Slitrk5 deficiency impairs corticostriatal circuitry and leads to obsessive-compulsive like behaviors in mice

Slitrk5 defiieny impairs ortiostriatal iruitry and leads to obsessive-ompulsive like behaviors in mie Sergey V Shmelkov 1,9, Adília Hormigo 1 3,9, Deqiang Jing,9, Catia C Proena,5,9, Kevin G Bath, Till Milde 1, Evgeny Shmelkov 1, Jared S Kushner 1, Muhamed Baljevi 1, Iva Dinheva,, Andrew J Murphy 7, David M Valenzuela 7, Niholas W Gale 7, George D Yanopoulos 7, Ipe Ninan, Franis S Lee, & Shahin Rafii 1 1 Nature Ameria, In. All rights reserved. Obsessive-ompulsive disorder (OCD) is a ommon psyhiatri disorder defined by the presene of obsessive thoughts and repetitive ompulsive ations, and it often enompasses anxiety and depressive symptoms 1,. Reently, the ortiostriatal iruitry has been impliated in the pathogenesis of OCD 3,. However, the etiology, pathophysiology and moleular basis of OCD remain unknown. Several studies indiate that the pathogenesis of OCD has a geneti omponent 5. Here we demonstrate that loss of a neuron-speifi transmembrane protein, SLIT and NTRK-like protein-5 (Slitrk5), leads to OCD-like behaviors in mie, whih manifests as exessive selfgrooming and inreased anxiety-like behaviors, and is alleviated by the seletive serotonin reuptake inhibitor fluoxetine. Slitrk5 / mie show seletive overativation of the orbitofrontal ortex, abnormalities in striatal anatomy and ell morphology and alterations in glutamate reeptor omposition, whih ontribute to defiient ortiostriatal neurotransmission. Thus, our studies identify Slitrk5 as an essential moleule at ortiostriatal synapses and provide a new mouse model of OCD-like behaviors. There are several disorders that have OCD-like linial manifestations 9, suh as obsessive-ompulsive disorder, Gilles de la Tourette s syndrome and trihotillomania. Reent human geneti analyses have linked the SLITRK1 gene to Tourette s syndrome 1, although the underlying mehanisms are not well understood. The Slitrk1 gene belongs to a new family of six members (Slitrk1 Slitrk) enoding onepass transmembrane proteins that ontain two extraellular leuinerih repeat domains, similar to Slit proteins, and a arboxy-terminal domain that is similar to Trk neurotrophin reeptors. These proteins have been shown to affet neuronal proess outgrowth 11,1. Slitrk1- knokout mie show inreased anxiety-like behaviors but do not show any other behavioral abnormalities 13. Although little is known about Slitrk1, the funtion of other members of the Slitrk family remains even more obsure. By gene expression fingerprinting, we have previously identified Slitrk5 in hematopoieti progenitors 1. Subsequently, we demonstrated that human SLITRK5 is expressed in leukemias, embryoni stem ells and subsets of endothelial ells 15. However, the Slitrk5 gene is expressed predominantly in neural tissues 1. 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 funtion of this protein and to delineate the expression pattern of the Slitrk5 gene in mouse tissues, we deided to reate a knokout mouse by replaing the Slitrk5 gene with a reporter gene. Analysis of the genomi struture of the Slitrk5 gene revealed that the oding region is loalized to a single exon. Using Veloigene tehnology 1, we replaed the entire enoding exon with the laz gene (Fig. 1a). Expression analysis of laz showed that Slitrk5 is widely expressed throughout the entral nervous system, inluding the ortex and striatum (Fig. 1b). Double staining for the neuronal marker NeuN showed that in the brain Slitrk5 expression is restrited to neurons and that the majority of neurons express Slitrk5 (Fig. 1). Slitrk5 / mie were born in aordane with Mendelian distribution. Gross anatomial and thorough histologial examination of young Slitrk5 / mie did not show any abnormalities. However, analysis of older Slitrk5 / mie revealed a behavioral phenotype. Starting at 3 months of age, Slitrk5 / mie developed faial hair loss and severe skin lesions. Over time, these lesions produed ulerations with hemorrhage (Fig. a). The penetrane of this phenotype inreased with age, and most of the knokout as well as the heterozygous mie were affeted. The lesions in heterozygous mie were similar to those in homozygous mie, but their emergene was delayed by 7 9 months. We