|Home | About | Journals | Submit | Contact Us | Français|
Neuropsychiatric disorders characterized by behavioral disinhibition, including disorders of compulsivity (e.g., obsessive-compulsive disorder; OCD) and impulse-control (e.g., impulsive aggression), are severe, highly prevalent and chronically disabling. Treatment options for these diseases are extremely limited. The pathophysiological bases of disorders of behavioral disinhibition are poorly understood but it has been suggested that serotonin dysfunction may play a role. Mice lacking the gene encoding brain tryptophan hydroxylase 2 (Tph2−/−), the initial and rate-limiting enzyme in the synthesis of serotonin, were tested in numerous behavioral assays that are well known for their utility in modeling human neuropsychiatric diseases. Mice lacking Tph2 (and brain 5HT) show intense compulsive and impulsive behaviors to include extreme aggression. The impulsivity is motor in form and not cognitive because Tph2−/− mice show normal acquisition and reversal learning on a spatial learning task. Restoration of 5HT levels by treatment of Tph2−/− mice with its immediate precursor 5-hydroxytryptophan attenuated compulsive and impulsive-aggressive behaviors. Surprisingly, in Tph2−/− mice, the lack of 5HT was not associated with anxiety-like behaviors. The results indicate that 5HT mediates behavioral disinhibition in the mammalian brain independent of anxiogenesis.
Behavioral disinhibition (BD) refers broadly to the inability to resist expressing inappropriate behaviors (Iacono et al. 2008). Evidence suggests that there may be a shared tendency toward BD that underlies disorders of compulsivity and impulsivity (Fineberg et al. 2010). Compulsivity refers to repetitive ritualistic behaviors or mental acts over which an individual has little or no control. An individual is driven to perform these in response to provocative stimuli to reduce stress or prevent a feared event from occurring (Fineberg et al. 2010, Winstanley et al. 2006). Impulsivity is a predisposition toward rapid, unplanned reactions to stimuli, decreased inhibitory control and a lack of consideration of negative consequences to self or others (Fineberg et al. 2010). Disorders of compulsivity, including obsessive-compulsive disorder (OCD) and impulsivity (impulse control disorders) including impulsive aggression, eating disorders, intermittent explosive disorder, substance abuse, and pathological gambling, constitute a significant portion of the global burden of disease (Collins et al. 2011). Unfortunately for affected individuals, these disorders are largely resistant to existing pharmacotherapies, establishing the need for additional research to improve the understanding of the neurochemical defects that contribute to BD.
Of all neurochemical systems implicated in neuropsychiatric disorders, including BD, alterations in serotonin (5HT) neurochemical function are perhaps the most often assigned causative roles. The study of 5HT roles in neuropsychiatric diseases has depended historically on the use of drugs that damage 5HT neurons or that act at one or more 5HT receptors. The difficulty in this approach can be appreciated for studies of impulsivity which show increases, decreases or no change in impulsive-like behaviors in animal models (see discussion in (Miyazaki et al. 2011)). Similarly, genetic approaches that deleted genes directly involved in the 5HT neuronal system (Saudou et al. 1994, Bevilacqua et al. 2010, Chou-Green et al. 2003, Bouwknecht et al. 2001, Thomas et al. 2010, Savelieva et al. 2008, Alenina et al. 2009, Gutknecht et al. 2008, Narboux-Neme et al. 2011, Hendricks et al. 2003, Waider et al. 2011, Yadav et al. 2009) have yielded conflicting results regarding a role for 5HT in BD and other psychiatric disorders. One approach that would surmount many of the difficulties associated with the study of 5HT dysfunction in neuropsychiatric diseases would be a genetic depletion of brain 5HT itself. We previously showed that mice lacking the gene encoding Tph2, the initial and rate-limiting enzyme in 5HT synthesis, are devoid of brain 5HT (Thomas et al. 2010). In the present study, we show that deletion of brain 5HT leads to a robust phenotype of BD, including compulsive and motor impulsive behaviors and extreme impulsive aggression. We also find that Tph2−/− mice do not display increased anxiety-like behaviors suggesting that the BD phenotype is not confounded by the presence of co-morbid anxiety.
