Inducible nitric oxide synthase (NOS2) knockout mice as a model of trichotillomania

Introduction

The cortico-striato-thalamo-cortical circuitry (CSTC) is a series of reverberatory loops that control motor programs. The glutamatergic neurotransmission plays a crucial role in both afferent and efferent pathways in prefrontal cortex and striatum, two core structures in the regulation of CSTC (Langen et al., 2011a; Langen et al., 2011b). In this scenario, nitric oxide (NO) plays a major part, not only regulating the release of glutamate but also in the migration and maturation of neuronal cells in the cortex. This later feature of NO is apparently mediated by the inducible isoform of nitric oxide synthase (NOS2) (Arnhold et al., 2002).

Malfunctions of specific parts of the CSTC are speculated to be involved in the neurobiology of several motor and impulse control disorders, such as obsessive-compulsive disorder, Tourette’s syndrome, and of interest to the present study, trichotillomania (Langen et al., 2011b; Langen et al., 2011a).

Trichotillomania (TTM) is an impulse control disorder characterized by a distressing urge for hair pulling or trimming, affecting 1–3.5% of the population (Christenson, Pyle & Mitchell, 1991; American Psychiatric Association, 2013). Nowadays, the pharmacotherapy for TTM is based on treatment with the serotonin reuptake inhibitor clomipramine (Swedo et al., 1989; Swedo, Lenane & Leonard, 1993; McGuire et al., 2014). Recent studies also indicate that the glutamatergic neurotransmission modulator, N-acetyl-cysteine, is effective in alleviating the symptoms of TTM (Grant, Odlaug & Kim, 2009). Other drugs, such as atypical antipsychotic olanzapine, have shown success in placebo-controlled trials (Van Ameringen et al., 2010).

Neuroimaging studies indicated increased gray matter volume in striatum, amygdalo-hippocampal formation, and multiple cortical regions (including the cingulate, supplemental motor cortex, and frontal cortex) of patients with TTM (Chamberlain et al., 2009). This increase in gray matter was associated with a relatively decreased activity in cortical structures, and hyperactivity in the striatal region (Fineberg et al., 2010). Despite these findings, the neurobiology of TTM is not completely understood.

Animal models for the investigation of neurobiological aspects of TTM, as well as the search for new therapies, rely on the observation of barbering behaviors (BB). In laboratory and captive animals, BB is characterized by pulling or trimming of fur or whiskers (Garner et al., 2004). Some transgenic strains, e.g., SAPAP3 and SLITRK5 knockouts, display compulsive/repetitive behavior, like excessive grooming (Welch et al., 2007; Shmelkov et al., 2010), however no specific BB has been described for them. Interestingly, SAPAP3 is a structural protein crucial for the proper docking of glutamatergic receptors (Welch et al., 2007).

In the present study, we describe the BB observed in NOS2 KO mice, and the effects of chronic treatment with clomipramine or memantine (a noncompetitive antagonist of NMDA receptors), as well as the partial rescue of NOS2 expression in haploinsufficient mice.

Methods

Experimental design

Results

Time course of drug treatment effect in BB

As shown in Figs. 1B–1D, there is a significant effect of clomipramine treatment on BB in both males [F(1, 18) = 27.03; p < 0.0001] and females [F(1, 8) = 29.04, p = 0.0007]. The post hoc analysis (Fisher’s LSD) indicates a significant difference between clomipramine and vehicle treated groups at days 10, 15, and 20 for both males (p < 0.0001 for all) and females (p = 0.0011; 0.0038 and 0.0038, respectively). The treatment with memantine, only performed in females, was also effective on BB [F(1, 12) = 38.69, p < 0.0001], with significant difference from controls observed at days 10, 15 and 20 (p = 0.0024 for all).

Repetitive behaviors in NOS2 KO and WT mice

Student’s t test indicates a significant effect of genotype, with NOS2.KO females exhibiting reduced total activity [t(11) = 4.79, p = 0.0006]. However, an increase in the percentage of repetitive movements was observed [t(11) = 2.69, p = 0.021], as found in Figs. 2A and 2B.

