Reciprocal moderation by Toxoplasma gondii seropositivity and blood phenylalanine – tyrosine ratio of their associations with trait aggression

Abstract

We previously reported that trait aggression, proposed as an endophenotype for suicidal behavior, is positively associated with Toxoplasma gondii (T. gondii) seropositivity in females, but not in males. Additionally, older males seropositive for T. gondii had lower scores on measures of trait aggression, including self-aggression. Trait aggression may be influenced by dopaminergic signaling, which is known to be moderated by gender and age, and potentially enhanced in T. gondii positives through the intrinsic production of dopamine by the microorganism. Therefore, we investigated associations between trait aggression and interactions between T. gondii enzyme-linked immunoabsorbant assay (ELISA) IgG titer-determined seropositivity and high-performance liquid chromatography- (HPLC-) measured blood levels of dopamine precursors phenylalanine (Phe), tyrosine (Tyr), and their ratio in a sample of 1000 psychiatrically healthy participants. Aggressive traits were assessed using the questionnaire for measuring factors of aggression (FAF), the German version of the Buss-Durkee hostility questionnaire. We found that 1) the decrease in trait aggression scores in T. gondii-positive older males was only present in individuals with a low Phe:Tyr ratio, and 2) that there was a positive correlation between Phe:Tyr ratio and total aggression and selected subscales of aggression in T. gondii-positive males, but not in T. gondii-negative males. These findings point toward a gender-specific reciprocal moderation by Phe:Tyr ratio and T. gondii seropositivity of their associations with aggression scores, and lead to experimental interventions geared to manipulating levels of dopamine precursors in selected T. gondii positive individuals with increased propensity for aggression.

Introduction

We previously reported [1] that trait aggression, proposed as an endophenotype for suicidal behavior, is positively associated with Toxoplasma gondii seropositivity in females, but not in males. Converging lines of evidence have catalyzed an interest in the link between inflammation and mental illness [2, 3]. While elevated proinflammatory cytokines and glial cell abnormalities represent trait markers of major mental illnesses like major depressive disorder [4, 5], bipolar disorder [5, 6] and schizophrenia [5, 7], there also appears to be a state-mediated effect of suicidal behavior [8–10], acuity of mood episode or psychosis [6, 11, 12] and treatment [11–14] on cytokine levels, microglial activation and other indicators of central inflammation. Trait aggression [15–18] and impulsivity [19, 20], behaviors often present in these illnesses, are also associated with elevations of inflammatory biomarkers as suggested by human [21–25] and animal studies [26–29]. Higher rates of latent infection with neurotropic pathogens like Toxoplasma gondii (T. gondii) and Herpesvidae family viruses [30–34], may contribute to the low-grade immune activation observed in these conditions. Latent toxoplasmosis, during which only bradyzoites (tissue cysts) are present, previously considered harmless, is linked to major psychiatric illnesses [30, 31, 33] as well as suicidal behavior [35–41].

We recently reported, in psychiatrically healthy individuals, that trait aggression, proposed as an endophenotype for suicide [42–44], was associated with T. gondii seropositivity [1] in females but not in males. We also noted that males over the age of 60 who were seropositive for T. gondii had significantly lower scores on measures of trait aggression including self-aggression. This is consistent with previous reports that the differences in personality measures associated with T. gondii seropositivity are sex-specific [45, 46]. Sex-specific differences in personality dimensions in relationship to T. gondii have been described in the past in subjects in whom mental illness was not ruled out in a systematic manner [47–50].

