Metabolomic profiles in individuals with negative affectivity and social inhibition: A...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

http://www.elsevier.com/authorsrights

Author's personal copy

Metabolomic profiles in individuals with negativeaffectivity and social inhibition: A population-basedstudy of Type D personality

Elisabeth Altmaier a,1, Rebecca T. Emeny b,1,*, Jan Krumsiek a,Maria E. Lacruz b, Karoline Lukaschek b, Sibylle Hafner b,c, Gabi Kastenmuller a,Werner Romisch-Margl a, Cornelia Prehn d, Robert P. Mohney e, Anne M. Evans e,Michael V. Milburn e, Thomas Illig f,g, Jerzy Adamski d,h, Fabian Theis a,i,Karsten Suhre a,j, Karl-Heinz Ladwig b,k

a Institute of Bioinformatics and Systems Biology, Helmholtz Zentrum Munchen, German Research Center for Environmental Health,85764 Neuherberg, Germanyb Institute of Epidemiology II, Helmholtz Zentrum Munchen, German Research Center for Environmental Health,85764 Neuherberg, GermanycDepartment of Psychiatry and Psychotherapy, Georg-August-University Gottingen, 37075 Gottingen, Germanyd Institute of Experimental Genetics, Genome Analysis Center, Helmholtz Zentrum Munchen, German Research Centerfor Environmental Health, 85764 Neuherberg, GermanyeMetabolon Inc., Durham, NC 27713, USAfResearch Unit of Molecular Epidemiology (AME), Helmholtz Zentrum Munchen, German Research Center for Environmental Health,85764 Neuherberg, GermanygHannover Unified Biobank, Hannover Medical School, 30625 Hannover, Germanyh Institute of Experimental Genetics, Life and Food Science Center Weihenstephan, Technische Universitat Munchen,85354 Freising, Germanyi Institute for Mathematical Sciences, Technische Universitat Munchen, 85748 Garching, GermanyjDepartment of Physiology and Biophysics, Weill Cornell Medical College in Qatar, Education City, Qatar Foundation,PO Box 24144, Doha, QatarkDepartment of Psychosomatic Medicine and Psychotherapy, Klinikum rechts der Isar, Technische Universitat Munchen,81644 Munchen, Germany

Received 6 August 2012; received in revised form 9 November 2012; accepted 10 November 2012

Psychoneuroendocrinology (2013) 38, 1299—1309

* Corresponding author at: EPI2, Helmholtz Zentrum Munchen, Ingolstadter Landstr. 1, D-85764 Neuherberg, Germany.Tel.: +49 089 3187 3631; fax: +49 089 3187 3667.

E-mail address: [email protected] (R.T. Emeny).1 These authors contributed equally to this work.

Available online at www.sciencedirect.com

j our na l h omepa g e: www.e l se v ie r.c om/l oca te/ psyne ue n

0306-4530/$ — see front matter # 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.psyneuen.2012.11.014


1. Introduction

The Type D personality describes a ‘‘general propensity topsychological distress’’ (Denollet and Conraads, 2011) and hasproven to have substantial significance in cardiovascular dis-ease prediction (Denollet et al., 1996; Kupper and Denollet,2007). While the global symptoms of social inhibition (SI) andnegative affectivity (NA) define the Type D individual (Denol-let, 2005), poorer mental (depression, anxiety, and increasedexhaustion) (Versteeg et al., 2012) and physical health (Molsand Denollet, 2010) are associated with this personality pat-tern. Prevalence rates between 13.5 and 35% have beenreported (Grande et al., 2012). A recent meta-analysis ofprospective studies demonstrated a significant associationbetween Type D and mortality, however this prognostic effectappears to be relevant in patients with coronary artery disease(CAD) but not congestive heart failure (CHF) (Grande et al.,2012), as indicated by large cohort studies of patients withheart failure and various cardiovascular diagnoses (Pelle et al.,2010; Coyne et al., 2011; Grande et al., 2011). Despite themounting evidence for a risk association with Type D, under-lying biologic mechanisms that link this personality type to apathological state are still being identified. While generalstress reactivity (increased heart rate, blood pressure, andcardiac output) is known to be associated with Type D (Einviket al., 2011), metabolic syndrome was also shown to be moreprevalent among healthy people with Type D personality(Mommersteeg et al., 2010). Interestingly, the SI subscale ofthe Type D construct was shown to be more strongly associatedwith night time systolic blood pressure than the NA component(Nyklicek et al., 2011). Oxidative stress (Kupper et al., 2009) aswell as increases in the inflammatory cytokine TNFa (Denolletet al., 2008, 2009b) were both observed in Type D individuals;

although more robust associations were shown between sTNFreceptors than with TNFa (Mommersteeg et al., 2012). Addi-tionally, Type D individuals suffering from acute coronarydisease are known to have dysregulated cortisol production(Whitehead et al., 2007; Molloy et al., 2008).

The elucidation of biochemical mechanisms in mentalhealth disorders is a new challenge within metabolomic stu-dies, in which numerous metabolites are simultaneously ana-lyzed. A metabolite profile captures a momentary glimpse intocellular metabolic pathways that are predominant in an indi-vidual, and therefore provide a functional readout of thephysiological state of the human body (Gieger et al., 2008).Metabolic profiles result from the interactions of individualgenetic predispositions, environmental, and behavioral fac-tors. Thus, perturbations in metabolic pathways may drivebiochemical pathways that are also associated with mentalhealth diseases such as major depressive (Paige et al., 2007)and bipolar disorder (Quinones and Kaddurah-Daouk, 2009).Comprehensive metabolic studies revealed associationsbetween changes in metabolic profiles and schizophrenia(Oresic et al., 2011b, 2012; Xuan et al., 2011) as well asAlzheimer’s disease (Oresic et al., 2011a). We thereforeemployed a metabolomic approach to identify a biochemicalsignature of the Type D personality in a population-based study.

2. Material and methods

2.1. Study participants and cross-sectional studydesign

Data are based on the KORA (Cooperative Health Research inthe Region of Augsburg) F4 study (2006—2008), a follow-up

KEYWORDSType D personality;Negative affectivity;Social inhibition;Metabolomics;Population study;Kynurenine

Summary

Background: Individuals with negative affectivity who are inhibited in social situations arecharacterized as distressed, or Type D, and have an increased risk of cardiovascular disease(CVD). The underlying biomechanisms that link this psychological affect to a pathological stateare not well understood. This study applied a metabolomic approach to explore biochemicalpathways that may contribute to the Type D personality.Methods: Type D personality was determined by the Type D Scale-14. Small molecule biochem-icals were measured using two complementary mass-spectrometry based metabolomics plat-forms. Metabolic profiles of Type D and non-Type D participants within a population-based study inSouthern Germany were compared in cross-sectional regression analyses. The PHQ-9 and GAD-7instruments were also used to assess symptoms of depression and anxiety, respectively, within thismetabolomic study.Results: 668 metabolites were identified in the serum of 1502 participants (age 32—77); 386 ofthese individuals were classified as Type D. While demographic and biomedical characteristicswere equally distributed between the groups, a higher level of depression and anxiety wasobserved in Type D individuals. Significantly lower levels of the tryptophan metabolite kynureninewere associated with Type D ( p-value corrected for multiple testing = 0.042), while no significantassociations could be found for depression and anxiety. A Gaussian graphical model analysisenabled the identification of four potentially interesting metabolite networks that are enrichedin metabolites (androsterone sulfate, tyrosine, indoxyl sulfate or caffeine) that associatenominally with Type D personality.Conclusions: This study identified novel biochemical pathways associated with Type D personalityand demonstrates that the application of metabolomic approaches in population studies canreveal mechanisms that may contribute to psychological health and disease.# 2012 Elsevier Ltd. All rights reserved.

