Moderation of the Relationship between Toxoplasma gondii Seropositivity and Trait Impulsivity in Younger Men by the Phenylalanine-Tyrosine Ratio

Abstract

Previously, we reported that Toxoplasma gondii (T. gondii)-seropositivity is associated with higher impulsive sensation seeking in younger men. As dopaminergic and serotonergic signaling regulates impulsivity, and as T. gondii directly and indirectly affects dopaminergic signaling and induces activation of the kynurenine pathway leading to diversion of tryptophan from serotonin production, we investigated if dopamine and serotonin precursors or the tryptophan metabolite kynurenine interact with the T. gondii – impulsivity association. In 951 psychiatrically healthy participants, trait impulsivity scores were related to T. gondii IgG seropositivity. Interactions were also identified between categorized levels of phenylalanine (Phe), tyrosine (Tyr), Phe:Tyr ratio, kynurenine (Kyn), tryptophan (Trp) and Kyn:Trp ratio and age and gender. Only younger T. gondii-positive men with a high Phe:Tyr ratio were found to have significantly higher impulsivity scores. There were no significant associations in other demographic groups including older men and younger or older women. No significant effects or interactions were identified for Phe, Tyr, Kyn, Trp or Kyn:Trp ratio. Phe:Tyr ratio, therefore, may play a moderating role in the association between T. gondii seropositivity and impulsivity in younger men. These results could potentially lead to individualized approaches to reduce impulsivity based on combined demographic, biochemical and serological factors.

1. Introduction

The relationship between mental illness and inflammation has garnered increasing attention (Haroon et al., 2012; Najjar et al., 2013). Across serious mental illness including schizophrenia (Cotter et al., 2001; Potvin et al., 2008; Elsheikha and Zhu, 2016; Abdollahian et al., 2017; Fuglewicz et al., 2017), bipolar disorder (Cotter et al., 2001; Brietzke et al., 2011; Del Grande et al., 2017) and major depressive disorder (Cotter et al., 2001; Miller et al., 2009; Miller and Raison, 2016), immune changes, such as higher levels of proinflammatory cytokines, are commonly encountered biological signatures. Furthermore, signs of central inflammation such as microglial activation and elevated proinflammatory cytokine levels in the cerebrospinal fluid are linked to suicidal behavior (Janelidze et al., 2011; Bay-Richter et al., 2015; Brundin et al., 2015; Lund-Sorensen et al., 2016) and psychosis- or mood-related episode (McNally et al., 2008; Brietzke et al., 2011; Miller et al., 2011), and are linked to treatment responses for these conditions as well (McNally et al., 2008; Kubera et al., 2000; Miller et al., 2011; de Witte et al., 2014). In both human (Suarez, 2003; Suarez et al., 2004; Coccaro, 2006; Graham et al., 2006; Coccaro et al., 2014) and animal studies (Zalcman and Siegel, 2006; Bhatt et al., 2008; Patel et al., 2010; Pesce et al., 2011), behaviors commonly seen among those with mental illness, such as trait impulsivity (Moeller et al., 2001; Swann et al., 2009) and aggression (Oquendo et al., 2004; Soyka, 2011; Volavka and Citrome, 2011; Ballester et al., 2012), are also linked to higher levels of inflammation.