hypothesized that this phenotype ould be the result of exessive grooming. We did not find the lesions in the wild-type littermates, even when they were housed in the same age with Slitrk5 / mie, indiating that this phenotype an be attributed to self-grooming. This type of behavior is similar to that previously observed in mie defiient for the Sapap3 gene 17. Targeted deletion of this gene, whih enodes a postsynapti saffold protein, leads to ompulsive 1 Howard Hughes Medial Institute, Ansary Stem Cell Institute and Department of Geneti Mediine, Weill Cornell Medial College, New York, New York, USA. Department of Neurology, Memorial Sloan-Kettering Caner Center, New York, New York, USA. 3 Department of Neurology, Weill Cornell Medial College, New York, New York, USA. Department of Psyhiatry, Weill Cornell Medial College, New York, New York, USA. 5 Gulbenkian PhD Programe in Biomediine, Instituto Gulbenkian de Ciênia, Oeiras, Portugal. Department of Pharmaology, Weill Cornell Medial College, New York, New York, USA. 7 Regeneron Pharmaeutials, Tarrytown, New York, USA. Department of Psyhiatry, New York University Langone Medial Center, New York, New York, USA. 9 These authors ontributed equally to this work. Correspondene should be addressed to F.S.L. (fslee@med.ornell.edu) or S.R. (srafii@med.ornell.edu). Reeived 9 Otober 9; aepted February 1; published online 5 April 1; doi:1.13/nm.15 NATURE MEDICINE ADVANCE ONLINE PUBLICATION 1

Figure 1 Targeted inativation of Slitrk5 in mie and its expression pattern in the mouse brain. (a) Genomi struture and the design of the Slitrk5-knokout, laz knok-in mouse. The entire open reading frame (ORF) is loalized to exon (Ex); exon 1 (Ex1) is nonoding. The Slitrk5-enoding region was replaed with laz downstream of the signal sequene leavage site., wild-type;, knokout. (b) X-gal staining of mouse brain tissue, showing ubiquitous expression of laz in the gray matter of the various parts of the brain, inluding ortex and striatum. Cx, ortex; St, striatum; Hp, hippoampus;, orpus allosum; Th, thalamus; Cbl, erebellum. The higher magnifiation image shows the distribution of laz-expressing ells in the striatum of the Slitrk5-knokout, laz knok-in mouse. () Immunostaining of ortex and striatum with antibodies to -galatosidase (anti -gal) and NeuN (anti-neun), indiating that the majority of neurons express Slitrk5. Ex1 Ex a Slitrk5 ORF laz b 1 3 5 7 9 1 Kb Cx St 5 m Th Hp Cbl 5 m Anti-NeuN + anti -gal Cortex Striatum 5 m 5 m 1 m 1 m 1 m 1 m 1 Nature Ameria, In. All rights reserved. overgrooming behavior and inreased anxiety, whih are ameliorated by seletive serotonin reuptake inhibitors 17. We assessed the grooming behavior of Slitrk5 / mie by ounting the number and duration of grooming events in the knokout and wild-type littermates before any lesions or hair loss developed to exlude the possibility that the overgrooming was due to irritation in a wound area. Our data show a signifiant inrease in the duration of grooming events in Slitrk5 / mie as ompared to their wild-type littermates (Fig. b). As OCD is linked to a defiit in serotonin prodution, and beause seletive serotonin reuptake inhibitors (SSRIs) are the major therapeuti agents for OCD, we sought to test the effet of hroni administration of the SSRI fluoxetine on overgrooming behavior in Slitrk5 / mie. Indeed, treatment of Slitrk5 / mie with fluoxetine led to a signifiant (P =.9) redution in the duration of grooming ompared to pretreated mie (Fig. b). The duration of grooming in Slitrk5 / mie after fluoxetine treatment was the same as in wild-type littermates (Fig. b). The duration of grooming events in wild-type mie was not affeted by fluoxetine (Fig. b). Thus, treatment of Slitrk5 / mie with an SSRI prevents their ompulsive behavior. To determine whether Slitrk5 / mie show additional behavioral phenotypes that also our in OCD-related onditions, we assessed anxiety-like behaviors in these mie. We performed the elevated-plus-maze and the open-field tests, standard measures of anxiety-like behavior that plae the mie in onflit situations. In omparison with wild-type littermate mie, Slitrk5 / mie showed a lower perentage of time spent in the enter ompartment and a lower number of entries into the enter ompartment in the