSF levels of HVA and 5-HIAA (Nilsson, Nordin, Jonsson, Engberg, Linderholm, Erhardt 2007). Neither study examined race differences and only a small number of CSF metabolites within the tryptophan and tyrosine pathways were measured.

3. OBJECTIVE

The purpose of this study was to examine sex and race differences in CSF metabolites, primarily in the tryptophan, tyrosine, and purine pathways in a healthy cohort using a targeted electrochemistry-based metabolomics platform.

4. METHODS

4.1. Study population

The participants were recruited to take part in a study designed to examine the moderating effects of genetic, behavioral, and environmental mechanisms on health disparities, and consisted of 163 healthy participants (Williams, Marchuk, Gadde, Barefoot, Grichnik, Helms, Kuhn et al. 2003). The final study sample consisted of 129 individuals who provided enough CSF for metabolomics analysis and included 37 African American women, 29 white women, 38 African American men, and 25 white men (see Table 1 for sample characteristics). Assignment of race and sex were based on subject self-report. Participants were recruited via advertisements in the local media, flyers posted in supermarkets and other public locations, and outreach at public events. For details on recruitment procedures, see a previous publication (Williams, Marchuk, Gadde, Barefoot, Grichnik, Helms, Kuhn et al. 2003). The present study was conducted at Duke University Medical Center, and all subjects gave informed consent prior to their participation in the study using a form approved by the Duke University Medical Center Institutional Review Board. Those enrolled in the study received $500 for their participation. Participants underwent a comprehensive psychological examination, as well as medical history, physical exam, electrocardiogram, chest radiograph, hemoglobin, hematocrit, white cell count, and blood chemistries to rule out current medical and psychiatric disorders. Use of prescription drugs as well as use of illegal drugs (as detected by a urine screen prior to entry into study) were grounds for exclusion.

Table 1.

Descriptive statistics of study population (means ± SD).

Total Sample
N 129
Sex (% Women) 51
Race (% White) 42
Age (years) 34.1 ± 8.4
Body mass index (kg/m2) 26.7 ± 5.7
Fasting Glucose (mg/dL) 82.0 ± 7.8
Fasting Insulin (μU/mL) 9.5 ± 7.9
Total cholesterol (mg/dL) 182.6 ± 31.3
Systolic Blood pressure (mm Hg) 109.6 ± 11.5
Diastolic Blood pressure (mm Hg) 66.7 ± 8.2
Heart rate (bpm) 69.6 ± 10.6
Education (years) 14.6 ± 2.5
Open in a new tab

4.2. Procedure

Participants were admitted to the General Clinical Research Unit at Duke University Medical Center during the early afternoon. After completing the admission procedures, lumbar puncture was performed by a board-certified anesthesiologist. A sample of 4- to 5-ml cerebrospinal fluid was obtained which was mixed (to abolish the expected gradient across successive samples during the collection) and separated into 2 ml aliquots and frozen at −80° C for later assay of monoamine metabolites.

4.2. Metabolomics Platform

Cerebrospinal fluid samples were thawed to 4°C in the dark with mild vortexing. For each sample, 50 uL was taken for a pool, and duplicate 80 uL aliquots were pipetted into autosampler vials. All autosampler vials were then re-frozen at −80 °C until assay. Cerebrospinal fluid levels of the metabolites were quantitated by a high-performance liquid chromatography coupled with electrochemical coulometric array detection (LCECA) (Matson, Langlais, Volicer, Gamache, Bird, Mark 1984). LCECA (Bouckoms, Sweet, Poletti, Lavori, Carr, Matson, Gamache et al. 1992; Kristal, Shurubor, Kaddurah-Daouk, Matson 2007; Kristal, Vigneau-Callahan, Matson 2002; Kristal, Vigneau-Callahan, Matson 1998; LeWitt, Galloway, Matson, Milbury, McDermott, Srivastava, Oakes 1992; Rozen, Cudkowicz, Bogdanov, Matson, Kristal, Beecher, Harrison et al. 2005; Shi, Vigneau-Callahan, Matson, Kristal 2002) employs gradient chromatography followed by a 16 coulometric sensor array in series set sequentially from 0 to 900mV in 60 mV increments. Essentially the chromatography separates compounds in time and the array by the characteristic sensor on which they respond. The elution time coupled with the signature of response in the array allows qualitative certainty and the ca.100% efficiency of the sensors allows resolution of co-eluting compounds as a function of their oxidation potential. Quantitation is typically against a mixed component standard at the approximate mean concentrations of the metabolites in the tissue investigated. The system is typically linear to 3 orders of magnitude and sensitive to high picogram/ml levels. Typically, a biological sample will yield 1000-1200 resolved responses. Details of mobile phase composition, gradient profiles, sensor settings, and factors affecting precision and accuracy are described in a number of studies ((Kristal, Shurubor, Kaddurah-Daouk, Matson 2007; Kristal, Vigneau-Callahan, Matson 2002; Shi, Vigneau-Callahan, Matson, Kristal 2002; Shurubor, Matson, Martin, Kristal 2005). The LCECA method utilizes a 110-min gradient from an A mobile phase of 1% (w/v) sodium pentane sulfonic acid with 0.5% (v/v) acetic acid to a B mobile phase of methanol/acetonitrile/isopropanol 80/10/10% with 0.05-M lithium acetate at pH 3.8. The sulfonic acid solubilizes any protein material in the sample, and the high organics in the B mobile phase remove any lipids allowing continuous repetitive runs without column degradation. Allowable time on the autosampler at 0°C was determined as 18 h from repetitive pool assay. Samples were thawed and loaded every 12 h. Aliquots of 50 uL were injected into the LCECA. The samples were randomized and run in the following sequence: mixed authentic reference standard of 80 compounds, 5 samples, sample pool, 5 samples, duplicate. Duplicates were spaced from 1 to 100 samples apart. Pools which represent the maximum expected complexity of the sample set were used for time normalization of all data and for assessment of precision of the database. Secondary assessment of precision was made from duplicate pairs for analytes with a high degree of biological variability. All values reported were less than ±15 % relative standard deviation (rsd). The least precise measures were 3-O-methyldopa and anthranilic acid with a rsd% of 14.2% and 12.8% respectively. All other compounds were between 3.8% and 8.7% rsd. The rsd of standards ranged between 2.8% and 7.4% and for pools between 3.1% and 9.7%.

The sensitivity of the assay is 0.5 ng/sample, and values are reported as nanogram per milliliter. This method has been used extensively in studies of ventricular CSF (Bouckoms, Sweet, Poletti, Lavori, Carr, Matson, Gamache et al. 1992; Goudas, Langlade, Serrie, Matson, Milbury, Thurel, Sandouk et al. 1999), post-mortem ventricular CSF (Kaddurah-Daouk, Rozen, Matson, Han, Hulette, Burke, Doraiswamy et al. 2011), rostro/caudal gradients in CSF (Volicer, Langlais, Matson, Mark, Gamache 1985) and lumbar CSF in a number of disorders and models (Kaddurah-Daouk, Yuan, Boyle, Matson, Wang, Zeng, Zhu et al. 2012; Kaddurah-Daouk, Zhu, Sharma, Bogdanov, Rozen, Matson, Oki et al. 2013; LeWitt, Galloway, Matson, Milbury, McDermott, Srivastava, Oakes 1992; Lucot, Crampton, Matson, Gamache 1989; Rozen, Cudkowicz, Bogdanov, Matson, Kristal, Beecher, Harrison et al. 2005). With this platform, we were able to quantitate 33 compounds, primarily from the tryptophan, tyrosine, and purine pathways.

4.3. Statistical Analyses

Values were classified as outliers and removed from analysis if concentrations were more than 3 standard deviations above or below the mean. One value was removed for 3,4-dihydroxyphenylacetic acid (DOPAC) and one value was removed for L-3,4-dihydroxyphenylalanine (L-Dopa). Visual inspection of the metabolite distributions (see Online histograms in Supplemental Figure 2) suggested that several metabolites were skewed. Therefore, we used rank ordered data in our analysis.