The associations between trait measures like aggression and T. gondii seropositivity may be mediated by dopaminergic neurotransmission. Previous studies have demonstrated an increase in dopamine synthesis and release, as well as elevated homovanillic acid (HVA), a metabolite of dopamine, in T. gondii-infected neurons in animals [51–54]. Additionally, anti-dopaminergic agents prevent behavioral alterations induced by T. gondii in animals [55, 56]. Of potential high relevance, two genes encoding tyrosine hydroxylase, the rate-limiting enzyme of dopamine synthesis, were identified in the genome of T. gondii [57]. Toxoplasma gondii and its influence on dopamine signaling is also suggested by a decrease in the psychobiological construct of novelty-seeking, one of the temperaments in the Temperament and Character Inventory (TCI) related to dopamine [58], in association with T. gondii seropositivity in humans [45, 59, 60].

Dopaminergic neurotransmission has also been implicated in the neurobiology of aggression [61–64]. Genetic polymorphisms involved in altered dopamine receptor and transporter activities as well as enzymes involved in the metabolism of dopamine have been identified in association with aggression in humans and animals [65, 66]. The direction of the relationship between dopamine and aggression is far from unequivocal. Some studies have reported a positive relationship between dopaminergic neurotransmission and aggression [67, 68], while others point to a dopaminergic deficit in the relationship with aggression, in particular impulsive aggression [64, 65, 69, 70]. These conflicting findings may be due in part to both context dependency of the relationship between dopamine and aggression, and genetic influences, such as functional allelic variants of monoamine oxidase A, which metabolizes dopamine [71].

Conversion of phenylalanine, an essential amino acid, to tyrosine via the activity of phenylalanine hydroxylase (PAH), is the first step towards the synthesis of dopamine. Tyrosine is further converted into dopamine via a two-step enzymatic reaction involving tyrosine hydroxylase (rate-limiting enzyme in the biosynthesis of catecholamines) and aromatic L-type amino acid decarboxylase [72, 73]. Type 1 T-helper cell immunity (Th1), which acts against intracellular pathogens like T. gondii [74], is linked to dysfunction of PAH, and subsequently, an elevated Phe:Tyr ratio [75, 76]. This is thought to be mediated by oxidative stress and reactive oxygen species- (ROS-) induced depletion of (6R)-L-erythro-5,6,7,8-tetrahydrobiopterin (BH4), an essential cofactor for PAH [77, 78]. The ratio between phenylalanine and tyrosine (Phe:Tyr) provides an estimate of PAH activity [78] and is elevated in proinflammatory conditions like cancer [79], trauma, sepsis [80], human immunodeficiency virus (HIV) infection [81] as well as depression [76, 82] and schizophrenia [83–85]. Additionally, children with phenylketonuria, a congenital decrease in activity of PAH leading to elevated phenylalanine levels, exhibit higher levels of aggression [86], with a positive correlation between aggression ratings and phenylalanine levels [87].

In this add-on project, we sought to investigate interactions among Phe, Tyr, and their ratio, T. gondii seropositivity, and aggression measures in a sample of psychiatrically healthy participants.

Materials and methods

Sample

This is an add-on study to a previous protocol on the genetics of mental illness. As a component of the “parent” protocol, 1000 healthy adults were recruited from the Munich area in Germany. All the healthy controls provided written informed consent to take part in the parent study after the study procedures were explained in detail. The local Ethics Committee of Ludwig Maximilians University, Munich, Germany, approved the study, and the University of Maryland Institutional Review Board (IRB) granted an exempt status considering that the data were deidentified and recontacting participants in the future was rendered impossible. Axis I or II diagnoses were ruled out by the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders 4th Edition Text Revision (DSM-IV TR) [88]. History of suicide attempts was also ruled out. Mental illness and suicidal behavior were ruled out in first degree-relatives as well. Blood samples were drawn without observing any fasting protocols or dietary restrictions from a forearm vein into ethylenediaminetetraacetic acid (EDTA)-containing tubes. Plasma obtained after centrifuging the sample for 10 min at 4°C was aliquoted into Eppendorf tubes and frozen immediately at −80°C and stored at this temperature until analysis of plasma Phe and Tyr.

Measurement of plasma phenylalanine and tyrosine

Using high performance liquid chromatography (HPLC), plasma phenylalanine and tyrosine levels were measured using 3-nitro-L-tyrosine as an internal standard as described elsewhere [84, 89].