1300 E. Altmaier et al.


study of the 4th KORA survey (S4) conducted in 1999—2001.The cross-sectional, population-based KORA S4 study samplewas recruited from the region of Augsburg and two adjacentcounties in the south of Germany (Holle et al., 2005). The F4study was restricted to subjects aged 32—81 years at follow-up. Of the 4261 participants of the S4 survey, 3080 wereexamined again in the F4 study, in average seven years afterthe baseline examination. From this group participantsbetween 32 and 77 years of age were chosen randomly formetabolomic characterization and analysis in previously pub-lished genome-wide association studies with metabolic traits(Illig et al., 2010; Suhre et al., 2011). Informed consent wasobtained from each study participant and all study protocolswere approved by the Local Ethics Committees.

2.2. Psychological characteristics

2.2.1. Type D personalityThe determination of Type D was performed using the Type DScale-14 (DS-14), a standard measure of negative affectivity(NA) and social inhibition (SI) (Denollet, 2005) as previouslyreported by Hausteiner et al. (2010). In total, 1502 of themetabolomic characterized study participants provided com-plete information for this instrument within a self-adminis-tered questionnaire. Non-Type D participants were comparedto Type D individuals who score high (>10) on both the NA andSI scales.

2.2.2. Depression (PHQ-9)The 9-item Public Health Questionnaire (PHQ-9) was admi-nistered in the personal interview. This is a short version ofthe PHQ that captures depressive symptoms on a scale of 0—27. An algorithm was used to define three categories ‘nodepressive symptom’ ‘other depressive symptoms’ and‘major depressive symptoms’ (Spitzer et al., 1999; Loweet al., 2004). This instrument provides a reliable and validassessment of depression severity (Kroenke et al., 2001).Participants with no depression were compared to those withother/major depressive symptoms.

2.2.3. Generalized anxiety disorder (GAD)The GAD-7, a 7-item instrument, that assesses fear andanxiety based symptoms, was administered in the personalinterview. This is a valid and reliable tool applied in bothclinical and research settings (Spitzer et al., 2006). A scale of0—21, with a cut-off of LE 4 is set to distinguish no to minimalanxiety from mild, moderate or heavy anxiety disorder. Thisscale was dichotomized to compare those who reported no tominimal anxiety to all others.

2.2.4. Intake of antidepressive medicationsThe use of antidepressive medication was assessed with theIDOM-software (Muhlberger et al., 2003) based on informa-tion given by the subjects and the packages they brought tothe personal interview. Medications with the AnatomicalTherapeutic Chemical Classification System (ATC)-codesN06AA (Non-selective monoamine reuptake inhibitors),N06AB (selective serotonin reuptake inhibitors), N06AF(monoamine oxidase inhibitors, and non-selective), N06AG(monoamine oxidase A inhibitors) or N06AX (other antide-pressants) were included.

2.3. Cardiometabolic risk factors

During the physical examination, standardized measure-ments of height, weight, waist circumference, blood pres-sure and serum lipids were performed as previously described(Rathmann et al., 2003). Previously known diabetes wasdefined based on self-reported physician diagnosis or useof antidiabetic agents. Hypertensive blood pressure wasdefined by systolic blood pressure more than 140 mm Hg ordiastolic blood pressure more than 90 mm Hg (with or with-out antihypertensive medications). Antihypertensive medi-cation was defined following the guidelines of the GermanHypertension Society.

2.4. Blood samples

For this study, we measured the metabolic panel in the bloodserum of study participants. The blood sampling was con-ducted by the KORA study staff in Augsburg, Germany. Bloodwas drawn in the morning between 8 and 10:30 am afterovernight fasting to avoid variation due to circadian rhythm.Material was drawn into serum gel tubes, gently invertedtwice and followed by 30 min resting at room temperature(18—25 8C) to obtain complete coagulation. The material wasthen centrifuged for 10 min (2750 � g at 15 8C). Serum wasaliquoted and stored for a maximum of 6 h at 4 8C, afterwhich it was deep frozen to �80 8C until analysis.

2.5. Metabolite profiling and metabolite panels

Two sets of metabolites were analyzed to produce a totalmeasured panel of 668 metabolites that include amino acids,carbohydrates, acylcarnitines, glycerophospholipids, sphingo-lipids, lipids, nucleotides, small peptides, xenobiotics, cofac-tors and vitamins. A third of these metabolites (n = 223),marked with an identifier consisting of the letter ‘‘X’’ andthe number, are currently identified by their mass, chromato-graphic retention time, and MS/MS fragmentation, but theirbiochemical identity is unknown. One set of metabolites(n = 517) was measured by Metabolon (Research Triangle Park,NC, USA), a commercial supplier of global metabolic analyses.The platform integrates the chemical analysis, including iden-tification and relative quantification, data reduction, andquality assurance components of the process. For every sampletwo separate ultrahigh performance liquid chromatography/tandem mass spectrometry (UHPLC/MS/MS; Thermo Scienti-fic, USA) injections and one GC/MS injection were conducted.This analytical platform is able to perform a relative quanti-tative analysis of analytical data in a high-throughput modeand identify a broad spectrum of molecules with a high degreeof confidence (Evans et al., 2009). This platform has beensuccessfully applied in studies of the adult human plasmametabolome (Lawton et al., 2008) and for the identificationof sarcosine as a biomarker for prostate cancer aggressiveness(Sreekumar et al., 2009).

A second metabolic set (n = 151) was performed at theHelmholtz Zentrum Munchen, Germany, based on electro-spray ionization (ESI) tandem mass spectrometry (MS/MS),which uses the AbsoluteIDQ kit p150 developed by BiocratesLife Sciences AG, Austria. Briefly, the assay is performedby an automated robotics system (Hamilton Bonaduz AG,

Metabolomics and Type D personality 1301


Switzerland) on a double-filter 96 well plate containing iso-tope-labeled internal standards. A 10 ml serum aliquot wasused for phenylisothiocyanate (PITC)-derivatisation of aminoacids, extraction with organic solvent and several liquid hand-ling steps. Metabolites were quantified with reference toappropriate internal standards using flow injection analysisand multiple reaction monitoring scans (FIA-MS/MS) on a 4000QTRAP instrument (AB Sciex, Germany) coupled to a HPLCPromincence (Shimadzu Deutschland GmbH, Germany). Themethod complies with the ‘‘Guidance for Industry — Bioana-lytical Method Validation’’ published by the US Food and DrugAdministration, which implies proof of reproducibility within agiven error range, and has been successfully applied in multi-ple academic and industrial settings (Altmaier et al., 2008;Mittelstrass et al., 2011; Romisch-Margl et al., 2012).

