|Home | About | Journals | Submit | Contact Us | Français|
Obsessive-compulsive disorder (OCD) is a common psychiatric disorder defined by the presence of obsessive thoughts and repetitive compulsive actions, and it often encompasses anxiety and depressive symptoms1,2. Recently, the corticostriatal circuitry has been implicated in the pathogenesis of OCD3,4. However, the etiology, pathophysiology and molecular basis of OCD remain unknown. Several studies indicate that the pathogenesis of OCD has a genetic component5–8. Here we demonstrate that loss of a neuron-specific transmembrane protein, SLIT and NTRK-like protein-5 (Slitrk5), leads to OCD-like behaviors in mice, which manifests as excessive self-grooming and increased anxiety-like behaviors, and is alleviated by the selective serotonin reuptake inhibitor fluoxetine. Slitrk5−/− mice show selective overactivation of the orbitofrontal cortex, abnormalities in striatal anatomy and cell morphology and alterations in glutamate receptor composition, which contribute to deficient corticostriatal neurotransmission. Thus, our studies identify Slitrk5 as an essential molecule at corticostriatal synapses and provide a new mouse model of OCD-like behaviors.
There are several disorders that have OCD-like clinical manifestations9, such as obsessive-compulsive disorder, Gilles de la Tourette’s syndrome and trichotillomania. Recent human genetic analyses have linked the SLITRK1 gene to Tourette’s syndrome10, although the underlying mechanisms are not well understood. The Slitrk1 gene belongs to a new family of six members (Slitrk1–Slitrk6) encoding one-pass transmembrane proteins that contain two extracellular leucine-rich repeat domains, similar to Slit proteins, and a carboxy-terminal domain that is similar to Trk neurotrophin receptors. These proteins have been shown to affect neuronal process outgrowth11,12. Slitrk1-knockout mice show increased anxiety-like behaviors but do not show any other behavioral abnormalities13. Although little is known about Slitrk1, the function of other members of the Slitrk family remains even more obscure. By gene expression fingerprinting, we have previously identified Slitrk5 in hematopoietic progenitors14. Subsequently, we demonstrated that human SLITRK5 is expressed in leukemias, embryonic stem cells and subsets of endothelial cells15. However, the Slitrk5 gene is expressed predominantly in neural tissues12.
We hypothesized that abnormal expression of Slitrk5 may lead to behavioral phenotypes similar to the involvement of SLITRK1 in Tourette’s syndrome. To investigate the function of this protein and to delineate the expression pattern of the Slitrk5 gene in mouse tissues, we decided to create a knockout mouse by replacing the Slitrk5 gene with a reporter gene. Analysis of the genomic structure of the Slitrk5 gene revealed that the coding region is localized to a single exon. Using Velocigene technology16, we replaced the entire encoding exon with the lacZ gene (Fig. 1a). Expression analysis of lacZ showed that Slitrk5 is widely expressed throughout the central nervous system, including the cortex and striatum (Fig. 1b). Double staining for the neuronal marker NeuN showed that in the brain Slitrk5 expression is restricted to neurons and that the majority of neurons express Slitrk5 (Fig. 1c).
Slitrk5−/− mice were born in accordance with Mendelian distribution. Gross anatomical and thorough histological examination of young Slitrk5−/− mice did not show any abnormalities. However, analysis of older Slitrk5−/− mice revealed a behavioral phenotype. Starting at 3 months of age, Slitrk5−/− mice developed facial hair loss and severe skin lesions. Over time, these lesions produced ulcerations with hemorrhage (Fig. 2a). The penetrance of this phenotype increased with age, and most of the knockout as well as the heterozygous mice were affected. The lesions in heterozygous mice were similar to those in homozygous mice, but their emergence was delayed by 7–9 months. We hypothesized that this phenotype could be the result of excessive grooming. We did not find the lesions in the wild-type littermates, even when they were housed in the same cage with Slitrk5−/− mice, indicating that this phenotype can be attributed to self-grooming. This type of behavior is similar to that previously observed in mice deficient for the Sapap3 gene17. Targeted deletion of this gene, which encodes a postsynaptic scaffold protein, leads to compulsive overgrooming behavior and increased anxiety, which are ameliorated by selective serotonin reuptake inhibitors17.