The output from the LCECA platform containing 33 metabolite levels were used for multivariate classification of data using Canonical Variate Analysis (CVA). CVA was performed using the Extended Canonical Variate Analysis toolbox (Nørgaard, Bro, Westad, Engelsen 2006)in MATLAB R2018a. The data was auto-scaled prior to CVA. Two CVA models were developed; (a) for classification of sex, and (b) for classification of race. CVA reduces the dimensionality of the data by developing canonical variates (CV). The CV are linear combinations of the original independent variables that maximize the separation between the different groups of data. For classification of data into two groups, dimension of CV is one. The group identities are represented as +1 and −1. Samples with CV > 0 are classified as group +1 and samples with CV < 0 are classified as group −1. The CV is computed such that maximum number of the samples are classified correctly with minimum classification errors.

Classification success rates were computed using CVA for all 129 individuals and all 33 metabolites. Cross-validation analysis was used for estimating classification error rates and testing the predictive ability of the CVA models for prediction of the samples not used for training the models. Cross-validation was performed using a stratified 3-fold split approach and thus three CVA models were done. We used 1/3rd of the samples for testing the CVA models that were obtained by training 2/3rd of the samples not in the test set. Every third sample was in the test set starting from; 1st sample for Set-1 (1st, 4th, 7th,…), 2nd sample for Set-2 (2nd, 5th, 8th,…), and 3rd sample for Set-3 (3rd, 6th, 9th,…).

IBM SPSS Statistics 25 was used to perform one-way analysis of variance (ANOVA) on rank ordered data to identify metabolites that were significantly different between men and women (Table 3) and between African Americans and white participants (Table 4). In an additional step, age and sex adjustments were made for race differences and age and race adjustments were made for sex differences. Type I error was controlled for by using a Bonferroni adjustment (.05/33 = .0015). Spearman’s correlation coefficients (rho) were calculated to determine the associations between metabolite level and age (Online Resource 1).

Table 3.

Sex Differences in CSF Metabolite Levels (ng/mL) and Ratios (medians and 25th and 75th quartiles).

Metabolite Women
Median (25th - 75th)		 Men
Median (25th - 75th)		 p-value N
2-Hydroxyphenylacetic acid 4.77 (4.21 - 5.30) 4.98 (4.52 - 5.62) 0.195 129
3-Hydroxykynurenine 1.03 (0.51 - 2.82) 0.94 (0.41 - 3.31) 0.909 129
3-O-methyldopa 1.09 (0.81 - 1.42) 1.28 (0.68 - 2.21) 0.386 105
4-Hydroxyphenyllactic acid 43.19 (35.75 - 58.08) 43.32 (32.49 - 56.26) 0.670 129
5-HIAA 13.60 (10.61 - 19.63) 10.26 (6.33 - 14.21) 0.0011 129
5-HTP 1.13 (0.84 - 1.39) 1.07 (0.79 - 1.33) 0.358 129
7-Methylxanthine 2.06 (1.65 - 2.57) 2.17 (1.67 - 2.46) 0.585 129
Alpha-Tocopherol 120.82 (100.49 - 141.71) 126.33 (95.64 - 152.58) 0.775 129
Anthranilic acid 2.09 (1.66 - 2.36) 2.16 (1.70 - 2.36) 0.563 123
Ascorbate 24127.02 (20230.46 - 30557.56) 23496.17 (17300.34 - 31353.01) 0.515 129
Cysteine 6.64 (5.36 - 8.71) 10.5 (7.90 - 13.62) <.0001 129
DOPAC 1.17 (0.70 - 2.03) 1.06 (0.58 - 1.49) 0.095 117
Guanosine 3.62 (3.06 - 4.32) 3.40 (2.98 - 4.01) 0.302 121
Glutathione 12.28 (10.81 - 13.34) 14.34 (9.88 - 15.69) 0.062 90
Glutathione Disulfide 10.52 (8.20 - 13.22) 9.95 (7.91 - 13.84) 0.797 129
Gamma-Tocopherol 22.63 (12.37 - 31.49) 24.56 (15.44 - 37.10) 0.208 125
HVA 37.76 (26.09 - 54.75) 34.06 (24.15 - 45.05) 0.107 129
Hypoxanthine 722.26 (550.38 - 909.68) 635.42 (508.92 - 815.96) 0.178 129
Indoleacetic acid 3.47 (2.43 - 5.67) 2.76 (2.15 - 5.08) 0.119 126
Indolelactic acid 1.86 (1.35 - 2.42) 1.91 (1.45 - 2.62) 0.369 111
Kynurenine 5.07 (3.39 - 6.30) 4.83 (3.33 - 6.90) 0.910 129
L-Dopa 0.95 (0.71 - 1.35) 0.94 (0.69 - 1.21) 0.473 127
Melatonin 0.67 (0.42 - 0.93) 0.65 (0.42 - 0.92) 0.892 128
Methionine 955.53 (796.19 - 1125.83) 1033.47 (849.31 - 1251.00) 0.118 129
MHPG 10.39 (8.88 - 11.76) 10.48 (9.46 - 11.55) 0.446 129
Acetylserotonin (N) 1.04 (0.69 - 1.27) 1.14 (0.88 - 1.55) 0.049 129
Tryptophol 11.67 (8.37 - 15.52) 13.10 (9.67 - 17.92) 0.176 129
Tryptophan 485.29 (417.20 - 550.93) 493.52 (442.82 - 566.98) 0.348 129
Tyrosine 2033.30 (1660.71 - 2283.83) 2023.74 (1744.40 - 2372.01) 0.298 129
Uric acid 5353.71 (4041.17 - 7102.65) 9008.48 (7033.92 - 11906.72) <.0001 129
Vanillylmandelic acid 0.76 (0.58 - 1.00) 0.71 (0.43 - 1.05) 0.268 128
Xanthine 324.28 (276.36 - 392.87) 320.78 (276.85 - 377.55) 0.870 129
Xanthosine 23.74 (20.25 - 28.26) 23.10 (19.41 - 28.82) 0.844 129
Ratios Women
Median (25th - 75th)		 Men
Median (25th - 75th)		 p-value N
5-HTP/Trp .0024 (.0018 - .0029) 0.0021 (0.0016 – 0.0027) 0.147 129
5-HIAA/5-HTP 12.37 (8.47 – 19.99) 9.49 (6.20 – 15.96) 0.030 129
Uric/Xanthine 16.34 (11.77 – 19.58) 30.20 (21.61 – 36.93) <.0001 129
Xanthine/Hypoxanthine 0.47 (0.40 – 0.57) 0.51 (0.41 – 0.64) 0.215 129
Open in a new tab

Table 4.

Race Differences in CSF Metabolite Levels (ng/mL) and Ratios (medians and 25th and 75th quartiles).

