ocd_PNAS-2013-Xu-10759-64

9

Click here to load reader

Transcript of ocd_PNAS-2013-Xu-10759-64

Page 1: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 1/9

Double deletion of melanocortin 4 receptorsand SAPAP3 corrects compulsive behaviorand obesity in micePin Xua, Brad A. Grueterb,1, Jeremiah K. Brittc, Latisha McDanielc, Paula J. Huntingtona, Rachel Hodgea,

Stephanie Trana

, Brittany L. Masond

, Charlotte Leed

, Linh Vonge

, Bradford B. Lowelle

, Robert C. Malenkab,2

,Michael Lutterc,2,3, and Andrew A. Pieperc,f,2,3

Departments of aNeuroscience and dInternal Medicine, University of Texas Southwestern Medical Center, Dallas, TX 75390; bNancy Pritzker Laboratory,Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94305; eDivision of Endocrinology, Departmentof Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA 02215; and Departments of cPsychiatry and fNeurology, CarverCollege of Medicine, University of Iowa, Iowa City, IA 52242

Contributed by Robert C. Malenka, May 2, 2013 (sent for review April 12, 2013)

Compulsive behavior is a debilitating clinical feature of many

forms of neuropsychiatric disease, including Tourette syndrome,obsessive-compulsive spectrum disorders, eating disorders, and au-

tism. Although several studies link striatal dysfunction to compulsiv-

ity, the pathophysiology remains poorly understood. Here, we show

that both constitutive and induced genetic deletion of the gene

encoding the melanocortin 4 receptor (MC4R), as well as pharmaco-logic inhibition of MC4R signaling, normalize compulsive grooming

and striatal electrophysiologic impairments in synapse-associated

protein 90/postsynaptic density protein 95-associated protein 3

(SAPAP3)-null mice, a model of human obsessive-compulsive disor-

der. Unexpectedly, genetic deletion of SAPAP3 restores normal

weight and metabolic features of MC4R-null mice, a model of humanobesity. Our findings offer insights into the pathophysiology and

treatment of both compulsive behavior and eating disorders.

anxiety disorders | conditional knockout mice | metabolism |synaptic transmission

I

nvestigation of the synapse-associated protein 90/postsynapticdensity protein 95-associated protein 3 (SAPAP3) and the mel-

anocortin 4 receptor (MC4R) have provided important insightsinto the pathophysiology of compulsivity and obesity, respectively.Genetic variations in SAPAP3 have been identified in somepatients with OCD-spectrum disorders, such as trichotillomania(1–3), and SAPAP3 is enriched in the striatum where it influencesexcitatory synaptic function (4). Importantly, SAPAP3-null micedisplay compulsive grooming that can be ameliorated by  fluoxe-tine, a mainstay treatment for patients with obsessive–compulsivedisorder (OCD) (4). Thus, SAPAP3-null mice are a valuablemodel of compulsive behavior in neuropsychiatric disease.

MC4R signaling, on the other hand, regulates feeding behaviorand energy expenditure (5). Mutations in the MC4R are the mostcommon monogenic cause of hyperphagia and morbid obesity inpeople, and MC4R-null mice thus provide a useful model of hu-

man obesity (6). MC4Rs have long been known to be highly activein the hypothalamus, and it has recently been shown that MC4Rsalso operate within D1 medium spiny neurons of the ventralstriatum to mediate procedural memory and affective responsesto stress (7, 8). In addition, previous results have suggested thatstimulation of MC4R-signaling induces compulsive grooming inrats (9). These findings, combined with the well-established andcritical role of striatal circuitry in mediating compulsive behavior(10–12), prompted us to hypothesize that loss of MC4R signalingmight ameliorate compulsive grooming in SAPAP3-null mice.

Results

Constitutive Genetic Elimination of MC4R Normalizes Grooming in

SAPAP3-Null Mice. After demonstrating that MC4R agonists in-duce compulsive grooming in wild-type mice (Fig. S1) as was

previously shown in rats (9), we addressed our hypothesis by crossing SAPAP3-null mice with MC4R-null mice to generatefour experimental groups for analysis of compulsive groomingbehavior: wild-type ( Mc4r + / +, Sapap3+ / +), SAPAP3-null ( Mc4r + / +,Sapap3− / −), MC4R-null ( Mc4r − / −, Sapap3+ / +), and double null( Mc4r − / −, Sapap3− / −). As expected (4), SAPAP3-null mice showed

increased grooming time in the standard 5-min spray test of induced grooming, due to increased bouts of grooming, both atbaseline and after stimulation by water spray (Fig. 1 A). However,mice lacking both SAPAP3 and MC4R displayed wild-type levelsof grooming time and bouts at both baseline and after spray,

 whereas MC4R-null mice showed no difference from wild type(Fig. 1 A).

We then tested grooming more rigorously using the Laborassystem, which allows 24-h automated collection of nonstimulated,basal grooming behavior using a vibration sensitive plate to mon-itor fine motor activity. Consistent with spray test results, SAPAP3-null mice exhibited significantly increased total grooming time andbouts over 24 h, whereas double null mice again showed normal-ized grooming time and bouts (Fig. 1 B). Grooming in MC4R-nullmice was the same as in wild-type mice. Loss of MC4R in SAPAP3null mice rescued compulsive grooming by normalizing the numberof grooming bouts rather than by shortening the duration of in-dividual bouts, which were the same across all four groups (averagetime of bouts: 7.01 ± 0.32 s for wild type, 7.23 ± 0.62 s for SAPAP3-null, 6.83 ± 0.51 s for MC4R-null, 6.23 ± 0.42 s for double null).We additionally observed that double deletion of MC4R andSAPAP3 prevented acquisition of self-induced facial lesions, whichare otherwise characteristic of SAPAP3-null mice (Fig. 1C).

