NCS-1 Deficiency Is Associated With Obesity and Diabetes Type 2 in Mice

Neuronal calcium sensor-1 (NCS-1) knockout (KO) in mice (NCS-1−/− mice) evokes behavioral phenotypes ranging from learning deficits to avolition and depressive-like behaviors. Here, we showed that with the onset of adulthood NCS-1−/− mice gain considerable weight. Adult NCS-1−/− mice are obese, especially when fed a high-fat diet (HFD), are hyperglycemic and hyperinsulinemic and thus develop a diabetes type 2 phenotype. In comparison to wild type (WT) NCS-1−/− mice display a significant increase in adipose tissue mass. NCS-1−/− adipocytes produce insufficient serum concentrations of resistin and adiponectin. In contrast to WT littermates, adipocytes of NCS-1−/− mice are incapable of up-regulating insulin receptor (IR) concentration in response to HFD. Thus, HFD-fed NCS-1−/− mice exhibit in comparison to WT littermates a significantly reduced IR expression, which may explain the pronounced insulin resistance observed especially with HFD-fed NCS-1−/− mice. We observed a direct correlation between NCS-1 and IR concentrations in the adipocyte membrane and that NCS-1 can be co-immunoprecipitated with IR indicating a direct interplay between NCS-1 and IR. We propose that NCS-1 plays an important role in adipocyte function and that NCS-1 deficiency gives rise to obesity and diabetes type 2 in adult mice. Given the association of altered NCS-1 expression with behaviorial abnormalities, NCS-1−/− mice may offer an interesting perspective for studying in a mouse model a potential genetic link between some psychiatric disorders and the risk of being obese.

In order to investigate the potential physiological roles of NCS-1, we used an NCS-1 knock-out (KO) mouse line (NCS-1 −/− ; Hermainski, 2012;Ng et al., 2016). Previous studies showed that young NCS-1 −/− mice are viable and generally healthy. Though NCS-1 −/− mice have a mild cardiac problem at a neonatal stage, i.e., a diminished systolic function, this contraction problem disappears in adulthood (Nakamura et al., 2011). NCS-1 −/− mice show little change in their physical activities, as determined via treadmillanalysis (Nakamura et al., 2017) and open field locomotion (Hermainski, 2012;Ng et al., 2016), but some behavioral deficits are notable. NCS-1 −/− mice showed impaired spatial learning and memory function in the Morris Water Maze Test (Nakamura et al., 2017) and a decreased willingness to work for food (Ng et al., 2016). These behavioral phenotypes were associated with a reduced release of dopamine and brain-derived neurotrophic factor (BDNF) in CA1 presynaptic neurons (Nakamura et al., 2017) and decreased presynaptic dopamine release in striatal neurons, respectively (Ng et al., 2016). Interestingly, NCS-1-deficiency also resulted in anxietyand depressive-like behaviors as demonstrated by elevated plus maze, large open field, forced swim and tail suspension tasks (De Rezende et al., 2014).
Though behavioral phenotypes of NCS-1 −/− mice have been investigated extensively, their obesity, which represents the most apparent adult NCS-1 −/− phenotype, remained uncharacterized so far. Here we show that adult NCS-1 −/− mice, especially when fed a high-fat diet (HFD), are hyperglycemic and hyperinsulinemic, typical symptoms associated with obesity (Modan et al., 1985;Mehran et al., 2012). HFD-fed NCS-1 −/− mice display a significant increase in fat body mass. NCS-1 −/− adipocytes are dysfunctional giving rise to lower serum concentrations of resistin and adipokinine than adipocytes of wild-type (WT) littermates. Importantly, insulin receptor (IR) concentration in NCS-1 −/− adipocyte membrane is severely reduced. Apparently, NCS-1 is required directly to up-regulate IR density in adipocyte membrane ensuring an adequate insulin response to changes in diet. Our data indicate that NCS-1 plays an important role in adipocyte function and that NCS-1 deficiency yields obesity and diabetes type 2 in adult NCS-1 −/− mice. Combining these data with the previously reported anxiety-and depressive-like behaviors of NCS-1 −/− mice (De Rezende et al., 2014;Ng et al., 2016) implicates NCS-1 in a relationship of diabetes type 2 and depression, frequently observed with human patients (Nouwen et al., 2010;Renn et al., 2011;Haljas et al., 2018).