Both metabolomics datasets have been previously used ingenome-wide association studies with metabolic traits.Further information on experimental characteristics of thesedatasets (CV%, statistical distribution of the data, qualitycontrol) can be found in the material and methods section ofIllig et al. (2010) and Suhre et al. (2011), respectively.

2.6. Statistical analysis

Descriptive analyses were run to determine the distributionof demographic, biomedical, and psychological characteris-tics of the Type D personality in the population. Differences infrequencies and means were tested using the Chi-squaredand T-tests, respectively. SAS 9.2 (NC, USA) for Windows wasused for these analyses and significance was determinedwhen p < 0.05.

For the identification of metabolites associating with TypeD personality (versus non-Type D), a linear regression test wasapplied to model each metabolite concentration. Age, gen-der, HDL, LDL, cholesterol, triglycerides, hypertension, BMI,diabetes and the intake of antidepressive medications wereincluded as confounders. This approach was also applied toindividuals dichotomized by the anxiety (GAD-7) and depres-sion (PHQ-9) instruments. Additionally, analyses were per-formed with both of the linear Type D subscales (NA or SI)entered into a model as independent predictors. To controlfor the effect of testing multiple hypotheses and, thus, taking

Table 1 Biomedical and psychological characteristics of the population according to Type D classification (n = 1502).

Personality type mean (SD) or n (%)

Type D (n = 386) Non-Type D (n = 1116) p-Value

Age 58.84 (7.27) 58.57 (7.53) 0.55

Gender 0.07Male 171 (44.3) 552 (49.5)Female 215 (55.7) 564 (50.5)

HDL-cholesterol (mmol/l) 1.48 (0.38) 1.47 (0.38) 0.63LDL-cholesterol (mmol/l) 3.65 (0.90) 3.62 (0.89) 0.49Total cholesterol (mmol/l) 5.76 (1.02) 5.73 (0.98) 0.55Triglycerides (mmol/l) 1.48 (1.04) 1.52 (1.11) 0.52BMI 27.80 (5.10) 28.02 (4.72) 0.46

Diabetes mellitus 0.33Yes 33 (8.6) 76 (6.8)No 353 (91.5) 1040 (93.2)

Hypertension 0.82Yes 126 (32.6) 364 (32.6)No 260 (67.4) 752 (67.4)

Psychological characteristicsDepression (PHQ-9, n = 1499) <0.0001

No depressive syndrome 347 (90.4) 1073 (96.2)Other depressive syndromes 24 (6.2) 31 (2.8)Major depressive syndrome 13 (3.4) 11 (1.0)

Generalized anxiety disorder scale (GAD-7, n = 1500) <0.0001Minimal anxiety 139 (36.1) 810 (72.7)Mild anxiety 185 (48.1) 254 (22.8)Moderate anxiety 46 (11.9) 44 (3.9)Heavy anxiety 15 (3.9) 7 (0.6)

Intake of antidepressants 0.009Yes 26 (6.7) 39 (3.5)No 360 (93.3) 1077 (96.5)

Differences in the distribution of categorical variables between Type D and non-Type D participants were tested with the Chi-square statisticand of continuous variables using the T-test.



possible correlations among metabolites into account, thepositive false discovery rate (q-value) was computed for eachmetabolite. The positive false discovery rate is defined byStorey (2003) as pFDR = E[V/RjR > 0], where ‘‘positive’’means that at least one rejection occurred. The pFDR canbe used to define the q-value, which is a natural pFDR analogto the p-value. Thus, the q-value is a measure for the fractionof false positives appearing even if the test itself was sig-nificant. In contrast to the p-value, the q-value is a multiplehypothesis testing quantity (Storey, 2003). The statisticalanalysis system R (http://www.r-project.org/) was usedfor this analysis.

For a comprehensive view of Type D-effects in this data-driven metabolic network, a Gaussian graphical model (GGM)approach was employed (Krumsiek et al., 2012). GGMs arebased on partial correlation coefficients: that is, pairwisePearson correlation coefficients conditioned against the cor-relation with all other metabolites. GGMs circumvent indirectassociation effects by evaluating conditional dependencies inmultivariate Gaussian distributions. A GGM is an undirectedgraph in which each edge represents the pairwise correlationbetween two variables conditioned against the correlationswith all other variables (also denoted as partial correlationcoefficients).

3. Results

3.1. Demographic, biomedical and psychologicalcharacteristics of Type D in the population

Approximately one quarter of the study participants, wereclassified with a Type D personality (25.7%, n = 386), yet theydid not differ from the rest of the population (non-Type Dn = 1116, 74.3%) with regards to age, gender or biomedicalcharacteristics. Type D individuals were, however, morelikely to be anxious or have a depressed mood comparedto non-Type D participants ( p < .0001 for both, Table 1). Thedistributions of anxiety (mild, moderate or heavy) anddepression (other and major depressive syndrome) were36.7% and 5.3%, respectively.

3.2. Kynurenine associates with Type Dpersonality

We analyzed the metabolic dataset to identify those metabo-lites that most strongly associate with the Type D personality(Table 2). The only metabolite that showed the strongestassociation with Type D and a positive false discovery rate

Table 2 Strongest associations of metabolite levels with Type D personality.

Metabolite Metabolitesub-network

Standarddeviation/mean

p-Value q-Value Relativedifference(b)# in %

Kynurenine+ 0.228 7.08 � 10�5 0.042 �2.1Caffeine+ 3 0.882 3.28 � 10�4 0.072 �9.0X12830+ 0.594 3.69 � 10�4 0.072 �6.5X18601+, * 1 0.620 6.26 � 10�4 0.092 5.2LysoPCaC26:1 0.114 9.70 � 10�4 0.102 0.9Tyrosine+ 2 0.171 1.04 � 10�3 0.102 �1.33-Indoxyl sulfate+ 4 0.393 1.61 � 10�3 0.115 �3.64-Androstene-3beta,

17beta-disulfate+1 0.466 1.70 � 10�3 0.115 3.8

X12244+ 0.301 1.76 � 10�3 0.115 �1.9PCaaC32:2 0.439 2.55 � 10�3 0.149 �3.7Androsterone sulfate+ 1 0.576 2.84 � 10�3 0.151 5.4PCaaC36:3 0.208 3.63 � 10�3 0.172 �1.3X02249+ 0.414 3.82 � 10�3 0.172 �3.2Disulfated steroid+,8 1 0.595 4.87 � 10�3 0.204 4.0C9 0.370 5.33 � 10�3 0.208 �2.5Paraxanthine+ 3 0.685 6.30 � 10�3 0.218 �6.0Gamma-glutamylphenylalanine+ 2 0.230 6.33 � 10�3 0.218 �1.6X10510+ 0.321 7.50 � 10�3 0.225 �2.0Piperine+ 3 0.757 7.84 � 10�3 0.225 �5.7X12405+ 4 0.481 8.51 � 10�3 0.225 �3.4X11485+ 3 0.556 9.27 � 10�3 0.225 �3.9Gamma-glutamyltyrosine+ 2 0.223 9.54 � 10�3 0.225 �1.5C10.2 0.208 9.70 � 10�3 0.225 �1.3PCaaC34:2 0.253 9.96 � 10�3 0.225 �1.4PCaaC38:6 0.280 9.99 � 10�3 0.225 �1.8

Results of the linear regression test (incl. the confounders age, gender, HDL, LDL, cholesterol, triglycerides, hypertension, BMI, diabetes andthe intake of antidepressive medications) with the strongest association ( p-value < 0.01) to Type D. The threshold for a significant q-value is0.05. The metabolites marked with ‘‘+’’ were measured by Metabolon; ‘‘*’’ most probably a monosulfated steroid; ‘‘8’’ either hydro-xypregnen-diol disulfate or pregnanolone-diol disulfate; ‘‘#’’ since the phenotype (Type D/non-Type D) is coded in ‘‘0’’ and ‘‘1’’ forprogramming beta can be used as relative difference.