Pharmacologic Inhibition of MC4R Signaling Normalizes Compulsive

Grooming in SAPAP3-Null Mice. We next tested whether normaliza-tion of grooming behavior in double null mice was due to inter-actions between MC4R signaling and SAPAP3 function in matureneuronal circuitry, rather than simply a developmental effect.

Author contributions: P.X., B.A.G., J.K.B., R.C.M., M.L., and A.A.P. designed research; P.X.,

B.A.G., J.K.B., L.M., P.J.H., R.H., S.T., B.L.M., C.L., M.L., and A.A.P. performed research; L.V.

and B.B.L. contributed new reagents/analytic tools; P.X., B.A.G., J.K.B., L.M., P.J.H., B.L.M.,

C.L., R.C.M., M.L., and A.A.P. analyzed data; and P.X., B.A.G., R.C.M., M.L., and A.A.P.

wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.

1Present address: Department of Anesthesiology, Vanderbilt University, Nashville,

TN 37232.

2To whom correspondence may be addressed. E-mail: [email protected] ,

[email protected] , or [email protected] .

3M.L. and A.A.P. contributed equally to this work.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.

1073/pnas.1308195110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1308195110 PNAS | June 25, 2013 | vol. 110 | no. 26 | 10759–10764

     M     E     D     I     C     A     L     S     C     I     E     N     C     E     S

Page 2: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 2/9

Specifically, we assessed the ef ficacy of pharmacologic inhibitionor genetic elimination of MC4R signaling in adult SAPAP3-nullmice. First, grooming was evaluated in the 24-h Laboras groomingassay following 2 wk of intracerebroventricular (ICV) delivery of the MC4R antagonist HS014 via Alzet osmotic minipumps intoadult SAPAP3 null mice. HS014 is a cyclic melanocyte stimulatinghormone analog that serves as an MC4R ligand ( K i of 3.2 nM)

 with 17-fold MC4/MC3 receptor selectivity (13). Without directly measuring brain tissue drug levels, it is impossible to determine

the brain concentration of HS014 after 2 wk of continuous ICVdelivery. Previously in the literature, however, others have de-termined that an ICV infusion rate of 0.17 nmol of HS014 perhour for 2 wk is suf ficient to induce hyperphagia and obesity inrodents (14). Accordingly, we conservatively infused at lower doses.The Alzet osmotic minipumps we used dispel their contents at arate of 0.11 μL/h. Thus, at a concentration of 1 mM HSO14 in thepump, drug delivery occurred at a rate of 0.11 nmol/h. Accordingly,0.3 mM concentration in the pump effected delivery of 0.033 nmol/h,and 0.1 mM concentration in the pump effected delivery at arate of 0.011 nmol/h. Although the lowest dose of 0.011 nmol/hof HS014 had no effect on grooming behavior, higher dosagesof 0.033 and 0.11 nmol/h significantly reduced grooming timeand bouts in SAPAP3-null mice (Fig. 2 A). Pumps and cannulae

 were then removed and mice were allowed to recover for 4 wk.

Subsequent reevaluation in the 24-h Laboras grooming assay at this later time point showed that grooming time and boutsreturned to the high levels typically seen in SAPAP3-null mice.

 Analogous experiments were conducted on wild-type littermatemice, and no grooming differences were seen at any time point.Thus, acute pharmacologic inhibition of MC4Rs significantly ameliorated compulsive grooming in SAPAP3-null mice, but didnot affect grooming in wild-type mice.

Induced Genetic Elimination of Mc4r Normalizes Grooming in SAPAP3-Null Mice. We next generated Sapap3− / − :Mc4r 

 lox/loxexperimental

mice by crossing SAPAP3-null mice with mice in which the singleexon of the mc4r gene was flanked by lox P sites (15). Sixteen- to17-wk-old Sapap3− / − :Mc4r 

 lox/loxmice were screened in the Laboras

24-h grooming assay to ensure that these new mice displayedcompulsive grooming (Fig. 2 B). Then, mice were randomly andequally divided into two groups, and key components of corti-costriatal circuitry were targeted stereotaxically and bilaterally by administering adeno-associated virus (AAV) vector encodingeither Cre recombinase and green fluorescent protein (GFP)(AAV-Cre-GFP), or GFP alone (AAV-GFP; “control group”) intothe orbitofrontal cortex and nucleus accumbens shell, a componentof the ventral striatum. Both regions were targeted for eliminationof MC4R to test our hypothesis of a role for MC4R signaling in

Fig. 1. Genetic elimination of both MC4R and SAPAP3 normalizes excessive grooming observed in singly deficient SAPAP3-null mice. ( A) SAPAP3-null mice

showed increased grooming time and number of grooming bouts at baseline (NS = “no spray”) and after grooming stimulus (S = “spray”) in the 5-min

grooming assay, compared with wild-type and MC4R-null mice. (B) SAPAP3-null mice showed increased grooming time and number of grooming boutsrelative to wild-type and MC4R-null mice in the 24-h home cage basal grooming assay. Time spent grooming and number of grooming bouts was normalized

in double null mice in all aspects of 5-min and 24-h assays (data presented as mean ± SEM). (C ) Double deletion of MC4R and SAPAP3 prevent acquisition of

self-induced facial lesions, which are characteristic of SAPAP3-null mice.