NCS-1 Knock-In and Knock-Out Mouse Lines
Generation of NCS-1-EGFP knockin mice and NCS-1 −/− mice (C57BL6/J background) has been described earlier (Hermainski, 2012;Ng et al., 2016). Briefly, NCS-1-EGFP knockin mice were generated that had the last four translated exons of Ncs-1 flanked by lox-P sites as well as an Egfp modification in exon seven to generate mice expressing NCS-1-EGFP fusion protein.
Knockin mice were then crossed with CMV-Cre transgenic mice (Schwenk et al., 1995), which had been backcrossed to C57BL/6 over 10 generations, producing mice with no detectable NCS-1. NCS-1 −/− mice were backcrossed to C57BL/6J over 10 generations and maintained on the C57BL/6J background. Genotyping was done by PCR as described (Hermainski, 2012;Ng et al., 2016). Animal protocols were in accordance with guidelines for humane treatment of animals and were reviewed and approved by the Animals Ethics Committee of the Saarland, Germany. Mice were kept at a regular 12 h day/night cycle. They had ad libitum access to a normal chow diet (ssniff TM R/M-H Pellets, ssniff-Spezialdiäten GmbH) and water. HFD (ssniff TM EF R/M D12492, ssniff-Spezialdiäten GmbH) was fed from 6 weeks of age until the termination of the experiment as indicated in figures. Throughout we used male mice for experiments. Mice were killed by cervical dislocation at the end of the study.

Measurement of Food and Water Consumption
Food and water consumption of mice was monitored using a nutrition monitoring system (Infra-e-Motion, Hamburg). The system also records every second cage activity of the mouse. Normal chow was provided as a powder (ssniff TM EF R/M-H, ssniff-Spezialdiäten GmbH). In case of HFD, the food (ssniff TM EF R/M D12492, ssniff-Spezialdiäten GmbH) was pulverized with a grinder and mixed with 25% ssniff TM EF R/M-H powder to make a pourable chow. After a 5-day trial, phase food and water supply were replenished, and data collection was started for a period of 4 days. Data were transmitted online to a central computer outside the animal facility.

Blood and Serum Analysis
Blood glucose levels were determined from whole venous tail blood using a glucose monitoring system (Roche Accu Chek). Serum was obtained by collecting blood from the submandibular vein. Blood samples were then incubated at room temperature for 60 min and centrifuged at 3,000 g for 10 min. Resulting serum was stored at −80 • C for further use.

Glucose and Insulin Measurements in Glucose Tolerance Tests
Glucose tolerance tests were carried out in 16-week-old animals following a 16 h fast. After measuring fasted blood glucose levels, mice were injected intraperitoneally with 2 mg glucose/g body weight (20% glucose solution, Sigma-Aldrich, St. Louis, MO, USA). Blood glucose levels were determined at 30, 60, 90, and 120 min postglucose injection. Insulin concentrations were measured after a 16 h fast and 30 min post glucose injection in blood samples taken from the vena facialis.

Preparation of Fat Body Lysate
Fat body (1 g per 1 mL lysis buffer) was transferred to a Potter homogenizer and homogenized on ice in 20 mM HEPES (pH 7.4), 125 mM KCl. 0.05% Tween 80, 100 nM CaCl 2 containing protease inhibitor cocktail according to directions of the supplier (Roche). Then 1 M sucrose was added to a final concentration of 300 mM. Lysis was continued for 45 min at 4 • C on a rotarod. It was centrifuged twice for 5 min at 2,000 g at 4 • C. The supernatant was centrifuged at 50,000 g for 1 h at 4 • C. The supernatant was saved as cytosol fraction and kept in aliquots at −80 • C until further use. The membranous pellet was resuspended in 300 mM sucrose, 10 mM HEPES (pH 7.4), 10 mM Tris-HCl (pH 8.0), 0.1 mM MgCl 2 , 100 nM CaCl 2 buffer containing 1 tablet/10 mL protease inhibitor cocktail (Roche). Final protein concentration was ≥2 mg/mL.

Densitometry Analysis
Scanned films of Western blots were analyzed and quantified using ImageJ software available in the public domain.