open-field test (Fig. ), and they showed redued time spent in open arms in the elevated-plus-maze test (Supplementary Fig. 1a). This redution in exploration ould not be explained by hanges in loomotor ativity, as there were no signifiant differenes in total distane traveled. To further assess the behavioral onsequenes of Slitrk5 inativation, we also tested Slitrk5 / mie in a marble-burying paradigm, a behavioral task that assesses both OCD-like and anxiety-related behaviors. We found that Slitrk5 / mie showed an inrease in marble-burying behavior (Supplementary Fig. 1b), whih is onsistent with our findings that this knokout mouse models ore symptoms in OCD spetrum disorders. We also assessed motor funtion in Slitrk5 / mie by using the ylinder test and by measuring the lateny to fall from a rotarod and found no differene in gross motor skills and no impairment in motor learning ompared to wild-type mie, indiating that these funtions are not affeted in Slitrk5 / mie (Supplementary Fig. and Supplementary Fig. 3). Beause ortiostriatal iruitry has been previously impliated in the pathogenesis of OCD, we performed detailed anatomial, histologial and funtional analyses of ortex and striatum in Slitrk5 / mie. Initially, we evaluated the differene in baseline ativity of seleted brain regions between wild-type and Slitrk5 / mie by assessing expression of FosB, an established marker for neural ativity 1. We found that FosB was upregulated exlusively in the orbitofrontal ortex of Slitrk5 / mie (Fig. 3a,b). Other brain regions suh as the audate putamen, hippoampus and thalamus did not show upregulation of FosB expression (Supplementary Fig. ). These findings are partiularly noteworthy, as it has been onsistently shown in funtional imaging studies that there is an inrease in orbitofrontal ortex ativity in individuals with OCD,19 1. Conversely, alterations in neural a b 1 P <.1 Time spent grooming (s per min) 1 1 1 1 Time spent grooming (s per min) 1 P =.995 1 1 1 1 Before fluoxetine After fluoxetine Perentage time in enter P =.1 7 5 3 1 Perentage entries to enter 1 P =.7 1 1 1 Figure Faial lesions, OCD-like behavior and its alleviation with fluoxetine treatment in Slitrk5- knokout mie. (a) Phenotypi harateristi of Slitrk5 / mie: exessive grooming leads to severe faial lesions. (b) Time spent grooming in Slitrk5 / mie (n = 9) ompared to their wildtype littermates (n = ) before and after treatment with fluoxetine. Error bars depit the s.e.m. () Anxiety-related behavior of Slitrk5 / and mie in the open-field test. Perentage of time spent in the enter and entries into the enter of the open field are shown. All open-field results are presented as means s.e.m. determined from analysis of mie per genotype. ADVANCE ONLINE PUBLICATION NATURE MEDICINE

1 Nature Ameria, In. All rights reserved. Figure 3 Metaboli hanges in the ortex and anatomial defets in the striatum of Slitrk5 / mie. (a) Expression of FosB in orbitofrontal ortex by immunostaining for FosB (red) and with DAPI (blue). The top images show the distribution of FosB expression in the various layers of orbitofrontal ortex. The bottom images show a higher magnifiation of layer II of FosB immunoreativity in nulei. (b) Quantifiation of FosB expression in all layers of the orbitofrontal ortex. () Cavalieri estimation of striatal volume in Slitrk5 / and mie. (d) Examples of Golgi staining and Neuroluida reonstrution of striatal medium spiny neurons in and Slitrk5 / mie. (e) Sholl analysis of striatal medium spiny neurons in and Slitrk5 / mie. All results are presented as means s.e.m.; neurons per genotype. (f) Fratal dimension analysis of striatal medium spiny neurons in Slitrk5 / and mie. All results are presented as means s.e.m.; neurons per genotype. a III III V VI ativity in the audate or thalamus have been less onsistently found in OCD 19,1,. 