10760 | www.pnas.org/cgi/doi/10.1073/pnas.1308195110 Xu et al.

Page 3: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 3/9

compulsive behavior secondary to loss of SAPAP3 in the stria-tum. Five weeks later, mice were reevaluated in the 24-h Lab-oras grooming assay. Mice that received AAV-Cre-GFP showed

reduced grooming time and bouts to wild-type levels (Fig. 2 B).Subsequently, brains were examined both histologically for GFPexpression to verify ef ficient viral delivery (Fig. S2) and by insitu hybridization for MC4R transcript expression to verify viral-mediated elimination of intact Mc4r  (Fig. S3) in both orbito-frontal cortex and nucleus accumbens shell. These results con-firm the specificity of our previously demonstrated effect of pharmacological inhibition of MC4Rs on compulsive grooming,and provide further evidence that inhibition of MC4R signalinghelps normalize grooming in adult SAPAP3-null mice. In futurestudies it will be informative to delineate pre- vs. postsynapticroles of MC4R signaling by selectively eliminating MC4R in thenucleus accumbens, while leaving expression intact in theorbitofrontal cortex.

Elimination of MC4R Restores Normal Ventral Striatal Synaptic

Transmission in SAPAP3-Null Mice. SAPAP3-null mice have reduced AMPA receptor (AMPAR)-mediated synaptic transmission and

an increased number of silent synapses in the dorsal striatum(16, 17). Furthermore, activation of MC4R signaling within ventralstriatal neurons reduces synaptic strength by increasing endocytosisof AMPARs (8). We thus hypothesized that normalization of grooming behavior in double null mice might correlate with res-toration of normal ventral striatal neuron AMPAR activity. Ac-cordingly, we recorded input-output curves of extracellular fieldpotential responses in acute ventral striatal slices from SAPAP3-null, MC4R-null, double null, and wild-type littermate mice.Consistent with previous results in the dorsal striatum, field po-tential amplitudes recorded in ventral striatal slices from SAPAP3-null mice exhibited a decreased, synaptically mediated N2 fieldpotential response relative to wild type. By contrast, peak N2amplitudes were the same in slices obtained from MC4R-nulland wild-type mice. However, N2 field potential amplitudes

   G   r   o   o   m   i   n   g

   T   i   m   e   i   n   2   4    h   r   s    (   s   e   c   o   n    d   s    ) 15000

10000

5000

0Before Pump

ImplantaƟon

2 Weeks AŌer

Pump ImplantaƟon

4 Weeks AŌer

Pump Removal

   G   r   o   o   m   i   n   g

   T   i   m   e   i   n   2   4    h   r   s    (   s   e   c   o   n    d   s    ) 15000

10000

5000

0Before Pump

ImplantaƟon

2 Weeks AŌer

Pump ImplantaƟon

4 Weeks AŌer

Pump Removal

SAPAP3-null wild type

3000

2500

2000

1500

1000

500

0

   T   o   t

   a    l   G   r   o   o   m   i   n   g   B   o   u   t   s   i   n   2   4    h   r   s

3000

2500

2000

1500

1000

500

0

   T   o   t

   a    l   G   r   o   o   m   i   n   g   B   o   u   t   s   i   n   2   4    h   r   s

Before Pump

ImplantaƟon

2 Weeks AŌer

Pump ImplantaƟon

4 Weeks AŌer

Pump Removal

Before Pump

ImplantaƟon

2 Weeks AŌer

Pump ImplantaƟon

4 Weeks AŌer

Pump Removal

SAPAP3-null wild type

A

3000

2000

1000

0

   T   o   t   a    l   G   r   o   o   m   i   n   g   B   o   u   t   s   i   n   2   4    h   r   s

Before AAV 5 Weeks AŌer AAV

   G   r   o   o   m   i   n   g   T   i   m   e   i   n   2   4    h   r   s    (   s   e   c   o   n    d   s    )15000

10000

5000

0

Before AAV 5 Weeks AŌer AAV

Sapap3-/-,Mc4rf/f AAV-Cre

Sapap3-/-,Mc4rf/f AAV-GFP

B

* *

* *

* *

vehicle

0.1 mM

0.3 mM

1 mM

vehicle

0.1 mM

0.3 mM

1 mM

Fig. 2. Pharmacologic inhibition and genetic elimination of MC4R in SAPAP3-null mice normalizes excessive grooming. ( A) Intracerebroventricular delivery

of the MC4R antagonist HS014 to SAPAP3-null mice significantly reduced grooming time and number of grooming bouts in the 24-h Laboras grooming assay

(*P < 0.01, Student’s t  test) at 0.3 and 1.0 mM concentrations, and this effect was reversed following cessation of HS014 delivery. Concentrations refer to

loading concentration of the drug in the osmotic minipump. Based on the delivery rate of the pump, the following rates of HS014 infusion into the left lateral

ventricle of the brain were implemented over a 2-wk time period: 1mM HS014 in the pump = 0.11 nmol/h, 0.3 mM HS014 in the pump = 0.033 nmol/h, and 0.1

mM HS014 in the pump = 0.011 nmol/h. (B) Cre-mediated deletion of Mc4r in Sapap3−/ −:Mc4r lox/lox mice normalizes grooming time and number of grooming

bouts (*P = 0.0002, Student’s t  test) in the 24-h Laboras grooming assay.

