kynurenine pathways abnormalities in a large cohort of aggressive inmates: markers for aggression

    Tryptophan via serotonin/kynurenine pathways abnormalities in a large cohort of aggressive inmates: Markers for aggression Stefano Comai, Antonella Bertazzo, Jeanne Vachon, Marc Daigle, Jean Toupin, Gilles Cˆot´e, Gustavo Turecki, Gabriella Gobbi PII: DOI: Reference:

S0278-5846(16)30058-6 doi: 10.1016/j.pnpbp.2016.04.012 PNP 8917

To appear in:

Progress in Neuropsychopharmacology & Biological Psychiatry

Received date: Revised date: Accepted date:

28 January 2016 20 April 2016 22 April 2016

Please cite this article as: Comai Stefano, Bertazzo Antonella, Vachon Jeanne, Daigle Marc, Toupin Jean, Cˆ ot´e Gilles, Turecki Gustavo, Gobbi Gabriella, Tryptophan via serotonin/kynurenine pathways abnormalities in a large cohort of aggressive inmates: Markers for aggression, Progress in Neuropsychopharmacology & Biological Psychiatry (2016), doi: 10.1016/j.pnpbp.2016.04.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

PT

Tryptophan via serotonin/kynurenine pathways abnormalities in a large cohort of aggressive inmates: markers for aggression

SC

Gilles Côtéd,e, Gustavo Tureckig, Gabriella Gobbia,*.

RI

Stefano Comaia,b, Antonella Bertazzoc, Jeanne Vachond, Marc Daigled,e, Jean Toupind,f,

a

NU

Neurobiological Psychiatry Unit, Department of Psychiatry, McGill University Health

Center, McGill University, Montreal, QC, Canada; bDivision of Neuroscience c

Department of

MA

San Raffaele Scientific Institute and Vita-Salute University, Milan, Italy;

Pharmaceutical and Pharmacological Sciences, University of Padova, Padova, Italy; d

Institut Philippe-Pinel, Montreal , Qc, Canada; eUniversité du Québec à Trois-Rivières,

D

Trois-Rivières, QC, Canada; fDepartment of Psychoeducation, Sherbrooke University,

TE

QC, Canada; gMcGill Group for Suicide Studies, Douglas Mental Health University

AC CE P

Institute, McGill University, Montreal, Canada. *Corresponding Author: Gabriella Gobbi, MD, PhD Neurobiological Psychiatry Unit Dept. of Psychiatry McGill University 1033 Pine Avenue West Montreal, Q.C. Canada H3A 1A1 Email: [email protected] Phone: 514-398-1290; Fax: 514-398-4866

Abstract: 250 words Text: 4767 Number of tables: 4 Number of figures: 4

1

ACCEPTED MANUSCRIPT Abstract Aggressive behaviour is one of the most challenging symptoms in psychiatry, and

PT

biological markers for aggression lack of large sample validations. Serotonin (5-HT) and

RI

other neuroactive compounds deriving from Tryptophan (Trp), including kynurenine (Kyn), have not yet been investigated in large cohorts of aggressive individuals to

SC

validate their potential as biomarkers of aggression.

NU

In 361 male inmates we measured serum levels of Trp, 5-hydroxytryptophan, 5-HT, Kyn, the ratios 5-HT/Trp*1000 and Kyn/Trp*1000, and performed Structured Clinical

MA

Interview for DSM-IV Axis-I and -II Disorders (SCID-I and -II), global assessment of functioning (GAF), and scales for aggressive behaviour, impulsivity, adult attention-

TE

D

deficit/hyperactivity disorder (ADHD), and intelligent quotient (IQ). Aggressive compared to non-aggressive inmates exhibited lower Trp and Kyn serum

AC CE P

levels but higher levels of 5-HT and 5-HT/Trp*1000, higher levels of impulsivity and ADHD indices, lower IQ and GAF, higher prevalence of mood disorders, drug abuse/dependence, and borderline, conduct and antisocial behaviours. Interestingly, Kyn/Trp*1000 was positively correlated to the number of severe aggressive acts (r=0.593, P<0.001). After adjusting for confounding factors, logistic regression analysis indicated that 5-HT/Trp*1000, antisocial behaviour, and GAF were predictors of aggressive behaviour. The model combining these three predictors had an area under the ROC curve of 0.851 (95% CI 0.806-0.895). This study indicates that while circulating Trp is reduced in aggressive individuals, the combination of biological (5-HT/Trp ratio) and psychopathological (antisocial behaviour and GAF) markers discriminates between aggressive and non-aggressive behaviour

2

ACCEPTED MANUSCRIPT suggesting the potential of a multi-marker approach in psychiatry given the heterogenic

PT

nature of mental diseases.

RI

Keywords: aggressive behaviour; tryptophan; serotonin; kynurenine; biomarkers.

SC

List of abbreviations:

NU

Trp, Tryptophan; 5-HTP, 5-hydroxytryptophan; 5-HT, serotonin; Kyn, kynurenine; BBB, blood-brain barrier; TDO, tryptophan 2,3-dioxygenase; IDO, indoleamine 2,3-

MA

dioxygenase; MacCVI, MacArthur Community Violence Instrument; LSARS, Lethality of Suicide Attempt Rating Scale; SCID-I, Structured Clinical Interview for DSM-IV Axis-I

TE

D

Disorders; SCID-II, Structured Clinical Interview for DSM-IV Axis-II Disorders; GAF, Global Assessment of Functioning; ASPD, antisocial personality disorder; BIS, Barratt

quotient.

AC CE P

Impulsiveness Scale; CAARS, Conners’ Adult ADHD Rating Scale; IQ, Intelligence

3

ACCEPTED MANUSCRIPT 1. Introduction According to the World Health Organization, injuries and violence are a threat to health

PT

in every country of the world (Krug et al., 2002). Aggressive behavior is therefore a

RI

major concern in social and criminal justice settings and in mental health (Cornaggia et al., 2011). Until now, while the psychopathological risk factors for aggression have been

SC

largely investigated (Comai et al., 2012a, Siever, 2008), the biological factors are still

NU

matter of debate. Most importantly, the relationship between neurobiological and psychosocial factors at the basis of aggressive behaviour is not yet elucidated by basic

MA

and clinical research (Comai et al., 2012a) due to a lack of large cohort studies and interdisciplinary approaches. The essential amino acid tryptophan (Trp) is not only the

TE

D

precursor of serotonin (5-HT) but is also degraded to other neuroactive compounds, including the neurotoxic quinolinic acid and the neuroprotective kynurenic acid, along

AC CE P

the enzymatic cascade known as kynurenine (Kyn) pathway (Schwarcz et al., 2012)(Fig. 1). The first step of this pathway is regulated by the tryptophan 2,3dioxygenase (TDO) and indoleamine 2,3-dioxygenase (IDO) enzymes (Schwarcz et al., 2012). The ratio Kyn/Trp is currently used as a good estimate of IDO enzymatic activity (Comai et al., 2011, Myint and Kim, 2014, Schwarcz et al., 2012). While 5-HT is not able to cross the blood-brain barrier (BBB) and therefore total 5-HT serum levels could not be used as indices of central 5-HT activity, its precursor Trp crosses the BBB via the large neutral amino acid transporter (Fernstrom and Wurtman, 1971). However, this issue needs to be clarified since it has been demonstrated that in case of augmented brain 5-HT levels, 5-HT can cross the BBB from the brain to the circulating blood through the 5-HT transporter (Nakatani et al., 2008). In any case, peripheral levels of 5-

