[1] 7w Estrogen Deficient Male Mice Develop Compulsive Behavior

[1] 200w Background: Aromatase converts androgen to estrogen. Thus, the aromatase knockout (ArKO) mouse is estrogen deficient. We investigated the compulsive behaviors of these animals and the protein levels of catechol-O-methyltransferase (COMT) in frontal cortex, hypothalamus and liver. Methods: Grooming was analyzed during the 20-min period immediately following a water-mist spray. Running wheel activity over two consecutive nights and barbering were analyzed. COMT protein levels were measured by Western analysis. Results: Six-month old male but not female ArKO mice develop compulsive behaviors such as excessive barbering, grooming and wheel-running. Excessive activities were reversed by 3 weeks of 17 ␤-estradiol replacement. Interestingly, the presentation of compulsive behaviors is accompanied by concomitant decreases (p Ͻ .05) in hypothalamic COMT protein levels in male ArKO mice. These values returned to normal upon 17 ␤-estradiol treatment. In contrast, hepatic and frontal cortex COMT levels were not affected by the estrogen status, indicating region-and tissue-specific regulation of COMT levels by estrogen. No differences in COMT levels were detectable between female animals of both genotypes. Conclusions: This study describes the novel observation of a possible link between estrogen, COMT and development of compulsive behaviors in male animals which may have therapeutic implications in obsessive compulsive disorder (OCD) patients.

[1] 119w The ambulatory study was completed prior to the commencement of the running wheel study. Male ArKO mice display significant (p Ͻ .05) increases in wheel-running activity when compared to WT controls (Figure 1A). This excessive running wheel activity was suppressed upon 17␤-estradiol replacement (Figure 1A). In contrast, general ambulatory activity was significantly decreased (p Ͻ .05) in the male ArKO compared to WT (Figure 1B), indicating a wheel-running specific increase in activity in male ArKO mice, instead of a general hyperactivity. No differences were observed in either wheel-running behavior or ambulatory activity in female ArKO when compared to female WT littermates (Figure 1C, 1D), indicating that the effect of estrogen on running wheel behaviors is specific to males only.

[2] 153w Adult male ArKO mice displayed extensive facial hair loss which may be an indicator of extreme grooming or barbering. When the extent of the facial hair loss was scored as shown in Figure 2A, there is a significant difference (p Ͻ .05) between the male ArKO and their WT littermates (Figure 2B). In addition, when the activity of each animal was analyzed for 20 min after water mist spray (a trigger for grooming), male ArKO mice exhibited significantly (p Ͻ .05) heightened grooming activity in terms of frequency of initiation and duration of grooming (Figure 2C, 2D). Three weeks of 17␤-estradiol replacement was sufficient to suppress the grooming activity to WT levels (Figure 2C, 2D), thus indicating that in adult males, induced grooming activity may be regulated by estrogen. No differences in either barbering score (Figure 2B) or grooming activity were observed in female ArKO compared to female WT littermates (Figure 2E, 2F).

[1] 161w The present study demonstrates that the absence of estrogen in adult male mice leads to excessive barbering, wheel-running and grooming activities, paralleled by a significant decrease in COMT protein expression in the hypothalamus. By contrast, we have not noticed stereotypy behaviors (such as barmouthing, jumping, somersaulting or route-tracing) in our ArKO colony during the behavior studies, day-to-day handling of the animals or after performing an observation analysis on our animals (first 5 min of every .5 hours, 10:00 -20:00 for 2 days; data not shown). This could be attributed to the housing conditions of our animals as stipulated by our Institute Animal Ethics Committee, as our mice have to be housed in an "enriched environment" instead of a barren cage. The animals were housed in groups of 2-5, and materials such as tissue papers and cardboard boxes were provided. It has been reported that stereotypy in laboratory mice could be prevented by provision of an enriched environment (Wurbel and Stauffacher 1996).