Tph2−/− mice were generated by deleting exon 1 of the Tph2 gene as described (Thomas et al. 2010). Wild-type and Tph2−/− mice used in this study were derived from matings of heterozygous (Tph2+/−) male and heterozygous (Tph2+/−) females and were on the same genetic background (C57BL/6J-129Sv). Mice (10–12 weeks of age) were acclimated to the behavioral testing room for 1hr prior to testing (10am–4pm daily). Independent groups were used for each behavioral test and experimenters scoring behaviors were blind to genotype. The intraclass correlation coefficients of the independent scorers always exceeded 0.90. In some experiments, Tph2−/− mice were injected with the peripheral L-amino acid decarboxylase inhibitor carbidopa (Sigma Aldrich, St. Louis, MO, USA; 25 mg/kg, ip) followed 30 min later by injections of either vehicle (water) or 5HTP (Sigma Aldrich; 10 mg/kg, ip) to replenish brain 5HT levels and were tested 30 min after 5HTP. The Institutional Care and Use Committee of Wayne State University approved the animal care and experimental procedures and all procedures are in compliance with the NIH Guide for the Care and Use of Laboratory Animals.
Tests for compulsive-like behaviors included the nestlet shredding and marble burying tests. Pre-weighted nestlets (Ancare, Bellmore, NY) were placed into a cage with a single mouse for 60 min. The unshredded remainder of the nestlet was weighted to calculate % nestlet shredded (Witkin 2008). For the marble burying test, opaque marbles were placed on top of fresh bedding and the number of marbles with two-thirds of their surface covered were scored as buried during the 30 min test session (Witkin 2008). In some experiments, the amount of time spent digging and the total distance traveled by mice in the marble burying test were recorded by video using EZ Video Software (AccuScan Instruments, Columbus, OH, USA). As a variant of the marble burying test, standard food pellets (i.e., familiar objects) from the subject’s home cage were used in place of marbles and the number of pellets with two-thirds of their surface covered and/or displaced more than 45° from their original position was scored in 30 min test sessions. Tests of motor impulsive behavior included novelty suppressed feeding, dark emergence and resident-intruder aggression. For novelty suppressed feeding (Bevilacqua et al. 2010), mice were fasted for 18 hours and placed into a novel box facing one corner (randomly selected). A single food pellet was placed in the center of the box and the time taken to start eating the pellet was recorded for up to 10 min. In the dark emergence test (Smith et al. 1998), mice were placed into the dark compartment of the AccuScan Instruments light-dark activity box facing away from the opening between chambers and the time taken to emerge into the lighted compartment (up to 10 min) was measured. Cognitive impulsive behavior was assessed by reversal learning of the Barnes maze spatial learning task as previously reported (O'Leary and Brown 2009). Mice are trained to locate a darkened goal box under one of 40 open holes on the perimeter of a brightly lit maze platform using external room cues. Mice were trained in sessions consisting of 4 trials with an inter-trial interval of 20 minutes. During acquisition, mice were trained in 2 sessions/day separated by 6 hr. For reversal learning, the position of the goal box was rotated 180° from its position during acquisition with respect to external room cues and mice were re-trained in the same protocol as during the acquisition phase. Impulsive-aggressive behavior was assessed using the resident-intruder paradigm (Miczek and O'Donnell 1978). Resident male mice were scored for time to first attack and number of attacks on intruder mice in 5 min test sessions. A modification of the resident-intruder test was also used. Briefly, group-housed (socialized, 10–12 weeks old for adults, 3–4 weeks old for weanlings), sexually naïve mice were placed individually into cages for 30 min (i.e., resident). A wild-type intruder mouse of similar age and size was then introduced into the cage and resident mice were scored aggression as described above. In the 5HTP rescue studies, adult male Tph2−/− mice were tested in the modified resident-intruder test. 24 hours later, mice were re-tested in the task following 5HTP treatment.
Anxiety-like behaviors were assessed using the light-dark test and the elevated plus maze. The light-dark box test was used as previously described (Popova et al. 2009). Mice were placed intothe light compartment facing away from the dark chamber. The number of entries into the dark chamber, total time spent in each compartment and total distance traveled was recorded for each mouse in 10 min sessions. In the elevated plus maze (Pellow and File 1986), the number of entries into each arm of the maze (AccuScan Instruments), time spent in each arm, number of crossings into each arm and total distance traveled was recorded for each mouse in 5 min sessions using a motion-sensitive digital video camera.