We assessed the grooming behavior of Slitrk5−/− mice by counting the number and duration of grooming events in the knockout and wild-type littermates before any lesions or hair loss developed to exclude the possibility that the overgrooming was due to irritation in a wound area. Our data show a significant increase in the duration of grooming events in Slitrk5−/− mice as compared to their wild-type littermates (Fig. 2b).
As OCD is linked to a deficit in serotonin production, and because selective serotonin reuptake inhibitors (SSRIs) are the major therapeutic agents for OCD, we sought to test the effect of chronic administration of the SSRI fluoxetine on overgrooming behavior in Slitrk5−/− mice. Indeed, treatment of Slitrk5−/− mice with fluoxetine led to a significant (P = 0.0009) reduction in the duration of grooming compared to pretreated mice (Fig. 2b). The duration of grooming in Slitrk5−/− mice after fluoxetine treatment was the same as in wild-type litter-mates (Fig. 2b). The duration of grooming events in wild-type mice was not affected by fluoxetine (Fig. 2b). Thus, treatment of Slitrk5−/− mice with an SSRI prevents their compulsive behavior.
To determine whether Slitrk5−/− mice show additional behavioral phenotypes that also occur in OCD-related conditions, we assessed anxiety-like behaviors in these mice. We performed the elevated-plus-maze and the open-field tests, standard measures of anxiety-like behavior that place the mice in conflict situations. In comparison with wild-type littermate mice, Slitrk5−/− mice showed a lower percentage of time spent in the center compartment and a lower number of entries into the center compartment in the open-field test (Fig. 2c), and they showed reduced time spent in open arms in the elevated-plus-maze test (Supplementary Fig. 1a). This reduction in exploration could not be explained by changes in locomotor activity, as there were no significant differences in total distance traveled. To further assess the behavioral consequences of Slitrk5 inactivation, we also tested Slitrk5−/− mice in a marble-burying paradigm, a behavioral task that assesses both OCD-like and anxiety-related behaviors. We found that Slitrk5−/− mice showed an increase in marble-burying behavior (Supplementary Fig. 1b), which is consistent with our findings that this knockout mouse models core symptoms in OCD spectrum disorders. We also assessed motor function in Slitrk5−/− mice by using the cylinder test and by measuring the latency to fall from a rotarod and found no difference in gross motor skills and no impairment in motor learning compared to wild-type mice, indicating that these functions are not affected in Slitrk5−/− mice (Supplementary Fig. 2 and Supplementary Fig. 3).
Because corticostriatal circuitry has been previously implicated in the pathogenesis of OCD, we performed detailed anatomical, histological and functional analyses of cortex and striatum in Slitrk5−/− mice. Initially, we evaluated the difference in baseline activity of selected brain regions between wild-type and Slitrk5−/− mice by assessing expression of FosB, an established marker for neural activity18. We found that FosB was upregulated exclusively in the orbitofrontal cortex of Slitrk5−/− mice (Fig. 3a,b). Other brain regions such as the caudate putamen, hippocampus and thalamus did not show upregulation of FosB expression (Supplementary Fig. 4). These findings are particularly noteworthy, as it has been consistently shown in functional imaging studies that there is an increase in orbitofrontal cortex activity in individuals with OCD4,19–21. Conversely, alterations in neural activity in the caudate or thalamus have been less consistently found in OCD19,21,22. Next, we measured the volume of the striatum relative to the whole-brain volume by Cavalieri estimation. Our data showed that the volume of striatum in Slitrk5−/− mice was significantly reduced compared to wild-type mice (Fig. 3c). In both young and aged Slitrk5−/− mice, the ratio of striatal volume to the total brain volume was decreased compared to wild-type mice, whereas volume ratios of other brain structures, such as the dorsal hippocampus, to the total brain volume were not changed, indicating that the anatomy of striatum is specifically affected by Slitrk5 deficiency (Fig. 3c and Supplementary Fig. 5). In line with these data, it has previously been reported that the volume of the striatum is decreased in some individuals with OCD23–25. However, this finding has not been consistent across all studies of individuals with OCD in which increased or no change in striatal volumes have been reported19,21,22.