Toxoplasma gondii serological analysis

Solid phase enzyme-linked immunosorbent assay (ELISA) was used to measure immunoglobulin G (IgG) to T. gondii [35] (T. gondii IgG ELISA, IBL Laboratories, Hamburg, Germany). After exposing diluted serum to immobilized T. gondii antigens on the wells of microtiter plates, specific antibodies binding to the antigen-coated wells were detected using a human IgG-specific secondary enzyme-conjugated antibody. Serointensity was determined by quantitative measurement of antibodies after substrate reaction, using optical density ratios between the blood sample and a standard containing 10 international units of anti-T. gondii antibody. A titer ≥0.8 was used to define seropositivity. All T. gondii antibody measurements were performed at the Stanley Laboratory of Developmental Neurovirology, Baltimore, MD and the technician was blind to the study hypothesis.

Measures of aggression

Questionnaire for measuring factors of aggression (FAF-Fragebogen zur Erfassung von Aggressivitätsfaktoren), a German version of the Buss-Durkee hostility questionnaire [90, 91], was used to assess trait-aggression in all participants. FAF is a self-administered inventory consisting of 76 items, 66 of which explore five components of aggressive behavior including FAF-self-aggression (11 items), FAF-total aggression (35 items) and its three component subscales: FAF-spontaneous aggression (19 items); FAF-reactive aggression (13 items); FAF-irritability (13 items). FAF demonstrates good internal consistency and test-retest reliability with Cronbach’s α-values range from 0.61 to 0.79 [90].

Statistical analysis

Analysis of covariance (ANCOVA) was used to determine differences in FAF-total aggression, its subscales and FAF-self aggression between T. gondii seropositive and seronegative individuals. Then subjects were stratified further based on age and sex to detect differences in aggression scores between seropositive and seronegative individuals. Phe:Tyr ratio was converted into a categorical variable by assigning all individuals with a ratio in the upper quartile as high and all individuals with a ratio in the lower three quartiles as low, a quartile-to-binary threshold we applied to monoamine precursors in our previous work [92].

Further analysis by stratification based on high vs. low ratio was performed.

Adjusted linear regression models, accounting for body mass index (BMI), age, education level and sex were then used to determine if Phe or Tyr levels or the Phe:Tyr ratio were related to specific subscales of aggression. All models were stratified by sex due to previous evidence of sex-dependent shifts in personality traits by T. gondii status.

Results

Sample characteristics

Approximately half the sample (n = 476, 50.1%) of patients were positive for T. gondii based on the criteria of seropositivity of IgG > 0.8 with higher seroprevalence among older subjects (60%) vs. younger subjects (34%) based on a median age split (Table 1). This seroprevalence pattern is similar to previously published estimates of seroprevalence of T. gondii from Germany [34, 93]. Mean age of the sample was 53.6 (±15.8) with T. gondii-positive participants being significantly older (59.1 ± 12.5) when compared to T. gondii-negative patients (48.1 ± 16.8). Close to half the sample was university educated (45.8%) with T. gondii-seropositive patients having lower educational attainment (p < 0.004). Mean Phe:Tyr ratio was significantly lower in T. gondii-positive individuals when compared to T. gondii-negative individuals. BMI was higher in T. gondii-positive vs. T. gondii-negative individuals.

Table 1.

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

	T. gondii IgG+a (n = 476)	T. gondii IgG− (n = 474)	Total (n = 950)	p-Valueb
Age, mean ± SD	59.1 ± 12.5	48.1 ± 16.8	53.6 ± 15.8	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)
Education, n (%)
Primary	145 (30.5)	86 (18.1)	230 (24.3)	p < 0.001
Secondary	147 (30.9)	137 (28.9)	284 (29.9)
University	184 (38.6)	251 (53.0)	435 (45.8)
BMIc, mean ± SD	25.3 ± 4.0	24.3 ± 3.7	24.8 ± 3.9	p < 0.001
Phe:Tyrd, mean SD	0.734 ± 0.31	0.821 ± 0.39	0.803 ± 0.57	p < 0.001

IgG, immunoglobulin G serum antibody status, tested by ELISA.

