Elsevier

Journal of Psychiatric Research

Volume 96, January 2018, Pages 23-32
Journal of Psychiatric Research

Plasma amino acid profile in major depressive disorder: Analyses in two independent case-control sample sets

https://doi.org/10.1016/j.jpsychires.2017.09.014Get rights and content

Abstract

Some amino acids act as neurotransmitters themselves, or are precursors of neurotransmitters. Previous studies reported inconsistent results regarding their changes in blood in major depressive disorder (MDD), which prompted us to examine plasma levels of amino acids and related molecules in two independent case-control sample sets. In total, 511 subjects were recruited. Sample set A consisted of 164 patients with MDD (147 currently depressed [dMDD]; 17 in remission, DSM-IV) and 217 healthy controls. Sample set B consisted of 65 patients (51 dMDD; 14 in remission) and 65 controls. Plasma amino acid levels were measured using high-performance liquid chromatography for set A and liquid chromatography/mass spectrometry for set B. We further analyzed the relationships between plasma amino acid levels and clinical variables. In sample set A, plasma asparagine, histidine+1-methylhistidine, methionine, phenylalanine, tryptophan, and tyrosine levels were decreased, while plasma glutamate and phosphoethanolamine were elevated in dMDD compared to controls (all P < 0.0005), even after correcting for multiple testing. Plasma leucine levels were associated with “psychic anxiety.” In sample set B, glutamate and methionine levels were also altered in the same direction to that in sample set A (both P < 0.05). In the integrative analysis, plasma glutamate and methionine levels were found to be significantly associated with the diagnosis of MDD with small to medium effect sizes (both P < 1.0E-6). In conclusion, several amino acids and related molecules were altered in patients with MDD. Decreased methionine and increased glutamate levels were found consistently in the two sample sets, suggesting their involvement in MDD. Further investigations are warranted on the possible role of amino acids in the pathophysiology of MDD.

Introduction

Major depressive disorder (MDD) is a common disease. Its adjusted global prevalence is estimated at 4.4% worldwide (Ferrari et al., 2013), and it is ranked as the third largest cause of years lived with disability (YLDs) (GBD 2015 Disease and Injury Incidence and Prevalence Collaborators, 2016). Although several promising biological mechanisms have been proposed, such as the monoamine hypothesis, hypothalamic-pituitary-adrenal (HPA) axis dysregulation, and chronic inflammation, the pathophysiology of MDD remains elusive, and there are no established biomarkers (Kunugi et al., 2015).

Alterations in serotonergic (Charney and Manji, 2004), noradrenergic (Charney and Manji, 2004), dopaminergic (Ogawa and Kunugi, 2015), glutamatergic (Hashimoto, 2009), and γ-amino-butyric acid (GABAergic) (Pehrson and Sanchez, 2015) systems are implicated in psychiatric disorders (Hasler, 2010), and these neurotransmitters are themselves amino acids or are synthesized from amino acids. For example, tryptophan is a precursor of serotonin, and has well been investigated in MDD. Our meta-analysis demonstrated decreased plasma tryptophan levels in patients with MDD (Ogawa et al., 2014). However, other amino acids and related molecules are of interest also. Several previous studies focused on blood levels of amino acids that are precursors to monoamine neurotransmitters in patients with MDD and controls (DeMyer et al., 1981, Hoekstra et al., 2001, Lucca et al., 1992, Maes et al., 1990, Moller, 1993, Russ et al., 1990). Other studies investigated the profiles of plasma amino acids and related molecules in patients and controls (Altamura et al., 1995, Maes et al., 1998, Mauri et al., 1998, Mauri et al., 2001, Mitani et al., 2006, Woo et al., 2015, Xu et al., 2012). However, the results are inconsistent. Some studies reported significantly increased alanine (Pinto et al., 2012, Woo et al., 2015), glutamate (Mauri et al., 1998, Mitani et al., 2006, Pinto et al., 2012, Woo et al., 2015), glutamine (Mitani et al., 2006, Xu et al., 2012), ornithine (Pinto et al., 2012), d-serine (Hashimoto et al., 2016), l-serine (Hashimoto et al., 2016, Pinto et al., 2012), taurine (Altamura et al., 1995, Mauri et al., 1998, Mitani et al., 2006, Pinto et al., 2012), and tyrosine (Pinto et al., 2012) levels, while others reported significantly decreased α-aminoadipic acid (Xu et al., 2012), β-aminoisobutyric acid (Woo et al., 2015), β-alanine (Woo et al., 2015), arginine (Pinto et al., 2012), aspartate (Pinto et al., 2012), cystathionine (Woo et al., 2015), GABA (Xu et al., 2012), homocysteine (Woo et al., 2015), isoleucine (Mauri et al., 2001), leucine (Xu et al., 2012), O-phospho-l-serine (Woo et al., 2015), sarcosine (Woo et al., 2015), and tryptophan (Xu et al., 2012) levels. The directions of change in blood ethanolamine (Woo et al., 2015, Xu et al., 2012), glycine (Altamura et al., 1995, Mitani et al., 2006), lysine (Mauri et al., 1998, Xu et al., 2012), methionine (Woo et al., 2015, Xu et al., 2012), and phenylalanine (Pinto et al., 2012, Xu et al., 2012) levels in patients with MDD were inconsistent. This may have arisen for several reasons. First, small sample sizes and low statistical power (Button et al., 2013) may have caused false-negative results (type II error). Second, not controlling for demographic factors, such as sex and age, may have yielded spurious results. Only a few studies (Altamura et al., 1995, Maes et al., 1998) controlled for these variables. Third, most studies did not correct for multiple testing, and so are subject to false-positive results (type I error).