Preparation of Paraffin-Embedded Adipose Tissue Sections
Mice were perfused with Zinc-Formal-Fixx TM (ThermoFisher Scientific, Waltham, MA, USA). Gonadal fat tissue was embedded in paraffin and sectioned at 8 µm. Deparaffinized and rehydrated tissue sections were stained with anti-NCS-1 or anti-GFP antibodies followed by incubation with biotinylated secondary anti-rabbit-and, respectively, anti-chicken IgG antibodies and staining with ABC/DAB solutions (Vector Laboratories).

Statistical Analysis
All numerical data are given as mean ± SEM. Repeated measurements MANOVA with Bonferroni post hoc test or ANOVA and Bonferroni post hoc test were used for statistical analysis with the help of the programme Prism5 (GraphPad). Statistical significance is indicated with * for P < 0.05, * * for P < 0.01, * * * for P < 0.001. Replicates and number of animals are indicated in Legends.

NCS-1 Deficiency Affects Adipokine Serum Concentration
Prime targets of insulin are muscle, liver and fat cells. NCS-1 expression is undetectable in liver and muscle (Nef et al., 1995), but prominant in adipose tissue. Western blots of fat body lysate and immunostaining of paraffin-embedded slices of gonadal adipose tissue demonstrated that both small and large adipocytes express NCS-1 (Supplementary Figures  S3A,B). Since obesity-linked insulin resistance may be associated with dysfunctional adipocytes (Friedman and Haalas, 1998;Steppan et al., 2001;Galic et al., 2010;Ouchi et al., 2011), we concentrated on adipocytes for further analysis of the NCS-1 −/− phenotype. Adipocytes are an endocrine organ and source of adipokines and cytokines, for example, of leptin, TNF-α, IL-6, adiponectin, and resistin, which have been implicated in modulating insulin sensitivity and energy balance (Friedman and Haalas, 1998;Berg et al., 2001;Steppan et al., 2001;Mojiminiyi et al., 2007;Galic et al., 2010;Ouchi et al., 2011). Normally, plasma concentration of leptin, an important hormone involved in the regulation of energy expenditure, positively correlates with body mass (Friedman and Haalas, 1998). Also, our data revealed, a positive correlation between body mass and leptin for normal chow-and HFD-fed NCS-1 −/− mice and WT littermates. Furthermore, TNF-α and IL-6 plasma concentrations were just as positively correlated with NCS-1 −/− and WT body weight (Supplementary Figures S4A-C). It follows that plasma concentrations of leptin were positively correlated with TNF-α and IL-6 (Supplementary Figure S4D). This is in good agreement with previous data that leptin increases the production of TNF-α and IL-6 (Galic et al., 2010;Ouchi et al., 2011). Thus, deficiency of NCS-1 has no effect on the secretion of leptin, TNF-α and IL-6, respectively.

NCS-1 Deficiency Affects Insulin-Receptor Concentration in Adipocyte Membranes
Next, we asked if NCS-1 deficiency affects the insulin response of NCS-1 −/− adipocytes. Unexpectedly, we found that the adipocyte membrane of NCS-1 −/− mice contained significantly less IR than WT littermates. Quantitative evaluation of Western blots of membrane fractions from fat body lysate of normal chow-fed NCS-1 −/− and WT mice showed that IR signal intensities normalized to those of actin were about twice as high in WT than in NCS-1 −/− mice (IR/β-actin-WT: 0.61 ± 0.04, n = 4; NCS-1 −/− : 0.30 ± 0.06, n = 4; P = 0.006; Figures 5A,B). Analysis of IR concentration in adipocyte membrane of HFD-fed WT mice revealed a significant rise that nearly tripled IR concentration (IR/β-actin: 1.60 ± 0.14, n = 5, P = 0.007). By contrast, there was no significant increase in IR concentration in adipocyte membrane of HFD-fed NCS-1 −/− mice (IR/β-actin: 0.43 ± 0.05, n = 6, P = 0.12; Figure 5B). The data indicates that IR concentration in the membrane of WT adipocytes is exquisitely sensitive to diet. In contrast, IR expression in NCS-1 −/− adipocytes is insensitive to HFD. In conclusion, NCS-1 deficiency leads to a default in diet-sensitive regulation of IR concentration in adipocyte membrane.