Xu et al. PNAS | June 25, 2013 | vol. 110 | no. 26 | 10761

     M     E     D     I     C     A     L     S     C     I     E     N     C     E     S

Page 4: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 4/9

in double null mice were restored to wild-type levels (Fig. 3 Aand B). No differences were observed in the N1 component of field potential recordings (Fig. S4), which reflects the direct ex-citability of axons and cells in the slice (18). These resultssuggest that similar to the dorsal striatum, decreased AMPAR-mediated synaptic transmission occurs in SAPAP3-null mice andthat this synaptic modification, like compulsive grooming, is res-cued by MC4R deletion. We also found a significant increase inpaired-pulse ratio (PPR) in SAPAP3-null mice, which was absentin both MC4R-null and double null mice (Fig. 3C). Although these

measures using field potential recordings from striatal slices reflecta mixture of presynaptic and postsynaptic events (18), these resultsprovide further evidence that deletion of MC4R signaling rescuessynaptic changes caused by loss of SAPAP3 in the striatum.

Elimination of SAPAP3 Restores Normal Weight and Metabolism in

MC4R-Null Mice. Because loss of function MC4R mutations causeobesity in humans (19) and rodents (20), we evaluated body com-position to ensure that changes in grooming were not secondary to changes in body habitus. Quite surprisingly, we observed thatalthough MC4R-null mice were obese as expected, double nullmice displayed body weight similar to wild-type and SAPAP3-nullmice (Fig. 4 A). Consistent with the rescue in body weight, deletionof SAPAP3 from MC4R-null mice also restored normal body 

length, adiposity, food intake and fasting blood glucose (Fig. 4 A–C). Taken together, our results confirm that the reduction ingrooming in double null mice relative to SAPAP3-null micedoes not result from nonspecific metabolic or locomotor aberra-tions, as double-null mice exhibit the same lean phenotype (Fig. 4)and locomotor activity (Fig. S5) as SAPAP3-null mice. Impor-tantly, our results also demonstrate an unanticipated interaction of SAPAP3–MC4R signaling in feeding behavior and metabolism.

Discussion

We have identified a biologically relevant interaction of MC4R–

SAPAP3 signaling in regulating compulsive behavior and body  weight. In the case of compulsive behavior as measured by excessivegrooming, this effect correlates well with modifications of AMPAR-mediated synaptic transmission in ventral striatal neurons. Our

findings also suggest that regulation of excitatory synapse strengthis a critical function of MC4R signaling in the ventral striatum.

On first glance, metabolism and compulsive behavior may ap-pear unlikely partners for a neuronally mediated mechanistic link.However, several clinically relevant intersections of compulsivity and food intake do occur. For example, a common obsessionin OCD is fear of contamination (21), typically associated withcompulsive cleansing rituals, such as washing of hands or objects.Nutrient-sensing pathways might thus regulate compulsivity tomaximize assessment of food quality and safety. MC4R signaling

is known to increase during periods of caloric abundance (22),and augmentation of compulsive concerns of cleanliness andfood safety would be beneficially adaptive during periods of plen-tiful food availability. Conversely, MC4R signaling is decreasedduring periods of starvation (22), which could diminish thesecompulsive concerns when it is more important to survive by consuming scarcely available nutrition. Symptoms centering onanxiety and food intake are also comorbid in eating disorders. Forexample, patients with anorexia nervosa have much higherrates of anxiety and compulsivity than the general population(23), and OCD-like symptoms generally precede the onset of aneating disorder (24).

Our findings also identify an unanticipated role for synapticplasticity in MC4R-signaling control of body weight. In the context

of reduced strength of excitatory synapses observed in SAPAP3-null mice, loss of MC4R signaling failed to elicit either hyperphagiaor obesity, indicating that modulation of glutamatergic signalingmay be a key factor in MC4R-mediated feeding behavior. Thisobservation extends the earlier finding that MC4R signaling actspresynaptically within the nucleus tractus solitarius to modulateglutamatergic synaptic transmission (5), and is consistent with therecent demonstration that MC4R signaling within the striatumreduces strength of excitatory synapses (8). In addition, othershave previously demonstrated that loss of SAPAP3 elicits post-synaptic endocytosis of AMPARs via increased metabotropicglutamate receptor 5 (mGluR5) activity (25), which ultimately reduces AMPAR-mediated synaptic transmission in the striatum.Here, we show that inhibition or genetic ablation of the MC4R-signaling system also normalizes levels of AMPAR-mediated

Fig. 3. MC4R deletion rescues ventral striatal synaptic transmission deficits in SAPAP3-null mice. ( A) Representative traces of field potential recordings at 10-,

30-, and 50-μA stimulus intensities in wild-type, SAPAP3-null, MC4R-null, and double null ventral striatum (N1 first negative peak representing axonal

stimulation, N2 second negative peak representing AMPAR antagonist sensitive potential). (B) Current-voltage relationship plot of ventral striatal N2fi

eldpotentials show reduced SAPAP3-null responses are rescued in double null acute slices. ( C ) Paired-pulse ratios at 50-ms interstimulus intervals are increased in

SAPAP3-null mice and rescued in double null mice.

10762 | www.pnas.org/cgi/doi/10.1073/pnas.1308195110 Xu et al.