Metabolite African American
Median (25th - 75th)		 White
Median (25th - 75th)		 p-value N
2-Hydroxyphenylacetic acid 4.94 (4.41 - 5.48) 4.84 (4.30 - 5.26) 0.643 129
3-Hydroxykynurenine 1.80 (0.49 - 3.51) 0.72 (0.50 - 1.37) 0.018 129
3-O-methyldopa 1.22 (0.82 - 1.88) 1.04 (0.69 - 1.45) 0.069 105
4-Hydroxyphenyllactic acid 43.32 (31.25 - 57.56) 43.97 (36.09 - 55.90) 0.543 129
5-HIAA 12.33 (8.99 - 17.95) 11.76 (8.41 - 16.00) 0.334 129
5-HTP 1.01 (0.79 - 1.36) 1.16 (0.85 - 1.34) 0.498 129
7-Methylxanthine 1.98 (1.61 - 2.38) 2.20 (1.84 - 2.80) 0.037 129
Alpha-Tocopherol 116.99 (94.60 - 137.53) 132.56 (111.66 - 151.31) 0.007 129
Anthranilic acid 2.13 (1.59 - 2.38) 2.12 (1.71 - 2.32) 0.907 123
Ascorbate 25197.53 (19613.42 - 30802.89) 23585.38 (15970.69 - 30411.53) 0.354 129
Cysteine 7.87 (5.60 - 9.57) 9.01 (6.67 - 12.91) 0.029 129
DOPAC 1.18 (0.74 - 1.91) 0.99 (0.59 - 1.45) 0.110 117
Guanosine 3.34 (2.96 - 4.18) 3.60 (3.16 - 4.27) 0.099 121
Glutathione 12.64 (9.75 - 15.63) 12.88 (11.69 - 14.89) 0.768 90
Glutathione Disulfide 9.60 (8.03 - 13.24) 11.19 (8.40 - 13.72) 0.166 129
Gamma-Tocopherol 21.19 (14.84 - 33.92) 25.20 (12.24 - 35.91) 0.519 125
HVA 36.54 (25.94 - 54.58) 34.41 (25.21 - 45.63) 0.503 129
Hypoxanthine 669.10 (530.06 - 835.36) 640.87 (513.24 - 911.57) 1.000 129
Indoleacetic acid 2.63 (2.01 - 4.08) 4.83 (2.76 - 6.41) <.0001 126
Indolelactic acid 1.87 (1.35 - 2.43) 1.91 (1.58 - 2.52) 0.406 111
Kynurenine 4.30 (3.13 - 5.92) 5.83 (4.84 - 7.31) <.0001 129
L-Dopa 0.95 (0.70 - 1.26) 0.94 (0.71 - 1.27) 0.861 127
Melatonin 0.59 (0.39 - 0.83) 0.80 (0.49 - 0.98) 0.036 128
Methionine 974.01 (806.38 - 1221.82) 982.47 (843.03 - 1169.90) 0.909 129
MHPG 10.58 (9.32 - 11.55) 9.95 (8.44 - 11.70) 0.182 129
Acetylserotonin (N) 1.06 (0.85 - 1.33) 1.10 (0.78 - 1.49) 0.610 129
Tryptophol 11.56 (8.03 - 16.43) 12.76 (10.34 - 17.61) 0.111 129
Tryptophan 484.05 (419.09 - 565.47) 491.69 (435.12 - 556.50) 0.531 129
Tyrosine 2025.29 (1696.14 - 2347.68) 2008.01 (1686.39 - 2363.03) 0.883 129
Uric acid 6749.83 (4619.79 - 8800.71) 7572.44 (5374.56 - 11723.86) 0.081 129
Vanillylmandelic acid 0.75 (0.46 - 1.05) 0.74 (0.52 - 0.98) 0.953 128
Xanthine 311.72 (266.49 - 355.13) 354.36 (305.15 - 412.95) 0.0012 129
Xanthosine 23.97 (20.59 - 29.34) 23.24 (17.83 - 25.84) 0.062 129
Metabolite Ratios African American
Median (25th - 75th)		 White
Median (25th - 75th)		 P-value N
Indoleacetic acid/tryptophan 0.0057 (0.0041 – 0.0086) 0.0090 (0.0056 - 0.0128) .0001 126
Anthranilic acid/ Indoleacetic acid 0.69 (0.45 – 1.03) 0.43 (0.29 – 0.70) .0002 121
Kynurenine/tryptophan 0.0088 (0.0065 – 0.0125) 0.0119 (0.0094 – 0.0149) .0002 129
Anthranilic acid/ Kynurenine 0.48 (0.33 – 0.68) 0.34 (0.25 – 0.45) .0002 123
3-Hydroxykynurenine/ Kynurenine 0.48 0.11 – 0.98) 0.12 (0.07 – 0.18) 0.0004 129
Uric/Xanthine 20.89 (14.48 – 30.61) 20.05 (14.84 – 33.76) 0.615 129
Xanthine/Hypoxanthine 0.47 (0.40 – 0.58) 0.53 (0.46 – 0.66) 0.016 129
Open in a new tab

For the metabolites that differed significantly by sex or race, the ratio of the metabolites to their immediate measured metabolic product, as well as the other precursor to product ratios in the same direct metabolic pathway, were calculated. Product to precursor ratios approximate the reaction rate through a biochemical pathway and can be informative of the enzymatic activity that accounts for the conversion of precursor to its product (Petersen, Krumsiek, Wagele, Theis, Wichmann, Gieger, Suhre 2012). While the analysis of metabolite data tells us what is different between the groups, the analysis of ratios attempts to tell us why they are different. Comparisons of those ratios between men and women (Table 3) and African American and white participants (Table 4) were examined using one-way ANOVA. In some cases (e.g. cysteine), metabolic products were not measured within the platform, and thus those ratios could not be examined. In other cases, the immediate metabolic product was not measured. For example, dopamine and serotonin levels were not measured by the platform and thus compounds, such as L-dopa and 5-hydroxytryptophan (5-HTP), were compared to the metabolic products in their respective pathways, i.e. DOPAC and 5-HIAA respectively. The analysis of ratios was used to support the interpretation of the metabolite results. Therefore, we did not apply any correction for false discovery rate.

5. RESULTS

The study sample was relatively young (34.1±8.4 years) and healthy based on extensive medical screening. These health factors and full demographic characteristics of the study sample are presented in Table 1. Histograms of the distribution of each metabolite are presented in Online Resource 2.

5.1. Sex Differences in Metabolites

Multivariate analyses

In the CVA analysis for classification of sex using all 33 metabolites, observed classification success rate (% correct) was 84.8% for women and 76.2% for men, with a total classification success rate of 80.6% (Table 2). The classification plot showing each individual participant represented in accuracy of classification of sex using all 33 metabolites is presented in Figure 1a. Percent error computed by running cross validation was 25.8% for women and 30.2% for men, with a total classification error of 27.9% . The CVA cross-validation tests done for sex with 3-fold split sets are presented in Online Resource 3 of supplemental material. The corresponding 3 classification plots are presented in Online Resources 4.

Table 2.

CVA classification of sex and race using the 33 CSF metabolite levels.

Sex-CVA no. samples no. correct % correct % error*
Female 66 56 84.8 25.8
Male 63 48 76.2 30.2
All 129 104 80.6 27.9
 
Race-CVA no. samples no. correct % correct % error*
African American 75 63 84.0 21.3
White 54 51 94.4 27.8
All 129 114 88.4 24.0
Open in a new tab
*

error computed using 3-fold cross-validation sets

Figure 1 a).

Figure 1 a)

Open in a new tab

CV scores of samples for classification of sex using 33 metabolites. Women are represented with a red 1 and men are represented with a blue 2.

Univariate analyses

Table 3 presents medians and 25th-75th percentile of metabolite concentrations (ng/mL) and p-values associated with comparisons of rank ordered scores between men and women. There were significant sex differences in 4 metabolites in the univariate analysis, where men had higher levels of cysteine (p<0.001), uric acid (p<0.001) and N-acetylserotonin (p = 0.049), while women had higher levels of 5-HIAA (p=0.001). Adjusting for age and race did not significantly change the effect of sex on cysteine (p<0.001), uric acid (p<0.001), 5-HIAA (p=0.001) and N-acetylserotonin (p = 0.048). Using the Bonferonni adjusted p-value (i.e. p = 0.0015), the effects for cysteine, uric acid, and 5-HIAA were still significant. Analysis of ratios showed that men had a significantly higher uric acid/xanthine ratio (p<0.0001) and women had a significantly higher 5-HIAA/5-HTP ratio (p=0.030) (Table 3).

6.2. Race Differences in Metabolites

Multivariate analyses

In a CVA analysis for classification of race using all 33 metabolites, we observed a classification success rate (% correct) of 84.0% for the African American participants and 94.4% for the white participants with a total classification success rate of 88.4% (Table 2). The classification of race for each individual participant is demonstrated in Figure 1b. Percent error computed by running cross validation was 21.3% for African American, 27.8% for white, and 24.0% the total. The CVA cross-validation tests done with 3-fold split sets are presented in Online Resource 3. The corresponding classification plots are presented in Online Resource 5.

Figure 1 b).

Figure 1 b)

Open in a new tab

CV Scores of Samples for Classification of Race using 33 metabolites. African American participants are represented with a red 1, white participants are represented with a blue 2.

Univariate analyses

Table 4 presents medians and 25th-75th percentile of all metabolite concentrations (ng/mL) and p-values associated with comparisons of rank ordered scores between African American and white participants. There were significant race differences in 9 metabolites in the univariate analysis, where African American participants had higher levels of 3-hydroxykynurenine (3-OHKY) (p=0.018) , while white participants had higher levels of kynurenine (p<0.001), indoleacetic acid (p<0.001), xanthine (p=0.001), xanthosine (p=0.001), alpha tocopherol (p=0.007), cysteine (p=0.029), melatonin (p=0.036), and 7-MX (p=0.037) . Adjusting for sex and age did not eliminate the race differences in levels of kynurenine (p<0.001), indoleacetic acid (p<0.001), xanthine (p=0.006), xanthosine (p=0.037), 3-OHKY (p=0.008), alpha tocopherol (p=0.031), cysteine (p=0.016). However, adjusting for age and sex did reduce the race effect for melatonin (p=0.064), and 7-MX (p=0.088). Using the Bonferonni adjusted p-value (i.e. p = 0.0015), the effects for kynurenine and indoleacetic acid were still significant. Analysis of ratios showed that white participants had higher ratios of kynurenine/tryptophan (p<0.001), indoleleacetic acid/tryptophan (p<0 .001), and xanthine/hypoxanthine (p=0.016), while African American participants had higher ratios of anthranilic acid/indoleacetic acid (p< 0.001) anthranilic acid/kynurenine (p< 0.001) and 3-hydroxykynurenine/kynurenine (p< 0.001) (Table 3).