Locomotor activity was measured for 30 min in an open field-locomotor activity apparatus (AccuScan Instruments). Total activity (horizontal + vertical) was recorded automatically and analyzed by Fusion software (AccuScan Instruments). In some tests (specified below), movement and stereotyped episodes were counted. Movement episodes are defined as locomotor activity that cause breaks in multiple infrared light beams before and after a rest period (i.e., no movement) ≥ 1 s. Stereotyped episodes are defined as repeated breaks of the same infrared light beam (e.g., grooming or head bobbing in place) before and after a rest period (i.e., no stereotypy) ≥ 1 s. Rest periods are defined as a period of inactivity (i.e., no beam breaks) ≥ 1 s.
Cayman EIA testosterone kits (Cayman Chemical, Ann Arbor, MI) were used to measure testosterone levels according to manufacturer specifications.
The levels of 5HT and metabolites were measured by HPLC, as described (Thomas et al. 2010).
All statistical analyses were performed using GraphPad Prism version 5.00 for Windows. (GraphPad Software, San Diego California USA, www.graphpad.com). Behavioral tests were analyzed by student’s 2-tailed t-test, Chi-square, or one-way ANOVA with Tukey’s post-hoc multiple comparison test (as specified). For the Barnes maze task, main effects of genotype and trial (i.e., total distance travelled and time to reach the goal box) were tested for significance using a two-way ANOVA with Bonferroni’s post-hoc multiple comparison test. For tests that used both males and females as subjects, main effects of genotype and sex were tested for significance using a two-way ANOVA with Bonferroni’s post-hoc multiple comparison test. When the main effect of sex was not significant, data from males and females was collapsed into a single group for graphical purposes. p values < 0.05 were deemed statistically significant.
We tested Tph2−/− mice for compulsive behavior in the nestlet shred and marble-burying tests. Tph2−/− mice shred significantly more nestlet material and buried significantly more marbles than their wild-type counterparts (Figs. 1a and 1b; Fig. S1). Because marbles are novel objects, the presence of which could stimulate digging due to novelty-induced anxiety (Thomas et al. 2009), they were replaced with food pellets from the subject’s home cage (i.e., familiar objects). Tph2−/− mice buried and displaced significantly more food pellets than wild-type mice (Fig. 1c) indicating that they dig compulsively regardless of the type of object placed in the test cage. To quantify this compulsive digging behavior, we measured the time spent digging during the marble-burying test. Tph2−/− mice spent significantly more time digging (~3 fold increase) than wild-type mice (Fig. 1d). Tph2−/− mice showed no evidence of hyperactivity in the marble burying test compared to wild-type mice (Fig. 1e), ruling this out as a cause of increased digging. We also noted that Tph2−/− mice compulsively dig in the absence of any object (Videos S1 and S2), a behavior commonly observed in the non-provoking home-cage environment as well.