Page 5: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 5/9

synaptic activity. Taken together, these results suggest that distinctmGluR5 and MC4R signaling pathways converge to affect thesame synaptic property within striatal neurons. Further dis-section of the relative contribution and molecular events of thesepathways to this common outcome could thus contribute toidentification of potential targets for therapeutic interventionin compulsive behavior.

In summary, we have found an interaction of striatal MC4Rand SAPAP3 signaling that may help foster development of newpharmacologic treatments for patients with pathologically com-pulsive behavior or eating disorders. It will be important to furtheranalyze the relationship between MC4R-SAPAP3 interactions andspecific subtypes of human illness. Loss of MC4R function modelshuman monogenic obesity, which is particularly sensitive to hy-perphagia of calorically dense foods (26). By contrast, humanmutations in SAPAP3 have been associated with trichotillomania,a rare disorder that sometimes involves ingestion of a non-nutrient object (hair) (2). These polar opposites in human be-havior suggest that the interaction of MC4R-SAPAP3 signalingmay be critical for the evaluation and selection of nutrients forconsumption. Although the SAPAP3 model of compulsive groom-ing has good face validity and predictive validity for OCD-like

behaviors (27), caution should be applied when interpretingthe results. Compulsive or stereotypic grooming has also beenlinked to numerous stress-inducing events, including stimulantexposure, feeding, sexual behavior, social interactions, and ex-ploration of a novel environment (28). Furthermore, repetitivemovements or behaviors are observed in multiple conditions,including intellectual disability, autism, dementia, Tourette syn-

drome, and temporal lobe epilepsy (29), indicating that multipledistinct forms of neural dysfunction may result in a commonbehavioral manifestation. Future work will need to delineatethe specificity of the interaction of SAPAP3-MC4R signalingin diverse disease states.

Materials and Methods

Animal procedures were performed in accordance with University of

Texas Southwestern Medical Center Institutional Animal Care and Use

Committee guidelines. Mice were handled in accordance with the Guide

for the Care and Use of Laboratory Animals as adopted by the US National

Institutes of Health. Specific protocols were approved by the Institutional

Animal Care and Use Committee. SAPAP3-null mice were provided by

Guoping Feng (Massachusetts Institute of Technology, Cambridge, MA),

and MC4R-null mice were provided by Joel Elmquist of University of Texas

Fig. 4. SAPAP3 is required for the metabolic deficits characteristic of MC4R-null mice. ( A) Loss of SAPAP3 causes a gene dosage-dependent rescue of body

weight (significant time × genotype interaction, F 40,679 = 12.27, P < 0.001), and body length (F 1,67 = 26.91, P < 0.001), whereas complete loss of SAPAP3 rescues

adiposity (F 1,72 = 83.46, P < 0.001). (B) Loss of SAPAP3 reduces food intake (significant time × genotype interaction, F 35,565 = 9.14, P < 0.001) and feed ef-

ficiency (F 1,76 = 22.88, P < 0.001) in MC4R-null mice. (C ) Loss of SAPAP3 causes a gene dosage-dependent rescue of fasting glucose ( F 1,71 = 19.56, P < 0.001;

data are presented as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001). (D) Representative pictures of 20-wk-old MC4R-null, wild-type, SAPAP3-null, and

double null mice illustrate that elimination of SAPAP3 restores normal weight in the MC4R-null background.

Xu et al. PNAS | June 25, 2013 | vol. 110 | no. 26 | 10763

     M     E     D     I     C     A     L     S     C     I     E     N     C     E     S

Page 6: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 6/9

Southwestern Medical Center. For a detailed description of methods, see

SI Materials and Methods.

ACKNOWLEDGMENTS. We thank Noelle Williams, Joel Elmquist, Héctor DeJesús-Cortés, James Potash, and John Wemmie for critical review of themanuscript. We thank Aaron Burket for technical assistance. We thankGuoping Feng for generously providing breeding pairs of SAPAP3-null mice.

This work was supported by funds from The Hartwell Foundation (to A.A.P.);National Institutes of Health (NIH) Grants DK081185-01, DK081182-01, andMH084058-01A1 (to M.L.);the Dylan TauberResearcher Award from the Brainand Behavior Foundation (to M.L.); the STARS Summer Research Program(Kathryn and Ashley H. Priddy Award for Young Scientists) (to R.H.); NIHgrants (to R.C.M.); NIH Grants R01DK075632, P30DK046200, and P30DK057521(to B.B.L.); and Grant F3DK078478 (to L.V.).

1. Boardman L, et al. (2011) Investigating SAPAP3 variants in the etiology of obsessive-

compulsive disorder and trichotillomania in the South African white population.

Compr Psychiatry  52(2):181–187.2. Züchner S, et al. (2009) Multiple rare SAPAP3 missense variants in trichotillomania

and OCD. Mol Psychiatry  14(1):6–9.

3. Bienvenu OJ, et al. (2009) Sapap3 and pathological grooming in humans: Results from

the OCD collaborative genetics study. Am J Med Genet  150B(5):710–720.

4. Welch JM, et al. (2007) Cortico-striatal synaptic defects and OCD-like behaviours in

Sapap3-mutant mice. Nature 448(7156):894–900.

5. Balthasar N, et al. (2005) Divergence of melanocortin pathways in the control of food

intake and energy expenditure. Cell  123(3):493–505.

6. Farooqi IS (2008) Monogenic human obesity. Front Horm Res 36:1–11.

7. Cui H, et al. (2012) Melanocortin 4 receptor signaling in dopamine 1 receptor neurons

is required for procedural memory learning. Physiol Behav  106(2):201–210.