Increase of IR and NCS-1 Concentrations in Adipocyte Membrane Are Correlated
Obviously, NCS-1 is involved in HFD-associated IR upregulation in adipocyte membrane. When we analyzed Western blots of fat body lysate of normal chow-and HFD-fed mice, we observed that NCS-1 expression was also HFD sensitive (Figures 5C,D). The data showed that in normal chow-fed mice two-thirds of NCS-1 (63.3 ± 6.5%, n = 4) and in HFD-fed mice almost all NCS-1 protein (91.2 ± 4.5%, n = 6) resided in the membrane fraction of fat body lysate. Importantly, fat-body membrane isolated from HFD-fed mice contained more than two-fold higher amounts of NCS-1 than those of mice fed with normal chow (HFD: NCS-1/β-actin-1.12 ± 0.07; normal chow: NCS-1/β-actin-0.48 ± 0.09; P < 0.001; Figures 5C,D). Thus, both the concentration of NCS-1 and IR are sensitive to HFD. Next, we constructed a plot of NCS-1 vs. IR-concentrations that were observed in fat body lysate of normal chow and resepectively, HFD-fed mice and normalized to β-actin. A linear fit to the data well described the correlation between NCS-1 and IR concentrations ( Figure 5E). The fit provided two important informations. First, if the NCS-1 concentration equals zero, the fit indicated a remaining normalized IR-concentration of 0.28. A very similar IR concentration was detected in the lysate of NCS-1 −/− adipocytes. These data indicate that there is a basal concentration of IR in the adipocyte membrane, which is independent of NCS-1 expression. Second, there is a diet-sensitive component of IR concentration, which is correlated with HFD-sensitive up-regulation of NCS-1 concentration. It suggests that NCS-1 is an important player in the regulation of IR membrane concentration in adipocytes in response to diet changes.
It has been reported that NCS-1 associates with D2 and adenosine A 2A receptors (Kabbani et al., 2002;Navarro et al., 2012). We investigated the possibility that NCS-1 likewise associates with IR like. NCS-1 in membrane fractions of HFD-fed mice was precipitated with anti-NCS-1 antibodies. NCS-1 immunoprecipitates were then analyzed by immunostaining respective Western-blots with anti-IR and anti-NCS-1 antibodies. For control, we used non-specific IgG. The results showed that anti-NCS-1 antibodies immunoprecipitated NCS-1 completely and co-immunoprecipitated a significant amount of IR with it ( Figure 5F). In the reverse experiment we immunoprecipitated IR and co-immunoprecipitated NCS-1 with it. In this case, anti-NCS-1 antibodies stained a band, which consistently migrated slightly more slowly in PAGE (n = 3) than NCS-1 of NCS-1 immuno-precipitates ( Figure 5F). The altered mobility of NCS-1, which may correspond to an additional band seen in overexpression experiments (e.g., Mora et al., 2002), is difficult to understand. Our anti-NCS-1 antibodies do not stain any other protein than NCS-1 ( Figure 5C, Supplementary  Figure S5). Also a spillover of stained light-chain IgG is very unlikely. It may reflect a NCS-1 conformation altered by binding other ions than Ca 2+ (Choudhary et al., 2018;Tsvetkov et al., 2018) or altered by an as yet unknown post-translational modification. In summary, the co-immunoprecipitation experiments suggest that IR and NCS-1 form a complex. Whether this reflects a direct or indirect interaction between IR and NCS-1, however, requires further experiments.