Tests of aggression, novelty suppressed feeding and dark emergence were used to assess motor impulsive behaviors. The latter two conflict paradigms measure the consequences of competing motivations (i.e., the drive to eat/explore versus fear of entering a novel, brightly lit arena (Bevilacqua et al. 2010)) and have been used to test impulsivity in 5HT 2b receptor knockout mice. Aggressive behavior was documented using the resident/intruder test. Tph2−/− males showed a significant decrease in the latency to attack intruders, as well as a significant increase in the number of attacks relative to wild-type males (Fig. 2a). In the course of routine colony management, we observed that socialized Tph2−/− animals (e.g. group-housed), regardless of age or sex, exhibited a “hard to handle” phenotype similar to that described for the Fierce mouse (Monaghan et al. 1997, Young et al. 2002). Based on this, we hypothesized that Tph2−/− mice would show impulsive aggression without the prolonged isolation period (i.e., 4 weeks) typically required to elicit this behavior in wild-type mice using the resident intruder test. To test this hypothesis, we developed a modified resident-intruder test (MRIT) in which socialized animals are individually housed for only 30 min prior to introduction of a socialized intruder. We found that all Tph2−/− animals, including adults and weanlings (3–4 weeks old) of both sexes, exhibited highly aggressive behavior in this test (Figs. 2b and 2c, Table S1, Videos S3–S6). As increased testosterone is associated with increased aggression (Bevilacqua et al. 2010), we measured plasma testosterone levels and observed that adult Tph2−/− males showed ~4-fold increases in plasma testosterone levels over wild-type males, whereas adult female Tph2−/− mice had levels similar to their wild-type counterparts (Fig. 2d). Tph2−/− mice also expressed increased motor impulsivity relative to wild-type mice in the novelty suppressed feeding and dark emergence tests (Fig. 3a and 3b). The behavior of Tph2−/− mice was particularly striking in the dark emergence test where they emerged from the dark enclosure in 9 sec by comparison to controls that took > 3 min (Fig. 3b). Reversal learning in a spatial learning task was used to assess cognitive impulsivity. Tph2−/− mice performed similarly to wild-types in acquisition of the Barnes spatial learning task (Fig. 3c and 3d) and on reversal learning of this task (Fig. 3e and 3f) indicating that these mice show motor and not cognitive impulsivity.
To determine if the BD phenotype could be rescued, brain 5HT levels were restored with injections of its immediate precursor 5HTP (10mg/kg, ip) prior to behavioral testing. Tph2−/− mice were also treated with carbidopa (25mg/kg, ip), an inhibitor of peripheral l-aromatic amino acid decarboxylase, 30 min prior to 5HTP administration, to prevent the behaviorally disruptive effects of elevated peripheral 5HT (e.g. diarrhea). 5HTP treatment significantly reduced marble burying (Fig. 4a) without changing overall motility (i.e., movement or stereotyped behavior) of Tph2−/− mice (Fig. 4b). Male Tph2−/− mice already identified as highly aggressive were injected with 5HTP (and carbidopa) and this treatment (but not carbidopa alone) also attenuated impulsive aggression in the MRIT (Fig. 4c). HPLC analyses confirmed that administration of 5HTP restored 5HT to wild-type levels in several brain regions implicated in compulsive and impulsive behaviors including hippocampus, frontal cortex, hypothalamus, and tegmentum (Fig. 4d–g, respectively), indicating both neurochemical and behavioral rescue in Tph2−/− mice.
Because of the nature of the behavioral tests used to score impulsivity, it is possible that the results can be confounded by co-morbid anxiety-like behavior. Therefore, Tph2−/− mice were tested in the light-dark box and the elevated plus maze, two extensively validated tests of anxiety-like behaviors (Treit et al. 2010). Tph2−/− mice actually showed less anxiety-like behavior in the light-dark test (Figs. 5a and 5b) while performing the same as wild-type mice on the elevated plus maze (Fig. 5c). In addition, periods of inactivity and total distance traveled on the elevated plus maze did not differ between genotypes, indicating that Tph2−/− mice were not hyperactive and did not exhibit increased exploratory behaviors during these tests (Fig. 5d).
To determine if there were any compensatory neuroarchitectural changes in the 5HT system in Tph2−/− animals, we performed a variety of neurochemical and functional analyses. We (Thomas et al. 2010) and others (Gutknecht et al. 2008) have previously shown that 5HT neurons are intact in the brains of Tph2−/− animals. Injections of 5HTP lead to the synthesis of 5HT in 5HT neurons of the dorsal raphe nuclei (Fig. S2a). The expression (Figs. S2b–e) and function (Fig. S2f) of the SERT and selected 5HT receptors (Figs. S3a–i) was unchanged in Tph2−/− mice by comparison to wild-type controls. While not expressed selectively within 5HT neurons or their projections, expression levels of monoamine oxidase A (MAO-A) and neuronal nitric oxide synthase (nNOS) were determined because reductions in their levels are associated with increased aggression (Cases et al. 1995, Nelson et al. 1995, Chiavegatto et al. 2001). We found that the levels of both MAO-A (Fig. S4) and nNOS (Fig. S5) were unchanged except in the frontal cortex where MAO-A levels were slightly but significantly reduced in Tph2−/− mice (Figs. S4e).