5.3. Correlations Between Metabolites and Age

Online Resource 1 presents correlations between age and metabolite concentrations. Age was significantly and positively associated with 4-hydroxyphenyllactic acid (Spearman’s rho (r)=0.29, p<.001), 7-methylxanthine (7-MX) (r=0.25, p<0.01), cysteine (r=0.30, p<0.001), guanosine (r=0.30, p<0.001, glutathione disulfide (r=0.20, p<0.05), gamma tocopherol (r=0.24, p<0.01), kynurenine (r=0.330, p<0.001), methionine (r=0.27, p<0.01), tryptophol (r=0.18, p<0.05), uric acid (r=0.31, p<0.001), xanthine (r=0.366, p<0.001), and negatively associated with 3-O-methyldopa (r=−0.270, p=0.005).

6. CONCLUSION

In this study we examined sex and race differences of CSF metabolites in apparently healthy African American and white men and women. Multivariate CVA analyses of the 33 measured metabolites using resampling techniques successfully and consistently separated both for sex and race. Univariate analyses revealed sex and race related differences in aspects of tryptophan, tyrosine, cysteine, and purine pathway metabolism. The meaning of these differences and their possible relevance for health are discussed in turn.

6.1. Sex Differences

We found that women had higher CSF levels of 5-HIAA, which is consistent with two previous studies (Blennow, Wallin, Gottfries, Karlsson, Mansson, Skoog, Wikkelso et al. 1993; Brewerton, Putnam, Lewine, Risch 2018). This may be due to sex related differences in serotonin transporter function. One of our previous studies, which utilized the current study’s population (Williams, Marchuk, Gadde, Barefoot, Grichnik, Helms, Kuhn et al. 2003), found that the 15 women with the 5HTTLPR SS genotype, a functional polymorphism for the gene that encodes for the serotonin transporter promoter, had higher CSF 5-HIAA levels than the women without this genotype, while the 8 men with the SS genotype had lower levels of 5-HIAA than the men without this genotype. There were no sex differences for participants with the LL or LS genotypes. Previous studies also reported increased monoamine oxidase (MAO) activity in the brain, as well as in platelets, in women compared to men (Bridge, Soldo, Phelps, Wise, Francak, Wyatt 1985; Murphy, Wright, Buchsbaum, Nichols, Costa, Wyatt 1976). Since MAO is involved in metabolism of serotonin to 5-HIAA in the tryptophan (Online Resource 6), the sex difference in 5-HIAA levels could reflect enhanced serotonin metabolism in the brain. The higher 5-HIAA/5-HTP ratio seen in women is consistent with this interpretation. So, too, is the trend toward higher levels of HVA and DOPAC among women in this study as levels of both of those metabolites are also influenced by MAO activity. Monoamine levels in CSF have previously been examined in studies on depressive disorders, as raising monoamine levels have been a target for drug development for depression. Several studies have measured CSF concentrations of HVA, 5-HIAA and 3-methoxy-4-hydroxyphenylglycol (MHPG) in depressed patients and compared them to healthy controls, but results are inconsistent (Asberg, Bertilsson, Martensson, Scalia-Tomba, Thoren, Traskman-Bendz 1984; Berger, Faull, Kilkowski, Anderson, Kraemer, Davis, Barchas 1980; De Bellis, Geracioti, Altemus, Kling 1993; Ehnvall, Sjogren, Zachrisson, Agren 2003; Geracioti, Loosen, Ekhator, Schmidt, Chambliss, Baker, Kasckow et al. 1997; Gerner, Fairbanks, Anderson, Young, Scheinin, Linnoila, Hare et al. 1984; Hou, Jia, Liu, Li 2006; Jones, Stanley, Mann, Frances, Guido, Traskmanbendz, Winchel et al. 1990; Kaddurah-Daouk, Yuan, Boyle, Matson, Wang, Zeng, Zhu et al. 2012; Kasa, Otsuki, Yamamoto, Sato, Kuroda, Ogawa 1982; Koslow, Maas, Bowden, Davis, Hanin, Javaid 1983; Little, Ketter, Mathe, Frye, Luckenbaugh, Post 1999; Molchan, Lawlor, Hill, Martinez, Davis, Mellow, Rubinow et al. 1991; Oreland, Wiberg, Asberg, Traskman, Sjostrand, Thoren, Bertilsson et al. 1981; Palaniappun, Ramachandran, Somasundaram 1991; Post, Gordon, Goodwin, Bunney 1973; Reddy, Khanna, Subhash, Channabasavanna, Rao 1992; Roy, Agren, Pickar, Linnoila, Doran, Cutler, Paul 1986; Roy, Pickar, De Jong, Karoum, Linnoila 1988; Roy, Pickar, De Jong, Karoum, Linnoila 1989; Sher, Mann, Traskman-Bendz, Winchel, Huang, Fertuck, Stanley 2006; Sher, Oquendo, Li, Burke, Grunebaum, Zalsman, Huang et al. 2005; Sher, Oquendo, Li, Huang, Grunebaum, Burke, Malone et al. 2003; Sullivan, Mann, Oquendo, Lo, Cooper, Gorman 2006; Sullivan, Oquendo, Huang, Mann 2006; Swann, Katz, Bowden, Berman, Stokes 1999; Westenberg, Verhoeven 1988; Widerlov, Bissette, Nemeroff 1988; Yoon, Hattori, Ogawa, Sasayama, Ota, Teraishi, Kunugi 2017). The gut microbiome has also been associated with serotonin and tryptophan metabolism through the brain-gut axis (O'Mahony, Clarke, Borre, Dinan, Cryan 2015). Additionally, a 2019 study found sex differences in gut microbiota (Kim, Unno, Kim, Park 2019), so future studies should explore potential sex differences of gut microbiota associated with tryptophan metabolism.

Men were found to have higher levels of CSF uric acid and cysteine, which to our knowledge, has not been previously reported. CSF uric acid is thought to originate from xanthine oxidase-enriched endothelial cells and astrocytes. Xanthine oxidase is the enzyme that converts xanthine into uric acid, as well as hypoxanthine to xanthine (Stover, Lowitzsch, Kempski 1997). Hypoxanthine enters the brain through a low affinity adenine transport system (Betz 1985). Xanthine and uric acid levels in CSF are thought to reflect activity of xanthine oxidase in the brain (Stover, Lowitzsch, Kempski 1997). Given the similar levels of xanthine and higher level of uric acid, men may have higher xanthine oxidase activity in the brain. This is supported by the significantly lower xanthine/uric acid ratio in men (Table 1). Increased xanthine oxidase activity could lead to higher generation of free radicals, as xanthine oxidase is a radical producing enzyme. This finding warrants further investigation, as uric acid levels have been associated with neurodegenerative disorders such as Amyotrophic lateral sclerosis, Alzheimer’s Disease, Parkinson’s Disease, and Multiple Sclerosis (Cascalheira, Joao, Pinhancos, Castro, Palmeira, Almeida, Faria et al. 2009; Liu, Shen, Xiao, Tang, Cen, Wei 2012; Paganoni, Zhang, Quiroz Zarate, Jaffa, Yu, Cudkowicz, Wills 2012; Tabunoki, Ono, Ode, Ishikawa, Kawana, Banno, Shimada et al. 2013), in which sex differences in prevalence have also been noted.