Toxoplasma gondii (T. gondii) (Dupont et al., 2012), similar with Herpesvidae family viruses (Nicholas, 2005; Jochum et al., 2012), is neurotropic pathogen that may result in low-grade immune activation due to increased rates of latent infection. Contrary to earlier beliefs that the sole presence of bradyzoites (i.e., slower growing forms) and tissue cysts during latent T. gondii was innocuous, research has shown a link between chronic toxoplasmosis and prominent psychiatric illnesses (Hinze-Selch et al., 2010; Tedla et al., 2011; Pearce et al., 2012; Torrey et al., 2012; Fond et al., 2013; Sutterland et al., 2015;) and suicidal behavior across diagnostic modalities (Arling et al., 2009; Yagmur et al., 2010; Ling et al., 2011; Okusaga et al., 2011; Pedersen et al., 2012; Zhang et al., 2012; Postolache and Cook, 2013; Ansari-Lari et al., 2017). As we have previously reported, trait impulsivity, considered an intermediate endophenotype for suicidal behavior (Turecki, 2005; Kovacsics et al., 2009; Mann et al., 2009), was positively associated with T. gondii seropositivity in younger (i.e., 20–59 years old) psychiatrically healthy men (Cook et al., 2015). Similarly, in psychiatric patients, T. gondii seropositivity has been associated with trait impulsivity and aggression, and with an Intermittent Explosive Disorder diagnosis (Coccaro et al., 2016). By contrast, a population-representative birth cohort study reported that suicide attempts were only marginally more frequent among individuals with T. gondii seropositivity (p = .06) (Sugden et al., 2016). The study finds no evidence of T. gondii relating to increased risk of psychiatric disorders or neurocognitive impairment either. Additionally, while most of the prior literature on psychiatric illnesses and T. gondii infection used T. gondii IgG antibodies for titer and seropositivity analysis, more recent literature has identified T. gondii IgM, but not IgG, antibody elevations in relationship with suicide attempts (Dickerson et al., 2017). A positive association between T. gondii IgG seropositivity and death by suicide vs. death due to other causes was identified postmortem, but only in the 38–58 age group (Samojlowicz et al., 2013). Of relevance, T.gondii seropositivity was higher in decedents with a positive blood alcohol test, both in those who died by suicide and those who died by other causes (Samojlowicz et al., 2013). Yet another study in Mexico found no association between T. gondii seropositivity and suicidal behavior but did confirm an association between the titers of IgG antibody and suicide attempts in psychiatric outpatients (Alvarado-Esquivel et al., 2013). Similar association of higher titers of T. gondii antibody and suicide attempt were reported too (Arling et al., 2009; Yagmur et al., 2010; Ling et al., 2011).

Dopaminergic neurotransmission is also involved in neurobiological aspects of impulsivity (van Gaalen et al., 2006; Dalley et al., 2007; Voon et al., 2010; Dalley and Roiser, 2012; Fernando et al., 2012; Costa et al., 2013; Djamshidian et al., 2013; Malloy-Diniz et al., 2013; Nandam et al., 2013; Simon et al., 2013; Mitchell and Potenza, 2014; Hamilton et al., 2015). Yet, there are conflicting findings regarding the direction of the association between impulsivity and dopaminergic function, with both direct associations reported by some (Cole and Robbins, 1989; Pezze et al., 2009; Eagle et al., 2011) and inverse associations presented by others (Oswald et al., 2007; Fernando et al., 2012; Kayser et al., 2012; Matuskey et al., 2013; Nandam et al., 2013; Chester et al., 2016). These heterogeneity may be a result of the complexity and specificity of dopamine receptor subtypes, and neural circuits involved in the dopamine neurotransmission (Oswald et al., 2007; Fernando et al., 2012; Kayser et al., 2012; Matuskey et al., 2013; Nandam et al., 2013; Chester et al., 2016). Additionally, impulsivity has been associated with genetic polymorphisms in enzymes associated with dopamine metabolic pathways, dopamine receptors and dopamine transporter activity (Ebstein et al., 1996; Retz et al., 2003; Bortolato and Shih, 2011; Malloy-Diniz et al., 2013; Chester et al., 2016). These moderating influences may also contribute to this variation.