(q-value) smaller than 0.05 ( p = 7.08 � 10�5) was kynurenine.The serum concentration of this tryptophan metabolite wasdecreased in individuals classified as Type D. It is important tonote that we performed the linear regression analysis foranxiety and depression as well; the results showed no signifi-cant associations between any metabolite concentration witheither anxiety (best p-value: 4.24 � 10�4, q-value: 0.24) ordepression (best p-value: 1.87 � 10�4, q-value: 0.12). Kynur-enine was the 8th best results for the depression variable butdid not appear in the top list of the anxiety analysis.

The NA and SI subscales of the Type D construct were alsotested for associations with metabolite concentrations; theonly significant association observed was between cortisoland NA ( p-value = 6.98 � 10�6, q-value = 4.66 � 10�3,b = �0.0039) as well as SI ( p-value = 1.98 � 10�5, q-value =0.012, b = 0.0036).

3.3. Metabolic sub-networks involvingandrosterone sulfate, tyrosine, caffeine andindoxylsulfate are linked to Type D personality

Using an undirected probabilistic graphical model to esti-mate the conditional dependence between variables, we

attempted to identify biochemical pathways that are linkedto Type D personality. A Gaussian graphical model (GGM) wasgenerated from the Metabolon panel. For this GGM thepartial correlation coefficients for all possible pairs of meta-bolites were calculated. A GGM is a graphical model whereeach edge corresponds to a significant partial correlationbetween the respective metabolites. A partial correlationwas included in the model as an edge if it was significantlydifferent from zero with a cut-off of 0.05 after Bonferronicorrection, yielding a corrected significance level of7.9 � 10�7 and an absolute partial correlation cut-off of0.178. Metabolites that were associated at nominal levels( p-value < 0.01; Fig. 1) were colored and the enriched sub-networks were visually identified. Four such specific sub-networks were identified in the GGM. The first sub-networkcomprised the metabolites androsterone sulfate, 4-andros-tene-3beta,17beta-disulfate, a disulfated steroid that islikely hydroxypregnen-diol disulfate or pregnanolone-dioldisulfate and an unknown metabolite (X18601) thatmost probably represents a monosulfated steroid. All meta-bolites in this sub-network were positively associated withhaving a Type D personality. Our second selected sub-network contained tyrosine, gamma-glutamyltyrosine and

Figure 1 Gaussian graphical model based on partial correlation coefficients using metabolic data from Metabolon. Metabolites thatassociate with Type D personality with a p-value smaller than 0.01 are marked. Red coloring: positive association with Type Dpersonality; blue coloring: negative association; (*) most probably a monosulfated steroid; (8) either hydroxypregnen-diol disulfate orpregnanolone-diol disulfate.



gamma-glutamylphenylalanine, and a third sub-networkincluded caffeine, paraxanthine, piperine and the unknownmetabolite X11485. Two metabolites 3-indoxyl sulfate andX12405 formed a smaller, fourth cluster. These latter threesub-networks were all negatively associated with Type D,suggesting that reduced levels of these metabolites weremore likely associated with Type D participants.

In summary four sub-networks around androsterone sul-fate, tyrosine, caffeine and 3-indoxylsulfate were identified,the first positively associated, the latter three negativelyassociated with Type D personality.

4. Discussion

The present investigation demonstrates associations betweenmetabolite profiles and the Type D personality. Most impor-tantly, the significant reduction of kynurenine in Type D parti-cipants supports the first insight into neuroprotectivefunctions of this metabolite in Type D individuals and suggestspotential therapeutic targets.

4.1. Type D phenotype is associated withdepression and anxiety

While all demographic and biomedical characteristics wereequally distributed among non-Type D and Type D studyparticipants, a significantly higher level of depressed oranxious phenotypes was observed among Type D individuals.This observation is supported by the profile of negativeemotions (anxiety, irritability, etc.) reported for this person-ality type (Denollet and Conraads, 2011).

4.2. Type D phenotype is associated withdecreased serum levels of kynurenine

Significantly lower levels of kynurenine, a tryptophan meta-bolite, were detected in Type D individuals. Despite theoverlapping phenotypes of Type D with depression and anxi-ety, this association appears to be specific for the Type Dphenotype since the analyses performed separately fordepression and anxiety showed no significant associations.The low frequency of participants with other/depressivesyndrome (79 = 5.3%), as well as the overlap of anxiety withdepressive syndrome (75.9%, 60 of 79) may explain the lack ofsignificance in the depressive group.

Kynurenine is the first stable intermediate formed alongthe tryptophan degradation pathway. Fig. 2 presents a gen-eral overview of mental health-related changes among somecomponents of the kynurenine pathway. Changes in serumlevels of neuroactive kynurenine pathway metabolites (bothprotective and toxic) have been implicated in the neuroim-muno-pathogenesis of depression and schizophrenia (Myintet al., 2007). An increase in the N-methyl-D-aspartic acid(NMDA) receptor antagonist kynurenic acid (KA) is reported inschizophrenia (Erhardt et al., 2001; Muller and Schwarz,2010; Wonodi et al., 2011), where as the accumulation ofthe excitotoxin quinolinic acid (QA), an NMDA receptor ago-nist, in depression is suggested by Leonard and Myint (2006),Muller and Schwarz (2008) and Muller et al. (2011) andevidenced by the report from Steiner et al. (2011) of

increased microglial QA in brain sections of severelydepressed suicide victims. Future studies are warranted todetermine the associations between serum concentrations ofkynurenine pathway enzymes and metabolites with Type D,depression and anxiety in the current study population.

Our results add to the growing evidence of the role of thekynurenine pathway in several mental health-related diseasestates and suggest that kynurenine levels may act as thefulcrum to controlled versus impulsive behavior. Such a linkto anxious behavior has been demonstrated in mice (Kanaiet al., 2009; Laugeray et al., 2011), and patients with anxietywere shown to have increased levels of kynurenine whereasdepressed patients had diminished levels compared tohealthy controls (Orlikov et al., 1994). Recently, increasedlevels of kynurenine were found in patients with violentimpulsive behavior and alcohol-related blackouts (Vignauet al., 2010) as well as in major depressive patients whoattempted suicide (Sublette et al., 2011). Both reports sug-gest an association between increased kynurenine and impul-sivity, which offers an intriguing juxtaposition to theobserved negative association between lower serum levelsof kynruenine and Type D individuals.