Because Slitrk family members have been shown to influence neuronal differentiation10,12, the decreased striatal volume in the Slitrk5−/− mice might be accounted for by altered neuronal morphology. We used Golgi staining to visualize individual medium spiny neurons of the striatum in Slitrk5−/− mice. There was no difference in striatal cell soma area between Slitrk5−/− mice and their wild-type littermates. Next, we analyzed dendritic complexity in the same neurons. Sholl analysis revealed a decrease in dendritic arbor complexity at 50-μm and greater distances from the soma in Slitrk5−/− mice (Fig. 3d,e). We also used fractal dimension analysis to quantify how completely a neuron fills its dendritic field. There was a significant decrease in dendritic complexity of striatal neurons in Slitrk5−/− mice (Fig. 3f). Although the striatum contains two equally abundant subpopulations of medium spiny neurons, which are classified on the basis of the neuropeptides that they produce and the dopamine receptors that they express (D1 and D2), distinguishing between these two types of cells is technically challenging26. However, in our detailed comparative analysis of 40 randomly selected medium spiny neurons in Slitrk5−/− mice, we found no evidence for a bimodal distribution in their dendritic complexity (Fig. 3e,f). These data suggest that there is no selective deficit of arborization in one subpopulation of medium spiny neurons but rather a general deficit in all medium spiny neurons. Sholl and fractal dimension analyses of neurons in other brain regions with high Slitrk5 expression such as dentate granular cells showed no difference in dendritic branching complexity (Supplementary Fig. 6).
We subsequently assessed the cellular localization of Slitrk5 in striatal neurons and found Slitrk5 in dendritic spines that are positive for post-synaptic density protein-95 (PSD95) in cocultures of cortical neurons, isolated from transgenic mice that express enhanced GFP under the control of the human ubiquitin C promoter, and rat striatal neurons infected with Flag-Slitrk5 lentivirus and transfected with PSD95 fused to mCherry (Fig. 4a). Next, we examined the expression of glutamate receptors in the striatum and found that they are downregulated in Slitrk5−/− mice (Fig. 4b). Indeed, protein amounts of glutamate receptor subunits NR2A, NR2B, GluR1, and GluR2 were decreased by 20–60%, with no significant changes in PSD95 amounts (Fig. 4b). We found these changes in both the total lysates (Fig. 4b) and in PSD-enriched fractions of synaptosomes (Supplementary Fig. 7).
Given these findings, we investigated whether Slitrk5−/− mice have deficits in corticostriatal neurotransmission by extracellular recordings in acute striatal slices. We recorded population spikes from striatum by stimulating the white matter between cortex and striatum. We found a significantly reduced population spike amplitude in Slitrk5−/− mice (Fig. 4c). We did not observe any difference in paired-pulse ratios of the population spike in Slitrk5−/− mice and their wild-type littermates, suggesting that the presynaptic mechanism involved in paired-pulse facilitation is not responsible for the observed difference in population spike amplitude (Fig. 4d). Also, we did not observe any significant difference in fiber volley amplitude, suggesting that cortical axon input is normal in Slitrk5−/− mice (Supplementary Fig. 8).