Dopaminergic neurotransmission may play an important role in modulating behavioral changes including impulsivity in T. gondii infection. A microRNA study found that microRNA-132, a non-coding RNA involved in control of gene expression, is the microRNA that is substantially upregulated by all three-prototype T. gondii strains (Xiao et al., 2014). Moreover, in the same study, upregulation of microRNA-132 was found associated with changes in dopamine receptor signaling by decreasing expression of D1-like dopamine receptors, a dopamine receptor involved in the negative feedback regulation of dopamine release in the brain (Saklayen et al., 2004), and decreasing metabolizing enzyme (i.e., MAO-A). As a consequence, an increased level of dopamine might be expected. Indeed, microRNA-132 is upregulated and dopamine levels are elevated (by 38%) in striatal tissue of mice infected with T. gondii (Xiao et al., 2014). Similarly, other experimental studies also found that dopaminergic neurons containing T. gondii have demonstrated an increased synthesis of dopamine and an elevated presence of homovanillic acid, a dopamine metabolite (Stibbs, 1985; Prandovszky et al., 2011; Martin et al., 2015). A recent study reported a higher acoustic startle response magnitude among T. gondii infected participants (Massa et al., 2017). The authors proposed that the finding might be the outcome of increased dopamine production (Swerdlow et al., 1986; Parlog et al., 2015). Additionally, two genes that encode tyrosine (Tyr) hydroxylase, a rate-limiting enzyme of dopamine synthesis, were identified in the genome of T. gondii (Gaskell et al., 2009). The encoded enzymes metabolize phenylalanine (Phe), as well as Tyr with substrate preference for Tyr. Thus, the enzymes catabolize Phe to Tyr and Tyr to L-DOPA. To further support a T. gondii and dopamine connection, anti-dopaminergic agents prevent behavioral alterations induced by T. gondii in animals (Skallova et al., 2006; Webster et al., 2006). On the contrary, rats that have lower dopamine D2/D3 receptor availability in the ventral striatum display greater response impulsivity on the 5-choice serial-reaction time task, a laboratory behavioral task used to assess motor impulsivity in animals (Dalley et al., 2007). It was also reported that infection with T. gondii increases impulsivity in male rats with concomitant T. gondii infection caused reduction in dopamine level and neuronal spine density in the nucleus accumbens core (Tan et al., 2015). In summary, dopamine appears to be implicated in the association between T. gondii seropositivity and impulsivity, but the specifics of this interaction necessary to be targeted experimentally or therapeutically remain poorly understood.

The first step towards the synthesis of dopamine involves the conversion of the amino acid (i.e., Phe to Tyr) by activating Phe hydroxylase (PAH). It is followed by a two-part enzymatic reaction where Tyr is converted into dopamine via Tyr hydroxylase (the rate-limiting enzyme for biosynthesis of catecholamines) and L-type amino acid decarboxylase (van Spronsen et al., 2009; Daubner et al., 2011). The Phe and Tyr (Phe:Tyr) ratio, an estimate of PAH activity (Anderson et al., 1994), is higher in proinflammatory conditions including cancer (Neurauter et al., 2008a), trauma and sepsis (Ploder et al., 2008), human immunodeficiency virus infection (Zangerle et al., 2010) as well as depression (Hoekstra et al., 2001; Capuron et al., 2011) and schizophrenia (Rao et al., 1990; Wei et al., 1995; Okusaga et al., 2014). Oxidative stress and reactive oxygen species (ROS) induced depletion of (6R)-L-erythro-5,6,7,8-tetrahydrobiopterin (BH4), an essential co-factor for PAH, may mediate, at least in part, the association between the Phe:Tyr ratio and Type 1 T-helper cell (Th1) immunity (Anderson et al., 1994; Werner et al., 2011). A high Phe:Tyr ratio, resulting from a dysfunction of PAH (Neurauter et al., 2008b; Capuron et al., 2011), can be the consequence of Th1 activation, which is known to be one of the protective immune mechanisms responsive to T. gondii infection (Spellberg and Edwards, 2001).

Based on the overriding hypothesis that dopamine neurotransmission may moderate the relationship between T. gondii seropositivity and impulsivity, we investigated how Phe, Tyr, and their ratio interact with T. gondii seropositivity in predicting impulsivity traits.

Low grade immune activation controls T. gondii and maintains the latency of chronic infection in the immunocompetent host, but a consequence of this is activation of the enzyme indoleamine 2,3-dioxygenase 1 (IDO-1) (Pfefferkorn, 1984; Nagineni et al., 1996; Fujigaki et al., 2002; Silva et al., 2002) resulting in a bias in the metabolism of the amino acid Trp toward the kynurenine (Kyn) pathway, at the detriment of serotonin synthesis. As Kyn inversely regulates dopamine function in the brain (Wu et al., 2007; Okuno et al., 2011; Linderholm et al., 2016;), and as T. gondii also activates the Kyn pathway leading to diverting tryptophan (Trp) from production of serotonin (Silva et al., 2002), another key neurotransmitter in impulsivity regulation (Kruesi et al., 1990; Dolan et al., 2001; Paaver et al., 2007; Xu et al., 2017), we also analyzed how Kyn, Trp, and their ratio interact with T. gondii seropositivity in predicting impulsivity traits.