Furthermore, the results of the subscale componentanalysis with cortisol (positively associated with SI while

Figure 2 A general overview of mental health-related changesamong components of the kynurenine pathway. NA = negativeaffectivity, SI = socially inhibited, tryptophan dioxygenase(TDO), indoleamine 2,3-dioxygenase (IDO), kynurenine amino-transferase (KAT) or kynurenine 3-monooxygenase (KMO).Adapted from McNally et al. (2008) and Dantzer et al. (2008).



negatively associated with NA), offer an insight into thecomplexity of the Type D construct. Increased cortisol (awa-kening response or daily output) was observed in Type D acutecoronary syndrome patients (Whitehead et al., 2007; Molloyet al., 2008). Cortisol is known to activate the rate limitingenzyme tryptophan dioxygenase (TDO) and increase kynur-enine production in the liver (Oxenkrug, 2010).

Associations between individuals with the Type D per-sonality and increased oxidative and inflammatory burden(TNFa) are also becoming increasingly clear (Conraadset al., 2006; Kupper et al., 2009; Mommersteeg et al.,2012). This mounting evidence provides a mechanistic linkto the proposed inflammation-induced increased degrada-tion of tryptophan via indoleamine 2,3-dioxygenase (IDO)that is thought to drive many disease states (Dantzer et al.,2008; Chen and Guillemin, 2009; Schwarcz et al., 2012;Maes et al., 2002). Nonetheless, without knowing thelevels of other kynurenic pathway enzymes (i.e. kynure-nine aminotransferase (KAT) or kynurenine 3-monooxygen-ase (KMO) that drive the production of KA or QA,respectively) the relevance of this hypothesis in Type Dremains speculative.

4.3. Metabolite networks that correlate withType D

The Gaussian graphical model analysis enabled the identifi-cation of four other, albeit based on nominal significancelevels, potentially interesting networks of metabolites,namely those associated with sulphated steroids, tyrosine,indoxyl sulfate, and caffeine. The increased observance of ametabolite (X18601) that is likely a monosulfated steroid,and its neighboring sulfated steroids (androsterone sulfate,4-androstene-3beta,17beta-disulfate and a disulfated ster-oid which is either hydroxypregnen-diol disulfate or pregna-nolone-diol disulfate) may provide an interesting linkbetween altered levels of steroid hormones and Type Dindividuals which has been previously reported in schizo-phrenic patients (Strous et al., 2003; Ritsner et al., 2007).

The decreased levels of the non-essential amino acidtyrosine and its stable metabolite precursors in Type Dindividuals may also be tied to biomechanisms underlyingpsychiatric disorders characterized by negative affectivity. Indopaminergic cells, tyrosine serves as the precursor to theneurotransmitter dopamine as well as the catecholaminesnorepinephrine and epinephrine. Low tyrosine levels may beindicative of low norepinephrine and thus support the cate-cholamine hypothesis of affective disorders. Additionally,low tyrosine may reflect low dopamine levels which areknown to be associated with a lack of attention, motivationand pleasure.

Decreased levels of indoxyl sulfate and an unidentifiedmetabolite X12405 provide insight into yet another interestingmechanistic link between Type D personality, kidney diseaseand heart failure (CHF) that has been observed in the clinicalsetting (Denollet et al., 2009b). Indoxyl sulfate is a uremictoxin that is altered in comorbid patients with both chronickidney disease (CKD) and CHF and in vitro studies have demon-strated direct pro-inflammatory and pro-fibrotic effects ofindoxyl sulfate on cardiac fibroblast and myocytes (Lekawan-vijit et al., 2010). Finally, the network of metabolites related

to caffeine might be linked to an increased activity of CYP1A2,the primary enzyme for caffeine metabolism. CYP1A2 belongsto the superfamily of cytochrome P450 monooxygenases thatcatalyze the metabolism of many different compounds includ-ing steroids and xenobiotics. Steroids, in turn, contribute toregulate the transcription of P450s (Monostory and Dvorak,2011). One may now speculate about a connection betweenincreased levels of disulfated steroids in Type D patients and anincrease of P450s activity.

Overall the finding that Type D was associated with meta-bolite profiles, and that depression or anxiety were not,provides important new information on the value of the TypeD construct as compared to more traditional measures ofpsychological distress. As previously mentioned, studies haveshown that Type D personality was associated with biologicalpathways of disease such as increased levels of oxidativestress (Kupper et al., 2009) and cortisol (Whitehead et al.,2007; Molloy et al., 2008), whereas depression was not. Thisfurther supports the growing evidence for the biologic dis-tinction between ‘stress’-associated, distressed, mentalhealth states (Rief et al., 2001; Denollet et al., 2009a).

4.4. Limitations and strengths

The analysis we present in this paper uses data from a popula-tion study. Individuals are influenced by many different envir-onmental factors like nutrition, medication or physicalactivity. Thus, it is remarkable to find metabolic profilesassociating with Type D personality in a population studywhere, in comparison to clinical short-term studies, the studyparticipants are not subject to specific control criteria. Weshowed that applying a metabolomic approach to populationstudies can provide hypotheses about metabolic mechanismsthat might be connected to mental diseases. Such hypothesescan then be used for a more focused analysis in further studies.

Because of the cross-sectional design of this study, we areunable to draw conclusions about the causal nature of theassociations observed. Our sampling population was com-posed of Caucasians of European ancestry, so the general-izability of these results to other ethnicities requires furtherinvestigation.

Despite these limitations, our study has several strengths,including the large number of study participants, the popula-tion-based design and particularly the detailed and struc-tured information on metabolic variables and behavioral andpsychosocial risk factors of study participants.

5. Conclusions

Our study demonstrates novel biochemical pathways that maybe associated with the Type D construct. This adds to themounting evidence of biologically distinct mechanisms thatunderlie the Type D personality, such as increased oxidativestress and cortisol levels in Type D cardiac patients, even afteradjustment for depressive symptoms (Sher, 2005; Whiteheadet al., 2007; Molloy et al., 2008; Kupper et al., 2009). Addi-tionally, the results demonstrate that applying a metabolomicapproach to population studies can help formulate hypothesesabout metabolic mechanisms that may contribute to affectivedisorders and subsequent diseases.



Role of funding source

The KORA research platform and the KORA Augsburg studiesare financed by the Helmholtz Zentrum Munchen, GermanResearch Center for Environmental Health, which is fundedby the German Federal Ministry of Education, Science,Research and Technology and by the State of Bavaria. Thisstudy was supported in part by a grant from the GermanFederal Ministry of Education and Research (BMBF) to theGerman Center for Diabetes Research (DZD e.V.) and to theproject ‘‘Systems Biology of Metabotypes (SysMBo)’’. K.S. issupported by ‘Biomedical Research Program’ funds at WeillCornell Medical College in Qatar, a program funded by theQatar Foundation. The statements made herein are solely theresponsibility of the authors.