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 / mie was signifiantly redued ompared to wild-type mie (Fig. 3). In both young and aged Slitrk5 / mie, the ratio of striatal volume to the total brain volume was dereased ompared to wild-type mie, whereas volume ratios of other brain strutures, suh as the dorsal hippoampus, to the total brain volume were not hanged, indiating that the anatomy of striatum is speifially affeted by Slitrk5 defiieny (Fig. 3 and Supplementary Fig. 5). In line with these data, it has previously been reported that the volume of the striatum is dereased in some individuals with OCD 3 5. However, this finding has not been onsistent aross all studies of individuals with OCD in whih inreased or no hange in striatal volumes have been reported 19,1,. Beause Slitrk family members have been shown to influene neuronal differentiation 1,1, the dereased striatal volume in the Slitrk5 / mie might be aounted for by altered neuronal morphology. We used Golgi staining to visualize individual medium spiny neurons of the striatum in Slitrk5 / mie. There was no differene in striatal ell soma area between Slitrk5 / mie and their wild-type littermates. Next, we analyzed dendriti omplexity in the same neurons. Sholl analysis revealed a derease in dendriti arbor omplexity at 5- m and greater distanes from the soma in Slitrk5 / mie (Fig. 3d,e). We also d III III V VI 5 µm 5 µm II 5 µm 5 µm Slitrk5 II e Intersetion points b FosB density (1 ells per mm 3 ) 5 15 1 5 P =.977 P =.1 P <.1 7 5 3 1 I II/III V/VI 5 5 75 P <.1 Radius (µm) Slitrk5 P <.1 1 15 15 175 5 used fratal dimension analysis to quantify how ompletely a neuron fills its dendriti field. There was a signifiant derease in dendriti omplexity of striatal neurons in Slitrk5 / mie (Fig. 3f). Although the striatum ontains two equally abundant subpopulations of medium spiny neurons, whih are lassified on the basis of the neuropeptides that they produe and the dopamine reeptors that they express (D 1 and D ), distinguishing between these two types of ells is tehnially hallenging. However, in our detailed omparative analysis of randomly seleted medium spiny neurons in Slitrk5 / mie, we found no evidene for a bimodal distribution in their dendriti omplexity (Fig. 3e,f). These data suggest that there is no seletive defiit of arborization in one subpopulation of medium spiny neurons but rather a general defiit in all medium spiny neurons. Sholl and fratal dimension analyses of neurons in other brain regions with high Slitrk5 expression suh as dentate granular ells showed no differene in dendriti branhing omplexity (Supplementary Fig. ). We subsequently assessed the ellular loalization of Slitrk5 in striatal neurons and found Slitrk5 in dendriti spines that are positive for postsynapti density protein-95 (PSD95) in oultures of ortial neurons, isolated from transgeni mie that express enhaned GFP under the ontrol of the human ubiquitin C promoter, and rat striatal neurons infeted with Flag-Slitrk5 lentivirus and transfeted with PSD95 fused to mcherry (Fig. a). Next, we examined the expression of glutamate reeptors in f Fratal dimension Striatum vol. / total brain vol..15.1.5 1. 1.15 1.1 1.5 1. P =.33 P =.95 weeks old 5 weeks old P <.1 Slitrk5 a 5 µm Anti-Flag Anti-Flag 1 µm 1 µm GFP 1 µm PSD95- herry PSD95-herry GFP 1 µm 5 µm 5 µm 5 µm b GluR1 GluR NRA NRB PSD95 Atin Relative expression of PSD proteins in striatum 1 5 GluR1 GluR NRA NRB PSD95 P =...1.7. Population spike amplitude (mv) 1. 1.......5 mv ms... Stimulation intensity (ma) 1. d Paired-pulse ratio 1. 1. 1. 1. 1... 1 Interstimulus interval (ms) Figure Defiieny in ortiostriatal transmission in Slitrk5 / mie is mediated by hanges in glutamate reeptor omposition. (a) Immunostaining of primary striatal rat neurons (infeted with Flag-Slitrk5 lentivirus and transfeted with PSD95 fused to mcherry (PSD95-herry)) in ulture with ortial neurons (isolated from transgeni mie that ubiquitously express green fluoresent protein) with Flag-speifi antibody (anti-flag). The arrow points to a magnified area (bottom images) that represents the synapses between ortial and striatal neurons. (b) Western blot analysis of NMDA and AMPA reeptor subunits in the striatum of 5-month-old Slitrk5 / and mie. The protein amounts are adjusted to the expression of atin. () Population spike amplitude in Slitrk5 / mie (n = 11, from four mie) and mathed mie (n = 9, from four mie). The population spike amplitude is signifiantly lower in Slitrk5 / mie, P <.1, repeated-measures analysis of variane. The inset shows examples of ortiostriatal population spike amplitudes in Slitrk5 / mie and mathed mie. (d) Average paired-pulse ratios of the population spike in Slitrk5 / mie (n = 17, from five mie) and mathed mie (n = 17, from five mie). There is no signifiant differene in the paired-pulse ratio between Slitrk5 / mie and wild-type mie. NATURE MEDICINE ADVANCE ONLINE PUBLICATION 3