[2] 351w Whilst adult male ArKO mice displayed a decrease in normal ambulatory activity when compared to WT (Figure 1B), running wheel activity was significantly increased in male ArKO compared to WT (Figure 1A). This specific behavior of excessive running wheel activity was returned to WT levels upon estrogen replacement, although this was not statistically significant (p ϭ .08) when compared to KO levels. Previous studies have consistently reported that running wheel activity increases with increasing levels of estrogen (Fahrbach et al 1985;King 1979;Morgan and Pfaff 2001;Roy and Wade 1975). However, most of these studies were carried out on females only. We found a decrease in running wheel activity in female ArKO mice compared to WT although this was not significant (Figure 1C). This concurs with previous reports that estrogen increases running wheel activity (Fahrbach et al 1985;King 1979;Morgan and Pfaff 2001;Roy and Wade 1975). In contrast, our male ArKO showed increased running wheel activity when compared to WT animals. Previous studies on both male and female ␣ERKO and ␤ERKO mice revealed that whilst there were no differences in running wheel activity in either untreated ␣ERKO or ␤ERKO compared to WT, estradiol benzoate treatment increased running wheel activity in both 10-12 week old male and female WT and ␤ERKO mice, but not in ␣ERKO, suggesting that ER␣ is the predominant receptor involved in estrogen stimulated running wheel activity (Ogawa et al 2003). In contrast, we did not see any difference in either running wheel or ambulatory activity in 12 week old animals (data not shown). Rather, the changes we observed did not occur until six months of age. Face validity of the running wheel experiment is weakened by the fact that although compulsive running/exercise occurs in humans, sometimes in association with eating disorders such as anorexia nervosa (for review refer to Beumont et al 1994), it has not been defined as a variant of OCD (DSM-IV, American Psychiatry Association 2000). However, it has been reported that excessive exercise is positively related to an obsessive-compulsive personality profile in men but not women (Davis et al 1993), which is analogous to our data presented here.

[3] 182w Excessive grooming behavior has commonly been referred to as an obsessive-compulsive related behavior in both animals and humans (Ferris et al 2001;Greer and Capecchi 2002;Nordstrom and Burton 2002;Rapoport 1991). Analysis of grooming behaviors revealed extreme grooming activities in the male ArKO mice that are at a significantly higher level than those of their WT littermates (Figure 2C, 2D). This excessive grooming behavior of the male ArKO was sufficiently restored to WT levels following three weeks of 17␤-estradiol replacement (Figure 2C, 2D), thus indicating that in adult males, induced grooming activity via a water mist spray may be regulated by levels of estrogen. No differences in grooming behavior were observed in female ArKO compared to WT (Figure 2E), indicating that estrogen actions on this OCD related behavior are specific to males only. Clinical trials using flutamide (androgen receptor antagonist) treatment for OCD patients proved no changes in measures of obsessions and compulsions over an eight-week trial period (Altemus et al 1999), suggesting that any effects of steroids on OCD symptoms are more likely to be mediated through estrogen receptors (Altemus et al 1999).

[4] 164w The enzyme catechol-O-methyltransferase (COMT) is involved in the inactivation of catecholaminergic neurotransmitters such as dopamine (Axelrod and Tomchick 1958). Here we show that the male ArKO mice show a decrease in COMT expression in the hypothalamic region (Figure 3A). Decreases in COMT enzyme activity may lead to increases in dopamine levels which are paralleled by increases in grooming and running wheel behaviors. Development of an excessive grooming phenotype has also been reported in the dopamine transporter knockdown mouse model (Berridge et al 2005) which is hyperdopaminergic (Zhuang et al 2001). In addition, low COMT activity may be a risk factor for obsessive-compulsive disorder in male patients (Karayiorgou et al 1997). Furthermore, it has been established that the hypothalamus is involved in grooming (for review see Kruk et al 1998) and running wheel activities (Rhodes et al 2003). Therefore, all these data support the correlation between the low hypothalamic COMT level and excessive grooming/wheel running activities we observed in the 6 month-old male ArKO mice.

[5] 154w Current literature strongly implicates the frontal cortex in the regulation of OCD behaviors (see review Graybiel and Rauch 2000), especially from evidence gathered through neuroimaging (Friedlander and Desrocher 2006). However, our Western blot data did not show any differences between the COMT protein levels in the frontal cortex of both genotypes. Given that in the COMT knockout (COMTϪ/Ϫ) mouse, dopamine accumulation in nerve endings has been found to be more evident in the striatum and hypothalamus than in the cortex (Huotari et al 2002), it would appear that the effects of decreased COMT on dopamine levels are more evident in the hypothalamic region, as this current paper has confirmed. Behavioral phenotypes arising from these increased dopamine levels in the hypothalamus may be due to direct effects on specific areas such as the medial preoptic area previously shown to regulate both grooming (Lumley et al 2001) and running wheel behaviors (Fahrbach et al 1985;King 1979).