8. Lim BK, Huang KW, Grueter BA, Rothwell PE, Malenka RC (2012) Anhedonia requires

MC4R-mediated synaptic adaptations in nucleus accumbens. Nature 487(7406):

183–189.

9. Alvaro JD, Taylor JR, Duman RS (2003) Molecular and behavioral interactions between

central melanocortins and cocaine. J Pharmacol Exp Ther  304(1):391–399.

10. Bourne SK, Eckhardt CA, Sheth SA, Eskandar EN (2012) Mechanisms of deep brain

stimulation for obsessive compulsive disorder: Effects upon cells and circuits. Front Int 

Neurosci  6:29.

11. Wu K, Hanna GL, Rosenberg DR, Arnold PD (2012) The role of glutamate signaling

in the pathogenesis and treatment of obsessive-compulsive disorder. Pharmacol 

Biochem Behav  100(4):726–735.

12. Taylor JL, Rajbhandari AK, Berridge KC, Aldridge JW (2010) Dopamine receptor

modulation of repetitive grooming actions in the rat: potential relevance for

Tourette syndrome. Brain Res 1322:92–101.

13. Schiöth HB, Mutulis F, Muceniece R, Prusis P, Wikberg JES (1998) Discovery of novel

melanocortin4 receptor selective MSH analogues. Br J Pharmacol  124(1):75–82.

14. Kask A, et al. (1999) Long-term administration of MC4 receptor antagonist HS014

causes hyperphagia and obesity in rats. Neuroreport 10(4):707–711.

15. Sohn JW, et al. (2013) Melanocortin 4 receptors reciprocally regulate sympathetic

and parasympathetic preganglionic neurons. Cell  152(3):612–619.

16. Wan Y, Feng G, Calakos N (2011) Sapap3 deletion causes mGluR5-dependent silencingof AMPAR synapses. J Neurosci  31(46):16685–16691.

17. Backs J, et al. (2011) Selective repression of MEF2 activity by PKA-dependent pro-

teolysis of HDAC4. J Cell Biol  195(3):403–415.

18. Malenka RC, Kocsis JD (1988) Presynaptic actions of carbachol and adenosine on

corticostriatal synaptic transmission studied in vitro. J Neurosci  8(10):3750–3756.

19. Farooqi IS, et al. (2003) Clinical spectrum of obesity and mutations in the melano-

cortin 4 receptor gene. N Engl J Med  348(12):1085–1095.

20. Cone RD (2006) Studies on the physiological functions of the melanocortin system.

Endocr Rev  27(7):736–749.

21. Bloch MH, Landeros-Weisenberger A, Rosario MC, Pittenger C, Leckman JF (2008) Meta-

analysis of the symptom structure of obsessive-compulsive disorder. Am J Psychiatry 

165(12):1532–1542.

22. Ellacott KL, Cone RD (2006) The role of the central melanocortin system in the reg-

ulation of food intake and energy homeostasis: lessons from mouse models. Philos

Trans R Soc Lond B Biol Sci  361(1471):1265–1274.

23. Kaye WH, Bulik CM, Thornton L, Barbarich N, Masters K (2004) Comorbidity of anxiety

disorders with anorexia and bulimia nervosa. Am J Psychiatry  161(12):2215–2221.

24. Bulik CM, Sullivan PF, Fear JL, Joyce PR (1997) Eating disorders and antecedent anxietydisorders: A controlled study. Acta Psychiatr Scand  96(2):101–107.

25. Wan S, et al. (2008) Presynaptic melanocortin-4 receptors on vagal afferent fibers

modulate the excitability of rat nucleus tractus solitarius neurons. J Neurosci  28(19):

4957–4966.

26. Butler AA, et al. (2001) Melanocortin-4 receptor is required for acute homeostatic

responses to increased dietary fat. Nat Neurosci  4(6):605–611.

27. Yang XW, Lu XH (2011) Molecular and cellular basis of obsessive-compulsive disorder-

like behaviors: emerging view from mouse models. Curr Opin Neurol  24(2):114–118.

28. Audet MC, Goulet S, Doré FY (2006) Repeated subchronic exposure to phencyclidine

elicits excessive atypical grooming in rats. Behav Brain Res 167(1):103–110.

29. Kelley AE (2001) Measurement of rodent stereotyped behavior. Curr Protoc Neurosci 

8:8.8.

10764 | www.pnas.org/cgi/doi/10.1073/pnas.1308195110 Xu et al.

Page 7: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 7/9

Supporting Information

Xu et al. 10.1073/pnas.1308195110

SI Materials and Methods

Animals. Mice were housed in the University of Texas SouthwesternMedical Center vivarium in a temperature-controlled environment

(lights on: 0600–

1800) with ad libitum access to water and stan-dard chow. Mice heterogyzous for synapse-associated protein 90/ postsynaptic density protein 95-associated protein 3 (SAPAP3)and melanocortin 4 receptor (MC4R)-null alleles were bredto generate the following experimental genotypes of littermatemice: wild-type ( Mc4r + / +, Sapap3+ / +), SAPAP3-null ( Mc4r + / +,Sapap3− / −), MC4R-null ( Mc4r − / −, Sapap3+ / +), and double null( Mc4r − / −, Sapap3− / −). Additionally, mice heterozygous for Sapap3( Mc4R+ / +, Sapap3+ / − and Mc4r − / −, Sapap3+ / −) were also usedfor body weight studies. Floxed-MC4R (MC4Rlox/lox ) mice

 were generated by inserting loxP sites to flank the single exon(5′loxP site into 5′UTR), provided by Brad Lowell (Beth IsraelDeaconess Medical Center). MC4Rlox/lox  mice were bred to EIIA-Cre mice to produce Mc4r null mice (Mc4rdelta/delta). Mc4rdelta/delta

mice were obese, with increased linear length and lean mass,

similar to what has been reported for Mc4r null mice and also forthe loxtb-Mc4r mice. MC4Rlox/lox  mice were crossed to Sapap3− / −

mice to generate Mc4r  f/f , Sapap3− / − breeding pairs. All mice arein C57BL6 background.