1 Nature Ameria, In. All rights reserved. the striatum and found that they are downregulated in Slitrk5 / mie (Fig. b). Indeed, protein amounts of glutamate reeptor subunits NRA, NRB, GluR1, and GluR were dereased by %, with no signifiant hanges in PSD95 amounts (Fig. b). We found these hanges in both the total lysates (Fig. b) and in PSD-enrihed frations of synaptosomes (Supplementary Fig. 7). Given these findings, we investigated whether Slitrk5 / mie have defiits in ortiostriatal neurotransmission by extraellular reordings in aute striatal slies. We reorded population spikes from striatum by stimulating the white matter between ortex and striatum. We found a signifiantly redued population spike amplitude in Slitrk5 / mie (Fig. ). We did not observe any differene in paired-pulse ratios of the population spike in Slitrk5 / mie and their wild-type littermates, suggesting that the presynapti mehanism involved in paired-pulse failitation is not responsible for the observed differene in population spike amplitude (Fig. d). Also, we did not observe any signifiant differene in fiber volley amplitude, suggesting that ortial axon input is normal in Slitrk5 / mie (Supplementary Fig. ). Taken together, our data demonstrate that targeted inativation of Slitrk5 in mie leads to OCD-like behavioral phenotypes, inluding overgrooming with elements of self-mutilation. Although Slitrk5 expression is widespread in the entral nervous system, we found inreased neuronal ativity speifially in the orbitofrontal ortex of Slitrk5 / mie, whih is onsistent with funtional imaging findings in humans with OCD that impliated dysregulation of ortiostriatal iruitry 3,7, and whih has not been reported in previous mouse models of OCD,9. In addition, Slitrk5 / mie have anatomial defiits in the striatum, suh as redued striatal volume, as well as dereased dendriti omplexity of striatal medium spiny neurons. Although this region has not been onsistently found to be altered anatomially in people with OCD 19,1,, emerging literature suggests that striatal dysfuntion may underlie behavioral defiits in individuals with OCD 7. In this ontext, it has reently been postulated that striatal dysfuntion, in the presene of orbitofrontal ortex overativation, ould lead to defiits in thalami filtering or imbalane in the diret and indiret pathways of the basal ganglia 3. Given the ubiquitous neuronal expression of Slitrk5, this seletive effet on the orbitofrontal ortex and on striatal neurons is hard to explain. On the one hand, it is reminisent of the effet of other proteins suh as huntingtin, whih is also widely expressed in the entral nervous system, but alterations in the huntingtin protein result in funtional defets predominantly in striatal neurons, diretly leading to Huntington s disease pathology 31. On the other hand, it is possible that Slitrk5 may form a signaling omplex with ortiostriatal-speifi proteins, whih may explain these region-speifi effets. Overall, our data suggest that Slitrk5 may have a entral role in the development of OCD-like behaviors. Although human geneti studies have impliated another Slitrk family member, SLITRK1, in Tourette s syndrome, these assoiations have not been onsistently repliated 3,33. In this ontext, our studies link Slitrk5 to the ore symptoms of OCD: self-injurious repetitive behavior and inreased anxiety. In all, we provide a new mouse model of OCD-like behaviors, involving a previously unharaterized neuronal transmembrane protein that modulates region-speifi glutamatergi neurotransmission. This model an be used to further disset the role of Slitrk5 in moleular pathways underlying the pathogenesis of obsessive-ompulsive behaviors. METHODS Methods and any assoiated referenes are available in the online version of the paper at http://www.nature.om/naturemediine/. Note: Supplementary information is available on the Nature Mediine website. ACKNOWLEDGMENTS We thank M. Flint Beal for providing expertise on behavioral experiments. We thank G. Thurston for ritial omments and suggestions. PSD95-herry