^aOptical density ratio between the blood sample and a standard containing 10 international units of anti-Toxoplasma gondii IgG ≥ 0.8.

^bStatistical tests using t-test for continuous and χ²-test for categorical variables.

^cBMI, body mass index (kg/m²).

^dPhe:Tyr, ratio between plasma phenylalanine and tyrosine levels.

Toxoplasma gondii seropositivity and trait aggression

There were no differences in aggression scores between seropositive and seronegative individuals. As previously reported [1], upon stratification by sex, females who were positive for T. gondii had higher FAF-reactive aggression scores when compared to T. gondii-negative females after adjustment for age, BMI and education [F₍₁₄₈₀₎ = 4.1, p = 0.04] but did not differ on any other scales of aggression including total aggression.

In males, T. gondii-seropositive individuals had decreased scores on FAF-total aggression [F₍₁₄₅₇₎ = 6.98, p = 0.009], FAF-spontaneous aggression [F₍₁₄₅₇₎ = 5.0, p = 0.03], FAF-irritability F₍₁₄₅₇₎ = 5.0, p = 0.03] and FAF-self-aggression [F₍₁₄₅₆₎ = 5.2, p = 0.03] when compared to T. gondii-negative males after adjusting for age, BMI and education. On further stratification by age, this strong effect was observed only in older males (≥60 years old) but not in younger males [F₍₁₂₇₄₎ = 10.5, p = 0.001]. Post hoc analysis showed that the average difference between seropositive and seronegative older males in FAF-total aggression was 2.5 points (CI = 1.1–4.0, p = 0.001). Similar observations of lower aggression scores in seropositive males over the age of 60 were observed with FAF-spontaneous aggression [F₍₁₂₇₄₎ = 6.2, p = 0.01]. FAF-reactive aggression [F₍₁₂₇₄₎ = 6.4, p = 0.01], FAF-irritability F₍₁₂₇₄₎ = 9.2, p = 0.003), and FAF-self-aggression [F₍₁₂₇₃₎ = 9.0, p = 0.003].

High vs. low Phe:Tyr ratio

Upon further stratification by high vs. low Phe:Tyr ratio, seropositive males over the age of 60 continued to demonstrate decreased FAF-total aggression scores [F₍₁₁₉₈₎ = 8.7, p = 0.004] and FAF-self-aggression [F₍₁₁₉₇₎ = 6.1, p = 0.02] if the Phe:Tyr ratio was low, but no differences were observed if the Phe:Tyr ratio was high (p-value was 0.24 and 0.90 respectively). This effect was also present for the subscale scores of FAF-reactive aggression [F₍₁₁₉₈₎ = 7.4, p = 0.007] and FAF-irritability [F₍₁₁₉₈₎ = 6.6, p = 0.01].

Associations between Phe, Tyr, Phe:Tyr ratio and trait aggression score

There were no relationships detected between Phe, Tyr or Phe:Tyr ratio and FAF-total aggression, self-aggression or the subscale scores (reactive aggression, spontaneous aggression and irritability) initially, after adjusting for age, sex, BMI and education (Table 2). After stratifying by T. gondii status and sex, there was a positive correlation between FAF-total aggression/FAF-spontaneous aggression and Phe:Tyr ratio in T. gondii-seropositive males but not in seronegative males.

Table 2.

Associations of FAF total score and selected subscale scores with Phe:Tyr rank in T. gondii positive vs. T. gondii negative individuals stratified by sex after adjustment for age, BMI and education.