[6] 393w COMT exists in mammals in two different forms: soluble (S-COMT), which is found more commonly in glial cells, and membrane bound (MB-COMT), localized predominantly in neurons (Rivett et al 1983). In humans, MB-COMT is reported to represent 70% of total COMT protein in the brain, however in the rat, the soluble form of COMT is the predominant form, representing 70% of total COMT protein (Reenila and Mannisto 2001). Our data show that the soluble form of COMT is the predominant form in the mouse hypothalamus (Figure 3A) and liver (Figure 3E), however in the frontal cortex levels of the soluble form were comparable to membrane bound COMT (Figure 3D). Both forms of COMT catalyze the O-methylation of catecholamines, but MB-COMT has been reported to have a much higher affinity for its substrates (Assicot and Bohuon 1971;Rivett and Roth 1982). In this current study, we found a more pronounced decrease in MB-COMT than S-COMT, indicating that the decreases in COMT activity may be occurring in neurons rather than glial cells. It has been reported that several putative half-palindromic estrogen response elements (ERE) are present in the proximal and distal promoters of the COMT gene (Xie et al 1999), suggesting a role for estrogen in regulating COMT expression. Xie et al (1999) went on to demonstrate that 17␤-estradiol could down-regulate COMT mRNA expression in MCF-7 cells, with the suppression mediated through ERE1 and ERE8. Incidentally, these two EREs are conserved in the murine Comt gene (data not shown). Further studies by Jiang et al (2003) then revealed that 17␤estradiol was able to significantly reduce COMT protein and activity in MCF-7 cells but not in a glial cell line (U138MG). In addition, these effects were blocked by a specific estrogen antagonist (Jiang et al 2003). In contrast to these in vitro studies on the breast cancer cell line MCF-7, our current data show that the lack of estrogen leads to a significant decrease in protein levels of COMT in male mouse hypothalamus and estrogen replacement actually increases COMT expression in vivo (Figure 3A). These in vitro studies, along with our current in vivo data which demonstrate that estrogen has no effects on COMT expression in the frontal cortex or liver (Figure 3D, 3E), suggest a tissue-specific regulation of COMT by estrogen. Our in vivo data show for the first time that estrogen up-regulates hypothalamic COMT levels.

[7] 165w Significant sexually dimorphic differences were found in COMT expression in WT hypothalamus -WT males expressing twice the level of COMT of that in WT females (Figure 3C). This observation paralleled that in humans with the COMT activity reported to be higher in the brains of men than that in women (Chen et al 2004). The difference in COMT levels could be attributed to the higher local concentration of estrogens as a result of higher expression of aromatase in the male brain. In fact, it has been reported that adult female rats have a lower number of aromatase mRNA expressing cells in each aromatase positive region compared to male rats (Wagner and Morrell 1996). It is interesting to note that male ArKO hypothalamic levels of COMT (Figure 3A) are approximately equal to those of the female WT and female ArKO levels (Figure 3B) and the grooming activity levels are similar between these groups (Figure 2C-F). Again, this shows a correlation between hypothalamic COMT and grooming activity.

[8] 265w No obsessive-compulsive behavior has been reported in the COMT knockout (COMTϪ/Ϫ) mouse although sexual dimorphisms have been reported. The female COMTϪ/Ϫ has been reported to display impaired emotional reactivity, whilst male heterozygous COMTϪ/ϩ mice tended to be significantly more aggressive (Gogos et al 1998). Attenuation of the startle response by prepulse inhibition (PPI) is a measure of sensorimotor gating, and it has been reported that OCD patients display PPI deficit (Kumari et al 2001). However, PPI experiments on the COMTϪ/Ϫ mice revealed no significant effect of genotype in either male or female animals (Gogos et al 1998). In contrast to the COMT Ϫ/Ϫ mice, our male ArKO exhibited PPI deficits in an age dependant manner (van den Buuse e t a l 2003). The lack of obsessivecompulsive related behaviors in the COMTϪ/Ϫ mice could be explained by the age of the animals that were studied since the OCD phenotype of our male ArKO did not become apparent until 6 months of age. Both grooming and running wheel studies were performed on 5 and 12 week ArKO and WT of both sexes but no significant differences between the two genotypes were observed (data not shown). Therefore, by the same token, no OCD behavior would be apparent in the reported behavioral studies (Gogos et al 1998) on the 11-16 week COMTϪ/Ϫ mice. We expect that any excessive or compulsive behaviors would not be noticeable until 6 months of age in these COMTϪ/Ϫ animals. In this current study, we observed that only 6-month male ArKO mice show OCD related behaviors parallelled by decreased COMT levels in the hypothalamus.