Taken together, our data demonstrate that targeted inactivation of Slitrk5 in mice leads to OCD-like behavioral phenotypes, including overgrooming with elements of self-mutilation. Although Slitrk5 expression is widespread in the central nervous system, we found increased neuronal activity specifically in the orbitofrontal cortex of Slitrk5−/− mice, which is consistent with functional imaging findings in humans with OCD that implicated dysregulation of corticostriatal circuitry23,27, and which has not been reported in previous mouse models of OCD28,29. In addition, Slitrk5−/− mice have anatomical deficits in the striatum, such as reduced striatal volume, as well as decreased dendritic complexity of striatal medium spiny neurons. Although this region has not been consistently found to be altered anatomically in people with OCD19,21,22, emerging literature suggests that striatal dysfunction may underlie behavioral deficits in individuals with OCD27. In this context, it has recently been postulated that striatal dysfunction, in the presence of orbitofrontal cortex over-activation, could lead to deficits in thalamic filtering or imbalance in the direct and indirect pathways of the basal ganglia30. Given the ubiquitous neuronal expression of Slitrk5, this selective effect on the orbitofrontal cortex and on striatal neurons is hard to explain. On the one hand, it is reminiscent of the effect of other proteins such as huntingtin, which is also widely expressed in the central nervous system, but alterations in the huntingtin protein result in functional defects predominantly in striatal neurons, directly leading to Huntington’s disease pathology31. On the other hand, it is possible that Slitrk5 may form a signaling complex with corticostriatal-specific proteins, which may explain these region-specific effects.
Overall, our data suggest that Slitrk5 may have a central role in the development of OCD-like behaviors. Although human genetic studies have implicated another Slitrk family member, SLITRK1, in Tourette’s syndrome, these associations have not been consistently replicated32,33. In this context, our studies link Slitrk5 to the core symptoms of OCD: self-injurious repetitive behavior and increased anxiety. In all, we provide a new mouse model of OCD-like behaviors, involving a previously uncharacterized neuronal transmembrane protein that modulates region-specific glutamatergic neurotransmission. This model can be used to further dissect the role of Slitrk5 in molecular pathways underlying the pathogenesis of obsessive-compulsive behaviors.
Methods and any associated references are available in the online version of the paper at http://www.nature.com/naturemedicine/.
We thank M. Flint Beal for providing expertise on behavioral experiments. We thank G. Thurston for critical comments and suggestions. PSD95-cherry was a kind gift from R.H. Edwards (University of California–San Francisco). We acknowledge support from US National Institutes of Health grants MH079513 (F.S.L.) and NS052819 (F.S.L.), HL66592 (S.R.), HL097797 (S.R.) and AI080309 (S.R.), Burroughs Wellcome Foundation (F.S.L.), International Mental Health Research Organization (F.S.L.), the Sackler Institute (K.G.B., F.S.L.), DeWitt-Wallace Fund of the New York Community Trust (F.S.L.), Pritzker Consortium (F.S.L.), National Alliance for Research on Schizophrenia and Depression (S.V.S.), Mildred-Scheel-Stiftung, Deutsche Krebshilfe (T.M.), Gulbenkian PhD Programe in Biomedicine (C.C.P.), Fundacao para Ciencia e Tecnologia (C.C.P.), Howard Hughes Medical Institute (S.R.), Ansary Stem Cell Institute (S.R.), Anbinder and Newmans Own Foundations (S.R.), Qatar National Priorities Research Program (S.R.), Empire State Stem Cell Board (S.R.) and the New York State Department of Health grant NYS C024180 (S.R.).
Note: Supplementary information is available on the Nature Medicine website.
AUTHOR CONTRIBUTIONSS.V.S. conceived of and designed the study, performed experiments, analyzed data and wrote the manuscript; A.H., D.J, C.C.P. and K.G.B. designed and performed experiments, analyzed data and assisted in writing the manuscript; T.M., E.S., J.S.K., M.B. and I.D. performed experiments and analyzed data; A.J.M., D.M.V., N.W.G. and G.D.Y. designed and generated the Slitrk5−/− mice; I.N. designed, performed and analyzed electrophysiology experiments; F.S.L. and S.R. conceived of and designed the study and wrote the manuscript.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.