FAF scores	T. gondii IgG+c			T. gondii IgG−			Total
	B	SE	p-Value	B	SE	p-Value	B	SE	p-Value
Total aggression
Male	2.51	1.12	0.026	0.55	1.23	0.6	1.23	0.83	0.1
Female	−1.5	1.17	0.2	−1.33	0.79	0.09	−1.32	0.68	0.06
Total	0.02	0.81	0.9	−0.07	0.73	0.9	−0.019	0.54	0.9
Self-aggression
Male	0.77	0.39	0.050	−0.31	0.36	0.4	0.03	0.26	0.9
Female	0.31	0.36	0.4	−0.22	0.29	0.4	0.07	0.22	0.7
Total	0.46	0.26	0.09	−0.22	0.23	0.3	0.06	0.17	0.7
Spontaneous aggression
Male	1.38	0.47	0.004	0.46	0.55	0.4	0.77	0.36	0.04
Female	−0.83	0.51	0.1	−0.29	0.28	0.3	−0.50	0.28	0.08
Total	0.024	0.35	0.95	0.20	0.30	0.5	0.14	0.23	0.5
Reactive aggression
Male	1.00	0.5	0.046	0.20	0.41	0.6	0.5	0.31	0.1
Female	0.28	0.42	0.5	−0.24	0.3	0.4	0.02	0.25	0.9
Total	0.54	0.32	0.09	0.07	0.25	0.8	0.26	0.2	0.2
Irritability
Male	0.12	0.48	0.8	−0.11	0.44	0.8	−0.03	0.32	0.9
Female	−0.98	0.48	0.042	−0.79	0.43	0.07	−0.85	0.32	0.008
Total	−0.54	0.34	0.1	−0.34	0.30	0.3	−0.42	0.22	0.06

FAF, Fragebogen zur Erfassung von Aggressivitätsfaktoren (German adaptation of the Buss-Durkee hostility inventory); Phe:Tyr, ratio between plasma phenylalanine and tyrosine; IgG, immunoglobulin G serum antibody status, tested by ELISA.

^aIndividual scale score differences are model-based estimates from linear regression model adjusted for age, education and BMI and stratified by sex/T. gondii status.

^bTotal FAF score and selected FAF component score (self-aggression, spontaneous aggression, reactive aggression and irritability).

^cOptical density ratio between the blood sample and a standard containing 10 international units of anti-Toxoplasma gondii IgG ≥ 0.8. Bold face values denote statistically significant associations.

Discussion

Our main findings were that; a) the decrease in trait aggression scores in T. gondii-positive older males and aggression was only present in individuals with low Phe:Tyr ratio, with no significant association observed in subjects with a high Phe:Tyr ratio, and b) that there was a positive correlation between Phe:Tyr ratio and Total aggression and selected subscales of aggression in T. gondii-positive males but not in seronegative males. To our knowledge, this is the first study reporting the moderating role by the Phe:Tyr ratio of the association between T. gondii-seropositivity and trait aggression.

Our findings point to Phe:Tyr ratio moderating the association between T. gondii-seropositivity and aggression in older males, with a lowering of aggression in T. gondii-positives only present in those who have a lower Phe:Tyr.

A higher Phe:Tyr ratio may lead to impedance in dopaminergic neurotransmission, which may negate the potentially dopamine-mediated protective link between T. gondii and aggression in older males. Supporting this hypothesis, decreased dopamine synthesis capacity in the midbrain was linked with aggression in humans using positron-emission tomography (PET) [64]. A negative relationship between cerebrospinal fluid HVA levels, a major metabolite of dopamine, and impulsive aggression also points to decreased dopaminergic function [94] in aggressive behaviors. Lower functioning dopamine receptor profiles were reported to predict aggressive responses in an analysis of dopamine receptor polymorphisms and their relationships to aggression [65]. Toxoplasma gondii-infected dopaminergic neurons show an increase in intracellular dopamine in rodents [52]. In rodent-models, an activation of dopamine-responsive sexual pathways, has been described in T. gondii-infected rodents upon exposure to cat urine [95]. Moreover, T. gondii genome contains two genes that code for tyrosine hydroxylase enzyme, one of which is exclusively activated in the bradyzoite or tissue cyst stage, which localizes to brain and muscle tissue in latent toxoplasmosis [57].