4

ACCEPTED MANUSCRIPT HT although not linked to brain 5-HT can be interesting markers of disease and/or response to treatment (Comai et al., 2010, Fontana et al., 2008, Freedman et al., 1981,

PT

Irwin et al., 1981). Research has shown that total peripheral Trp concentrations

RI

accurately reflect the rate of influx of Trp and thus 5-HT levels into the brain (Comai et al., 2011, Fernstrom and Wurtman, 1971), and that fluctuations in the blood levels of

SC

both Trp and Kyn, which also crosses the BBB, directly affect the metabolism of the Kyn

NU

pathway in the brain (Fukui et al., 1991). Dysregulation of both 5-HT and Kyn pathways, resulting in hyper- or hypo-function of active metabolites, has been associated to

MA

several psychiatric and neurodegenerative disorders (Brown et al., 1982, Comai et al., 2012a, Hanley et al., 1977, Schwarcz et al., 2012). 5-HT is the most studied

TE

D

neurotransmitter related to the pathophysiology of aggression (Brown et al., 1982, Brown et al., 1979, Coccaro, 1992, Linnoila and Virkkunen, 1992) and one of the targets

AC CE P

of anti-aggressive medications (Comai et al., 2012b). In general, the dogmatic view of a low 5-HT activity linked to aggression (Brown et al., 1979, Coccaro, 1992, Linnoila and Virkkunen, 1992) appears true only for certain type of aggressive behaviour such as impulsive aggression, defined as a disproportionate reaction to any provocation, real or perceived (Coccaro, 1992). Indeed, functional aggression, a natural form of social behavior aimed at the establishment of a territory, social dominance and defense of resources, seems positively associated to 5-HT function (de Boer and Koolhaas, 2005). Similarly, the question of whether the precursor of 5-HT, Trp, is altered in aggressive or suicidal individuals is still controversial. Several studies have indicated that a Trp depletion may lead to increased aggression (Cleare and Bond, 1995, LeMarquand et al., 1999, LeMarquand et al., 1998), but other research examining the plasmatic levels

5

ACCEPTED MANUSCRIPT of Trp found elevated levels of the amino acid associated with the presence of antisocial (Tiihonen et al., 2001, Virkkunen and Narvanen, 1987) and conduct (Virkkunen et al.,

PT

2003) violent behaviours.

RI

Therefore, the complex neurobiology of aggression, 5-HT and Trp needs clarifications

SC

and especially lacks of large cohort human studies. The role of Kyn in mood disorders and neurodegenerative diseases has been already investigated (Schwarcz et al., 2012),

NU

but no studies have yet examined its possible involvement in aggression. Here, we

MA

studied for the first time both Trp metabolic pathways in aggression, in particular severe aggression meant as the intent to seriously harm or kill others or themselves, because

D

by sharing the same precursor, impairments in the 5-HT pathway may be the cause or

TE

the consequence of impairments in the Kyn pathway and viceversa. In this study, we thus filled this gap in knowledge by showing in a large cohort of severe aggressive and

AC CE P

non-aggressive inmates that peripheral levels of Trp and its metabolites via 5-HT and Kyn were different comparing severe aggressive versus non-aggressive individuals and most importantly, that 1) they correlated to psychopathological features previously associated to aggression and 2) combined with psychopathological features they became markers of aggressive behaviour.

6

ACCEPTED MANUSCRIPT 2. Materials and methods 2.1.

Participants

PT

Participants were recruited in a federal penitentiary in Montreal, Quebec, Canada,

RI

between October 2007 and November 2011. Venous blood samples were withdrawn

SC

between 8:00 and 10:00 a.m. after an overnight fasting, allowed to clot at room temperature and centrifuged at 3000 g for 10 min. The obtained serum was stored at -

NU

80°C. The study was approved by the Ethical Review Boards of the Institut PhilippePinel de Montréal, McGill University and Université du Québec à Trois-Rivières, and by

MA

Correctional Service Canada, and followed the principles of the Helsinki Declaration. All

Biological assessment

TE

2.2.

D

participants, after adequate information, gave their written consent.

AC CE P

2.2.1. Serum levels of Trp, 5-hydroxytryptophan (5-HTP), 5-HT and Kyn Compounds were quantified following the method of Comai et al. (2011) Trp, 5-HTP and 5-HT were analyzed using a high performance liquid chromatography (HPLC) system equipped with a Shimadzu RF-10 AXL fluorometric detector set at excitation and emission wavelengths of 285 and 345 nm, respectively, and an analytical Platinum EPS-C18 100A column (5m; 250mm×4.6mm; Alltech, Deerfield, IL). Kyn was determined using a HPLC equipped with a Shimadzu SPD-10A UV–Vis detector set at 360 nm and an analytical Grace Smart RP-18 column (5m; 250mm×4.6mm; Alltech). The chromatographic separation was conducted using an isocratic gradient of acetonitrile-phosphate buffer 0.004 M, pH 3.5.

7

ACCEPTED MANUSCRIPT 2.3.

Psychiatric assessment

2.3.1. Severe aggression against others (MacCVI)

PT

Violence and aggression were evaluated using the MacArthur Community Violence

RI

Instrument (MacCVI) (Monahan et al., 2001). This self-reported instrument reports on

SC

lifetime prevalence of aggressive behavior at two severity levels. In this study, we focused our attention on severely aggressive behaviour such as that causing injury with

NU

an object, a knife or a gun, wounding someone seriously, or causing death. This

MA

instrument has been previously validated by contacting collateral informants including prison, arrest and police records (Steadman et al., 1998).

TE

D

2.3.2. Severe aggression against self (LSARS) The rate and severity of self-injurious behaviors were estimated using the Lethality of

AC CE P

Suicide Attempt Rating Scale (LSARS) (Smith et al., 1984). When a participant mentioned that there was a suicide attempt in his past or any kind of harm inflicted upon himself, the questionnaire on suicide and suicide-related behaviors was completed regardless of what the participant may have indicated in terms of intention of the act. Hence, any kind of injury to the self was scored using LSARS. Lifetime prevalence of suicide attempts was considered.

2.3.3. AXIS-I disorders DSM-IV-TR AXIS-I disorders were evaluated using the Structured Clinical Interview for DSM-IV Axis-I Disorders (SCID-I)(First et al., 1996) administered by professional raters who showed very good inter-rater agreement (K values from 0.918 to 1.000). We 8

ACCEPTED MANUSCRIPT assessed life-time presence of 1) mood disorders (major depressive disorder, depressive disorder not otherwise specified, dysthymia, bipolar disorder-I or -II, mood

PT

disorder due to drugs of abuse); 2) anxiety disorders (panic disorder, generalized

posttraumatic

stress

disorder,

RI

anxiety disorder, agoraphobia without panic disorder, specific phobia, social phobia, obsessive-compulsive

disorder);

3)

psychosis

SC

(schizophrenia, schizophreniform disorder, schizoaffective disorder, brief psychotic

NU

disorder, psychotic disorders due to drugs of abuse); 4) pathological gambling; 5) substance abuse/dependence to alcohol, cannabis, stimulants, opioids, cocaine, and

MA

hallucinogens.