Conflict of interest statement

M.V.M., R.P.M. and A.M.E. are employees of Metabolon Inc.All other authors declare that they have no conflicts ofinterest.

Acknowledgements

The authors are grateful for the commitment and involve-ment of all the study participants and for the work anddedication of the MONICA/KORA Augsburg Study staff. Wethank Julia Scarpa, Tamara Halex and Arsin Sabunchi formetabolomic measurements performed at the HelmholtzZentrum Munchen, Genome Analysis Center, MetabolomicsCore Facility.

References

Altmaier, E., Ramsay, S.L., Graber, A., Mewes, H.W., Weinberger,K.M., Suhre, K., 2008. Bioinformatics analysis of targeted meta-bolomics — uncovering old and new tales of diabetic mice undermedication. Endocrinology 149, 3478—3489.

Chen, Y., Guillemin, G.J., 2009. Kynurenine pathway metabolites inhumans: disease and healthy states. Int. J. Tryp. Res. 2, 1—19.

Conraads, V.M., Denollet, J., De Clerck, L.S., Stevens, W.J., Bridts, C.,Vrints, C.J., 2006. Type D personality is associated with increasedlevels of tumour necrosis factor (TNF)-alpha and TNF-alpha recep-tors in chronic heart failure. Int. J. Cardiol. 113, 34—38.

Coyne, J.C., Jaarsma, T., Luttik, M.L., van Sonderen, E., van Veld-huisen, D.J., Sanderman, R., 2011. Lack of prognostic value oftype D personality for mortality in a large sample of heart failurepatients. Psychosom. Med. 73, 557—562.

Dantzer, R., O’Connor, J.C., Freund, G.G., Johnson, R.W., Kelley,K.W., 2008. From inflammation to sickness and depression: whenthe immune system subjugates the brain. Nat. Rev. Neurosci. 9,46—57.

Denollet, J., Sys, S.U., Stroobant, N., Rombouts, H., Gillebert, T.C.,Brutsaert, D.L., 1996. Personality as independent predictor oflong-term mortality in patients with coronary heart disease.Lancet 347, 417—421.

Denollet, J., 2005. DS14: standard assessment of negative affectivi-ty, social inhibition, and Type D personality. Psychosom. Med. 67,89—97.

Denollet, J., Vrints, C.J., Conraads, V.M., 2008. Comparing Type Dpersonality and older age as correlates of tumor necrosis factor-alpha dysregulation in chronic heart failure. Brain Behav. Immun.22, 736—743.

Denollet, J., de Jonge, P., Kuyper, A., Schene, A.H., van Melle, J.P.,Ormel, J., Honig, A., 2009a. Depression and Type D personalityrepresent different forms of distress in the Myocardial INfarctionand Depression — Intervention Trial (MIND-IT). Psychol. Med. 39,749—756.

Denollet, J., Schiffer, A.A., Kwaijtaal, M., Hooijkaas, H., Hendriks,E.H., Widdershoven, J.W., Kupper, N., 2009b. Usefulness of TypeD personality and kidney dysfunction as predictors of interpatientvariability in inflammatory activation in chronic heart failure.Am. J. Cardiol. 103, 399—404.

Denollet, J., Conraads, V.M., 2011. Type D personality and vulnera-bility to adverse outcomes in heart disease. Cleve. Clin. J. Med.78 (Suppl. 1), S13—S19.

Einvik, G., Dammen, T., Hrubos-Strom, H., Namtvedt, S.K., Randby,A., Kristiansen, H.A., Somers, V.K., Nordhus, I.H., Omland, T.,2011. Prevalence of cardiovascular risk factors and concentrationof C-reactive protein in Type D personality persons withoutcardiovascular disease. Eur. J. Cardiovasc. Prev. Rehabil. 18,504—509.

Erhardt, S., Blennow, K., Nordin, C., Skogh, E., Lindstrom, L.H.,Engberg, G., 2001. Kynurenic acid levels are elevated in thecerebrospinal fluid of patients with schizophrenia. Neurosci. Lett.313, 96—98.

Evans, A.M., DeHaven, C.D., Barrett, T., Mitchell, M., Milgram, E.,2009. Integrated, nontargeted ultrahigh performance liquid chro-matography/electrospray ionization tandem mass spectrometryplatform for the identification and relative quantification of thesmall-molecule complement of biological systems. Anal. Chem.81, 6656—6667.

Gieger, C., Geistlinger, L., Altmaier, E., Hrabe de Angelis, M.,Kronenberg, F., Meitinger, T., Mewes, H.W., Wichmann, H.E.,Weinberger, K.M., Adamski, J., Illig, T., Suhre, K., 2008. Genet-ics meets metabolomics: a genome-wide association studyof metabolite profiles in human serum. PLoS Genet. 4,e1000282.

Grande, G., Romppel, M., Vesper, J.M., Schubmann, R., Glaesmer,H., Herrmann-Lingen, C., 2011. Type D personality and all-causemortality in cardiac patients — data from a German cohort study.Psychosom. Med. 73, 548—556.

Grande, G., Romppel, M., Barth, J., 2012. Association between TypeD personality and prognosis in patients with cardiovascular dis-eases: a systematic review and meta-analysis. Ann. Behav. Med.43, 299—310.

Hausteiner, C., Klupsch, D., Emeny, R., Baumert, J., Ladwig, K.H.,2010. Clustering of negative affectivity and social inhibition in thecommunity: prevalence of type D personality as a cardiovascularrisk marker. Psychosom. Med. 72, 163—171.

Holle, R., Happich, M., Lowel, H., Wichmann, H.E., 2005. KORA — aresearch platform for population based health research. Gesund-heitswesen 67 (Suppl. 1), S19—S25.

Illig, T., Gieger, C., Zhai, G., Romisch-Margl, W., Wang-Sattler, R.,Prehn, C., Altmaier, E., Kastenmuller, G., Kato, B.S., Mewes,H.W., Meitinger, T., de Angelis, M.H., Kronenberg, F., Soranzo, N.,Wichmann, H.E., Spector, T.D., Adamski, J., Suhre, K., 2010. Agenome-wide perspective of genetic variation in human metabo-lism. Nat. Genet. 42, 137—141.

Kanai, M., Funakoshi, H., Takahashi, H., Hayakawa, T., Mizuno, S.,Matsumoto, K., Nakamura, T., 2009. Tryptophan 2,3-dioxygenaseis a key modulator of physiological neurogenesis and anxiety-related behavior in mice. Mol. Brain 2, 8.

Kroenke, K., Spitzer, R.L., Williams, J.B., 2001. The PHQ-9: validityof a brief depression severity measure. J. Gen. Intern. Med. 16,606—613.

Krumsiek, J., Suhre, K., Mitchell, M.W., Mohney, R.P., Milburn, M.V.,Wagele, B., Romisch-Margl, W., Illig, T., Adamski, J., Gieger, C.,Theis, F.J., Kastenmuller, G., 2012. Mining the unknown: a sys-tems approach to metabolite identification. PLoS Genet. 8,e1003005.



Kupper, N., Denollet, J., 2007. Type D personality as a prognosticfactor in heart disease: assessment and mediating mechanisms. J.Pers. Assess. 89, 265—276.