[9] 55w As estrogen is required during neuronal development, we cannot rule out the possibility that the OCD related phenotype presented in the male ArKO mice is a result of estrogen deficiency during development. Nonetheless, just 3 weeks of 17␤estradiol replacement did restore the COMT levels in the ArKO hypothalamus with concomitant amelioration of the compulsive behaviors.

[10] 44w In summary, we have established for the first time that estrogen up-regulates COMT protein levels in the male hypothalamus. Our study also showed that decreased estrogen levels correlate with a specific decrease in hypothalamic COMT expression and development of compulsive behaviors in male mice.

[1] 105w ArKO mice (C57B6J X J129) were generated by disruption of the Cyp19 gene. Homologous null or wild type (WT) offspring were bred by crossing heterozygous ArKO, and genotyped by PCR. Animals were housed under SPF and environmentally enriched conditions and had ad libitum access to water and soy free mouse chow (Glen Forrest Stockfeeders, Glen Forrest, Western Australia, Australia). Mice were killed by cervical dislocation, the brains removed and dissected in RNAlater TM (Ambion Inc, Austin, Texas) and snap frozen in liquid nitrogen for protein extraction. All mice were treated according to Monash Medical Centre Animal Ethics Committee policies of animal welfare and ethical use.

[2] 108w Each animal (6 month-old male and female ArKO or WT) was anesthetized with a Ketamine (Pfizer, Auckland, New Zealand)/ Xylazil (Troy Laboratories PTY Ltd, Smithfield, New South Wales, Australia) concoction, (.6 ml Ketamine ϩ .4 ml Xylazil in 4ml Phosphate Buffer Saline (PBS)). A 5mm incision was made between the shoulder blades and a placebo or 17␤-estradiol pellet (.05 mg released over a 21-day period; Innovative Research of America, Sarasota, Florida) was placed just under the skin. The incision was then closed with a Michel clip, and Bupivicaine (.5% in 1ϫ PBS) was applied to the wound. Further behavioral experiments then recommenced at three weeks post pellet implantation.

[3] 127w Running Wheel Activity. Mice were placed in a cage with a running wheel with a recording device attached to measure the number of revolutions of the wheel. The animals had voluntary access to the wheel, food and water over a 16 hour night period (17:00 -09:00). Mice were placed in the wheel for four consecutive nights. The first two nights were training nights to allow the mice to become accustomed to the new environment and therefore were not recorded. The number of revolutions ran by the mice on the running wheel over the third and fourth nights were recorded. Values for the number of revolutions over two nights were analyzed statistically using Mann-Whitney test (nonparametric test; GraphPad PRISM version 3.0 for Windows, GraphPad Software, San Diego, California).

[4] 74w Ambulatory Activity. In order to measure ambulatory activity, mice were individually placed in cages equipped with four infra red sensor devices to record their movements over four continuous days/ nights. Recordings of number of times the mice passed through the sensors were taken every 12 hours. Total recordings over the four days and four nights were then added and analyzed statistically by Mann-Whitney test (nonparametric test; GraphPad PRISM version 3.0 for Windows, GraphPad Software).

[5] 72w Barbering Score. Whilst harvesting animals, each mouse was graded according to extent of facial barbering as illustrated in Figure 2A. A score of 0 was given for no loss of facial hair, 1 for clipped whiskers, 2 for a shaved snout and a score of 3 if they had shaved their entire face. These scores were then analyzed statistically by Mann-Whitney test (nonparametric test; GraphPad PRISM version 3.0 for Windows, GraphPad Software).

[6] 58w Grooming after Water Mist Spray. Each mouse was subjected to two squirts of sterile water mist spray and its grooming activity was recorded for 20 min immediately after the mist spray. Grooming duration and frequency of grooming initiations were then analyzed blind, and analyzed statistically by Mann-Whitney test (nonparametric test; GraphPad PRISM version 3.0 for Windows, GraphPad Software).