Discussion

The present study suggests that the increased barbering behavior (BB) of NOS2 KO mice could be a putative model of trichotillomania (TTM). NOS2 KO exhibited increased BB after weaning (significantly higher at 4th week-of-age), and this phenotype increases until adulthood. TTM is an impulse control disorder, characterized by hair pulling or trimming. To our knowledge, there is no established animal model to study this disorder, which reflects in the poor pharmacotherapy and interventions available (Johnson & El-Alfy, 2016). In animals, BB is a rare phenotype, with an incidence of approximately 2.5% in laboratory mice (Garner et al., 2004). In our facilities, BB frequency was even lower (less than 0.5% out of 500 animals checked). However, all NOS2 KO animals exhibited some degree of BB. This phenotype was described in the manufacturer’s webpage (https://www.jax.org/strain/002596), and all attempts to eradicate BB in the colony resulted in loss of the genotype.

One of the few pharmacological options to treat TTM is clomipramine, a tricyclic antidepressant drug that preferentially inhibits the serotonin reuptake system (Swedo et al., 1989; Swedo, Lenane & Leonard, 1993). In line with currently available data, our results indicate that repeated clomipramine administration was able to prevent the evolution of BB in NOS2 KO. Both males and females were equally responsive to the drug treatment. A possible confounding factor in the analysis of BB is the allobarbering instead of putatively selfbarbering. In the present case, we observed an incidence of 100% of BB in NOS2 KO in our colony; however, parts of the animal body only accessible by a conspecific, such as the back, were found barbered. This led us to conclude that the BB observed in these animals is a mixture of self- and allobarbering.

TTM is comorbid to several other psychiatric disorders, such as anxiety and obsessive-compulsive disorder (Grant et al., 2017). To investigate this association in NOS2 KO, the animals were submitted to the actimeter test, where both the total activity and the presence of spontaneous repetitive behavior were analyzed. Our results indicate a decreased activity of NOS2 KO associated with an increase in repetitive movements compared to WT animals. Interestingly, clomipramine was only effective in reducing the last parameter, with no effect on total activity. Despite reduced activity, this strain exhibits an antidepressant-like phenotype (Montezuma et al., 2012), and a deficit in extinction of conditioned fear (Lisboa et al., 2015), but no changes in repetitive behaviors was described. This mouse strain also presents decreased lung tumorigenesis induced by urethane (Kisley et al., 2002). On the other hand, the lack of NOS2 does not change the animal’s survival rate to bacterial lipopolysaccharide (LPS) (Laubach et al., 1995).

Compulsive-like behaviors have been observed in many transgenic animals. For example, Francis Lee’s group described increased grooming behavior in SLITRK KO mice (Shmelkov et al., 2010). Rodents lacking the adaptor protein SAPAP3 also presented increased grooming (Welch et al., 2007). Notwithstanding, this protein is responsible for anchoring glutamatergic receptors in the cell membrane and multiple rare missense variants in its gene sequence were found in OCD- and TTM-suffering patients (Züchner et al., 2009). Two other studies pioneered the analysis of BB in transgenic mice, and described increased grooming and BB in animals lacking the transcription factor HOXB8 (Greer & Capecchi, 2002; Chen et al., 2010). These two articles indicates that HOXB8 plays a major role in the excessive grooming phenotype, and is mainly expressed in components of CSTC. This is also an interesting possibility for NOS2. This enzyme level is highly modulated in cultured glial cells after several stimuli, such as LPS and interferon (Galea, Feinstein & Reis, 1992; Garthwaite & Boulton, 1995). Interestingly, the studies from Mario Capecchi’s group points that the HOXB8 is expressed in bone-marrow-derived microglial cells. Thus in an intriguing speculative scenario, glial cells play a central role in the maturation and function of the CSTC components, leading to BB or excessive grooming. However, it is important to highlight that knocking out of NOS2 apparently does not lead to excessive grooming at the same extension of HOXB8 deletion. It is clear in the studies from Greer & Capecchi (2002) as well as in Chen and colleagues (2010) that the animals exhibited extensive lesions in the snout area. This feature was absent in the NOS2 KO, and the test used in the present study (actimeter) does not address if the increase in repetitive movements is a consequence of increased grooming behavior.