Thus, we aimed to determine plasma levels of amino acids and related molecules in two independent case-control sample sets (N = 511) taking demographic, clinical variables, and multiple comparisons into account. We also examined whether plasma amino acids levels were related to clinical variables, such as psychotropic drug doses and MDD symptoms.

We examined two independent sample sets, A and B. Set A consisted of 164 patients with MDD (147 currently depressed [dMDD] and 17 remitted [rMDD]) and 217 healthy controls (Table 1), and was obtained from our bio-resource repository for biomarker research projects, collected between October 2003 and October 2015. Set B consisted of 65 patients with MDD (51 dMDD and 14 remitted) and 65 controls (Supplemental Table S1) who participated in a nutrition study between March 2011 and July 2012. Subjects were recruited through advertisements in free local magazines, our website, and the National Center of Neurology and Psychiatry hospital (Tokyo, Japan). Thirty-four participants who were included in both sets A and B were omitted from set B. Trained psychologists/psychiatrists screened all subjects using the structured interview M.I.N.I. (Mini International Neuropsychiatric Interview) (Sheehan et al., 1998), Japanese edition (Otsubo et al., 2005). A consensus diagnosis was made by at least two psychiatrists according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) criteria (American Psychiatric Association, 1994) and information from medical records, if available. Individuals with prior medical history of central nervous system disease, severe head injury, or substance abuse/dependence were excluded. Individuals who had a history of past or current regular contact with psychiatric services, or a history of psychotropic drug use were excluded from the healthy controls. Depressive symptoms were assessed using the Japanese version of the 17-item Hamilton Rating Scale for Depression (Hamilton, 1960, Tabuse et al., 2007) and the cut-off score for remission was ≤ 7 (Zimmerman et al., 2013). Daily doses of benzodiazepine derivatives and antidepressants were converted to equivalent diazepam and imipramine doses, respectively, using published guidelines (Inagaki et al., 2013).

This study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the ethics committee at NCNP (No. 305). All subjects provided written informed consent.

Blood collection was performed in a “real world setting” without fasting. Venous blood was drawn into ethylenediamine tetra-acetic acid sodium (EDTA-2Na) vacutainer tubes (Terumo, Tokyo, Japan). The blood was immediately placed on ice and then centrifuged at 4 °C at 3000 rpm for 15 min, and the supernatant was aliquoted into microtubes and stored at −80 °C.

Researchers who handled the blood samples were blind to clinical information. For set A, samples were analyzed using high-performance liquid chromatography (HPLC) as described previously (Ogawa et al., 2015). We prepared 40 standards for the absolute calibration curve using the Amino Acids Mixture Standard Solution, Type B and Type AN-2 (Wako Pure Chemical Industries, Ltd. Osaka, Japan), and an additional standard solution mixture for tryptophan, glutamine, theanine, proline, and asparagine (all purchased from Wako). Since histidine and 1-methylhistidine cannot be separated by their peak in our system, we used combined plasma histidine+1-methylhistidine values. For set B, samples were analyzed using liquid chromatography/mass spectrometry (LC/MS) (SRL Co. Inc. Tokyo, Japan). Briefly, plasma samples deproteinized using acetonitrile were centrifuged twice at 4 °C at 12,000 rpm for 10 min and the supernatant was aliquoted. The prepared samples were separated on an octadecyl-silica (ODS) column (Inertsil ODS-3, 2.1 mm ID × 100 mm, GL Sciences Inc. Tokyo, Japan) during the HPLC analysis and amino acid levels were determined via the LC-MS2020 system (Shimadzu Corp., Kyoto, Japan) at SRL.