The biological significance of the higher CSF levels of cysteine in men is unclear. Cysteine is a non-essential amino acid that is metabolized from the essential amino acid, methionine. Methionine undergoes transmethylation to form homocysteine, which can be remethylated back to methionine if methionine levels are low or can undergo transsulfuration and interact with vitamin B6 to form cysteine if methionine levels are high or cysteine is needed (Fukagawa, Martin, Wurthmann, Prue, Ebenstein, O'Rourke 2000). Cysteine can then be integrated into glutathione, an important antioxidant and free radical scavenger (Droge 2005), incorporated into other proteins, cleared from the body, or converted into its oxidized form (cystine), which plays a role in excitatory neuronal signaling through glutamate (Lewerenz, Hewett, Huang, Lambros, Gout, Kalivas, Massie et al. 2013). This pathway plays an essential role in cellular metabolism, through its ability to donate methyl groups and the incorporation of cysteine and other amino acids into various proteins (Blom, Smulders 2011). Given the different branches of this pathway, it is not entirely clear which branch is being differentially metabolized in men and women, but a few things can be noted. This difference could be partially explained by increased dietary intake of cysteine, but given the magnitude of the difference, it is unlikely that diet would fully explain it. Methionine and glutathione levels did not differ significantly between the sexes. This indicates that the difference likely lies either in the transsulfuration of homocysteine to cysteine, the clearance of excess cysteine, the conversion of cysteine to cystine, or the incorporation of cysteine into other proteins. The measurement of CSF homocysteine and cystine could provide further insight into the cause of this difference. Cysteine and other products of this metabolic pathway have been associated with various neurodegenerative diseases and psychiatric disorders, including Parkinson’s disease and Alzheimer’s disease (Blom, Smulders 2011; Heafield, Fearn, Steventon, Waring, Williams, Sturman 1990). Given the importance of cysteine as an amino acid and its association with various diseases, this finding warrants further research.

6.2. Race Differences

There were several race related differences in kynurenine metabolism. In the kynurenine branch of the tryptophan pathway, tryptophan is metabolized to kynurenine and then either kynurenic acid, anthranilic acid, or 3-OHKY, where anthranilic acid and 3-OHKY can then be metabolized into 3-hydroxyanthranilic acid (Colin-Gonzalez, Maldonado, Santamaria 2013; Krause, Suh, Tarassishin, Cui, Durafourt, Choi, Bauman et al. 2011) (See Online Resource 6). 3-OHKY can alternatively be degraded to quinolinic acid (Condray, Dougherty, Keshavan, Reddy, Haas, Montrose, Matson et al. 2011). White participants had higher levels of kynurenine than African Americans. The higher kynurenine/tryptophan ratio seen in whites suggests that they have higher indoleamine 2,3-dioxygenase and formamidase activity, which are the enzymes responsible for converting tryptophan to kynurenine. Despite having much lower kynurenine levels, African Americans had significantly higher levels of 3-OHKY. This suggests that kynurenine is being metabolized to 3-OHKY at a higher rate in African Americans, which is supported by the significantly higher 3-OHKY/kynurenine ratio found in the African American participants (Table 4). If the rate of metabolism of kynurenine to 3-OHKY is higher in African Americans then higher enzymatic activity should be seen for kynurenine 3-monooxygenase, the enzyme that catalyzes this metabolism. However, it is unclear whether 3-OHKY in the brain is metabolized locally or derived from circulatory kynurenine and 3-OHKY (Colin-Gonzalez, Maldonado, Santamaria 2013). Determining whether the same race difference in 3-OHKY levels is evident in blood could help identify the cause of this difference. African Americans and whites had similar levels of anthranilic acid, but African Americans had a higher anthranilic acid/kynurenine ratio. It appears that the greater availability of the precursor in white participants is offset by the higher level of kynureninase activity in African American participants. Taken together, our results suggest an overall pattern of enzymatic activity that favors the production of kynurenine in whites and 3-OHKY and anthranilic acid in African Americans. We did not assay kynurenic acid in this study, so it is difficult to comment on possible race differences in that branch of kynurenine metabolism. However, it would be of interest to know, for example, if the higher levels of kynurenine seen in our white participants is being metabolized towards the production of kynurenic acid. Future studies should assay additional kynurenine pathway metabolites to more fully delineate race differences in this important branch of metabolism.

These race differences could have implications for disease development and treatment, given the importance of kynurenine in inflammatory pathways (Krause, Suh, Tarassishin, Cui, Durafourt, Choi, Bauman et al. 2011; Reus, Jansen, Titus, Carvalho, Gabbay, Quevedo 2015) and the link of this pathway to major depressive disorder, schizophrenia, bipolar disorder, Alzheimer’s disease, ALS, Huntington’s disease, and cancer (Chen, Guillemin 2009; Kegel, Bhat, Skogh, Samuelsson, Lundberg, Dahl, Sellgren et al. 2014; Reus, Jansen, Titus, Carvalho, Gabbay, Quevedo 2015; Savitz, Dantzer, Wurfel, Victor, Ford, Bodurka, Bellgowan et al. 2015). In addition, kynurenic acid has been shown to be neuroprotective, while 3-OHKY and quinolinic acid have been shown to be neurotoxic (Condray, Dougherty, Keshavan, Reddy, Haas, Montrose, Matson et al. 2011). Previous research has found 3-OHKY levels to be positively associated with symptom severity in unmedicated schizophrenic patients (Condray, Dougherty, Keshavan, Reddy, Haas, Montrose, Matson et al. 2011). These researchers also found that baseline 3-OHKY levels predicted clinical improvement, where patients with the lowest 3-OHKY levels saw the greatest degree of improvement 4 weeks after beginning treatment. It is not known whether increased levels of 3-OHKY are directly involved in disease development or progression. However, given the potential impact of the tryptophan pathway, examining race differences of metabolites within the kynurenine branch could be an important area of research.

Indoleacetic acid levels were significantly higher in white participants. Indoleacetic acid in the CSF derives from two major intermediate pathways from tryptophan, through indole pyruvate and through tryptamine. Indole lactic acid the second major direct product of indole pyruvate is virtually identical in black (1.92 ng/ml) vs. white (2.00ng/ml) individuals (p=0.42) indicating that the pathway through indole pyruvate is not likely to cause the effect of the higher levels of indole acetic acid. Thus, the higher levels of indoleacetic acid in white participants probably to reflects CNS tryptamine metabolism (Young, Anderson, Gauthier, Purdy 1980)(Young, Anderson, Purdy 1980). Tryptamine is a low concentration amine that was initially believed to a be non-active by-product of serotonin synthesis. More recent thinking is that tryptamine plays a role as a neurotransmitter or neuromodulator. The evidence suggesting tryptamine is a neurotransmitter comes from studies that have identified specific post-synaptic receptors for tryptamine (Jones 1982). There is also evidence that tryptamine is a neuromodulator of serotonin neurotransmission, exhibiting both antagonistic and potentiating effects(Jones 1982). Elevated levels of indoleacetic acid in plasma have also been associated with an elevated risk of stroke and secondary heart attacks ((Dou, Sallee, Cerini, Poitevin, Gondouin, Jourde-Chiche, Fallague et al. 2015; Vanholder, Pletinck, Schepers, Glorieux 2018). However, we are not aware of any studies linking CSF levels of indoleacetic acid levels to health outcome in humans. Given the stark race related differences in indoleacetic acid, a metabolite reflecting CNS tryptamine turnover, obtaining a better understanding the role of tryptamine in CNS functionality may be of importance.

6.3. Limitations

Several caveats must be considered when evaluating these results. First, “race” is a complex concept and although categorizations are often made based on phenotypical characteristics, they are not clear biological concepts. In the present study, we relied on self-report of race which likely reflected a more heterogeneous biological background than we assigned by using “African American” and “white” as our group categorization. Second, the cellular origin of metabolites in CSF is likely heterogeneous as well, reflecting metabolism in neurons, astrocytes, microglia, and endothelial cells. Research in the future should address the biologic relevance of the differences identified in this paper.

Finally, the sample size was relatively small, although larger than many such studies. In addition, the relatively small sample may have resulted in parameter estimates that may not be stable across additional independent samples. Also, the use of the more conservative p-value (i.e. p < 0.0015) to judge significance within the context of relatively low power may have contributed to one or more type II errors. Several metabolites had p-values of <.05 and >.0015 including 3-OHKY, alpha tocopherol, xanthine, xanthosine, melatonin, N-acetylserotonin. Many of these compounds have been linked to important areas of function and should be considered for inclusion in future studies investigating sex and race differences in CSF metabolites. Thus, our findings should be validated in independent samples. Given the challenges involved in collecting CSF, this sample provides a rare opportunity to simultaneously examine multiple metabolites and their relationships within the tryptophan, tyrosine and purine pathways.