Our understanding of the neurobiology of TTM relies on few studies using brain image techniques suggesting a decrease in frontal cortical areas (mainly inferior frontal cortex) associated with an increase in striatal activity (Chamberlain et al., 2009; Fineberg et al., 2010). As previously mentioned, the cortico-striatal-thalamic circuitry (CSTC) is a series of reverberatory loops that integrate external signals and trigger motor programs in response. One of the integrative structures in this circuitry is striatum, including its multiple subdivisions, and disturbances in particular “hubs” in the CSTC are associated with different pathologies, from Huntington’s disease to obsessive-compulsive disorder, including trichotillomania (for review see Langen et al., 2011b; Langen et al., 2011a). As a proof of concept, the treatment with the NMDA antagonist memantine was equally effective as clomipramine, in preventing the evolution of BB in NOS2 KO mice. Taken together, our results suggest a potential therapeutic approach for TTM with compounds acting on the glutamatergic system. In fact, drugs acting on decreasing glutamatergic neurotransmission have been tested in TTM patients successfully. For example, N-acetyl-cysteine was found to be effective in alleviating TTM symptoms in both double blind placebo-controlled and case studies (Coric et al., 2007; Grant, Odlaug & Kim, 2009; Rodrigues-Barata et al., 2012). This compound also exhibited promising results in OCD suffering patients (for review see Oliver et al., 2015). Riluzole, another compound with antiglutamatergic properties, was also found effective in OCD-related disorders (Emamzadehfard et al., 2016).

Data from literature linking NOS2 to putative neuronal developmental processes that could lead to malfunctions in CSTC is also scarce. One of the few pieces of evidence about this topic was provided by Arnhold and colleagues (2002) describing the role of NOS2 in the early stages of neuronal maturation. They depicted a functional part for this enzyme in the cell differentiation and migration in embryonic cortex, although not through the canonical pathway of NO acting on soluble guanylate cyclase. Cultured cells from E14 rat embryo, when incubated with selective NOS2 inhibitors, showed a lower number of microtubule-associated protein (MAP2)-labeled neurons. Similarly, our data indicates that silencing NOS2 expression, leads to a reduction in the levels of MAP2 in activity-dependent differentiation of PC12 cells Banno et al. 2008. The role of NO in activity-dependent differentiation of PC12 cells has been described (for example Nakagawa, Yoshida & Miyamoto, 2000); however, the NOS1 isoform is implicated as being responsible for this process. This discrepancy may be due to the culture conditions between the studies. For example, Nakagawa and colleagues used a serum-free medium and 45 mM of KCl to differentiate the cells, while we rely on 0.5% FBS and 25 mM of KCl. Moreover, in the former study the authors cultivated a low density of PC12 cells (approximately 50 cells/cm²; in a 96-well plate), while we used a high density from the beginning of the experiment—125,000 cells/cm² (in a 12-well plate).

Conclusion

Based on our data and the evidence in literature, we speculate that increased BB observed in NOS2.KO is putatively a resultant of inefficient inhibition of striatal motor programs by PFC. The lack of NOS2 in the early stages of cortical development could affect the maturation of cortical neurons, which in turn would lead to the described phenotype. The BB observed in these animals is sensitive to treatment with clomipramine and memantine, strengthening its role as a putative model of TTM. Further experiments will be necessary to clarify the role of NOS2 in the development of the BB phenotype and its functional consequences in the CSTC; as well as validate its sensitivity to other drugs.