DISCUSSION
We have made two major observations in our characterization of the obese phenotype of NCS-1 −/− mice, which has a late onset and only emerges in adulthood. First, the NCS-1 −/− mice appear to have diabetes type 2. Second, NCS-1 deficiency impairs the up-regulation of IR in adipocyte membrane in response to HFD.
The obese phenotype of HFD-fed NCS-1 −/− mice, which exhibit abnormally increased fat body mass, is associated with hyperglycemia, hyperinsulinemia and dysfunctional adipocytes, in good agreement with a previous report that hyperinsulinemia and obesity are associated with dysfunctional adipocytes (Ouchi et al., 2011). We showed that NCS-1 deficiency specifically affects adipokine secretion of adipocytes. HFD-fed NCS-1 −/− mice have significantly lower serum levels of adiponectin and resistin than HFD-fed WT littermates, whereas serum levels of leptin, IL-6, and TNF-α were unaffected. Previous data suggest that NCS-1 has a regulatory role in dopamine and hormone secretion. Thus, NCS-1 −/− mice exhibit decreased presynaptic dopamine release in striatal neurons (Ng et al., 2016) and reduced release of dopamine and BDNF in CA1 presynaptic neurons, respectively (Nakamura et al., 2017). Conversely, overexpression of NCS-1 in PC12 cells increased evoked growth hormone release (McFerran et al., 1998). In this context, it is of note that NCS-1 interacts with IP 3 R in cardiac tissue and that NCS-1 overexpression enhances IP 3 R-stimulated phosphorylation of CaMKII-δ (Nakamura et al., 2011;Nakamura and Wakabayashi, 2012). Thus, NCS-1 may affect global Ca 2+ signals. It is likely that this effect is pertinent to the proposal that NCS-1 exerts regulatory functions in exocytosis of dense core vesicles (McFerran et al., 1998). A regulatory effect of NCS-1 on regulatory Ca 2+ -signaling may provide a common molecular basis for NCS-1 effects on hormone secretion in neurons as well as in adipocytes. Note that this does not concern insulin secretion of beta cells, which do not express NCS-1 (Supplementary Figure S5).
Adipocyte membrane of HFD-fed WT mice contained significantly more NCS-1 than WT mice fed with normal chow. In the light of reports that NCS-1 is up-regulated in the prefrontal cortex of schizophrenic and bipolar patients (Koh et al., 2003;Bai et al., 2004), it is interesting to see that a metabolic insult like HFD up-regulates NCS-1. The HFD-induced up-regulation of NCS-1 in adipocyte membrane was directly correlated with that of IR. By contrast, HFD-induced up-regulation of IR is impaired in NCS-1 −/− adipocytes. We conclude that NCS-1 deficiency impedes HFD-sensitive up-regulation of IR density in adipocyte membrane. Our data also indicate that adipocytes require NCS-1 for HFD-sensitive IR up-regulation. The relatively low IR concentration in NCS-1 deficient adipocyte membrane may be insufficient for a normal adipocyte response to insulin and, thus, lead to the obese phenotype in NCS-1 −/− mice (Soli et al., 1975). The consequences of insulin signaling in adipocytes will be discussed below.
IR activation by insulin induces cytoplasmic binding of IR substrates IRS-1 and IRS-2. They play key roles in transmitting signals from the IR to intracellular pathways. Phosphorylation on multiple sites upon activation of IR regulates IRS-1 and IRS-2 distribution between membrane and cytosol and their interaction with downstream partners. Down-regulation of IRS-1 and IRS-2 is associated with an obese phenotype (Kaburagi et al., 1997;Boucher et al., 2016;Kubota et al., 2017). Our Western-blot data indicated that IRS-1 concentrations in adipocytes of HFD-fed NCS-1 −/− mice are significantly reduced in comparison to WT littermates, both in cytosol and membrane fractions of fat body lysate (Supplementary Figures S6A,B). This result is in good agreement with the observation that IRS-1 is in adipocytes a major target in insulin signaling. IRS-2 concentration, on the other hand, was similar in cytosol fractions of NCS-1 −/− and WT fat body lysate. Only membrane fractions of NCS-1 −/− contained a significantly reduced IRS-2 concentration in comparison to WT (Supplementary Figures S6C,D). The data shows that NCS-1 deficiency affects down-stream of IR IRS activity and signaling, most likely as a consequence of reduced IR expression in NCS-1 −/− adipocyte membrane. In this context, it will be important to see whether the NCS-1 effect on IR concentration and IR signaling is specific for IR or whether NCS-1 affects also other receptors, for example, insulin growth factor-1. The activation of insulin signaling induces translocation of the glucose transporter GLUT4 to the plasma membrane. We observed that GLUT4 is expressed in NCS-1 −/− adipocytes at a lower level in comparison to WT (Supplementary Figures  S6E,F). The data suggest that adipocytes of HFD-fed NCS-1 −/− mice have a lower rate of glucose-uptake than adipocytes of HFD-fed WT littermates. Previously it has been reported that overexpression of NCS-1 in 3T3L1 tissue culture cells inhibits the translocation of GLUT4 to the plasma membrane involving a phosphatidyl-4 kinase-dependent pathway (Mora et al., 2002). How NCS-1 in combination with phosphatidyl-4 kinase influences the sorting and/or translocation process of GLUT4 remained however unclear. At this stage, it is difficult to compare data obtained by overexpression of NCS-1 in tissue culture cells with ours, which were obtained with NCS-1 −/− adipocytes in HFD-fed adult animals.
Our data indicate that there are at least two pathways for controlling IR concentration in the adipocyte membrane; one is NCS-1 independent, the other is NCS-1 dependent. The former provides a basic IR level observed in WT and NCS-1 −/− mice alike. This situation provides a likely explanation for phenotypic differences observed with NCS-1 −/− and FIRKO mice, where IR expression in adipose tissue is completely absent (Blüher et al., 2002;Entingh-Pearsall and Kahn, 2004;Boucher et al., 2014Boucher et al., , 2016Kubota et al., 2017). The latter involves a direct correlation between NCS-1 and IR concentrations in adipocytes. Whether this reflects a direct association between NCS-1 and IR, as suggested by our co-immunoprecipitation results, requires further study along the lines previously shown for D2-and adenosine A 2A receptors (Kabbani et al., 2002;Woll et al., 2011;Navarro et al., 2012). This also includes the possibility that NCS-1 attenuates IR internalization, as was shown for D2 receptor internalization when NCS-1 is overexpressed in HEK 293 cells (Kabbani et al., 2002).
Insulin resistance, decreased plasma levels of adiponectin (hypoadiponectinemia) and decreased plasma levels of resistin are reportedly related to obesity and type 2 diabetes mellitus in patients (Way et al., 2001;Fasshauer et al., 2002;Möhlig et al., 2002;Heilbronn et al., 2004;Mojiminiyi et al., 2007;Chen et al., 2014). The observed effects of NCS-1 deficiency on IR concentration and on adipokinine secretion in combination with the observed hyperinsulinemia, hyperglycemia and obesity strongly suggest that the phenotype of NCS-1 −/− mice resembles type 2 diabetes. Therefore, we propose NCS-1 as a novel player in the development of type 2 diabetes. In light of the potential clinical implications, it will be important to further explore insulin responses in NCS-1 −/− mice.
Our study extends the phenotypic characterization of NCS-1 −/− mice in an important aspect. It demonstrates that NCS-1 deficiency leads to diabetes type 2 and to behavioral phenotypes reminiscent of psychiatric disorders. Thus, NCS-1 −/− mice display depressive-like behavior in forced swim and tail suspension tests (Friedman and Haalas, 1998), and as we previously showed, NCS-1 −/− mice display decreased motivation to work for food and choose the less effortful option for obtaining food (Ng et al., 2016). NCS-1 −/− mice, therefore, exhibit a genetic link for diabetes type 2 and mood-related behaviors. In human patients, depressive symptoms and diabetes type 2 are frequently associated (Nouwen et al., 2010;Renn et al., 2011;Haljas et al., 2018). In future studies, the generation of mouse lines with specific tissue-specific NCS-1-deficiencies may provide further insights into a potential genetic link between some psychiatric disorders and the risk of being obese.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript and/or the Supplementary Files.