6.4. Overall Conclusions

Overall, this study of healthy individuals revealed significant sex and race differences, especially within the tryptophan pathway. The study is unique in that it applied targeted metabolomics platform to examine CSF metabolites in a healthy sample. While this study does not fully cover all metabolites within the tryptophan, tyrosine, and purine pathways, it serves as a good starting point for future exploration into these pathways. Also, these findings could potentially provide a starting point for linkage of metabolomic to genomic information, as we have previously done in studies of major depression with this platform (Liu, Ray, Neavin, Zhang, Athreya, Biernacka, Bobo et al. 2018). Our findings support the importance of understanding underlying sex and race differences, some of which could influence the health trajectory of individuals before awareness is raised clinically. In addition, the race and sex differences seen in this study may have implications for the effectiveness or side effects of drugs that modify those pathways as part of their therapeutic action.

Supplementary Material

1

7.1.

Funding

The research in this manuscript was funded by the NHLBI grant number P01-HL036587.

7.9. PubChem CID

Metabolite PubChem CID
2-Hydroxyphenylacetic acid 11970
3-Hydroxykynurenine 89
3-O-methyldopa 9307
4-Hydroxyphenyllactic acid 9378
5-HIAA 1826
5-HTP 144
7-Methylxanthine 68374
Alpha-Tocopherol 14985
Anthranilic acid 227
Ascorbate 54670067
Cysteine 5862
DOPAC 547
Guanosine 135398635
Glutathione 124886
Glutathione Disulfide 65359
Gamma-Tocopherol 92729
HVA 1738
Hypoxanthine 135398638
Indoleacetic acid 802
Indolelactic acid 92904
Kynurenine 846
L-Dopa 6047
Melatonin 896
Methionine 6137
MHPG 10805
Acetylserotonin (N) 903
Tryptophol 10685
Tryptophan 6305
Tyrosine 6057
Uric acid 1175
Vanillylmandelic acid 1245
Xanthine 1188
Xanthosine 64959
Open in a new tab

Footnotes

7.2

Conflicts of Interest/Competing Interests

Redford Williams holds a U.S. patent on the use of 5HTTLPR L allele as a marker of increased cardiovascular disease risk. The remaining authors have nothing to disclose.

7.3

Ethics Approval

The study was approved by the Duke University Medical Center Institutional Review Board.

7.4

Consent to Participate

The present study was conducted at Duke University Medical Center, and all subjects gave informed consent prior to their participation in the study using a consent form approved by the Duke University Medical Center Institutional Review Board.

7.5

Consent for Publication

Not applicable (the results presented contain no identifiable information)

7.6

Availability of Data and Material

The metabolomics data reported in this study is available upon request to the corresponding author, Anastasia Georgiades.