was a kind gift from R.H. Edwards (University of California San Franiso). We aknowledge support from US National Institutes of Health grants MH79513 (F.S.L.) and NS519 (F.S.L.), HL59 (S.R.), HL97797 (S.R.) and AI39 (S.R.), Burroughs Wellome Foundation (F.S.L.), International Mental Health Researh Organization (F.S.L.), the Sakler Institute (K.G.B., F.S.L.), DeWitt-Wallae Fund of the New York Community Trust (F.S.L.), Pritzker Consortium (F.S.L.), National Alliane for Researh on Shizophrenia and Depression (S.V.S.), Mildred-Sheel-Stiftung, Deutshe Krebshilfe (T.M.), Gulbenkian PhD Programe in Biomediine (C.C.P.), Fundaao para Cienia e Tenologia (C.C.P.), Howard Hughes Medial Institute (S.R.), Ansary Stem Cell Institute (S.R.), Anbinder and Newmans Own Foundations (S.R.), Qatar National Priorities Researh Program (S.R.), Empire State Stem Cell Board (S.R.) and the New York State Department of Health grant NYS C1 (S.R.). AUTHOR CONTRIBUTIONS S.V.S. oneived of and designed the study, performed experiments, analyzed data and wrote the manusript; A.H., D.J, C.C.P. and K.G.B. designed and performed experiments, analyzed data and assisted in writing the manusript; 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 / mie; I.N. designed, performed and analyzed eletrophysiology experiments; F.S.L. and S.R. oneived of and designed the study and wrote the manusript. COMPETING FINANCIAL INTERESTS The authors delare no ompeting finanial interests. Published online at http://www.nature.om/naturemediine/. Reprints and permissions information is available online at http://npg.nature.om/ reprintsandpermissions/. 1. Miguel, E.C. et al. Obsessive-ompulsive disorder phenotypes: impliations for geneti studies. Mol. Psyhiatry 1, 5 75 (5).. Karno, M., Golding, J.M., Sorenson, S.B. & Burnam, M.A. The epidemiology of obsessive-ompulsive disorder in five US ommunities. Arh. Gen. Psyhiatry 5, 19 199 (19). 3. Graybiel, A.M. & Rauh, S.L. Toward a neurobiology of obsessive-ompulsive disorder. Neuron, 33 37 ().. Menzies, L. et al. Integrating evidene from neuroimaging and neuropsyhologial studies of obsessive-ompulsive disorder: the orbitofronto-striatal model revisited. Neurosi. Biobehav. Rev. 3, 55 59 (). 5. Clifford, C.A., Murray, R.M. & Fulker, D.W. Geneti and environmental influenes on obsessional traits and symptoms. Psyhol. Med. 1, 791 (19).. Rasmussen, S.A. & Tsuang, M.T. The epidemiology of obsessive ompulsive disorder. J. Clin. Psyhiatry 5, 5 57 (19). 7. Pauls, D.L., Alsobrook, J.P. II, Goodman, W., Rasmussen, S. & Lekman, J.F. A family study of obsessive-ompulsive disorder. Am. J. Psyhiatry 15, 7 (1995).. Nestadt, G. et al. A family study of obsessive-ompulsive disorder. Arh. Gen. Psyhiatry 57, 35 33 (). 9. Hollander, E., Kim, S., Khanna, S. & Pallanti, S. Obsessive-ompulsive disorder and obsessive-ompulsive spetrum disorders: diagnosti and dimensional issues. CNS Spetr. 1, 5 13 (7). 1. Abelson, J.F. et al. Sequene variants in SLITRK1 are assoiated with Tourette s syndrome. Siene 31, 317 3 (5). 11. Aruga, J. & Mikoshiba, K. Identifiation and haraterization of Slitrk, a novel neuronal transmembrane protein family ontrolling neurite outgrowth. Mol. Cell. Neurosi., 117 19 (3). 1. Aruga, J., Yokota, N. & Mikoshiba, K. Human SLITRK family genes: genomi organization and expression profiling in normal brain and brain tumor tissue. Gene 315, 7 9 (3). 13. Katayama, K. et al. Slitrk1-defiient mie display elevated anxiety-like behavior and noradrenergi abnormalities. Mol. Psyhiatry 15, 177 1 (1). 1. Shmelkov, S.V., Visser, J.W. & Belyavsky, A.V. Two-dimensional gene expression fingerprinting. Anal. Biohem. 9, 35 (1). 15. Milde, T. et al. A novel family of slitrk genes is expressed on hematopoieti stem ells and leukemias. Leukemia 1, 7 (7). 1. Valenzuela, D.M. et al. High-throughput engineering of the mouse genome oupled with high-resolution expression analysis. Nat. Biotehnol. 1, 5 59 (3). 17. Welh, J.M. et al. Cortio-striatal synapti defets and OCD-like behaviours in Sapap3-mutant mie. Nature, 9 9 (7). 1. MClung, C.A. et al. DeltaFosB: a moleular swith for long-term adaptation in the brain. Brain Res. Mol. Brain Res. 13, 1 15 (). 19. Saxena, S., Bota, R.G. & Brody, A.L. Brain-behavior relationships in obsessiveompulsive disorder. Semin. Clin. Neuropsyhiatry, 11 (1). ADVANCE ONLINE PUBLICATION NATURE MEDICINE