D

2.3.4. AXIS-II disorders

TE

Cluster B personality disorders (antisocial, conduct, borderline, histrionic and narcissistic) were evaluated using the Structured Clinical Interview for DSM-IV Axis-II

AC CE P

Disorders (SCID-II)(First and Gibbon, 1997). Inter-rater agreement for Axis-II Disorders was also very high (K values from 0.868 to 0.932). 2.3.5. Global Assessment of Functioning (GAF)-AXIS V GAF was evaluated by professional raters following the Axis-V of the DSM-IV-TR found in the SCID-I. The scale includes 10 sets of anchor descriptions spaced at 10-point intervals. 2.3.6. Impulsivity The Barratt Impulsiveness Scale version-11 (BIS-11) was used to measure impulsivity (Barratt, 1985). The three a priori defined components of impulsiveness were examined: cognitive, motor, and non-planning impulsiveness. 9

ACCEPTED MANUSCRIPT 2.3.7. Attention-deficit status Conners’ Adult ADHD Rating Scale (CAARS) was used to assess adult ADHD-related

PT

symptoms and behaviors (Conners et al., 1999). Here, we used the CAARS Self-

the

following

subscales:

A)

Inattention/Memory

Problems,

B)

SC

assess

RI

Report: Short Version (Conners, 2004) that is comprised of 26 items and allowed us to

Hyperactivity/Restlessness, C) Impulsivity/Emotional Lability, D) Problems with Self-

NU

Concept, E) ADHD Index.

MA

2.3.8. Intelligence quotient (IQ)

IQ was estimated using the Wechsler adult intelligence scale (WAIS-III)(Wechsler,

Statistical analysis

AC CE P

2.4.

TE

D

1939).

Statistical analysis was conducted using SPSS 20 (Chicago, IL). Continuous variables were reported as mean ± SEM. Comparisons in the prevalence of Axis-I and -II disorders between aggressive and non-aggressive inmates were calculated using Pearson's chi-square test. Differences between the two groups for continuous variables were analyzed using multivariate analysis of covariance (MANCOVA) with Pillai's criterion as the significance test, age as covariate, and the Bonferroni post-hoc correction for multiple comparisons. The relationship between biological variables and the number of severe aggressive acts was determined by Spearman’s correlations controlling for age. Predicting factors of aggression were examined by logistic regression analysis (enter-method). The multicollinearity (variance inflation factor (VIF)) was calculated prior entering the predictors into the logistic regression models. To avoid 10

ACCEPTED MANUSCRIPT multicollinearity, variables with VIF exceeding 10 should be used with caution (O’brien, 2007). In these models, sever multicollinearity corresponding to VIF > 10 was not

PT

detected. Statistical significance was accepted when p<0.05. The overall predictive

RI

ability of the final model was assessed using the area under the ROC curve. Fig. 2 summarizes the statistical methodology employed in this study. Among 361 inmates, 23

AC CE P

TE

D

MA

NU

SC

had missing data for 1 or more variables.

11

ACCEPTED MANUSCRIPT 3. RESULTS Between October 2007 and November 2011 we approached 731 male inmates. Of

PT

these inmates, 579 initially accepted the invitation, and among them 361 agreed to

RI

undergo both a psychiatric assessment and a blood withdrawal. The major health

SC

problems found in the 361 inmates were: allergies (11%), arthritis (4%), cancer (1%), chronic pain (2%), diabetes (6%), epilepsy (0.9%), hepatitis B or C positive (14%), HIV

NU

positive (0.5%), hypercholesterolemia (3%), hypertension (15%), traumatic brain injury

and three morbid obese (BMI>40).

MA

in the past (3%), urinary diseases (5%). Moreover, four inmates were obese (BMI>30)

28% of the inmates were receiving psychoactive drugs as monotherapy or as

TE

D

combination including selective serotonin re-uptake inhibitors (SSRI, 5%), serotonin– norepinephrine reuptake inhibitors (SNRI, 5%), trazodone (2%), mirtazapine (5%),

AC CE P

tricyclic antidepressants (4%), bupropion (0.3%), methylphenidate (0.5%), gabapentin (3%), carbamazepine (0.8%), valproic acid (0.5%), and atypical antipsychotics (8%). 258 inmates out of 361 showed aggressive behavior against others and/or themselves according to MacCVI and LSARS instruments. Importantly, 65.1% of them displayed a severe aggressive behaviour according to MacCVI, and 39.1% had at least one suicide attempt during their life. Aggressive inmates were older than non-aggressive inmates (mean ± S.D.: 47.8±14.1 vs. 37.7±11.6 years, p<0.001). Therefore, the potential confounding effect of age was considered in all the analyses.

3.1.

Clinical characteristics of aggressive and non-aggressive inmates

12

ACCEPTED MANUSCRIPT First, we compared the clinical characteristics of aggressive vs. non-aggressive inmates using Pearson's chi-square test (Fig. 2, Step 1 left). As noted previously (Appelbaum,

PT

2013, Fazel and Seewald, 2012), we confirmed that aggressive inmates display

RI

significantly higher rate of mood disorders, psychosis, and alcohol, cannabis, sedative, stimulants, opioids, cocaine and hallucinogens abuse/dependence than non-aggressive

SC

inmates (Table 1). Moreover, a tendency to a higher prevalence of anxiety disorders

NU

(p=0.058) was found in the former group. The two groups did not differ regarding the presence of pathological gambling.

MA

We also found that aggressive inmates compared to non-aggressive inmates had higher rate of borderline, conduct and antisocial personality (ASPD) disorders as previously

3.2.

AC CE P

disorders (Table 1).

TE

D

reported (Fazel and Seewald, 2012), but not of histrionic and narcissistic personality

Trp, 5-HTP, 5-HT and Kyn levels and behavioral variables in aggressive and non-aggressive inmates

Second, we compared biological and behavioural variables between aggressive and non-aggressive inmates. To control for the probability that a Type 1 error could occur due to the large number of variables, a MANCOVA analysis correcting for the factor age was conducted, with all biological and behavioral measures (see Table 2) entered as dependent variables and group membership (aggression and no-aggression) entered as factor (Fig. 2, Step 1 right). An overall significant difference between aggressive and non-aggressive inmates was found (Pillai's trace=0.155; F18,317=3.23, P<0.001). Subsequent univariate analyses of covariance (Table 2) indicated that Trp (P=0.021)

13

ACCEPTED MANUSCRIPT and Kyn (P=0.015) serum levels were lower in aggressive than in non-aggressive inmates, whereas 5-HT serum levels were higher (P=0.047) in the former group. 5-

PT

HT/Trp*1000 ratio was higher in the aggressive group (P=0.006) and no differences

RI

were detected for 5-HTP and 5-HTP/Trp*1000 and Kyn/Trp*1000 ratios. Aggressive inmates also displayed higher levels of impulsivity (BIS scores, P<0.001), and of all

SC

items of attention-deficit/hyperactivity (CAARS scores, P<0.05), but lower IQ (P=0.041)

NU

and GAF (P<0.001) compared to non-aggressive inmates (Table 2). Interestingly, in the sub-group of inmates who displayed a severe aggressive behaviour against others

MA

according to MacCVI (n=157), the number of severe aggressive acts was positively correlated to the ratio Kyn/Trp*1000 (r=0.332, P<0.001) after controlling for age (Fig.