Alterations in 5HT neurochemical function have long been linked to numerous psychiatric conditions including depression, OCD, suicide and anxiety (Gingrich and Hen 2001, Perez-Rodriguez et al. 2010), to list but a few. Strategies to broaden the understanding of how 5HT dysfunction contributes to psychiatric disorders have depended historically on pharmacological modification of the 5HT neuronal system in animal models. While the pharmacological approach has yielded several important advances, the use of non-selective drugs and neurotoxins can make interpretation of the data challenging (see in particular discussion in Miyazaki (Miyazaki et al. 2011) with regard to the role of 5HT in impulsivity). Similarly, a number of groups working independently have produced mouse strains with primary alterations in the 5HT neuronal system to include knockouts of Tph2 (Savelieva et al. 2008, Alenina et al. 2009, Liu et al. 2011, Gutknecht et al. 2008, Yadav et al. 2009, Thomas et al. 2010), VMAT2 (Narboux-Neme et al. 2011) and Pet-1 (Hendricks et al. 2003). The behavioral phenotypes emerging in these mice are not always consistent adding to the complexity in defining the role of 5HT in neuropsychiatric disorders. For example, Pet-1 knockout mice lack ~80% of their 5HT neurons (Hendricks et al. 2003) and show an aggressive phenotype as documented for Tph2−/− mice presently and by others (Alenina et al. 2009), but they reportedly show increased (Hendricks et al. 2003), decreased (Kiyasova et al. 2011) or no anxiety-like behavior (Schaefer et al. 2009).
In light of the difficulty in isolating the 5HT system for investigation using drugs or indirect genetic models, we have approached the study of this neurochemical system by targeting for deletion the initial and rate-limiting enzyme in 5HT synthesis, Tph2. We have shown previously that mice lacking the gene for Tph2 do not have Tph2 protein or enzymatic activity in brain. Consequently, they lack the ability to synthesize 5HT and their brains are devoid of this important neurotransmitter (Thomas et al. 2010). Therefore, the Tph2−/− mouse serves as a powerful model in which the behavioral consequences of a lack of brain 5HT on behavior can be investigated. The resulting behavioral phenotype of the Tph2−/− mouse manifests itself in intense compulsive and impulsive behaviors in a context-dependent manner. Tph2−/− mice engage in compulsive shredding and digging when presented with relevant objects (i.e., nestlets, marbles). These mice even bury familiar objects such as food pellets from their home cages. The behavior that penetrates into each of the above-mentioned tasks is a purposeless digging that is part of the “natural” behavioral repertoire of these mice. In line with the postulation that impulsivity can exist in two distinct forms (Brunner and Hen 1997) reflecting an inability to withhold a motor response (motor impulsivity) or the inability to delay gratification (cognitive impulsivity), we tested mice for both forms. It is very clear that the Tph2−/− mouse exhibits extreme motor but not cognitive impulsivity, manifested as impulsive-aggression and the near-immediate approach of an object in a novel environment (novelty suppressed feeding test) and emergence from a preferred enclosed dark location into a lighted compartment without delay (dark emergence test). These behaviors are highly uncharacteristic of wild-type mice. Importantly, these behaviors are not confounded by co-morbid anxiety-like behaviors as Tph2−/− mice score less anxious than wild-type controls on independent and well-accepted animal tests of anxiety.