AUTHOR CONTRIBUTIONS
OP conceived the project and wrote together with OR the manuscript. OP, OR, JH and KR designed the experiments, analyzed and interpreted the data. OR carried out the Western blot analyses and the immunoprecipitations. JH constructed the knock-out mouse line and measured glucose tolerance and insulin resistance. KR measured hormone levels in mouse serum.

ACKNOWLEDGMENTS
OP is grateful to Jens Rettig for hosting him in his CIPPM institute and his generous support and interest in this work. We especially thank Anja Ludes for her help in breeding mice.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnmol. 2019.00078/full#supplementary-material  and WT littermates (black trace; n = 10) fed normal chow. Measurements were started at an age of 26-31 weeks. Activity was measured every second and is plotted as number of events per minute. (B) Four days average of 24 h profile of cage activity of NCS-1 −/− mice (red trace; n = 9) and WT littermates (black trace; n = 9). Measurements were started at an age of 26-31 weeks. Mice were fed high-fat diet from an age of week 6 onwards. Activity was measured every second and is plotted as the number of events per minute. Black bar indicates light/dark cycle. (B) Eight micrometer paraffin embedded cryosections of gonadal fat tissue of NCS-1-EGFP (KI), NCS-1 −/− (KO) and WT mice were immunostained either with anti-GFP antibodies (EGFP -panels in upper row) or with anti-NCS-1 antibodies (NCS-1-panels in lower row). Antigen-antibody complexes were visualized with secondary biotinylated antibodies followed by staining with 3,3 -diaminobenzidine (DAB). Black scale bar-100 µm.