7.7

Code Availability

Not applicable

REFERENCES

  1. Asberg M, Bertilsson L, Martensson B, Scalia-Tomba GP, Thoren P, Traskman-Bendz L (1984). CSF monoamine metabolites in melancholia. Acta Psychiatr Scand 69, 201–19 doi: 10.1111/j.1600-0447.1984.tb02488.x [DOI] [PubMed] [Google Scholar]
  2. Berger PA, et al. (1980). CSF monoamine metabolites in depression and schizophrenia. Am J Psychiatry 137, 174–80 doi: 10.1176/ajp.137.2.174 [DOI] [PubMed] [Google Scholar]
  3. Betz AL (1985). Identification of hypoxanthine transport and xanthine oxidase activity in brain capillaries. J Neurochem 44, 574–9 doi: 10.1111/j.1471-4159.1985.tb05451.x [DOI] [PubMed] [Google Scholar]
  4. Blennow K, et al. (1993). Cerebrospinal fluid monoamine metabolites in 114 healthy individuals 18-88 years of age. Eur Neuropsychopharmacol 3, 55–61 doi: 10.1016/0924-977x(93)90295-w [DOI] [PubMed] [Google Scholar]
  5. Blom HJ, Smulders Y (2011). Overview of homocysteine and folate metabolism. With special references to cardiovascular disease and neural tube defects. J Inherit Metab Dis 34, 75–81 doi: 10.1007/s10545-010-9177-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bouckoms AJ, et al. (1992). Monoamines in the brain cerebrospinal fluid of facial pain patients. Anesth Prog 39, 201–8 [PMC free article] [PubMed] [Google Scholar]
  7. Brewerton TD, Putnam KT, Lewine RRJ, Risch SC (2018). Seasonality of cerebrospinal fluid monoamine metabolite concentrations and their associations with meteorological variables in humans. J Psychiatr Res 99, 76–82 doi: 10.1016/j.jpsychires.2018.01.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bridge TP, Soldo BJ, Phelps BH, Wise CD, Francak MJ, Wyatt RJ (1985). Platelet monoamine oxidase activity: demographic characteristics contribute to enzyme activity variability. J Gerontol 40, 23–8 doi: 10.1093/geronj/40.1.23 [DOI] [PubMed] [Google Scholar]
  9. Cascalheira JF, et al. (2009). Serum homocysteine: interplay with other circulating and genetic factors in association to Alzheimer's type dementia. Clin Biochem 42, 783–90 doi: 10.1016/j.clinbiochem.2009.02.006 [DOI] [PubMed] [Google Scholar]
  10. Chen Y, Guillemin GJ (2009). Kynurenine pathway metabolites in humans: disease and healthy States. Int J Tryptophan Res 2, 1–19 doi: 10.4137/ijtr.s2097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Colin-Gonzalez AL, Maldonado PD, Santamaria A (2013). 3-Hydroxykynurenine: an intriguing molecule exerting dual actions in the central nervous system. Neurotoxicology 34, 189–204 doi: 10.1016/j.neuro.2012.11.007 [DOI] [PubMed] [Google Scholar]
  12. Condray R, et al. (2011). 3-Hydroxykynurenine and clinical symptoms in first-episode neuroleptic-naive patients with schizophrenia. Int J Neuropsychopharmacol 14, 756–67 doi: 10.1017/S1461145710001689 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. De Bellis MD, Geracioti TD Jr., Altemus M, Kling MA (1993). Cerebrospinal fluid monoamine metabolites in fluoxetine-treated patients with major depression and in healthy volunteers. Biol Psychiatry 33, 636–41 doi: 10.1016/0006-3223(93)90103-k [DOI] [PubMed] [Google Scholar]
  14. Dou L, et al. (2015). The cardiovascular effect of the uremic solute indole-3 acetic acid. J Am Soc Nephrol 26, 876–87 doi: 10.1681/ASN.2013121283 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Droge W (2005). Oxidative stress and ageing: is ageing a cysteine deficiency syndrome? Philosophical Transactions of the Royal Society B-Biological Sciences 360, 2355–2372 doi: 10.1098/rstb.2005.1770 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ehnvall A, Sjogren M, Zachrisson OC, Agren H (2003). Lifetime burden of mood swings and activation of brain norepinephrine turnover in patients with treatment-refractory depressive illness. J Affect Disord 74, 185–9 doi: 10.1016/s0165-0327(02)00011-3 [DOI] [PubMed] [Google Scholar]
  17. Fukagawa NK, Martin JM, Wurthmann A, Prue AH, Ebenstein D, O'Rourke B (2000). Sex-related differences in methionine metabolism and plasma homocysteine concentrations. Am J Clin Nutr 72, 22–9 doi: 10.1093/ajcn/72.1.22 [DOI] [PubMed] [Google Scholar]
  18. Geracioti TD Jr., et al. (1997). Uncoupling of serotonergic and noradrenergic systems in depression: preliminary evidence from continuous cerebrospinal fluid sampling. Depress Anxiety 6, 89–94 [PubMed] [Google Scholar]
  19. Gerner RH, et al. (1984). Csf Neurochemistry in Depressed, Manic, and Schizophrenic-Patients Compared with That of Normal Controls. American Journal of Psychiatry 141, 1533–1540 [DOI] [PubMed] [Google Scholar]
  20. Goudas LC, et al. (1999). Acute decreases in cerebrospinal fluid glutathione levels after intracerebroventricular morphine for cancer pain. Anesth Analg 89, 1209–15 [PubMed] [Google Scholar]
  21. Heafield MT, Fearn S, Steventon GB, Waring RH, Williams AC, Sturman SG (1990). Plasma cysteine and sulphate levels in patients with motor neurone, Parkinson's and Alzheimer's disease. Neuroscience Letters 110, 216–20 doi: 10.1016/0304-3940(90)90814-p [DOI] [PubMed] [Google Scholar]
  22. Hou CL, Jia FJ, Liu Y, Li LJ (2006). CSF serotonin, 5-hydroxyindolacetic acid and neuropeptide Y levels in severe major depressive disorder. Brain Research 1095, 154–158 doi: 10.1016/j.brainres.2006.04.026 [DOI] [PubMed] [Google Scholar]
  23. Jones JS, et al. (1990). Csf 5-Hiaa and Hva Concentrations in Elderly Depressed-Patients Who Attempted-Suicide. American Journal of Psychiatry 147, 1225–1227 [DOI] [PubMed] [Google Scholar]
  24. Jones RS (1982). Tryptamine: a neuromodulator or neurotransmitter in mammalian brain? Prog Neurobiol 19, 117–39 doi: 10.1016/0301-0082(82)90023-5 [DOI] [PubMed] [Google Scholar]
  25. Kaddurah-Daouk R, et al. (2011). Metabolomic changes in autopsy-confirmed Alzheimer's disease. Alzheimers Dement 7, 309–17 doi: 10.1016/j.jalz.2010.06.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kaddurah-Daouk R, et al. (2012). Cerebrospinal fluid metabolome in mood disorders-remission state has a unique metabolic profile. Sci Rep 2, 667 doi: 10.1038/srep00667 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kaddurah-Daouk R, et al. (2013). Alterations in metabolic pathways and networks in Alzheimer's disease. Transl Psychiatry 3, e244 doi: 10.1038/tp.2013.18 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kasa K, Otsuki S, Yamamoto M, Sato M, Kuroda H, Ogawa N (1982). Cerebrospinal-Fluid Gamma-Aminobutyric Acid and Homovanillic-Acid in Depressive-Disorders. Biological Psychiatry 17, 877–883 [PubMed] [Google Scholar]
  29. Kegel ME, et al. (2014). Imbalanced kynurenine pathway in schizophrenia. Int J Tryptophan Res 7, 15–22 doi: 10.4137/IJTR.S16800 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kim YS, Unno T, Kim BY, Park MS (2019). Sex Differences in Gut Microbiota. World J Mens Health, doi: 10.5534/wjmh.190009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Koslow SH, Maas JW, Bowden CL, Davis JM, Hanin I, Javaid J (1983). Csf and Urinary Biogenic-Amines and Metabolites in Depression and Mania - a Controlled, Univariate Analysis. Archives of General Psychiatry 40, 999–1010 [DOI] [PubMed] [Google Scholar]
  32. Krause D, et al. (2011). The tryptophan metabolite 3-hydroxyanthranilic acid plays anti-inflammatory and neuroprotective roles during inflammation: role of hemeoxygenase-1. Am J Pathol 179, 1360–72 doi: 10.1016/j.ajpath.2011.05.048 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kristal BS, Shurubor YI, Kaddurah-Daouk R, Matson WR (2007). High-performance liquid chromatography separations coupled with coulometric electrode array detectors: a unique approach to metabolomics. Methods Mol Biol 358, 159–74 doi: 10.1007/978-1-59745-244-1_10 [DOI] [PubMed] [Google Scholar]
  34. Kristal BS, Vigneau-Callahan K, Matson WR (2002). Simultaneous analysis of multiple redox-active metabolites from biological matrices. Methods Mol Biol 186, 185–94 doi: 10.1385/1-59259-173-6:185 [DOI] [PubMed] [Google Scholar]
  35. Kristal BS, Vigneau-Callahan KE, Matson WR (1998). Simultaneous analysis of the majority of low-molecular-weight, redox-active compounds from mitochondria. Anal Biochem 263, 18–25 doi: 10.1006/abio.1998.2831 [DOI] [PubMed] [Google Scholar]
  36. Lewerenz J, et al. (2013). The cystine/glutamate antiporter system x(c)(−) in health and disease: from molecular mechanisms to novel therapeutic opportunities. Antioxid Redox Signal 18, 522–55 doi: 10.1089/ars.2011.4391 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. LeWitt PA, et al. (1992). Markers of dopamine metabolism in Parkinson's disease. The Parkinson Study Group. Neurology 42, 2111–7 doi: 10.1212/wnl.42.11.2111 [DOI] [PubMed] [Google Scholar]
  38. Little JT, Ketter TA, Mathe AA, Frye MA, Luckenbaugh D, Post RM (1999). Venlafaxine but not bupropion decreases cerebrospinal fluid 5-hydroxyindoleacetic acid in unipolar depression. Biol Psychiatry 45, 285–9 doi: 10.1016/s0006-3223(98)00078-x [DOI] [PubMed] [Google Scholar]
  39. Liu B, Shen Y, Xiao K, Tang Y, Cen L, Wei J (2012). Serum uric acid levels in patients with multiple sclerosis: a meta-analysis. Neurol Res 34, 163–71 doi: 10.1179/1743132811Y.0000000074 [DOI] [PubMed] [Google Scholar]
  40. Liu D, et al. (2018). Beta-defensin 1, aryl hydrocarbon receptor and plasma kynurenine in major depressive disorder: metabolomics-informed genomics. Transl Psychiatry 8, 10 doi: 10.1038/s41398-017-0056-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lucot JB, Crampton GH, Matson WR, Gamache PH (1989). Cerebrospinal fluid constituents of cat vary with susceptibility to motion sickness. Life Sci 44, 1239–45 doi: 10.1016/0024-3205(89)90359-7 [DOI] [PubMed] [Google Scholar]