2. Materials and methods

2.1. Sample

As part of a case-control study of schizophrenia, 951 healthy adults were recruited from Munich, Germany. After presenting participants with the study procedures, each participant completed a written informed consent form. The study was approved by the committee of local ethics of Ludwig Maximilians University, Munich, Germany. Additionally, the study was given exempt status from the University of Maryland IRB as data were de-identified and no further contact with participants was planned. Participants had no prior history of suicide attempts. Moreover, the Structured Clinical Interview or Diagnostic and Statistical Manual of Mental Disorders 4^th Edition Text Revision (DSM-IV-TR) (First MB, 2002) was used to ascertain the exclusion of Axis I or II diagnoses in the sample. Blood samples were collected from a forearm vein into ethylenediaminetetraacetic acid (EDTA)-containing flasks, with no fasting protocols or dietary restrictions observed. The plasma acquired after 10 minutes of centrifuging samples at 4°C was aliquoted into Eppendorf tubes and stored at −80°C.

2.2. Measurement of plasma phenylalanine, tyrosine, kynurenine, and tryptophan

Plasma Phe, Tyr, Kyn and Trp levels were measured using high performance liquid chromatography (HPLC) with fluorescence detection, using 3-nitro-L-tyrosine as an internal standard, as described elsewhere (Neurauter et al., 2013; Okusaga et al., 2014).

2.3. T. gondii serological analysis

We used solid phase enzyme-linked imunosorbent assay (ELISA) to measure Immunoglobulin G (IgG) to T. gondii (Arling et al., 2009) (T. gondii IgG ELISA, IBL Laboratories, Hamburg, Germany). Using a human IgG-specific secondary enzyme-conjugated antibody, we identified particular antibodies bound to T. gondii antigens on the wells of the microtiter plates in the diluted serum once it had been exposed to the restrained antigens on the wells. A quantitative measurement of antibodies, involving optical density ratios of blood sample to a standard of 10 international units of anti-T. gondii antibody determined serointensity following the substrate reaction. Seropositivity was described by a titer greater than or equal to 0.8.

2.4. Measures of impulsivity

The Disinhibition subscale of the Sensation Seeking Scale-V [SSS-V(DIS)] (Zuckerman and Neeb, 1979), a valid and reliable measure consisting of 40 items of forced-choice, was used to estimate trait sensation-seeking impulsivity. The SSS-V was designed to assess individual differences in arousal and stimulation needs (Zuckerman, 1994; Roberti et al., 2003). Of the four SSS-V subscales, SSS-V(DIS) has shown the strongest correlation with reckless behavior (Trimpop et al., 1998; Jonah et al., 2001; Roberti, 2004) and risky driving (Constantinou et al., 2011). In previous studies, the SSS-V(DIS) subscale was used to measure impulsivity in the association between T. gondii and the risk of accidents caused by motor vehicle (Flegr et al., 2002; Yereli et al., 2006; Kocazeybek et al., 2009). Additionally, the sub-scale SSS-V(DIS) has been linked to elevated aggression (Wilson and Scarpa, 2013) and recurrent suicide attempts (Laget et al., 2006). Furthermore, the SSS-V(DIS) sub-scale has been associated with repeated suicide attempts (Laget et al., 2006) and higher levels of aggression (Wilson and Scarpa, 2013).

2.5. Variable transformations

The subjects were stratified into two age categories (i.e., older and younger) based on the median age (60 years old). Individuals who were 20–59 years old fell into the younger category, while individuals 60 years old and above were in the older category. We transformed levels of the molecular markers of the Phe-Tyr and the Trp-Kyn pathways into categorical variables (i.e., LOW and HIGH) according to prior published protocols (Okusaga et al., 2016). For the Phe:Tyr ratio, Phe, Kyn:Trp and Kyn, we considered all individuals with values in the lower 3 quartiles (i.e., lower 75%) as LOW and all individuals with a ratio falling in the upper quartile (i.e., top 25%) as HIGH. For Trp and Tyr, we considered all individuals with levels in the higher 3 quartiles (i.e., top 75%) as HIGH and all individuals with a ratio falling in the lowest quartile (i.e., lower 25%) as LOW.