TE

D

3A). This relationship was stronger (r=0.593, P<0.001) when in this same sub-group, we considered only those inmates who had at least one lifetime suicide attempt according

AC CE P

to LSARS (n=51) (Fig. 3B). Due to the potential confounding effects of the health problems of the inmates on the metabolism of Trp, in particular obesity, hypertension, and infections with HIV and/or hepatitis B and C virus, and of the psychopharmacotherapy which might act directly or indirectly on the 5-HT neurotransmission, we excluded these inmates and examined the difference in the serum levels of Trp and its metabolites via 5-HT and Kyn between aggressive and non-aggressive inmates. Even excluding these confounders, we reached the same results on both biological and behavioral variables between aggressive and non-aggressive inmates (data not shown). For this reason, we decided to do not exclude any inmate in the following analyses.

3.3.

Analysis of markers of aggression

14

ACCEPTED MANUSCRIPT Given the multifactorial pathophysiology of aggressive behavior, we then assessed which psychiatric, behavioral and/or biological factors predicted severe aggression.

PT

Because of the high number of factors under consideration and that differed between

RI

aggressive and non-aggressive inmates, we could not enter all the variables in the model predicting aggressive behavior. Therefore, we decided to follow a stepwise

SC

strategy to reduce the number of variables by initially separating psychiatric from

NU

behavioral/biological factors, and then by combining in the final model only those factors significantly associated to aggression. Since this is the first study analyzing together a

MA

high number of possible correlates of aggression, factors were excluded from the subsequent step only for analytical/statistical reasons and not because of a priori

TE

D

theoretical association and/or known link between factors. Details of the steps undertaken in the analysis are the following (Fig. 2): 1) only psychiatric variables in step

AC CE P

1 having a significant difference between aggressive and non-aggressive inmates were included in step 2 (Fig. 2, Step 2 left); 2) only biological/behavioral variables that were significant in step 1 were included in step 2 (Fig. 2, Step 2 right); 3) a final model combining significant psychiatric and biological/behavioral predictors of aggressive inmates in step 2 was then examined in step 3 (Fig. 2, Step 3).

3.3.1. AXIS-I and -II Cluster B disorders predicting aggressive behavior Logistic regression analysis revealed that aggressive inmates were 2 times more likely to have a mood disorder and 3 times more likely to have a conduct or ASPD (Table 3) in accordance with previous literature (Fazel and Seewald, 2012). Mean VIF of the variables entered in the model was 1.272. Please see Table S1 in supplementary

15

ACCEPTED MANUSCRIPT information for all significant and non-significant predictors. ASPD, although by definition is highly correlated to the presence of aggressive behavior (American

PT

Psychiatric Association, 2000), has been included in the logistic analysis because 37%

RI

of non-aggressive individuals displayed ASPD and 20 % of aggressive individuals did not display ASPD. It was therefore important in our opinion to evaluate in our study

SC

population the specific impact of ASPD on aggressive behavior among the different

NU

psychiatric disorders.

MA

3.3.2. Biological markers and behavioral variables predicting aggressive behavior

TE

D

Among biological and behavioural variables (mean VIF=4.587), 5-HT levels, 5HT/Trp*1000 ratio, CAARS-C Impulsivity/Emotional Lability and GAF levels were

AC CE P

significant predictors of aggressive behaviour (Table 3). Please see Table S2 in supplementary information for all significant and non-significant predictors.

3.3.3. Combination of 5-HT/Trp*1000, ASPD and GAF represents a biomarker of aggressive behaviour Finally, we combined psychiatric, behavioural and biological variables to establish the best predictors of aggressive behaviour (Table 4)(Fig. 2, Step 3). The logistic regression model revealed that ASPD, 5-HT/Trp*1000 and GAF-Axis V were good predictors of aggressive inmates. Mean VIF of the predictors entered into this final model was 2.736. To further validate these findings, we conducted a ROC curve analysis to measure the goodness-of-fit for the final model including these three significant predictors of

16

ACCEPTED MANUSCRIPT aggressive behaviour (Fig. 2, Step 4). The area under the ROC curve was 0.851 with a 95% confidence interval of 0.806–0.895, thus indicating that only combination of 5-

PT

HT/Trp*1000, ASPD, and GAF generates a good predictive power of aggressive

RI

behaviour (Fig. 4A). As a negative control, we have calculated the ROC for the biological index 5-HT/Trp*1000 alone, and indeed it had no predicting validity for the

SC

presence of aggressive behavior (Fig. 4B). We then calculated the precise 5-

NU

HT/Trp*1000 threshold discriminating between aggressive and non-aggressive inmates and we found that a 5-HT/Trp*1000 ratio greater than 30 paired with the presence of

MA

ASPD and low GAF predicted aggressive behaviour with a very good power (AUC =

AC CE P

TE

D

84.7; 95% Confidence interval: 0.801-0.893; Fig. 4C).

17

ACCEPTED MANUSCRIPT 4. DISCUSSION In this study, we examined in an inmate population whether peripheral levels of 5-HT,

PT

Kyn and their precursor Trp are altered in aggressive individuals, and whether they are

RI

markers of aggression together with/without psychopathological features. Results

SC

showed that regardless of age, aggressive individuals displayed lower serum levels of Trp and Kyn and higher 5-HT serum levels. Notably, logistic regression analysis

NU

followed by a ROC curve indicated that the integration of 1) high serum 5-HT/Trp ratio (>30), 2) presence of ASPD, and 3) low GAF-Axis V score is a marker of aggressive

MA

behaviour.

Previous research on aggression and 5-HT (Brown et al., 1982, Brown et al., 1979,

TE

D

Coccaro, 1992, Linnoila and Virkkunen, 1992) attempted to evaluate the levels of its metabolite 5-hydroxyindoleacetic acid (5-HIAA) in the cerebrospinal fluid (CSF). These

AC CE P

studies found reduced 5-HIAA CSF levels in suicidal (Linnoila and Virkkunen, 1992) and impulsive aggression (Coccaro, 1992) thus suggesting that a reduction in 5-HT central function may enhance impulsivity and disinhibition, thereby increasing the likelihood of engaging in aggressive acts.

Trp depletion studies, another method to examine central 5-HT function (Hood et al., 2005), showed that Trp depletion increases violent and impulsive behaviours (Cleare and Bond, 1995, LeMarquand et al., 1999, LeMarquand et al., 1998). Given the strict link between peripheral and central Trp levels (Comai et al., 2011, Fernstrom and Wurtman, 1971), the lower Trp levels we found in aggressive inmates suggest that its availability for the synthesis of 5-HT in the brain is very likely reduced in aggressive individuals, thus supporting a reduced central 5-HT activity associated to aggressive

18

ACCEPTED MANUSCRIPT behaviour. Tryptophan supplementation which increases peripheral Trp levels has indeed demonstrated significant effects in reducing aggressive behaviour in patients

PT

with schizophrenia (Morand et al., 1983) and in the need for antipsychotics and

RI

sedatives to control agitated or violent behavior in psychiatric inpatients (Volavka et al., 1990). Some studies, although conducted with small sample sizes or even case-reports,

SC

reported a somehow contrasting finding: violent individuals display increased plasma

NU

levels of Trp (Tiihonen et al., 2001, Virkkunen et al., 2003, Virkkunen and Narvanen, 1987). These contradictions might be explained by methodological limitations, and

MA

similarly to the 5-HT and aggression story, by different types of aggressive behaviour under consideration.