[7] 213w Western blot analysis was performed on protein lysates extracted from the hypothalamus, frontal cortex and liver of 6 month old male and female, ArKO and WT animals. Sample volume required for 50 g of protein was added and made up to 10l with lysis buffer along with an equal volume of sample buffer (125 mM TRIS, pH 6.8, 4% SDS, .2 M dithiothreitol, .02% bromophenol blue, 20% glycerol). Samples were then denatured at 100°C, 5 min, before SDS-PAGE (15% acrylamide gel, 150 V, 2 hours). Primary antibodies, COMT (1:5,000; BD Biosciences Pharmingen, San Diego, California) and ␤-tubulin (1:10,000; Chemicon International, Temecula, California) were incubated with the membrane overnight at 4°C. The next day, the membrane was incubated with fluorescent secondary antibodies: IR Dye 800 conjugated affinity purified anti-mouse IgG (Rockland Immunochemicals, Gilbertsville, Pennsylvania), Alexa fluor 680 goat anti-rabbit IgG (Molecular Probes, Eugene, Oregon). Image was captured and analyzed using the Odyssey infra red imaging system (LI-COR Biosciences, Lincoln, Nebraska). COMT expression levels were then normalized against the house keeping gene, ␤-tubulin levels. Each Western blot assay was repeated three times, using randomly selected animals from each group if there were more than four animals. Statistical analysis was performed by Mann-Whitney test (nonparametric test) between groups (GraphPad PRISM version 3.0 for Windows, GraphPad Software).

[8] 46w Data were presented as mean Ϯ standard error of the mean (SEM), except for the barbering score data. All statistical analyses were performed by Mann-Whitney test (nonparametric test) between groups (GraphPad PRISM version 3.0 for Windows, GraphPad Software). Differences were considered significant when p Ͻ .05.

[9] 234w As low COMT activity in humans has been associated with OCD, we investigated protein expression levels of COMT. Data presented are representative of three repeated analyses and each assay showed the same trend. Results indeed revealed significant decreases (p Ͻ .05) in membrane bound and total COMT protein expression, whilst the soluble form of COMT was also decreased in the hypothalamus of male ArKO compared to male WT littermates (Figure 3A). Estrogen replacement in male ArKO mice restored membrane bound and soluble COMT expression to normal WT levels (Figure 3A). No differences in COMT expression were observed between female ArKO and WT hypothalami (Figure 3B). When we compared the hypothalamic COMT protein levels between female and male WT animals, a significant gender difference was revealed. The level of COMT protein expression in male WT was double that in female WT mice (Figure 3C), consistent with the difference in WT grooming levels between male and female animals (Figure 2C, 2E). In addition, protein expression of COMT was measured in the frontal cortex and liver of male ArKO and WT animals (Figure 3D, 3E). Results revealed no differences in membrane bound or soluble COMT expression between all groups (Figure 3D, 3E). The soluble form of COMT was the predominant form in the hypothalamus and liver (Figure 3A, 3E), however in the frontal cortex levels of soluble COMT were comparable to the membrane bound form (Figure 3D).

[1] 94w O bsessive-compulsive disorder (OCD) has a prevalence rate of 1%-3% (Weissman et al 1994) and has been ranked as one of the top ten leading causes of disability worldwide of all medical disorders (Lopez and Murray 1998). The burden of OCD straddles different domains, including higher rates of divorce and separation than in individuals without OCD (Regier et al 1990), and significantly impaired instrumental functioning (work, school, home-making and family life) (Amir et al 2000). OCD is characterized by anxiety-provoking and intrusive thoughts and repetitive behavior such as compulsive checking and grooming (Stein 2002).

[2] 169w The first line of treatment for OCD is serotonin reuptake inhibitors (SRIs) but 40 -60% of OCD patients do not respond to SRI-monotherapy (Pallanti et al 2002). Considerable symptomatic improvement may be seen when SRIs are prescribed with a dopaminergic antagonist (McDougle et al 2000). These pharmacological data imply that the serotonergic as well as the dopaminergic systems are involved in mediating OCD. The involvement of the dopaminergic system in OCD is further supported by studies which showed that dopamine receptor agonists could induce OCD behaviors in rat (Szechtman et al 2001). Hormonal influences on OCD have also been implicated, as female patients have reported that their symptoms worsen during menstruation or begin after menopause or within one month of childbirth (Lochner et al 2004) -suggesting that estrogen withdrawal may be involved. Furthermore, male OCD patients have been re-ported to display symptoms at a younger age and with more severe outcomes (Castle et al 1995), supporting the notion that gender may play a role in the development of OCD.