2.6. Statistical analyses

An analysis of co-variance (ANCOVA) was performed in SYSTAT 13 to estimate the interactions of Phe level, Tyr level, and Phe:Tyr ratio on one hand, and Kyn, Trp and Kyn:Trp ratio on the other hand, with T. gondii seropositivity status in predicting impulsivity scores [i.e., SSS-V(DIS) scores). Covariates included age and gender. We also considered using BMI and education as covariates. BMI, however, did not correlate with impulsivity scores. Additionally, multivariate models with only age, gender and education as independent variables and impulsivity scores rendered education non-significant. The final model, thus, included T. gondii, monoamine precursors or ratios, age and gender. Further pairwise comparisons between means were performed using the post hoc Tukey’s Honestly-Significant-Difference test only after identifying statistically significant overall interactions of the multivariable models.

3. Results

Similar to previously published seroprevalence patterns of T. gondii in German populations (Pappas et al., 2009; Hinze-Selch et al., 2010), almost half of our sample of 950 psychiatrically normal individuals (n=476, 52.2%) were T. gondii seropositive (Table 1). The mean age of the sample was 53.6 (±15.8) years old. T. gondii positive participants were significantly older (59.1± 12.5) than T.gondii negative patients (48.1 ± 16.8, p<0.001). The mean Phe:Tyr ratio of T. gondii positive individuals was significantly lower than that of individuals who were T. gondii negative (p<0.001). The mean Phe level of T. gondii positive individuals was significantly lower than that of individuals who were T. gondii negative (p=0.019) but the mean Tyr level of T. gondii positive individuals had no significant difference from that of individuals who were T. gondii negative (p=0.531).

Table 1:

Characteristics of study sample (n=950) by T. gondii status

	T. gondii IgG +a n=476 (52.2)	T. gondii IgG − n=474 (47.8)	Total n=950	p valueb
Age mean± SD	59.10± 12.50	48.10±16.80	53.60±15.80	p=0.001
Sex n (%)
Male	243 (51.0)	220 (46.4)	463 (48.7)	p=0.153
Female	233 (48.9)	254 (53.6)	487 (51.3)
SSS-V(DIS) meaniSD	12.02±1.75	12.34±1.88	12.18±1.82	p=0.007
cPhe:Tyr mean±SD	0.73±0.31	0.82±0.39	0.80 ±0.57	p=0.001
Phe mean±SD (micromolar)	67.24±49.21	75.05±53.78	71.20±51.70	p=0.019
Tyr mean±SD (micromolar)	87.66±51.35	89.56±42.49	88.62±47.05	p=0.531
dKyn:Trp mean±SD	44.05±30.96	37.82±22.71	40.93±27.31	p<0.001
Kyn mean±SD (micromolar)	2.48±1.12	2.30±1.28	2.39±1.21	p=0.022
Trp mean±SD (micromolar)	65.28±31.88	65.58±24.01	65.43±28.20	p=0.87

SSS-V(DIS): The scores of disinhibition subscale of the Sensation Seeking Scale-V

T. gondii = Toxoplasma gondii

The numbers in the parentheses in the table are percentages

IgG = immunoglobulin G serum antibody status, tested by ELISA

^aT. gondii IgG+: Toxoplasma gondii seropositivity was defined by an IgG titer greater than or equal to 0.8

^bStatistical tests using t-test for continuous and Chi-square test for categorical variables

^cPhe:Tyr = ratio between plasma phenylalanine (Phe) and tyrosine (Tyr) levels

^dKyn:Trp = ratio between plasma kynurenine (Kyn) and tryptophan (Trp) levels

The mean Kyn:Trp ratio (p<0.001) and Kyn level (p=0.022) of T. gondii positive individuals were significantly different from these of individuals who were T. gondii negative. No significant difference was found in the mean Trp level between T. gondii positive and negative groups (p=0.87).