TE

D

Despite these controversies, the high serum 5-HT/Trp ratio, which provides indirect evidence on the peripheral functioning of the 5-HT metabolic pathway and that was

AC CE P

found as one of the significant predictors of aggressive behaviour, confirms a main role of the Trp to 5-HT pathway in aggression. The decreased levels of total peripheral Trp also caused a reduction of its availability for the Kyn pathway; indeed, aggressive inmates displayed lower serum levels of Kyn. The ratio Kyn/Trp is used as an indirect measure of TDO/IDO enzyme activities (Comai et al., 2011). Although this ratio did not vary between aggressive and non-aggressive inmates, we found that in the former group, the number of severe aggressive acts was positively correlated to the Kyn/Trp ratio, indicating that a greater degradation of Trp to Kyn occurs in more severe aggressive inmates. Similarly, Bradley et al. (2015) recently reported a significant positive correlation between Kyn/Trp and suicidality in adolescents with major depressive disorder. Remarkably, we found that the correlation between the

19

ACCEPTED MANUSCRIPT number of aggressive acts and the Kyn\Trp ration was stronger when we considered only those inmates who had at least one suicide attempt. This novel finding may be

PT

linked to the recent report indicating that a perinatal immune challenge in rodents leads

RI

to neurodevelopmental dysfunction and a cascade of other events including activation of the Kyn pathway (Kubesova et al., 2015), and the fact that aggressive and suicidal

SC

behaviours might be neurodevelopmental in origin (Fountoulakis et al., 2008, Lesch et

NU

al., 2012, Turecki et al., 2012). Many studies have shown activation of the Kyn pathway leading to an altered ratio between the neurotoxic quinolinic acid and neuroprotective acid

in

several

neurodegenerative

MA

kynurenic

diseases, mood

disorders and

schizophrenia (Chen and Guillemin, 2009, Schwarcz et al., 2012). Interestingly, we

TE

D

found higher rate of mood disorders in aggressive inmates. Therefore, further studies are needed to evaluate 1) Kyn and its downstream metabolites in aggression and

AC CE P

suicide, 2) if the Kyn pathway activation is linked to aggression directly or because of mood disorders or neurodevelopmental dysfunctions that are associated to aggression, and 3) if altered Kyn pathway is more strictly associated to self-harm\suicide rather than to aggression against others.

Similar to Moffitt et al. (1998) who previously reported elevated whole blood serotonin in violent men but not in violent females, we found higher 5-HT serum levels in aggressive inmates. Elevated circulating levels of 5-HT are also displayed by patients suffering of chronic schizophrenia (Freedman et al., 1981), ADHD (Irwin et al., 1981), retardation of various etiologies and autism (Hanley et al., 1977). In addition, hyperserotonemia was roughly correlated with low IQ values in children (Hanley et al., 1977). Interestingly,

20

ACCEPTED MANUSCRIPT aggressive inmates showed not only hyperserotonemia, but also lower IQ and higher adult ADHD indices.

PT

In a subgroup of our study population composed by 86 inmates with ASPD, we recently

RI

found MAOA gene promoter hypermethylation that caused a downregulation of MAOA activity; importantly, MAOA gene promoter hypermethylation was positively associated

SC

to serum 5-HT levels (Checknita et al., 2014). It is therefore very likely that the

NU

increased peripheral 5-HT levels we observed in aggressive inmates who also showed high prevalence of ASPD, are due to a downregulation of MAO-A activity.

MA

Although we examined several biological/psychopathological correlates of aggression, the large number of individuals in this study allowed us to evaluate possible markers of

TE

D

aggression. A high 5-HT/Trp ratio along with the presence of ASPD and low GAF resulted a good marker of aggression. While it was not surprising the association

AC CE P

between ASPD and aggression (aggressiveness is a core symptoms of ASPD (American Psychiatric Association, 2000)), for the first time a link between the overall functioning and aggression has been demonstrated. In this study, we further confirmed the high levels of impulsivity present in aggressive individuals (Comai et al., 2012a, Seroczynski et al., 1999), but we found that adult ADHD symptoms levels are also associated to aggression. In particular, the symptoms of adult ADHD associated to aggression are related to impulsivity and to the ability to control emotions, rather than to difficulty in concentrating, planning or completing tasks, and hyperactivity. In keeping with our data, Gordon et al. (2014) recently reported that criminals in the symptomatic and ‘at-risk’ groups for adult ADHD had a trend to higher

21

ACCEPTED MANUSCRIPT number of total, violent, but also non-violent criminal convictions compared to the nonsymptomatic and ‘no-risk’ groups.

PT

One of the main limitations of our study is that these findings may not generalize to a

RI

normal population and to aggression in females. A confirmatory study in non-inmate populations and in females is thus required. Nonetheless, our combined biomarker of

SC

aggression needs validation in an independent sample of aggressive individuals. Other

NU

limitations are that we did not analyze whether the reduced Kyn in aggressive inmates yielded to an imbalance in downstream neuroactive metabolites of the Kyn pathway,

MA

and thus we did not provide their functional involvement in aggression, and that collectively we did not specifically examined the impact of systemic diseases or

TE

D

pharmacotherapies known for their possible effects on the metabolism of Trp via 5-HT and/or Kyn. Future research on these specific topics may address these unanswered

AC CE P

questions. However, biological differences between aggressive and non-aggressive inmates were confirmed even if inmates with these potential confounding factors were excluded from the sample. Remarkably, logistic regression analysis conducted in the group of inmates without obesity, hepatitis C or B, hypertension and any psychotropic medication, confirmed 5-HT/Trp ratio, GAF and ASPD as the three significant predictors of aggressive behaviour, and the AUC (0.821) of the model was still good (data non shown). Two important novel findings related to the neurobiology of aggression have been here reported. First, biochemical measures related to the metabolism of Trp via 5-HT and Kyn are significantly altered in aggressive individuals as indicated by an increased

22

ACCEPTED MANUSCRIPT serum 5-HT/Trp ratio in aggressive inmates and the correlation between the Kyn/Trp ratio and the number of severe aggressive acts.

PT

Second, given the heterogeneous construct of aggression, only a combination of

RI

biological and psychopathological markers, namely high 5-HT/Trp ratio in the serum, presence of ASPD, and low GAF score, may better help predicting aggressive

SC

behaviour. Indeed, a single biological marker such as the ratio 5-HT/Trp has basically

NU

no power to distinguish between aggressive and non-aggressive individuals (Fig. 4B). Similarly, a single psychopathological marker such as the presence of ASPD, although

MA

highly associated to aggression, cannot be used as single predictor of aggressive behavior. Indeed, as indicated in our study and others (Fazel and Danesh, 2002, Petras

TE

D

et al., 2008), not all individuals with ASPD display aggressive behavior and not all aggressive acts are committed in the presence of ASPD (Table 1). Importantly, we

AC CE P

found a peripheral marker linked to 5-HT that is associated to aggression. Of note, previous 5-HT markers of aggression were determined in the cerebrospinal fluid (Brown et al., 1982, Brown et al., 1979, Coccaro, 1992, Linnoila and Virkkunen, 1992) that is collected through a lumbar puncture, an invasive technique that cannot be used routinely as compared to a blood withdrawal. A similar approach consisting of a combination of biomarkers has already proven its higher predicting validity over a single biomarker in other branches of medicine (Castelli, 1984, Imbert-Bismut et al., 2001). Remarkably, in keeping with our findings, the integration of the top serum biomarkers and the clinical information has recently gained attention also in psychiatry for its very high predicting capability of suicidal ideation across different psychiatric diagnoses (Niculescu et al., 2015). This approach may thus become a future milestone also in

23

ACCEPTED MANUSCRIPT psychiatry given the heterogenic nature of mental diseases. Up to date medications used to treat aggressive patients are still prescribed as ‘‘off-label use” (Comai et al.,

PT

2012a). According to our data, the conversion of Trp into 5-HT and Kyn may be a novel

RI

potential target for therapeutic intervention in psychiatric disorders associated to aggression. In particular, inhibitors of IDO may reduce Trp to Kyn degradation occurring

SC

in most aggressive individuals, thus increasing peripheral Trp levels and consequently

MA

NU

the availability of the amino acid for the synthesis of 5-HT in the brain.