Quantification of Grooming. Two methods for quantification of grooming were used: a rapid grooming assay and a comprehensive24-h Laboras-based assay.

 Rapid grooming assay. Between 0800 and 1700 hours, test animals were acclimated for 30 min in clear housing and then exposed toa small water spray bottle without being sprayed, followed by 

 video recording for 5 min. Immediately following this period, testanimals were sprayed 4 times with water, with the same spray bottle, near the head to induce grooming behavior, and again

 were videotaped for 5 min. This method provided a means forrapid quantification to compare baseline grooming to inducedgrooming. Time spent grooming was then manually determined

 with a stopwatch while observing the video, with the observer blindto genotype and treatment group. Number of grooming bouts wasalso manually determined by observing the video.

 LABORAS, a comprehensive grooming assay. LABORAS (Metris) is asystem that uses a carbon fiber plate to detect behavior-specific

 vibration patterns created by animals. Various behavioral param-eters are determined by LABORAS software processing of the

 vibration pattern. Data were collected uninterrupted over a 24-hperiod, enabling comprehensive quantification of basal groomingtime, bouts and locomotor activity in the home cage environmentthroughout the light–dark cycle. Before data collection, test ani-mals were acclimated in the test room for 1 wk. Then, test animals

 were placed in a standard cage atop the carbon fiber platforms.Vibrations were recorded for 24 h, and then the animals wereremoved. Vibration data were processed via LABORAS 2 soft-

 ware. Further assessment of locomotor activity (shown above asdistance in meters traveled) indicated that changes in grooming

 were not due a nonspecific effect on locomotor activity, asSAPAP3-null and double null mice display similar levels of locomotor activity (Fig. S5).

Surgery Methods. Intracerebroventricular (i.c.v.) infusion of com-pounds was performed as per our established methods (1). Briefl y,male mice were anesthetized and implanted with osmotic mini-pumps (Alzet 1004) with cannulas placed in the left lateral ven-tricle 3 mm deep to the pial surface, −0.3 mm anteroposteriorrelative to bregma, and 1.3 mm lateral to midline. HS014 (Tocris)

 was delivered for 4 wk at the following concentrations: 0.1 mM,0.3 mM, 1 mM, with artificial cerebrospinal fluid as control.HS014 is a cyclic melanocyte stimulating hormone analog with∼300-fold higher af finity for the MC4R than α-melanocyte stim-ulating hormone. These concentrations of HS104 were selectedbased on previous publications that reported an effect in thisrange on feeding and anxiety-like behaviors (2). Grooming wastested in the LABORAS 24-h assay before surgery and 2 wk aftersurgery. Pumps were then surgically removed, and grooming wastested again 4 wk after pump removal.

Viral Delivery. Sixteen- to 17-wk-old male Sapap3− / − , Mc4r  f/f  mice were evaluated in the LABORAS system for 24 h before surgery to ensure expression of the excessive grooming phenotype. Allmice showing obsessive grooming were then randomly assignedinto two groups to receive either adeno-associated virus (AAV)encoding both Cre recombinase and GFP: AAV-Cre-GFP (AAV2/ 1.CMV.HI.GFP-Cre.WPRE.SV40, 2.43 × 1013) or AAV encoding

only GFP: AAV-GFP (AAV2/1.CMV.PI.EGFP.WPRE.bGH,3.69 × 1013) as control. Both AAV viruses were obtained fromthe University of Pennsylvania virus core facility. Mice wereanesthetized using ketamine/xylazine and placed into a stereotaxicapparatus (Kopf Instruments) for targeted delivery of virus tothe cortico-striatal pathway (orbitofrontal cortex and nucleusaccumbens shell). A total of 0.5 μL of virus was delivered bi-laterally in nucleus accumbens shell to achieve maximum cov-erage of this circuit with a 30-gauge needle [anteroposterior(AP) +1.5 mm, mediolateral (ML) +1.5 mm, dorsoventral(DV) −4.2 mm, 10 degree angle and AP +2.2 mm, ML +0.6 mm,DV from −2.3 mm to −1.8 mm, 10 degree angle). For orbito-frontal cortex, the needle was inserted to −2.3 mm depth, and0.1 μL of virus was injected every 0.1 mm of withdrawal untilpositioned at −1.8 mm below dura, for a total of 0.6 μL (AP +

2.2 mm, ML +0.6 mm, DV from −2.3 mm to −1.8 mm, 10° angle). After surgery, mice were allowed to recover on a heating padand were provided with moistened food in their home cage.Five weeks after surgery, assessment of grooming was repeatedin the LABORAS system. Mice were then transcardially per-fused with 4% paraformaldehyde in DEPC-PBS (pH 7.4). Brains

 were dissected, sliced at 30 μm on a microtome and stored inantifreeze solution (30% glycerol in DEPC-PBS) at −20 °C.Slices were mounted and accurate placement was confirmed by location of the GFP signal by fluorescent microscopy (Olympus).The Cre-recombinase mediated loss of MC4R was checked withradioactive in-situ hybridization. The following primers wereused to generate a mouse-specific Mc4r  cDNA probe: forward,5′-ATTACCTTGACCATCCTGAT-3′, and reverse, 5′-ATGT-

CAATTCATAACGCCCA-3′.