An ANCOVA on the impulsivity scores showed significant interaction among T. gondii seropositivity, Phe:Tyr ratio categories, and gender and age categories (F_(1,896) = 7.772, p = 0.007). We therefore further ran an ANCOVA on the impulsivity scores in different gender and age groups. The result yielded a significant interaction between T. gondii seropositivity status and Phe:Tyr ratio category (F_(1,173) = 10.635, p = 0.001) that was evident only in younger men.

A post-hoc Tukey’s Honestly-Significant-Difference test indicated that statistically significant differences in the impulsivity scores between seropositive and negative individuals were only present in younger men (aged 20 to 59 years) who also had high Phe:Tyr ratios (i.e., ratios that fell into the top quartile). Specifically in the high Phe:Tyr ratio group, the mean adjusted difference in impulsivity scores was 2.50 points higher in seropositive as compared to seronegative younger men (CI = 0.98 to 4.02, p < 0.001) (Figure 1A). Further, in seropositive younger men, mean adjusted impulsivity scores in those who had high Phe:Tyr ratios were also 1.87 points higher than in individuals with low Phe:Tyr ratios (95% CI = 0.43 to 3.31, p = 0.006). In individuals categorized as having LOW Phe:Tyr ratios, no significant differences were observed in impulsivity scores between individuals in the two seropositivity categories (p = 0.847, 95% CI = −0.555 to 1.105) (Figure 1A).

Fig. 1.

Comparisons of impulsivity scores between T. gondii IgG-seropositive versus seronegative subjects, stratified by age groups, gender, and blood Phenylalanine/Tyrosine ratio (Phe: Tyr) categories.

Impulsivity scores in seropositive vs. seronegatives participants stratified by gender, age and Phe: Tyr categories. Strata include younger women (aged 20 to 59 years), older women (aged 60 years and up), younger men (aged 20 to 59 years), and older men (aged 60 years and up). Phe:Tyr ratio was categorized as HIGH if the ratio fell in top 25th percentile, and was categorized as LOW if the ratio fell in lower 75th percentile. Impulsivity Scores on the Disinhibition subscale of the Sensation Seeking Scale-V [SSS-V(DIS)] are presented as least square means and standard errors (SEs), stratified by categorical variable and adjusted for age within the age category. ANCOVA analysis of impulsivity scores showed statistically significant interactions between T. gondii seropositivity, Phe:Tyr ratio, gender and age category (F_(1,896) = 7.772, p = 0.007). A significant interaction between T. gondii seropositivity status and Phe:Tyr ratio was significant in younger men (F_(1,173) = 10.635, p = 0.001), but not in older men (F_(1,256) = 0.593, p = 0.442), younger women (F_(1,280) = 0.516, p = 0.473), and older women (F_(1,184) = 1.868, p = 0.173).

* Tukey’s Honestly Significant Difference Test showed that the impulsivity score in younger men, who were also T. gondii seropositive and had HIGH Phe:Tyr ratios, were signifificantly higher than in all other subgroups (p < 0.01 for all), with no other significant differences between subgroups.

There were no significant interactions found between T. gondii seropositivity and Phe:Tyr ratio category in all other gender and age groups including older men (F_(1,256) = 0.593, p = 0.442; Figure 1B), younger women (F_(1,280) = 0.516, p = 0.473), and older women (F_(1,184) = 1.868, p = 0.173).

There were no significant interactions of categorical Phe (F_(1,896) = 1.901, p = 0.168) or Tyr (F_(1,896) = 0.695, p = 0.405) and T. gondii seropositivity status, gender and age category.

There were no significant interactions between T. gondii seropositivity and categorical Kyn, Trp and their ratio. Specifically, categorical Kyn:Trp ratio (F_(1,913) = 0.39, p = 0.532), Kyn (F_(1,913) = 1.913, p = 0.167), and Trp (F_(1,913) = 0.166, p = 0.684) showed no significant interactions in a model including T. gondii seropositivity, gender and age category.

4. Discussion

Our findings demonstrated increased trait impulsivity scores only in individuals who met four criteria - being men, younger (i.e., 20–59 years old), T. gondii seropositive, and having Phe:Tyr ratios in the upper quartile. To our knowledge, this is the first study regarding the moderation by Phe:Tyr ratio of the association between T. gondii seropositivity and trait impulsivity. This complements our recent publication on a positive correlation between Phe:Tyr ratio and trait aggression in T. gondii-positive men (Mathai et al., 2016). Changes in dopaminergic signaling may mediate current results.