ACKNOWLEDGMENTS

D

This study was supported by Canadian Institutes of Health Research (MOP 83133, to

TE

Drs. Cote, Daigle, Toupin, Gobbi, Turecki), MUHC postdoctoral Fellowship (to Dr. Comai), Fonds de recherche du Québec en Santé (salary awards to Dr Gobbi and Dr

AC CE P

Turecki) and the Quebec Network on Suicide, Mood Disorders and Related Disorders (to Dr Turecki and Dr Gobbi).

Conflict of interest

The authors report no potential conflicts of interest.

24

ACCEPTED MANUSCRIPT REFERENCES American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders

PT

Fourth Edition Text Revision (DSM-IV-TR). Washington, DC2000.

RI

Appelbaum PS. Public safety, mental disorders, and guns. JAMA Psychiatry 2013;70:565-566.

NU

emotion and personality. 1985:137-146.

SC

Barratt ES. Impulsiveness subtraits: arousal and information processing. Motivation,

Bradley KA, Case JA, Khan O, Ricart T, Hanna A, Alonso CM, Gabbay V. The role of

MA

the kynurenine pathway in suicidality in adolescent major depressive disorder. Psychiatry Res. 2015;227:206-212.

TE

D

Brown GL, Ebert MH, Goyer PF, Jimerson DC, Klein WJ, Bunney WE, Goodwin FK. Aggression, suicide, and serotonin: relationships to CSF amine metabolites. Am. J.

AC CE P

Psychiatry 1982;139:741-746.

Brown GL, Goodwin FK, Ballenger JC, Goyer PF, Major LF. Aggression in humans correlates with cerebrospinal fluid amine metabolites. Psychiatry Res. 1979;1:131139.

Castelli WP. Epidemiology of coronary heart disease: the Framingham study. Am. J. Med. 1984;76:4-12. Checknita D, Maussion G, Labonte B, Comai S, Tremblay RE, Vitaro F, Turecki N, Bertazzo A, Gobbi G, Côté G, Turecki G. Monoamine oxidase A gene promoter methylation and transcriptional downregulation in an offender population with antisocial personality disorder. Br. J. Psychiatry 2015; 206:216-222.

25

ACCEPTED MANUSCRIPT Chen Y, Guillemin GJ. Kynurenine pathway metabolites in humans: disease and healthy States. Int. J. Tryptophan Res. 2009;2:1-19.

PT

Cleare AJ, Bond AJ. The effect of tryptophan depletion and enhancement on subjective

RI

and behavioural aggression in normal male subjects. Psychopharmacology 1995;118:72-81.

SC

Coccaro EF. Impulsive aggression and central serotonergic system function in humans:

NU

an example of a dimensional brain-behavior relationship. Int. Clin. Psychopharmacol. 1992.

MA

Comai S, Bertazzo A, Carretti N, Podfigurna-Stopa A, Luisi S, Costa CV. Serum levels of tryptophan, 5-hydroxytryptophan and serotonin in patients affected with different

TE

D

forms of amenorrhea. Int. J. Tryptophan Res. 2010;3:69-75. Comai S, Cavalletto L, Chemello L, Bernardinello E, Ragazzi E, Costa CV, Bertazzo A.

AC CE P

Effects of PEG-interferon alpha plus ribavirin on tryptophan metabolism in patients with chronic hepatitis C. Pharmacol. Res. 2011;63:85-92. Comai S, Tau M, Gobbi G. The psychopharmacology of aggressive behavior: a translational approach: part 1: neurobiology. J. Clin. Psychopharmacol. 2012a;32:83-94.

Comai S, Tau M, Pavlovic Z, Gobbi G. The psychopharmacology of aggressive behavior: a translational approach: part 2: clinical studies using atypical antipsychotics, anticonvulsants, and lithium. J. Clin. Psychopharmacol. 2012b;32:237-260.

26

ACCEPTED MANUSCRIPT Conners C, Erhardt D, Epstein J, Parker J, Sitarenios G, Sparrow E. Self-ratings of ADHD symptoms in adults I: Factor structure and normative data. J Atten Disord.

PT

1999;3:141-151.

settings. . Toronto: Multi-Health Systems; 2004.

RI

Conners K. Conners' Adult ADHD Rating Scales (CAARS): For use in correctional

SC

Cornaggia CM, Beghi M, Pavone F, Barale F. Aggression in psychiatry wards: a

NU

systematic review. Psychiatry Res. 2011;189:10-20.

de Boer SF, Koolhaas JM. 5-HT 1A and 5-HT 1B receptor agonists and aggression: a

Pharmacol. 2005;526:125-139.

MA

pharmacological challenge of the serotonin deficiency hypothesis. Eur. J.

TE

D

Fazel S, Danesh J. Serious mental disorder in 23000 prisoners: a systematic review of 62 surveys. Lancet. 2002;359:545-550.

AC CE P

Fazel S, Seewald K. Severe mental illness in 33,588 prisoners worldwide: systematic review and meta-regression analysis. Br. J. Psychiatry 2012;200:364-373. Fernstrom JD, Wurtman R. Brain serotonin content: physiological dependence on plasma tryptophan levels. Science. 1971;173:149-152. First MB, Gibbon M. User's guide for the structured clinical interview for DSM-IV axis II personality disorders: SCID-II: Amer Psychiatric Pub Incorporated; 1997. First MB, Gibbon M, Spitzer R, Williams J. User’s Guide for the Structured Clinical Interview for DSM-IV Axis I Disorders—Research Version. SCID-I, version. 1996;2. Fontana RJ, Kronfol Z, Lindsay KL, Bieliauskas LA, Padmanabhan L, Back-Madruga C, Lok AS, Stoddard AM; HALT-C Trial Group. Changes in mood states and

27

ACCEPTED MANUSCRIPT biomarkers during peginterferon and ribavirin treatment of chronic hepatitis C. Am. J. Gastroenterol. 2008;103:2766-2775.

PT

Fountoulakis KN, Leucht S, Kaprinis GS. Personality disorders and violence. Curr. Opin.

RI

Psychiatry 2008;21:84-92.

Freedman DX, Belendiuk K, Belendiuk GW, Crayton JW. Blood tryptophan metabolism

SC

in chronic schizophrenics. Arch. Gen. Psychiatry 1981;38:655-659.

NU

Fukui S, Schwarcz R, Rapoport SI, Takada Y, Smith QR. Blood-brain barrier transport of kynurenines: implications for brain synthesis and metabolism. J. Neurochem.

MA

1991;56:2007-2017.

Gordon V, Donnelly PD, Williams DJ. Relationship between ADHD symptoms and anti-

Dif. 2014;58:116-121.

TE

D

social behaviour in a sample of older youths in adult Scottish prisons. Pers. Individ.