. Whiteside, S.P., Port, J.D. & Abramowitz, J.S. A meta-analysis of funtional neuroimaging in obsessive-ompulsive disorder. Psyhiatry Res. 13, 9 79 (). 1. Saxena, S. & Rauh, S.L. Funtional neuroimaging and the neuroanatomy of obsessive-ompulsive disorder. Psyhiatr. Clin. North Am. 3, 53 5 ().. Aylward, E.H. et al. Normal audate nuleus in obsessive-ompulsive disorder assessed by quantitative neuroimaging. Arh. Gen. Psyhiatry 53, 577 5 (199). 3. Robinson, D. et al. Redued audate nuleus volume in obsessive-ompulsive disorder. Arh. Gen. Psyhiatry 5, 393 39 (1995).. Rosenberg, D.R. et al. Frontostriatal measurement in treatment-naive hildren with obsessive-ompulsive disorder. Arh. Gen. Psyhiatry 5, 3 (1997). 5. Szeszko, P.R. et al. Brain strutural abnormalities in psyhotropi drug-naive pediatri patients with obsessive-ompulsive disorder. Am. J. Psyhiatry 11, 19 15 ().. Surmeier, D.J., Ding, J., Day, M., Wang, Z. & Shen, W. D1 and D dopamine-reeptor modulation of striatal glutamatergi signaling in striatal medium spiny neurons. Trends Neurosi. 3, 35 (7). 7. Rauh, S.L. et al. Funtional magneti resonane imaging study of regional brain ativation during impliit sequene learning in obsessive-ompulsive disorder. Biol. Psyhiatry 1, 33 33 (7).. Wang, L., Simpson, H.B. & Dulawa, S.C. Assessing the validity of urrent mouse geneti models of obsessive-ompulsive disorder. Behav. Pharmaol., 119 133 (9). 9. Joel, D. Current animal models of obsessive ompulsive disorder: a ritial review. Prog. Neuropsyhopharmaol. Biol. Psyhiatry 3, 37 3 (). 3. Rauh, S.L. Neuroimaging and neuroiruitry models pertaining to the neurosurgial treatment of psyhiatri disorders. Neurosurg. Clin. N. Am. 1, 13 3 vii viii (3). 31. Cattaneo, E. et al. Loss of normal huntingtin funtion: new developments in Huntington s disease researh. Trends Neurosi., 1 1 (1). 3. Deng, H., Le, W.D., Xie, W.J. & Jankovi, J. Examination of the SLITRK1 gene in Cauasian patients with Tourette syndrome. Ata Neurol. Sand. 11, (). 33. Keen-Kim, D. et al. Overrepresentation of rare variants in a speifi ethni group may onfuse interpretation of assoiation analyses. Hum. Mol. Genet. 15, 33 33 (). 1 Nature Ameria, In. All rights reserved. NATURE MEDICINE ADVANCE ONLINE PUBLICATION 5

1 Nature Ameria, In. All rights reserved. ONLINE METHODS Mie. All experiments were approved by the Animal Care and Use Committee of the Weill Cornell Medial College. We generated Slitrk5 / mie with the previously desribed high-throughput VeloiGene 1. The expression pattern of Slitrk5 was determined by the detetion of -galatosidase ativity. Detailed methods are loated in the Supplementary Methods. All experiments were performed blinded to genotype. Tissue proessing and immunostaining. We transardially perfused mie and postfixed the brains with % paraformaldehyde and obtained - m serial oronal frozen setions. For immunofluoresene, we stained free-floating setions with antibodies to -galatosidase ( g ml 1, Chemion) and to NeuN (1 g ml 1, Chemion) and deteted with Alexa-onjugated seondary antibodies ( g ml 1, Invitrogen). Nissl staining was performed as desribed previously 3. For staining of neurons in ulture, striatal neurons grown on glass overslips ( d in vitro) were fixed in 3.7% paraformaldehyde and permeabilized with.% Triton X-1. We used Flag-speifi antibody (1 in 1,, Sigma) and Alexa-onjugated seondary antibodies (Invitrogen) for the detetion in oimmunofluoresene experiments. We used lipofetamine (Invitrogen) to transfet striatal neurons with PSD95 fused to mcherry (a kind gift from Dr. R.H. Edwards). Behavioral analysis and fluoxetine treatment. We performed -min videotaping sessions on 3-month-old mie for onseutive days at 9:3, 13:3, 1:3 and 19:. We treated the mie for 1 d with fluoxetine (Sigma) delivered in the drinking water as previously desribed 3. Detailed methodology an be found in the Supplementary Methods. Open-field test. The open-field test was performed as previously desribed 3. Detailed methodology an be found in the Supplementary Methods. Stereologial volume and ell density estimations. Volume and ell density estimations were performed with the Stereoinvestigator System (Mirobrightfield). After