Supporting a direct role of dopamine in modulating impulsivity, systemic use of dopamine D2-like receptor agonists has been reported to reduce response impulsivity in rats (Fernando et al., 2012) and in human healthy controls (Nandam et al., 2013). Additionally, administration of the dopamine D2/D3 receptor antagonist haloperidol in smokers results in increased motor impulsivity (i.e., reduced accuracy) and reduced activation in the prefrontal regions associated with inhibitory control (Luijten et al., 2013). Administration of the dopamine D2/D3 receptor antagonist nafadotride into the accumbens shell in rats also enhances premature responding in high impulsive rats (Besson et al., 2010). Moreover, highly impulsive rats display low dopamine D2-like receptor mRNA expression in the mesolimbic pathway (Besson et al., 2013). In human subjects, dopamine D2/D3 receptor availability also negatively correlates with impulsivity and positively correlates with inhibition-related fMRI activation in frontostriatal circuitry (Ghahremani et al., 2012). Specifically, rats that have lower dopamine D2/D3 receptor availability in the ventral striatum (more specifically the nucleus accumbens) display greater response impulsivity on a motor impulsivity animal model (Dalley et al., 2007). Furthermore, cerebrospinal fluid homovanillic acid, the main metabolite of dopamine, correlates negatively with impulsive aggression (Coccaro and Lee, 2010), which is supportive of decreased levels of dopaminergic activity in impulsive behaviors. Beside those findings supporting an inverse association between impulsivity and dopaminergic function, there are findings suggesting a direct association between them as well (Cole and Robbins, 1989; Pezze et al., 2009; Eagle et al., 2011). The explanation for the variation was discussed in introduction. Nonetheless, those findings support that dopaminergic neurotransmission plays an important role in the neurobiological aspects of impulsivity.

The rate limiting step of dopamine formation is the conversion of Tyr into dopamine via Tyr hydroxylase, which is followed by L-type amino acid decarboxylase (van Spronsen et al., 2009; Daubner et al., 2011). Interestingly, the genome of T. gondii includes genes that encode Tyr hydroxylase is also activated in the tissue cyst (bradyzoite, chronic) stage (Gaskell et al., 2009). Correspondingly, dopaminergic neurons containing T. gondii have demonstrated an increased synthesis of dopamine (Stibbs, 1985; Prandovszky et al., 2011; Xiao et al., 2014; Martin et al., 2015). Our study found that a higher host Phe:Tyr ratio, but not significantly changed individual Tyr or Phe level, is associated with an increased impulsivity in younger men who are T. gondii seropositive. The higher host Phe:Tyr ratio may result from a decreased PAH activity (Neurauter et al., 2008; Capuron et al., 2011). The implicated decreases in PAH activity can be secondary to Th1 immune activation in the brain that is necessary to contain T. gondii infection in the immunocompetent host (Spellberg and Edwards, 2001). Therefore, the relatively lower Tyr availability can be a consequence of low grade immune activation induced and perpetuated by T. gondii chronic infection. Moreover, while dopaminergic neurons infected by T. gondii have demonstrated increased synthesis of dopamine (Stibbs, 1985; Prandovszky et al., 2011; Xiao et al., 2014; Martin et al., 2015), an in vivo study demonstrated that the spread of T. gondii does not have a preference towards a particular functional system nor towards the dopaminergic brain tissue (Berenreiterova et al., 2011; McConkey et al., 2013). This may cause aberrant dopamine production in non-dopaminergic brain loci. Consequently, the relatively lower Tyr availability would be further accentuated by parasite through diversion of Tyr from normal dopaminergic neurons to the non-dopaminergic neurons. This may worsen the relative Tyr deficits in the dopaminergic brain tissue and dopaminergic tissues outside of central nervous system which are not directly infected with T. gondii (Kaufman, 1971; Di Cristina et al., 2008).