We assigned several dummy variables to each batch for HPLC in set A (three batches) to control for batch effects. Continuous variables are reported as means ± standard deviation (SD). We compared demographic and clinical data using a t-test or analysis of variance (ANOVA), and categorical distributions using a chi-squared test. An analysis of covariance (ANCOVA) was used to compare plasma amino acid and related molecule levels between dMDD and controls, controlling for age, sex, and batch effects. Plasma amino acid levels of the rMDD group were not used in the comparison of patients and controls. However, they were used in the multiple regression analyses to examine the association between plasma amino acid levels and depression severity. To estimate the association of plasma amino acids with diagnosis, a logistic regression was performed using diagnosis as the dependent variable, age, sex, and batch numbers as the explanatory variables with forced entry, and plasma amino acid levels as explanatory variables with forward stepwise selection. The relationships between plasma amino acid levels and MDD symptoms were assessed using multiple regressions in all patients (dMDD and rMDD). In the multiple regression analysis, the HAMD-17 total score and subscale scores were used as dependent variables, age, sex, and batch numbers were used as explanatory variables with forced entry, and plasma amino acid levels were used as explanatory variables with stepwise selection. The HAMD-17 items were assigned to the following subscales: “core” (items 1, 2, 7, 8, 10, 13), “sleep” (items 4, 5, 6), “activity” (items 7, 8), “psychic anxiety” (items 9, 10), and “somatic anxiety” (items 11, 12, 13) according to the study by Serretti et al. (1999). A partial correlation analysis, controlling for age, sex, and batch, was performed for all patients to examine the association between plasma amino acid levels and daily psychotropic drug doses. We used non-parametric receiver operating characteristic (ROC) curves to estimate the ability of an integrative value of several amino acid plasma levels to discriminate between the dMDD and control groups, adjusting for sex, age, and batch effects by residuals. Differences with a two-tailed P-value < 0.05 were deemed statistically significant, and Bonferroni correction was used for multiple comparisons. Analyses were performed using IBM SPSS Statistics 22.0 (IBM Japan, Tokyo, Japan).

Section snippets

Results

Demographic and clinical characteristics of sample set A are described in Table 1. There were no significant differences in sex or age between the groups. Although the ratio of smokers in the rMDD group was significantly different from that in the dMDD (χ2 = 4.72, df = 1, P = 0.030) and in the healthy control (χ2 = 7.90, df = 1, P = 0.0049) groups, there was no significant difference between the dMDD and control groups (χ2 = 1.12, df = 1, P = 0.29).

Discussion

The main findings of this study are as follows: plasma asparagine, histidine+1-methylhistidine, methionine, phenylalanine, tryptophan, and tyrosine levels were significantly decreased, and glutamate and phosphoethanolamine levels were significantly increased in the dMDD group compared to controls in set A, even after Bonferroni correction. There were significant associations between plasma glutamate levels, with a “small” effect size, and phosphoethanolamine levels, with a “small-to-medium”

Contributions

S. Ogawa performed the data and statistical analyses and wrote the draft of the manuscript. J. Matsuo assisted in manuscript composition and provided critical advice. N. Koga, K. Hattori, M. Ota, D. Sasayama, H. Hori, I. Ishida, and T. Teraishi performed psychiatric assessments and obtained demographic and clinical information from the study subjects. K. Hattori, M. Ota, H. Hori, D. Sasayama, and T. Teraishi collected the blood samples. N. Koga, K. Hattori, I. Ishida, S. Yoshida, T. Noda, and

Funding

This study was supported by an Intramural Research Grant (24-11, 27-1, 27-6) from the Neurological and Psychiatric Disorders of National Center of Neurology and Psychiatry (H.K.), Grant-in-Aid for Scientific Research (A) (No. 25253075) from the Ministry of Education, Culture, Sports, Science and Technology (H.K.), and Health and Labour Sciences Research Grants for Comprehensive Research on Persons with Disabilities from the Japan Agency for Medical Research and Development, AMED (

Conflicts of interest

All authors declare no conflicts of interest.

Acknowledgements

We wish to thank all volunteers who participated in the study. We express our gratitude to Ms. Tomoko Miyakawa, Noriko Yamamoto, Misao Nakano, Yukie Kuchiki, Mariko Iwaki, and Mr. Ryo Matsumura for their technical assistance. We extend sincere thanks to Ms. Yuki Yokota, Tomoko Kurashimo, Chie Kimizuka, Moeko Hiraishi, and Anna Nagashima for their assistance in the recruitment of participants and psychiatric assessments.