Feeding Studies. Mice were weaned and genotyped at 4 wk of age, and then individually housed for feeding studies. Mice were

 weighed at baseline and provided a measured amount of normalmouse chow (Global Diet 2016, Harlan-Teklad). Food intake andbody weight were monitored weekly. Body length (from nose toanus) and body composition (Bruker Minispec q10, Bruker) weremeasured at week 16. Tail vein blood was collected after anovernight fast at week 16, and glucose was measured (One TouchFastTake, Lifescan).

Statistics. Data are presented as mean ± SEM. GraphPad Prism 6software (GraphPad Software) was used to perform all statisticalanalyses. For long-term feeding studies, comparisons between

Xu et al. www.pnas.org/cgi/content/short/1308195110 1 of 3

Page 8: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 8/9

groups were made by two-way ANOVA followed by a Bonfer-roni’s post hoc analysis to determine group differences. One-way 

 ANOVAs followed by Tukey ’s post hoc analyses were used forgroup comparisons of grooming, body length, body composition,feed ef ficiency, and glucose. P ≤ 0.05 was considered to be sta-tistically significant.

Electrophysiology. Parasagittal slices (250 μm thick) containing thenucleus accumbens (NAc) were prepared from SAPAP3-null,

MC4R-null, double null and wild-type littermates on a C57BL/6background (12–16 wk of age). Briefl y, foll owing euthanasi aunder isoflurane, brains were quickly removed and placed in anice-cold low sodium/high sucrose dissecting solution. Slices werecut by adhering the two sagittal hemispheres of brain containingthe NAc core to the stage of a Leica vibroslicer. Slices were allowedto recover for a minimum of 60 min in a submerged holdingchamber (∼25 °C) containing artificial cerebrospinal fluid (aCSF)consisting of: 124 mM NaCl, 4.4 mM KCl, 2.5 mM CaCl2, 1.3 mM

MgSO4, 1 mM NaH2PO4, 11 mM glucose, and 26 mM NaHCO3.Slices were then removed from the holding chamber and placedin the recording chamber, where they were continuously perfused

 with oxygenated (95% O2 /5% CO2) aCSF at a rate of 2 mL/minat 30 ± 2 °C. Picrotoxin (50 μM) was added to the aCSF to block GABA  A -receptor mediated inhibitory synaptic currents.

Field potential recordings were obtained by stimulating ex-citatory afferents with a biopolar nichrome wire electrode placed

 within the NAc, roughly 100 μm dorsal and rostral to the re-cording site. Recordings were performed at 0.33 Hz at varyingstimulus intensities using a Multiclamp 700B (Molecular Devices).The peak amplitude of the N2 or AMPAR antagonist sensitivecomponent of the recordings was measured. Paired-pulse ratios(PPR) were acquired by applying a second afferent stimulus of equal intensity 50 ms after the first stimulus and then calculatingN2:1/N2:2. For a given intensity for each recording, the peak amplitudes and PPR of four responses were averaged.

1. Pieper AA, et al. (2010) Discovery of a proneurogenic, neuroprotective chemical. Cell 

142(1):39–51.

2. Kask A, et al. (1999) Long-term administration of MC4 receptor antagonist HS014

causes hyperphagia and obesity in rats. Neuroreport 10(4):707–711.

Fig. S1. Pharmacologic agonism of MC4R signaling induces compulsive grooming in wild-type mice. i.p. delivery of the MC4R agonist melantonan II (MTII) at 2

mg/kg significantly increased both basal and spray-induced grooming in the 5-min spray test 1 hr later, both in terms of total time spent grooming and total

number of grooming bouts. MTII had no effect on MC4R-null mice.

Fig. S2. GFP expression pattern after stereotaxic viral delivery demonstrates accurate targeting of medial orbital cortex and nucleus accumbens shell.

Xu et al. www.pnas.org/cgi/content/short/1308195110 2 of 3

Page 9: ocd_PNAS-2013-Xu-10759-64

7/28/2019 ocd_PNAS-2013-Xu-10759-64

http://slidepdf.com/reader/full/ocdpnas-2013-xu-10759-64 9/9

Fig. S3. Radioactive in situ hybridization for MC4R transcript 5 wk after stereotaxic delivery of AAV-Cre-GFP (Cre-GFP) demonstrates speci fic loss of MC4R

in medial orbital cortex and nucleus accumbens shell, compared with delivery of AAV-GFP (GFP) control. Boxed region indicates relevant targeted regions.

Results shown are representative of no less than five sections for five mice each.

Fig. S4. No genotype-specific differences were noted in the N1 (axonal) component of field potential recordings.

Fig. S5. SAPAP3-null and double null mice display similar levels of locomotor activity. Thisfinding indicates that changes in grooming were not due to a nonspecific

effect on locomotor activity. Distance is shown in meters traveled.

Xu et al. www.pnas.org/cgi/content/short/1308195110 3 of 3