We speculated that immune activation may be the key ingredient contributing to our results. However, if this were so, considering the upregulation of IDO-1 by immune activation, we would have expected to see similarities between Phe:Tyr and Kyn:Trp ratios in interacting with T. gondii in predicting impulsivity (Capuron et al., 2011). This however was not the case as there were no significant effects in impulsivity scores and interactions in a model including T. gondii seropositivity and Kyn:Trp. Furthermore, arguing against an immune hypothesis, in our recent study, we have identified an association between IED (and trait impulsivity in IED) and T. gondii seropositivity, but the association did not appear to be mediated by levels of IL-6 (Coccaro et al., 2016). However, we cannot rule out the role of other cytokines including interferon gamma and tumor necrosis factor, which act to restrain T. gondii (Pfefferkorn, 1984; Nagineni et al., 1996; Fujigaki et al., 2002; Silva et al., 2002). It is also possible that the proinflammatory processes keeping T. gondii latent are limited to the central nervous system, so that the changes in IL-6 levels are not observed in peripheral blood circulation. Our finding also suggests dopaminergic activity may play a more impactful role than serotonergic activity in increased trait impulsivity scores in T. gondii seropositive younger men.

Sex differences in the effects of dopamine activity on impulsivity have been recently uncovered in rodents. Specifically, dopamine D2 receptor antagonism decreases advantageous responding (i.e, increased impulsivity) in male rats, whereas dopamine D2 receptor agonist decreases advantageous responding in female rats (Georgiou et al. 2018). In our earlier work, we identified higher reactive aggression scores in T. gondii positive women and higher impulsivity scores in younger men (Cook et al., 2015), which were further expanded by our current finding. These are consistent with previously reported gender differences in linking T. gondii seropositivity and measures of personality (Flegr et al., 1996; Flegr et al., 2000; Lindova et al., 2006; Flegr, 2007; Flegr et al., 2008; Lindova et al., 2010; Fond et al., 2013; Tan et al., 2015; Tan and Vyas, 2016).

Longitudinal studies will have the capability to elucidate directions of causality, mediations, moderation and effect modifications by demographic, clinical, serological and biochemical markers presented in our manuscript. In particular, clinical trials with interventions targeting T. gondii infection, inflammation and Phe and Tyr levels (particularly Phe:Tyr ratio) could be tailored for specific subgroups. The subgroups should include younger men with a variety of conditions associated with high impulsivity (Moeller et al., 2001; Sharma et al., 2014), potentially leading to violence towards self or others. Specifically, alcohol use disorder can be one of such conditions, given there are a robust associations between alcohol use and impulsivity (Allen et al., 1998; Brady et al., 1998; Bechara, 2005; Perry and Carroll, 2008; Czapla et al., 2015), self-directed violence (Pompili et al., 2010), as well as T. gondii seropositivity (Samojlowicz et al., 2013).

Our study has multiple limitations. No causal relationships between impulsivity in T. gondii positive individuals and catecholamine precursors can be concluded because of the cross-sectional design of this study and the absence of immune marker measurements. Opportunities to measure dopamine metabolites were not pursued, considering that the primary study had other, unrelated, aims. Fasting protocols, dietary regimens and timing of meals, which may be relevant (Wurtman et al., 1968), were not obtained. Moreover, trait impulsivity responses on questionnaires were not corroborated with personal history and neuropsychological testing of impulsivity.

Despite the aforementioned limitations, this study possessed important strengths. We studied a large sample size of healthy individuals, in which psychopathology was ruled out methodically using SCID for DSM-IV-TR. Thus, we avoided confounding variables related to mental illness (e.g., their symptoms, functional impairments, ongoing treatment and treatment changes).

In conclusion, as trait impulsivity is considered an endophenotype of suicidal behavior (Mann et al., 2009), our data suggest potential for individualized interventions to prevent suicide based on certain intersections of clinical, demographic, psychological, and laboratory markers.

Highlights.

• Toxoplasma gondii (T. gondii) was previously associated with higher impulsivity.

• Dopaminergic system regulates impulsivity, and T. gondii secretes dopamine (DA).

• DA precursors (phenylalanine [Phe]; tyrosine [Tyr]) change with mild inflammation.

• We measured and related T. gondii, Phe, Tyr and impulsivity in healthy subjects.

• Only young men with both high Phe:Tyr and T.gondii (+) reported high impulsivity.