systemi random sampling, we traed brain strutures using the Allen Mouse Brain Map as a referene. Striatal and total brain volumes were alulated based on traed ontours by the Cavalieri estimation method. Frationator probe was used for the stereologial estimation of ell density (Supplementary Fig. 9). Detailed methodology an be found in the Supplementary Methods. Golgi impregnation and traing. We impregnated fresh brains in Golgi-ox using the FD Rapid GolgiStain Kit (Neurodigiteh) solution for 1 d at 5 C in the dark and then transferred them to 3% surose at C for 7 h. We prepared 15- m oronal serial setions with a vibrotome; slides were soaked in 5% surose and air-dried for 7 h in the dark. Quantitative mirosopy was performed on a Mirobrightfield imaging system (Mirobrightfield). Two hundred striatal neurons were hosen by systemi random sampling, and traeable neurons for eah genotype were reonstruted three dimensionally with the Neuroluida system. The morphologial traits of ells were analyzed with Neuroexplorer. Eletrophysiology. We killed -month-old mie by pentobarbital anesthesia to obtain ortiostriatal slies for eletrophysiologial reordings. Coronal brain slies ( m) were made on a vibrotome (Campden Instruments) and submerged in artifiial erebrospinal fluid in a brain-slie keeper (Sientifi Systems Design) for 9 min at 5 C and gassed with 95% O, 5% CO before transfer to the reording hamber 35. The artifiial erebrospinal fluid ontained 11 mm NaCl,.5 mm KCl, 1 mm gluose, 1 mm NaH PO, 3 mm CaCl, mm MgCl and 5 mm NaHCO 3. 1 M pirotoxin was inluded in the reording solution. Reording eletrodes were filled with M NaCl solution, and population spikes were reorded from the striatum with the IE-1 amplifier (Warner Instruments) using Digidata 1A and pclamp 1 software (Moleular Devies) at 3 C. For synapti stimulation, we plaed bipolar eletrodes in the white matter between the ortex and the striatum to ativate ortiostriatal fibers 35. Population spike amplitude was alulated by the mean of the amplitude from the first peak positivity to the peak negativity of the population spike and the amplitude of the peak negativity of the population spike to the seond peak positivity 3,37. The presynapti fiber volley amplitude was measured as a differene between the initial positive and the following negative peak. Three onseutive responses were averaged for measuring the spike and fiber volley amplitude. Paired-pulse responses were evoked at interstimulus intervals of,,, 1 and ms using a stimulation intensity of.9 ma. Paired-pulse ratio is defined as the ratio of seond population spike amplitude to the first population spike amplitude. Population spike amplitudes were analyzed by Clampfit software (Moleular Devies). Neuronal ulture. Rat primary striatal neuron ultures were prepared as desribed previously, with a few modifiations 3. Detailed methodology an be found in the Supplementary Methods. Immunoblotting. Immunoblotting was performed as previously desribed 3 with antibodies to the following proteins: NRA (Covane), NRB (Invitrogen), GluR1, GluR, NR1, PSD95 (all Chemion) and atin (Santa Cruz). Detailed methodology an be found in the Supplementary Methods. Statistial analyses. Data are shown as means s.e.m. We used Student s t test to alulate statistial signifiane between group differenes. Signifiane was set at P <.5. We used two-way ANOVA with repeated measures to analyze behavioral performane in the rotarod test and eletrophysiologial studies. We performed data analysis using Prism. (GraphPad software) for neuron traing. 3. Chen, Z.Y. et al. Geneti variant BDNF (ValMet) polymorphism alters anxietyrelated behavior. Siene 31, 1 13 (). 35. Pioni, B. et al. Loss of bidiretional striatal synapti plastiity in l-dopa indued dyskinesia. Nat. Neurosi., 51 5 (3). 3. Lovinger, D.M., Tyler, E.C. & Merritt, A. Short- and long-term synapti depression in rat neostriatum. J. Neurophysiol. 7, 1937 199 (1993). 37. Alger, B.E. & Teyler, T.J. Long-term and short-term plastiity in the CA1, CA3 and dentate regions of the rat hippoampal slie. Brain Res. 11, 3 (197). 3. Pereira, D.B. & Chao, M.V. The tyrosine kinase Fyn determines the loalization of TrkB reeptors in lipid rafts. J. Neurosi. 7, 59 9 (7). NATURE MEDICINE doi:1.13/nm.15