Kupper, N., Gidron, Y., Winter, J., Denollet, J., 2009. Associationbetween type D personality, depression, and oxidative stressin patients with chronic heart failure. Psychosom. Med. 71,973—980.

Laugeray, A., Launay, J.M., Callebert, J., Surget, A., Belzung, C.,Barone, P.R., 2011. Evidence for a key role of the peripheralkynurenine pathway in the modulation of anxiety- and depres-sion-like behaviours in mice: focus on individual differences.Pharmacol. Biochem. Behav. 98, 161—168.

Lawton, K.A., Berger, A., Mitchell, M., Milgram, K.E., Evans, A.M.,Guo, L., Hanson, R.W., Kalhan, S.C., Ryals, J.A., Milburn, M.V.,2008. Analysis of the adult human plasma metabolome. Pharma-cogenomics 9, 383—397.

Lekawanvijit, S., Adrahtas, A., Kelly, D.J., Kompa, A.R., Wang, B.H.,Krum, H., 2010. Does indoxyl sulfate, a uraemic toxin, have directeffects on cardiac fibroblasts and myocytes? Eur. Heart J. 31,1771—1779.

Leonard, B.E., Myint, A., 2006. Inflammation and depression: isthere a causal connection with dementia? Neurotoxicol. Res.10, 149—160.

Lowe, B., Kroenke, K., Herzog, W., Grafe, K., 2004. Measuringdepression outcome with a brief self-report instrument: sensitiv-ity to change of the Patient Health Questionnaire (PHQ-9). J.Affect. Disord. 81, 61—66.

Maes, M., Verkerk, R., Bonaccorso, S., Ombelet, W., Bosmans, E.,Scharpe, S., 2002. Depressive and anxiety symptoms in the earlypuerperium are related to increased degradation of tryptophaninto kynurenine, a phenomenon which is related to immuneactivation. Life Sci. 71, 1837—1848.

McNally, L., Bhagwagar, Z., Hannestad, J., 2008. Inflammation,glutamate, and glia in depression: a literature review. CNS Spectr.13, 501—510.

Mittelstrass, K., Ried, J.S., Yu, Z., Krumsiek, J., Gieger, C., Prehn,C., Roemisch-Margl, W., Polonikov, A., Peters, A., Theis, F.J.,Meitinger, T., Kronenberg, F., Weidinger, S., Wichmann, H.E.,Suhre, K., Wang-Sattler, R., Adamski, J., Illig, T., 2011. Discoveryof sexual dimorphisms in metabolic and genetic biomarkers. PLoSGenet. 7, e1002215.

Molloy, G.J., Perkins-Porras, L., Strike, P.C., Steptoe, A., 2008. Type-D personality and cortisol in survivors of acute coronary syn-drome. Psychosom. Med. 70, 863—868.

Mols, F., Denollet, J., 2010. Type D personality in the generalpopulation: a systematic review of health status, mechanismsof disease, and work-related problems. Health Qual. Life Out-comes 8, 9.

Mommersteeg, P.M., Kupper, N., Denollet, J., 2010. Type D person-ality is associated with increased metabolic syndrome prevalenceand an unhealthy lifestyle in a cross-sectional Dutch communitysample. BMC Public Health 10, 714.

Mommersteeg, P.M., Pelle, A.J., Ramakers, C., Szabo, B.M., Denol-let, J., Kupper, N., 2012. Type D personality and course of healthstatus over 18 months in outpatients with heart failure: Multiplemediating inflammatory biomarkers. Brain Behav. Immun. 26,301—310.

Monostory, K., Dvorak, Z., 2011. Steroid regulation of drug-metabo-lizing cytochromes P450. Curr. Drug Metab. 12, 154—172.

Muhlberger, N., Behrend, C., Stark, R., Holle, R., 2003. Datenbank-gestutzte Online-Erfassung von Arzneimitteln im Rahmengesundheitswissenschaftlicher Studien — Erfahrungen mit derIDOM-software. Inform. Biometrie Epidemiol. Med. Biol. 34,601—611.

Muller, N., Schwarz, M.J., 2008. A psychoneuroimmunological per-spective to Emil Kraepelins dichotomy: schizophrenia and majordepression as inflammatory CNS disorders. Eur. Arch. PsychiatryClin. Neurosci. 258 (Suppl. 2), 97—106.

Muller, N., Schwarz, M.J., 2010. Immune system and schizophrenia.Curr. Immunol. Rev. 6, 213—220.

Muller, N., Myint, A.M., Schwarz, M.J., 2011. Inflammatory biomark-ers and depression. Neurotox. Res. 19, 308—318.

Myint, A.M., Kim, Y.K., Verkerk, R., Scharpe, S., Steinbusch, H.,Leonard, B., 2007. Kynurenine pathway in major depression: evi-dence of impaired neuroprotection. J. Affect. Disord. 98, 143—151.

Nyklicek, I., Vorselaars, A., Denollet, J., 2011. Type D personality andcardiovascular function in daily life of people without documen-ted cardiovascular disease. Int. J. Psychophysiol. 80, 139—142.

Oresic, M., Hyotylainen, T., Herukka, S.-K., Sysi-Aho, M., Mattila, I.,Seppanan-Laakso, T., Julkunen, V., Gopalacharyulu, P.V., Halli-kainen, M., Koikkalainen, J., Kivipelto, M., Helisalmi, S., Lotjo-nen, J., Soininen, H., 2011a. Metabolome in progression toAlzheimer’s disease. Transl. Psychiatry 1, e57.

Oresic, M., Tang, J., Seppanen-Laakso, T., Mattila, I., Saarni, S.E.,Saarni, S.I., Lonnqvist, J., Sysi-Aho, M., Hyotylainen, T., Perala,J., Suvisaari, J., 2011b. Metabolome in schizophrenia and otherpsychotic disorders: a general population-based study. GenomeMed. 3, 19.

Oresic, M., Seppanen-Laakso, T., Sun, D., Tang, J., Therman, S.,Viehman, R., Mustonen, U., van Erp, T.G., Hyotylainen, T.,Thompson, P., Toga, A.W., Huttunen, M.O., Suvisaari, J., Kaprio,J., Lonnqvist, J., Cannon, T.D., 2012. Phospholipids and insulinresistance in psychosis: a lipidomics study of twin pairs discordantfor schizophrenia. Genome Med. 4, 1.

Orlikov, A.B., Prakhye, I.B., Ryzov, I.V., 1994. Kynurenine in bloodplasma and DST in patients with endogenous anxiety and endog-enous depression. Biol. Psychiatry 36, 97—102.

Oxenkrug, G.F., 2010. Tryptophan kynurenine metabolism as a com-mon mediator of genetic and environmental impacts in majordepressive disorder: the serotonin hypothesis revisited 40 yearslater. Isr. J. Psychiatry Relat. Sci. 47, 56—63.

Paige, L.A., Mitchell, M.W., Krishnan, K.R., Kaddurah-Daouk, R.,Steffens, D.C., 2007. A preliminary metabolomic analysis of olderadults with and without depression. Int. J. Geriatr. Psychiatry 22,418—423.