As part of their impulsive phenotype, the Tph2−/− mouse also shows intense aggression that emerges as early as 3–4 weeks of age and increases in intensity as the animals mature to adulthood. Aggression as seen in male-male pairings is expected but is very unusual in females and weanlings (Pinna et al. 2005). Use of the resident/intruder test, which involves prolonged isolation housing to elicit aggression in wild-type male mice, readily results in aggression in male Tph2−/− mice. However, socialized male Tph2−/− mice (i.e., housed with other males) caged singly for only 30 min also show intense aggression toward socialized wild-type intruders. Aggression can serve in a defensive and survival capacity, which is normal, or it can be elicited in an impulsive and pathological form that is very destructive, particularly in humans (Nelson and Trainor 2007, Popova 2006). The impulsive aggression shown by Tph2−/− mice of both sexes is also context dependent and is part of a set of maladaptive behaviors. Other knockout models such as the 5HT1B receptor (Saudou et al. 1994), MAO-A (Cases et al. 1995), nNOS (Chiavegatto et al. 2001, Nelson et al. 1995) and Pet-1 (Hendricks et al. 2003), and a knockin expressing a low-activity form of Tph2 (Beaulieu et al. 2008) also show increased aggression which was linked to perturbations in the 5HT neuronal system. However, these mice do not express other compulsive and impulsive behaviors seen in Tph2−/− mice. Tph2−/− mice express normal levels of at least MAO-A, nNOS and the 5HT1B receptor, ruling out these factors in the impulsive aggression seen presently. We did find that testosterone levels were increased in adult male Tph2−/− mice showing the highest degree of aggressive behavior. Studies have linked increased testosterone with impulsive aggression, as orchiectomized males exhibit low levels of aggression (Pinna et al. 2005), supporting a potential interaction between testosterone and 5HT in mediating impulsive aggressive behavior, albeit through an unknown molecular mechanism (Chiavegatto et al. 2001). Taken together, the phenotype of the Tph2−/− mouse can be characterized most accurately as BD.
The BD phenotype of Tph2−/− mice can be attributed largely to a lack of brain 5HT. Restoration of 5HT by treating mice with 5HTP significantly reduces both compulsive and impulsive behaviors. Parallel neurochemical and immunohistochemical studies establish that the 5HT neuronal system is fully intact in Tph2−/− mice. The SERT (levels and function) and 5HT receptor subtypes 1A, 1B, 2A and 2C are expressed at wild-type levels throughout brain. Therefore, acute restoration of brain 5HT in the Tph2−/− mouse is necessary and sufficient to rescue the BD phenotype and suggests that developmental deficiencies are not responsible. This finding is surprising and challenges the long-standing hypothesis that 5HT is important for brain development (Whitaker-Azmitia et al. 1996, Daubert and Condron 2010), in particular at the time when the source of forebrain 5HT switches from an exogenous (i.e., placental) to an endogenous source around embryonic day 16.5 (Bonnin et al. 2011). Others have also recognized the fact that genetic depletion of 5HT results in minimal effects on the developing brain (Trowbridge et al. 2011).
The present finding that mice lacking brain 5HT exhibit a profound phenotype of BD is consistent with existing evidence suggesting involvement of 5HT in the pathophysiology of OCD (Albelda and Joel 2012) and impulse control disorders (Evenden 1999, Cardinal 2006). To the best of our knowledge, the Tph2−/− mouse is the first genetic model to express the complex phenotype of BD with the combination of robust compulsivity and motor impulsivity, reflecting the high co-morbidity of OCD and impulse control disorders in humans (Fineberg et al. 2010, Grant and Potenza 2006). Other constitutive knockout mouse models such as Slitrk5 (Shmelkov et al. 2010), Sapap3 (Welch et al. 2007), Hoxb8 (Greer and Capecchi 2002) and Shank3 (Peca et al. 2011) express compulsive behaviors but primarily in the form of excessive, self-injurious grooming. In at least the case of the Hoxb8 knockout (Greer and Capecchi 2002), the compulsive grooming is attributable to spinal cord sensory abnormalities. The other models mentioned above involve transmembrane, scaffolding and postsynaptic proteins that are expressed throughout the CNS, making it hard to pinpoint a specific pathological defect behind compulsive behaviors in these mice. Therefore, the Tph2−/− mouse has high face, construct and predictive validity as a translational model for studying the neuropathological mechanisms underlying BD and in the development of novel therapies.
In summary, the present study shows that mice with genetic deletion of Tph2 exhibit a complex phenotype of BD characterized by compulsive, impulsive, and extreme aggressive behaviors. Our findings should stimulate future studies into the role of brain 5HT in the development and maintenance of the neural circuitry mediating this broad and debilitating category of neuropsychiatric dysfunction.
This work was supported by the Department of Veterans Affairs and by National Institutes of Health grants R01-DA010756 and R01-DA017327.
The authors declare that they have no conflict of interest related to the publication of this article.