  42. Matson WR, Langlais P, Volicer L, Gamache PH, Bird E, Mark KA (1984). n-Electrode three-dimensional liquid chromatography with electrochemical detection for determination of neurotransmitters. Clin Chem 30, 1477–88 [PubMed] [Google Scholar]
  43. Molchan SE, et al. (1991). CSF monoamine metabolites and somatostatin in Alzheimer's disease and major depression. Biol Psychiatry 29, 1110–8 doi: 10.1016/0006-3223(91)90253-i [DOI] [PubMed] [Google Scholar]
  44. Murphy DL, Wright C, Buchsbaum M, Nichols A, Costa JL, Wyatt RJ (1976). Platelet and Plasma Amine Oxidase Activity in 680 Normals - Sex and Age-Differences and Stability over Time. Biochemical Medicine 16, 254–265 doi:Doi 10.1016/0006-2944(76)90031-4 [DOI] [Google Scholar]
  45. Nagata Y, et al. (2018). Comparative analysis of cerebrospinal fluid metabolites in Alzheimer's disease and idiopathic normal pressure hydrocephalus in a Japanese cohort. Biomark Res 6, 5 doi: 10.1186/s40364-018-0119-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Nilsson LK, Nordin C, Jonsson EG, Engberg G, Linderholm KR, Erhardt S (2007). Cerebrospinal fluid kynurenic acid in male and female controls - correlation with monoamine metabolites and influences of confounding factors. J Psychiatr Res 41, 144–51 doi: 10.1016/j.jpsychires.2005.12.001 [DOI] [PubMed] [Google Scholar]
  47. Nørgaard L, Bro R, Westad F, Engelsen SB (2006). A modification of canonical variates analysis to handle highly collinear multivariate data. Journal of Chemometrics 20, 425–435 [Google Scholar]
  48. O'Mahony SM, Clarke G, Borre YE, Dinan TG, Cryan JF (2015). Serotonin, tryptophan metabolism and the brain-gut-microbiome axis. Behav Brain Res 277, 32–48 doi: 10.1016/j.bbr.2014.07.027 [DOI] [PubMed] [Google Scholar]
  49. Olivola E, et al. (2014). Serotonin impairment in CSF of PD patients, without an apparent clinical counterpart. PLoS One 9, e101763 doi: 10.1371/journal.pone.0101763 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Oreland L, et al. (1981). Platelet MAO activity and monoamine metabolites in cerebrospinal fluid in depressed and suicidal patients and in healthy controls. Psychiatry Res 4, 21–9 doi: 10.1016/0165-1781(81)90004-4 [DOI] [PubMed] [Google Scholar]
  51. Paganoni S, et al. (2012). Uric acid levels predict survival in men with amyotrophic lateral sclerosis. J Neurol 259, 1923–8 doi: 10.1007/s00415-012-6440-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Palaniappun V, Ramachandran V, Somasundaram O (1991). Norepinephrine and serotonin metabolism and clinical response to combined imipramine and amitriptyline therapy in depression. Indian J Psychiatry 33, 193–9 [PMC free article] [PubMed] [Google Scholar]
  53. Petersen AK, et al. (2012). On the hypothesis-free testing of metabolite ratios in genome-wide and metabolome-wide association studies. BMC Bioinformatics 13, 120 doi: 10.1186/1471-2105-13-120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Post RM, Gordon EK, Goodwin FK, Bunney WE Jr. (1973). Central norepinephrine metabolism in affective illness: MHPG in the cerebrospinal fluid. Science 179, 1002–3 doi: 10.1126/science.179.4077.1002 [DOI] [PubMed] [Google Scholar]
  55. Reddy PL, Khanna S, Subhash MN, Channabasavanna SM, Rao BS (1992). CSF amine metabolites in depression. Biol Psychiatry 31, 112–8 doi: 10.1016/0006-3223(92)90198-9 [DOI] [PubMed] [Google Scholar]
  56. Reus GZ, Jansen K, Titus S, Carvalho AF, Gabbay V, Quevedo J (2015). Kynurenine pathway dysfunction in the pathophysiology and treatment of depression: Evidences from animal and human studies. J Psychiatr Res 68, 316–28 doi: 10.1016/j.jpsychires.2015.05.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Roy A, et al. (1986). Reduced CSF concentrations of homovanillic acid and homovanillic acid to 5-hydroxyindoleacetic acid ratios in depressed patients: relationship to suicidal behavior and dexamethasone nonsuppression. Am J Psychiatry 143, 1539–45 doi: 10.1176/ajp.143.12.1539 [DOI] [PubMed] [Google Scholar]
  58. Roy A, Pickar D, De Jong J, Karoum F, Linnoila M (1988). Norepinephrine and its metabolites in cerebrospinal fluid, plasma, and urine. Relationship to hypothalamic-pituitary-adrenal axis function in depression. Arch Gen Psychiatry 45, 849–57 doi: 10.1001/archpsyc.1988.01800330081010 [DOI] [PubMed] [Google Scholar]
  59. Roy A, Pickar D, De Jong J, Karoum F, Linnoila M (1989). Suicidal behavior in depression: relationship to noradrenergic function. Biol Psychiatry 25, 341–50 doi: 10.1016/0006-3223(89)90181-9 [DOI] [PubMed] [Google Scholar]
  60. Rozen S, et al. (2005). Metabolomic analysis and signatures in motor neuron disease. Metabolomics 1, 101–108 doi: 10.1007/s11306-005-4810-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Savitz J, et al. (2015). Neuroprotective kynurenine metabolite indices are abnormally reduced and positively associated with hippocampal and amygdalar volume in bipolar disorder. Psychoneuroendocrinology 52, 200–11 doi: 10.1016/j.psyneuen.2014.11.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Sher L, et al. (2006). Lower cerebrospinal fluid homovanillic acid levels in depressed suicide attempters. J Affect Disord 90, 83–9 doi: 10.1016/j.jad.2005.10.002 [DOI] [PubMed] [Google Scholar]
  63. Sher L, et al. (2005). Higher cerebrospinal fluid homovanillic acid levels in depressed patients with comorbid posttraumatic stress disorder. Eur Neuropsychopharmacol 15, 203–9 doi: 10.1016/j.euroneuro.2004.09.009 [DOI] [PubMed] [Google Scholar]
  64. Sher L, et al. (2003). Lower CSF homovanillic acid levels in depressed patients with a history of alcoholism. Neuropsychopharmacology 28, 1712–9 doi: 10.1038/sj.npp.1300231 [DOI] [PubMed] [Google Scholar]
  65. Shi H, Vigneau-Callahan KE, Matson WR, Kristal BS (2002). Attention to relative response across sequential electrodes improves quantitation of coulometric array. Anal Biochem 302, 239–45 doi: 10.1006/abio.2001.5568 [DOI] [PubMed] [Google Scholar]
  66. Shurubor YI, Matson WR, Martin RJ, Kristal BS (2005). Relative contribution of specific sources of systematic errors and analytical imprecision to metabolite analysis by HPLC–ECD. Metabolomics 1, 159–168 doi: 10.1007/s11306-005-4431-8 [DOI] [Google Scholar]
  67. Stover JF, Lowitzsch K, Kempski OS (1997). Cerebrospinal fluid hypoxanthine, xanthine and uric acid levels may reflect glutamate-mediated excitotoxicity in different neurological diseases. Neuroscience Letters 238, 25–28 doi:Doi 10.1016/S0304-3940(97)00840-9 [DOI] [PubMed] [Google Scholar]
  68. Sullivan GM, Mann JJ, Oquendo MA, Lo ES, Cooper TB, Gorman JM (2006). Low cerebrospinal fluid transthyretin levels in depression: correlations with suicidal ideation and low serotonin function. Biol Psychiatry 60, 500–6 doi: 10.1016/j.biopsych.2005.11.022 [DOI] [PubMed] [Google Scholar]
  69. Sullivan GM, Oquendo MA, Huang YY, Mann JJ (2006). Elevated cerebrospinal fluid 5-hydroxyindoleacetic acid levels in women with comorbid depression and panic disorder. Int J Neuropsychopharmacol 9, 547–56 doi: 10.1017/S1461145705006231 [DOI] [PubMed] [Google Scholar]
  70. Swann AC, Katz MM, Bowden CL, Berman NG, Stokes PE (1999). Psychomotor performance and monoamine function in bipolar and unipolar affective disorders. Biol Psychiatry 45, 979–88 doi: 10.1016/s0006-3223(98)00172-3 [DOI] [PubMed] [Google Scholar]
  71. Tabunoki H, et al. (2013). Identification of key uric acid synthesis pathway in a unique mutant silkworm Bombyx mori model of Parkinson's disease. PLoS One 8, e69130 doi: 10.1371/journal.pone.0069130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Vanholder R, Pletinck A, Schepers E, Glorieux G (2018). Biochemical and Clinical Impact of Organic Uremic Retention Solutes: A Comprehensive Update. Toxins (Basel) 10, doi: 10.3390/toxins10010033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Volicer L, Langlais PJ, Matson WR, Mark KA, Gamache PH (1985). Serotoninergic System in Dementia of the Alzheimer Type: Abnormal Forms of 5-Hydroxytryptophan and Serotonin in Cerebrospinal Fluid. Archives of Neurology 42, 1158–1161 doi: 10.1001/archneur.1985.04060110040013 [DOI] [PubMed] [Google Scholar]
  74. Westenberg HG, Verhoeven WM (1988). CSF monoamine metabolites in patients and controls: support for a bimodal distribution in major affective disorders. Acta Psychiatr Scand 78, 541–9 doi: 10.1111/j.1600-0447.1988.tb06382.x [DOI] [PubMed] [Google Scholar]
  75. Widerlov E, Bissette G, Nemeroff CB (1988). Monoamine Metabolites, Corticotropin Releasing-Factor and Somatostatin as Csf Markers in Depressed-Patients. Journal of Affective Disorders 14, 99–107 doi:Doi 10.1016/0165-0327(88)90051-1 [DOI] [PubMed] [Google Scholar]
  76. Williams RB, et al. (2003). Serotonin-related gene polymorphisms and central nervous system serotonin function. Neuropsychopharmacology 28, 533–41 doi: 10.1038/sj.npp.1300054 [DOI] [PubMed] [Google Scholar]
  77. Yoon HS, et al. (2017). Relationships of Cerebrospinal Fluid Monoamine Metabolite Levels With Clinical Variables in Major Depressive Disorder. J Clin Psychiatry 78, e947–e956 doi: 10.4088/JCP.16m11144 [DOI] [PubMed] [Google Scholar]
  78. Young SN, Anderson GM, Gauthier S, Purdy WC (1980). The origin of indoleacetic acid and indolepropionic acid in rat and human cerebrospinal fluid. J Neurochem 34, 1087–92 doi: 10.1111/j.1471-4159.1980.tb09944.x [DOI] [PubMed] [Google Scholar]
  79. Young SN, Anderson GM, Purdy WC (1980). Indoleamine metabolism in rat brain studied through measurements of tryptophan, 5-hydroxyindoleacetic acid, and indoleacetic acid in cerebrospinal fluid. J Neurochem 34, 309–15 doi: 10.1111/j.1471-4159.1980.tb06598.x [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1679875-supplement-1.pdf (1.2MB, pdf)

RESOURCES