AC CE P

Hanley HG, Stahl SM, Freedman DX. Hyperserotonemia and amine metabolites in autistic and retarded children. Arch. Gen. Psychiatry 1977;34:521. Hood SD, Bell CJ, Nutt DJ. Acute tryptophan depletion. Part I: rationale and methodology. Aust. N. Z. J. Psychiatry 2005;39:558-564. Imbert-Bismut F, Ratziu V, Pieroni L, Charlotte F, Benhamou Y, Poynard T. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. Lancet 2001;357:1069-1075. Irwin M, Belendiuk K, McCloskey K, Freedman DX. Tryptophan metabolism in children with attentional deficit disorder. Am. J. Psychiatry 1981;138:1082-1085. Krug EG, Mercy JA, Dahlberg LL, Zwi AB. The world report on violence and health. Lancet 2002;360:1083-1088.

28

ACCEPTED MANUSCRIPT Kubesova A, Tejkalova H, Syslova K, Kacer P, Vondrousova J, Tyls F, et al. Biochemical, histopathological and morphological profiling of a rat model of early

PT

immune stimulation: relation to psychopathology. PLoS One 2015;10:e0115439.

RI

LeMarquand DG, Benkelfat C, Pihl RO, Palmour RM, Young SN. Behavioral disinhibition induced by tryptophan depletion in nonalcoholic young men with

SC

multigenerational family histories of paternal alcoholism. Am. J. Psychiatry

NU

1999;156:1771-1779.

LeMarquand DG, Pihl RO, Young SN, Tremblay RE, Séguin JR, Palmour RM, Benkelfat

MA

C. Tryptophan depletion, executive functions, and disinhibition in aggressive, adolescent males. Neuropsychopharmacology 1998;19:333-341.

TE

D

Lesch K-P, Araragi N, Waider J, van den Hove D, Gutknecht L. Targeting brain serotonin synthesis: insights into neurodevelopmental disorders with long-term

AC CE P

outcomes related to negative emotionality, aggression and antisocial behaviour. Philos. Trans. R Soc. Lond. B Biol. Sci. 2012;367:2426-2443. Linnoila VM, Virkkunen M. Aggression, suicidality, and serotonin. J. Clin. Psychiatry 1992;53 Suppl:46-51.

Moffitt TE, Brammer GL, Caspi A, Fawcett JP, Raleigh M, Yuwiler A, Silva P. Whole blood serotonin relates to violence in an epidemiological study. Biol. Psychiatry 1998;43:446-457. Monahan J, Steadman HJ, Silver E, Appelbaum PS, Robbins PC, Mulvey EP, Roth, LH, Grisso T, Banks, S. Rethinking risk assessment: The MacArthur study of mental disorder and violence: Oxford University Press, USA; 2001.

29

ACCEPTED MANUSCRIPT Morand C, Young SN, Ervin FR. Clinical response of aggressive schizophrenics to oral tryptophan. Biol. Psychiatry 1983.

PT

Myint AM, Kim YK. Network beyond IDO in psychiatric disorders: revisiting

RI

neurodegeneration hypothesis. Prog. Neuropsychopharmacol. Biol. Psychiatry 2014;48:304-313.

SC

Nakatani Y, Sato-Suzuki I, Tsujino N, Nakasato A, Seki Y, Fumoto M, Arita H.

NU

Augmented brain 5-HT crosses the blood-brain barrier through the 5-HT transporter in rat. Eur. J. Neurosci. 2008;27:2466-2472.

MA

Niculescu AB, Levey DF, Phalen PL, Le-Niculescu H, Dainton HD, Jain N, et al. Understanding and predicting suicidality using a combined genomic and clinical

D

risk assessment approach. Mol. Psychiatry 2015;20:1266-1285.

TE

O’brien RM. A caution regarding rules of thumb for variance inflation factors. Quality &

AC CE P

Quantity. 2007;41:673-690.

Petras H, Kellam SG, Brown CH, Muthén BO, Ialongo NS, Poduska JM. Developmental epidemiological courses leading to antisocial personality disorder and violent and criminal behavior: Effects by young adulthood of a universal preventive intervention in first-and second-grade classrooms. Drug Alcohol Depend. 2008;95:S45-S59. Schwarcz R, Bruno JP, Muchowski PJ, Wu HQ. Kynurenines in the mammalian brain: when physiology meets pathology. Nat. Rev. Neurosci. 2012;13:465-477. Seroczynski AD, Bergeman CS, Coccaro EF. Etiology of the impulsivity/aggression relationship: genes or environment? Psychiatry Res. 1999;86:41-57.

30

ACCEPTED MANUSCRIPT Siever LJ. Neurobiology of aggression and violence. Am. J. Psychiatry. 2008;165:429442.

PT

Smith K, Conroy RW, Ehler BD. Lethality of suicide attempt rating scale. Suicide Life

RI

Threat. Behav. 1984;14:215-242.

Steadman HJ, Mulvey EP, Monahan J, Robbins PC, Appelbaum PS, Grisso T, Roth LH,

SC

Silver E. Violence by people discharged from acute psychiatric inpatient facilities

NU

and by others in the same neighborhoods. Arch. Gen. Psychiatry 1998;55:393401.

MA

Tiihonen J, Virkkunen M, Räsänen P, Pennanen S, Sainio EL, Callaway J, Halonen P, Liesivuori J. Free L-tryptophan plasma levels in antisocial violent offenders.

TE

D

Psychopharmacology 2001;157:395-400. Turecki G, Ernst C, Jollant F, Labonté B, Mechawar N. The neurodevelopmental origins

AC CE P

of suicidal behavior. Trends Neurosci. 2012;35:14-23. Virkkunen M1, Ebeling H, Moilanen I, Tani P, Pennanen S, Liesivuori J, Tiihonen J. Total plasma l-tryptophan, free l-tryptophan and competing amino acid levels in a homicidal male adolescent with conduct disorder. Acta Psychiatr. Scand. 2003;108:244-246.

Virkkunen M, Narvanen S. Plasma insulin, tryptophan and serotonin levels during the glucose tolerance test among habitually violent and impulsive offenders. Neuropsychobiology 1987;17:19-23. Volavka J, Crowner M, Brizer D, Convit A, Van Praag H, Suckow RF. Tryptophan treatment of aggressive psychiatric inpatients. Biol. Psychiatry 1990;28:728-732. Wechsler D. The measurement of adult intelligence. 1939.

31

ACCEPTED MANUSCRIPT Figure and Table Legends

Fig. 2. Flow Diagram of Statistical Analysis Procedure.

PT

Fig. 1. Schematic Diagram of the tryptophan metabolism via serotonin and kynurenine.

RI

Fig. 3. Correlation between the number of severe aggressive acts and the ratio

SC

KYN/Trp*1000. (A) Correlation between the log transformed number of severe

NU

aggressive acts and the ratio KYN/Trp*1000 in inmates showing severe aggression (n=157). (B) Correlation between the log transformed number of severe aggressive acts

suicide attempt in their life (n=51).

MA

and the ratio KYN/Trp*1000 in inmates showing severe aggression and at least one

TE

D

Fig. 4. (A) ROC curve of the model including 5-HT/Trp*1000 ratio, ASPD and GAF-Axis V between non-aggressive behavior and aggressive behavior. (B) ROC curve of 5-

AC CE P

HT/Trp*1000 ratio between non-aggressive behavior and aggressive behavior. (C) ROC curve of the model including 5-HT/Trp*1000 ratio >30, ASPD and GAF-Axis V between non-aggressive behavior and aggressive behavior.