Alternatively, Phe:Tyr ratio, a proxy for PAH activity, and thereby proinflammatory states [75], may also reflect the moderation of aggression by inflammation in T. gondii-positive individuals. Patients with intermittent explosive disorder (IED), a psychiatric condition, with recurrent impulsive aggression, have been reported to manifest elevated blood markers of inflammation like C-reactive protein (CRP) and interleukin (IL)-6 [23, 24], and very recently, increased rates of T. gondii seropositivity [96]. Moreover, in these patients with “rage” traits, and documented episodes of aggression, the aggression scores were higher in T. gondii-positive individuals. Our findings of a positive correlation between aggression and Phe:Tyr in T. gondii-positive males may lend credence to the possibility of low grade inflammation, triggered and perpetuated by the parasitic infection, leads to aggression either directly, or, in part, through a higher Phe:Tyr ratio.

Sex-specific differences in personality profile with regards to T. gondii have been described in the past and our results suggest that the correlations between Phe:Tyr ratio and aggression are specific to male subjects positive for T. gondii. We have previously reported age- and sex-dependent differences in trait aggression and impulsivity in T. gondii-positive individuals, with higher reactive aggression in T. gondii positive females and higher impulsivity in T. gondii-positive young males [1]. Despite the possibility of correlations among personality factors that overlap and co-vary by sex, animal studies point to the parasite inducing sex-specific changes in neurotransmitter activity [97], mating behavior [49, 98, 99] and production of sex hormones [100].

One of the limitations of the study is that we did not obtain fasting levels of Phe and Tyr as there is evidence for heterogeneity in plasma levels of amino acids based on timing of meals and dietary preferences [101]. Moreover, the cross-sectional nature of the study precludes estimating causal relationships between dopamine precursors and aggression in T. gondii-positive individuals. The effect of stress and the influence of stress-related hormones were not analyzed, thus not statistically accounting for these potentially relevant confounders. We did not correct the level of significance for multiple comparisons, and thus our results and interpretation are prone to Type I error. Yet, we position our study as, primarily, hypothesis generating rather than testing, leading to future hypothesis testing study designs. Direct markers of immune response like cytokines and acute phase reactants were not measured due to the insufficient volume of plasma samples, considering that this was an add-on study. Additionally, we only used self-report questionnaires instead of observed aggressive events, or collateral informants. For all the above reasons, the results of this study must be considered preliminary and, thus, requiring replication.

Despite these limitations, our study had several strengths; we studied a large sample of normal individuals without any overt psychopathology, family history of mental illness suicidal behavior or substance use [1]. The majority of studies exploring the relationship between personality and T. gondii involved subjects in whom mental illnesses was not ruled out using a structured and validated tool [45, 47, 49, 59, 60]. This previous approach runs the risk of spurious associations between T. gondii seropositivity and personality changes, as they may be confounded by unmeasured mental illness. This study supports the possibility that aggressive traits are directly linked to T. gondii seropositivity, and not as a mere byproduct of the established link between mental illness, and implicitly, its treatment, and T. gondii infection [30, 33]. These findings, as was the case in our previous work [1], persisted after post-hoc controlling for cytomegalovirus (CMV) and herpes simplex virus 1 (HSV1) titers and were independent of CMV/HSV1 seropositivity (for methodology see our previous article Cook et al. (2015) [1].

In conclusion, if replicated and confirmed by further longitudinal and, ideally, experimental designs, our findings may pave the way for individualized interventions geared towards optimizing blood levels of dopamine precursors in T. gondii-positive individuals with aggressive personality traits, and in particular those with overt history of aggression and suicidal behavior.