[3] 314w It is becoming clear that OCD is not simply a uniform or homogenous disorder, but a rather heterogeneous condition (Miguel et al 2005) mediated by a range of factors (Lochner et al 2005;Zohar et al 2000). One of the factors that has been associated with higher risk of developing OCD is low catechol-O-methyltransferase (COMT) activity, particularly in male patients (Karayiorgou et al 1997). COMT and monoamine oxidases constitute the major enzymes involved in the metabolic degradation of catechol neurotransmitters such as dopamine, norepinephrine and epinephrine (reviewed by Kopin 1994). Patients with micro deletions of the 22q11 region including the COMT gene have been observed to have an increased prevalence of psychosis and obsessive behavior (Karayiorgou et al 1997;Lachman et al 1996;Schindler et al 2000). Karayiorgou et al has reported that the low activity allele for COMT (COMT-L) is significantly associated in a recessive manner with susceptibility to OCD, particularly in males (Karayiorgou et al 1999). By contrast, other studies have also shown no association between COMT and OCD (Erdal et al 2003;Meira-Lima et al 2004;Ohara et al 1998), or associations only in females (Alsobrook et al 2002). In addition, an overall meta-analysis (Azzam and Mathews 2003) on the data from case-control studies and family-based studies found little evidence to support an association of the COMT gene and OCD, however, there was a statistically significant association (using the fixed effects model) between OCD and the L COMT allele in case control studies. Differences in the literature may be dependent upon the particular ethnic group studied or result from differences in OCD behavioral measures (Azzam and Mathews 2003). Furthermore, the presence of several putative half-palindromic estrogen response elements in the proximal and distal promoters of the human COMT gene (Xie et al 1999), may imply a role of estrogen in the regulation of the COMT gene, adding another layer of complexity in the development of OCD.

[4] 158w Aromatase is the enzyme that converts androgen to estrogen and is encoded by the Cyp19 gene. Aromatase is expressed in the brain, bone and adipose tissue in addition to the gonads (see review by Simpson 2004). Hence, estrogen can be produced locally in the brain and acts in a paracrine, intracrine or autocrine fashion as estrogen receptors are also expressed in the brain (Mitra et al 2003). An ovariectomized or castrated WT animal may still have estrogen production in the brain. The aromatase knockout mouse (ArKO) lacks a functioning aromatase enzyme and therefore is deficient in endogenous estrogens (Fisher et al 1998). Previous analysis of the brain phenotype of the ArKO mouse revealed that the male but not female ArKO mice display apoptosis in the medial preoptic area (MPOA) by one year of age (Hill et al 2004). The MPOA has been reported to be involved in the regulation of grooming behavior in mice (Lumley et al 2001).

[5] 101w This study aims to determine whether estrogen has a role in regulating COMT expression in vivo, and further, to determine a possible role for estrogen in regulating OCD-like behaviors in mice. OCD patients can be afflicted by compulsive repetition of movements, washing and checking (DSM-IV; American Psychiatry Association 2000). To test such behaviors in mice, repetitive behaviors such as running wheel activity were measured, as well as repetitive washing/grooming or excessive barbering, in ArKO and wildtype (WT) control and 17␤-estradiol treated mice of both sexes. In addition, the expression levels of the OCD related gene, COMT, were compared in these animals.

[6] 143w While immersed in RNALater™, the frontal cortex and hypothalamus (including the MPOA), were dissected from whole brain using the optic chiasm, fornix and third ventricle as anatomical landmarks. Tissue samples were weighed and the appropriate amount of 2ϫ lysis buffer (50 mM HEPES, 137 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM NaF, 2 mM EDTA, 10 mM sodium pyrophosphate, 1% NP40, 10% glycerol) was added according to the weight (1000 l per .1 g). Samples were homogenized with a hand held homogenizer and incubated at 4°C for 10 min. Samples were then solubilized on a rotator for 45-60 min at 4°C. After centrifugation (15 min, 14000ϫ g) the resulting lysate was transferred to a new eppendorf tube. Protein assay was then performed on the lysates (10 l) using the BCA protein kit (Pierce Chemical, Rockford, Illinois).