Pelle, A.J., Pedersen, S.S., Schiffer, A.A., Szabo, B., Widdershoven,J.W., Denollet, J., 2010. Psychological distress and mortality insystolic heart failure. Circ. Heart Fail. 3, 261—267.

Quinones, M.P., Kaddurah-Daouk, R., 2009. Metabolomics tools foridentifying biomarkers for neuropsychiatric diseases. Neurobiol.Dis. 35, 165—176.

Rathmann, W., Haastert, B., Icks, A., Lowel, H., Meisinger, C.,Holle, R., Giani, G., 2003. High prevalence of undiagnoseddiabetes mellitus in Southern Germany: target populationsfor efficient screening. The KORA survey 2000. Diabetologia46, 182—189.

Rief, W., Pilger, F., Ihle, D., Bosmans, E., Egyed, B., Maes, M., 2001.Immunological differences between patients with major depres-sion and somatization syndrome. Psychiatry Res. 105, 165—174.

Ritsner, M., Gibel, A., Maayan, R., Ratner, Y., Ram, E., Modai, I.,Weizman, A., 2007. State and trait related predictors of serumcortisol to DHEA(S) molar ratios and hormone concentrations inschizophrenia patients. Eur. Neuropsychopharmacol. 17, 257—264.

Romisch-Margl, W., Prehn, C., Bogumil, R., Rohring, C., Suhre, K.,Adamski, J., 2012. Procedure for tissue sample preparation andmetabolite extraction for high-throughput targeted metabolo-mics. Metabolomics 8, 133—142.

Schwarcz, R., Bruno, J.P., Muchowski, P.J., Wu, H.Q., 2012. Kynur-enines in the mammalian brain: when physiology meets patholo-gy. Nat. Rev. Neurosci. 13, 465—477.

Sher, L., 2005. Type D personality: the heart, stress, and cortisol. QJM98, 323—329.

Spitzer, R.L., Kroenke, K., Williams, J.B., 1999. Validation and utilityof a self-report version of PRIME-MD: the PHQ primary care study.Primary Care Evaluation of Mental Disorders. Patient HealthQuestionnaire. JAMA 282, 1737—1744.



Spitzer, R.L., Kroenke, K., Williams, J.B., Lowe, B., 2006. A briefmeasure for assessing generalized anxiety disorder: the GAD-7.Arch. Intern. Med. 166, 1092—1097.

Sreekumar, A., Poisson, L.M., Rajendiran, T.M., Khan, A.P., Cao, Q.,Yu, J., Laxman, B., Mehra, R., Lonigro, R.J., Li, Y., Nyati, M.K.,Ahsan, A., Kalyana-Sundaram, S., Han, B., Cao, X., Byun, J.,Omenn, G.S., Ghosh, D., Pennathur, S., Alexander, D.C., Berger,A., Shuster, J.R., Wei, J.T., Varambally, S., Beecher, C., Chinnai-yan, A.M., 2009. Metabolomic profiles delineate potential role forsarcosine in prostate cancer progression. Nature 457, 910—914.

Steiner, J., Walter, M., Gos, T., Guillemin, G.J., Bernstein, H.G.,Sarnyai, Z., Mawrin, C., Brisch, R., Bielau, H., Meyer zuSchwabedissen, L., Bogerts, B., Myint, A.M., 2011. Severedepression is associated with increased microglial quinolinicacid in subregions of the anterior cingulate gyrus: evidencefor an immune-modulated glutamatergic neurotransmission?J. Neuroinflamm. 8, 94.

Storey, J.D., 2003. The positive false discovery rate: a Bayesianinterpretation and the q-value. Ann. Statist. 31, 2013—2035.

Strous, R.D., Maayan, R., Lapidus, R., Stryjer, R., Lustig, M., Kotler,M., Weizman, A., 2003. Dehydroepiandrosterone augmentation inthe management of negative, depressive, and anxiety symptomsin schizophrenia. Arch. Gen. Psychiatry 60, 133—141.

Sublette, M.E., Galfalvy, H.C., Fuchs, D., Lapidus, M., Grunebaum,M.F., Oquendo, M.A., Mann, J.J., Postolache, T.T., 2011. Plasmakynurenine levels are elevated in suicide attempters with majordepressive disorder. Brain Behav. Immun. 25, 1272—1278.

Suhre, K., Shin, S.Y., Petersen, A.K., Mohney, R.P., Meredith, D.,Wagele, B., Altmaier, E., Deloukas, P., Erdmann, J., Grundberg,E., Hammond, C.J., de Angelis, M.H., Kastenmuller, G., Kottgen,A., Kronenberg, F., Mangino, M., Meisinger, C., Meitinger, T.,

Mewes, H.W., Milburn, M.V., Prehn, C., Raffler, J., Ried, J.S.,Romisch-Margl, W., Samani, N.J., Small, K.S., Wichmann, H.E.,Zhai, G., Illig, T., Spector, T.D., Adamski, J., Soranzo, N., Gieger,C., Assimes, T.L., Deloukas, P., Erdmann, J., Holm, H., Kathir-esan, S., Konig, I.R., McPherson, R., Reilly, M.P., Roberts, R.,Samani, N.J., Schunkert, H., Stewart, A.F., 2011. Human meta-bolic individuality in biomedical and pharmaceutical research.Nature 477, 54—60.

Versteeg, H., Spek, V.R.M., Pedersen, S.S., Denollet, J., 2012. Type Dpersonality and health status in cardiovascular disease popula-tions: a meta-analysis of prospective studies. Eur. J. Prev. Car-diol., http://dx.doi.org/10.1177/1741826711425338.

Vignau, J., Soichot, M., Imbenotte, M., Jacquemont, M.C., Danel, T.,Vandamme, M., Lhermitte, M., Allorge, D., 2010. Impact oftryptophan metabolism on the vulnerability to alcohol-relatedblackouts and violent impulsive behaviours. Alcohol Alcohol. 45,79—88.

Whitehead, D.L., Perkins-Porras, L., Strike, P.C., Magid, K., Steptoe,A., 2007. Cortisol awakening response is elevated in acute coro-nary syndrome patients with type-D personality. J. Psychosom.Res. 62, 419—425.

Wonodi, I., Stine, O.C., Sathyasaikumar, K.V., Roberts, R.C., Mitch-ell, B.D., Hong, L.E., Kajii, Y., Thaker, G.K., Schwarcz, R., 2011.Downregulated kynurenine 3-monooxygenase gene expressionand enzyme activity in schizophrenia and genetic associationwith schizophrenia endophenotypes. Arch. Gen. Psychiatry 68,665—674.

Xuan, J., Pan, G., Qiu, Y., Yang, L., Su, M., Liu, Y., Chen, J., Feng, G.,Fang, Y., Jia, W., Xing, Q., He, L., 2011. Metabolomic profiling toidentify potential serum biomarkers for schizophrenia and risper-idone action. J. Proteome Res. 10, 5433—5443.


Metabolomic profiles in individuals with negative affectivity and social inhibition: A...

Documents

Transcript of Metabolomic profiles in individuals with negative affectivity and social inhibition: A...