Table 1. Clinical characteristics of aggressive and non-aggressive inmates.

Table 2. Tryptophan via serotonin and kynurenine markers and behavioral variables scores in aggressive and non-aggressive inmates. Data are reported as mean ± SEM. Multiple analysis of covariance (MANCOVA) corrected for age plus Bonferroni post-hoc correction for multiple comparisons.

32

ACCEPTED MANUSCRIPT Table 3. Predicting factors for aggressive behaviour among AXIS-I and -II disorders and among tryptophan via serotonin and kynurenine markers and behavioral variables.

PT

Please see Table S1 (AXIS-I and -II disorders) and Table S2 (biological and behavioural

RI

markers) in supplementary information for all significant and non-significant predictors

NU

SC

for the logistic regression coefficients for all the tested variables.

AC CE P

TE

D

MA

Table 4. Psychopathological and biological variables predictors of aggressive behavior.

33

ACCEPTED MANUSCRIPT Table 1. Clinical characteristics of aggressive and non-aggressive inmates. Pearson Chi-Square test

No. of Inmates

p

145 (56) 100 (39) 19 (7) 28 (11) 171 (66) 151(59) 30 (11) 80 (31) 39 (15) 131 (51) 77 (30)

17.19 3.60 3.95 0.35 16.96 20.66 10.65 14.52 4.58 19.52 13.11

<0.001 0.058 0.047 0.550 <0.001 <0.001 0.001 <0.001 0.032 <0.001 <0.001

25 (10) 2 (1) 17 (7) 168 (65) 205 (80)

8.41 0.031 0.077 49.93 61.61

0.004 0.861 0.782 <0.001 <0.001

SC

RI

PT

χ2 (df=1)

NU

TE

Cocaine abuse/dependence Hallucinogens AXISabuse/dependence II disorders, No. (%)

33 (32) 29 (28) 2 (2) 9 (9) 44 (42) 33 (32) 1 (1) 12 (12) 7 (7) 26 (25) 12 (12)

MA

Stimulants abuse/dependence Opioids abuse/dependence

Aggressive (258)

D

AXIS I disorders, No. (%) Mood disorders Anxiety disorders Psychosis Pathological gambling Alcohol abuse/dependence Cannabis abuse/dependence Sedative abuse/dependence

Nonaggressive (103)

AC CE P

Borderline personality 1 (1) disorder Histrionic personality 1 (1) disorder Narcissistic personality 6 (6) disorder Conduct disorder 25 (24) Antisocial personality 38 (37) disorder

34

ACCEPTED MANUSCRIPT Table 2. Tryptophan via serotonin and kynurenine markers and behavioral variables scores in aggressive and non-aggressive inmates.

TE

AC CE P

MANCOVA F

p

11.70±0.16 58.05±1.48 173.69±9.55 445.90±9.24 39.01±0.92 5.21±0.16

5.38 0.23 3.97 6.03 0.12 1.61

0.021 0.62 0.047 0.015 0.72 0.21

16.06±0.97 66.76±0.68 23.59±0.27 16.92±0.25 26.16±0.28 4.17±0.19

7.70 12.92 5.56 11.58 10.59 9.59

0.006 <0.001 0.019 0.001 0.001 0.002

4.74±0.36

6.81±0.23

5.36

0.021

1.99±0.21

3.81±0.19

17.46

<0.001

2.88±0.32

4.60±0.26

12.91

<0.001

7.77±0.58 87.85±1.48 8.26±0.09

11.39±0.40 85.36±0.65 7.64±0.07

12.80 4.21 28.88

<0.001 0.041 <0.001

NU

SC

RI

PT

Aggressive (n = 258)

MA

11.59±0.93 59.21±1.14 21.26±0.46 14.32±0.39 23.81±0.45 2.95±0.23

D

Tryptophan (μg/mL) 5-Hydroxytryptophan (ng/mL) Serotonin (ng/mL) Kynurenine (ng/mL) Kynurenine/tryptophan*1000 5Hydroxytryptophan/tryptophan*1000 Serotonin/tryptophan*1000 BIS-11 Total score BIS-11 Motor impulsiveness BIS-11 Cognitive impulsiveness BIS-11 Non planning impulsiveness CAARS-A Inattention/Memory Problems CAARS-B Hyperactivity/Restlessness CAARS-C Impulsivity/Emotional Lability CAARS-D Problems with SelfConcept CAARS-E ADHD Index IQ GAF - Axis V

Inmates Nonaggressive (n = 103) 11.87±0.23 64.18±3.03 135.05±10.56 488.75±14.42 41.81±1.48 5.50±0.29

35

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

Data are reported as mean ± SEM. Multiple analysis of covariance (MANCOVA) corrected for age plus Bonferroni post-hoc correction for multiple comparisons.

36

ACCEPTED MANUSCRIPT Table 3. Predicting factors for aggressive behaviour among AXIS-I and -II disorders and among tryptophan via serotonin and kynurenine markers and behavioral variables.

8.873 5.019

0.003 0.025

0.342

PT

0.296 0.337

2.413 2.127

1.352 1.099

4.309 4.117

0.002

2.953

1.509

5.777

4.658 6.032

0.000 1.177

0.000 1.033

0.321 1.340

1.222

1.038

1.438

0.536

0.385

0.746

9.997

0.031 0.014

0.200

0.083

5.780

0.016

-0.623

0.169

13.679 <0.001

5.736 0.066

MA

-12.379 0.163

D

Biological and behavioural variables Serotonin Serotonin/tryptophan*1000 CAARS-C Impulsivity/Emotional Lability GAF- Axis V

P value OR OR 95% C.I. Exp(B) Lower Upper

NU

AXIS I and II disorders Mood disorders 0.881 Conduct disorder 0.755 Antisocial personality 1.083 disorder

Wald

RI

Coefficient S.E. B

SC

Predictors

AC CE P

TE

Please see Table S1 (AXIS-I and -II disorders) and Table S2 (biological and behavioural markers) in supplementary information for all significant and non-significant predictors for the logistic regression coefficients for all the tested variables.

37

ACCEPTED MANUSCRIPT Table 4. Psychopathological and biological variables predictors of aggressive behavior.

Mood disorders

Coefficient S.E. B 0.597 0.321

Wald 3.460

P value OR Exp(B) 0.063 1.816

OR 95% C.I. Lower Upper 0.968 3.407

0.134

PT

Predictors

0.353

2.249

1.698

0.850

3.391

0.343

18.920 <0.001 4.455

2.272

8.734

4.089

5.778

0.000

0.000

0.163

Serotonin/tryptophan*1000 0.133 GAF- Axis V -0.554 CAARS-C Impulsivity/Emotional 0.081 Lability

0.046 0.179

8.490 9.542

0.004 0.002

1.143 0.575

1.045 0.404

1.250 0.817

0.211

1.085

0.955

1.232

38

0.106

SC

NU

AC CE P

TE

D

MA

0.065

RI

Conduct disorder 0.529 Antisocial personality 1.494 disorder Serotonin -9.829

1.567

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

39

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

40

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

41

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

42

ACCEPTED MANUSCRIPT Highlights 

Altered metabolism of tryptophan via serotonin and kynurenine in aggressive

PT

inmates; Aggressive inmates display increased serum 5-HT/Trp ratio;



The serum Kyn/Trp ratio is correlated to the number of severe aggressive acts;



Only a combination of biological and psychopathological markers predicts

SC

RI



AC CE P

TE

D

MA

NU

aggression.

43