References (65)

  • S.E. Moller

    Plasma amino acid profiles in relation to clinical response to moclobemide in patients with major depression. Danish University Antidepressant Group

    J. Affect Disord.

    (1993)
  • T. Murakami et al.

    Forced swimming and imipramine modify plasma and brain amino acid concentrations in mice

    Eur. J. Pharmacol.

    (2009)
  • K.J. Parker et al.

    Neuroendocrine aspects of hypercortisolism in major depression

    Horm. Behav.

    (2003)
  • V.L. Pinto et al.

    Low plasma levels of L-arginine, impaired intraplatelet nitric oxide and platelet hyperaggregability: implications for cardiovascular disease in depressive patients

    J. Affect Disord.

    (2012)
  • M.J. Russ et al.

    Plasma tryptophan to large neutral amino acid ratios in depressed and normal subjects

    J. Affect Disord.

    (1990)
  • P. She et al.

    Disruption of BCATm in mice leads to increased energy expenditure associated with the activation of a futile protein turnover cycle

    Cell Metab.

    (2007)
  • H. Tabuse et al.

    The new GRID Hamilton Rating Scale for Depression demonstrates excellent inter-rater reliability for inexperienced and experienced raters before and after training

    Psychiatry Res.

    (2007)
  • H.B. Xu et al.

    Potential clinical utility of plasma amino acid profiling in the detection of major depressive disorder

    Psychiatry Res.

    (2012)
  • M. Zimmerman et al.

    Severity classification on the Hamilton depression rating scale

    J. Affect Disord.

    (2013)
  • American Psychiatric Association

    Diagnostic and Statistical Manual of Mental Disorders: DSM-IV

    (1994)
  • C.L. Beasley et al.

    Proteomic analysis of the anterior cingulate cortex in the major psychiatric disorders: evidence for disease-associated changes

    Proteomics

    (2006)
  • K.M. Bell et al.

    S-adenosylmethionine blood levels in major depression: changes with drug treatment

    Acta Neurol. Scand. Suppl.

    (1994)
  • T. Bottiglieri et al.

    Cerebrospinal fluid S-adenosylmethionine in depression and dementia: effects of treatment with parenteral and oral S-adenosylmethionine

    J. Neurol. Neurosurg. Psychiatry

    (1990)
  • T. Bottiglieri et al.

    Homocysteine, folate, methylation, and monoamine metabolism in depression

    J. Neurol. Neurosurg. Psychiatry

    (2000)
  • K.S. Button et al.

    Power failure: why small sample size undermines the reliability of neuroscience

    Nat. Rev. Neurosci.

    (2013)
  • D.S. Charney et al.

    Life stress, genes, and depression: multiple pathways lead to increased risk and new opportunities for intervention

    Sci. STKE

    (2004)
  • S. Chinn

    A simple method for converting an odds ratio to effect size for use in meta-analysis

    Stat. Med.

    (2000)
  • J. Cohen

    Statistical Power Analysis for the Behavioral Sciences

    (1988)
  • D. De Berardis et al.

    A comprehensive review on the efficacy of S-Adenosyl-L-methionine in Major Depressive Disorder

    CNS Neurol. Disord. Drug Targets

    (2016)
  • M.K. DeMyer et al.

    Plasma tryptophan and five other amino acids in depressed and normal subjects

    Arch. Gen. Psychiatry

    (1981)
  • A.J. Ferrari et al.

    The epidemiological modelling of major depressive disorder: application for the Global Burden of Disease Study 2010

    PLoS One

    (2013)
  • Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015

    Lancet

    (2016)
  • Cited by (56)

    • Endogenous hydrocortisone caused metabolic perturbation and nutritional deterioration of animal-derived food in a dose-dependent manner

      2023, Food Chemistry
      Citation Excerpt :

      It also synthesized hormones and neurotransmitters together with l-Tyrosine, and involved in the metabolism of collective fat and sugar (Huang et al., 2022). The lower l-Phenylalanine and l-Tyrosine levels (Fig. 1c) in hydrocortisone-interfered pork observed here was consistent with the results from the study by Ogawa et al., 2018 that the reduction in l-Phenylalanine may be caused in part by elevated glucocorticoid (hydrocortisone) levels. Creatine has been reported to affect the meat quality such as water-holding capacity and tenderness (Jia, Fan, Shi, Zhang, Wang, & Shi, 2021).

    View all citing articles on Scopus
    View full text