Casein kinase 1α is required to maintain murine hypothalamic pro-opiomelanocortin expression

Summary Hypothalamic pro-opiomelanocortin (POMC) neuron development is considered to play an essential role in the development of obesity. However, the underlying mechanisms remain unclear. Casein kinase 1α (CK1α) was expressed in the embryonic mouse hypothalamus at high levels and colocalized with POMC neurons. CK1α deletion in POMC neurons caused weight gain, metabolic defects, and increased food intake. The number of POMC-expressing cells was considerably decreased in Csnk1a1fl/fl;POMCcre (PKO) mice from embryonic day 15.5 to postnatal day 60, while apoptosis of POMC neurons was not affected. Furthermore, unchanged POMC progenitor cells and a decreased POMC phenotype established CK1α function in hypothalamic POMC neuron development. CK1α deletion led to elevated Notch intracellular domain (NICD) protein expression, and NICD inhibition rescued the PKO mouse phenotype. In summary, CK1α is involved in hypothalamic POMC expression via NICD-POMC signaling, deepening our understanding of POMC neuron development and control of systemic metabolic functions.


INTRODUCTION
Hypothalamic melanocortin system development is involved in regulating energy metabolism. In the melanocortin system, POMC neurons develop through two stages: proliferation, migration, and differentiation at the embryonic stage, followed by axonal growth and synaptogenesis after birth. 1-3 POMC neurons can be recognized in the anterior basal region at embryonic day (E) 10.5, and the Pomc mRNA expression level reaches its peek at E13. 4 Pomc progenitors further differentiate into at least three mature neural subtypes: POMC, NPY, 5 and kisspeptin. 6 After birth, axonal growth of POMC occurs under the regulation of leptin 7 or BDNF. 8 POMC neurons in the hypothalamic arcuate nucleus (ARC) release a-melanocyte-stimulating hormone (a-MSH) to paraventricular hypothalamic nucleus (PVN) via melanocortin 4 receptors (MC4R). 9 In the hypothalamic ARC melanocortin system, POMC, 10 which suppresses appetite and increases energy expenditure, and neuropeptide Y/agouti-related neuropeptide (NPY/AgRP), 11 which enhances appetite and reduces energy expenditure, work together to regulate energy balance in animals.
POMC neuron development is regulated by several factors, including Notch, 12 Shh, 12 Nkx2-1, 13 and Rax 14 during embryonic development. Notch signaling is evolutionarily conserved and regulates POMC expression and neurogenesis through intercellular communication in the embryonic development stage. 12 When Notch receptors in neurons interact with Notch ligands (delta or serrate/jagged) in neighboring cells, the Notch intracellular domain (NICD) is released. 15 Recombinant signal-binding protein for the immunoglobulin kappa J region (RBPJ) forms a nuclear complex with NICD, which activates the transcription of the target gene Hes1, 16 thereby preventing the onset of POMC expression. 17 Activation of NICD leads to an increasing hypothalamic progenitor cell count and decreasing differentiated POMC neurons, resulting in the decrease of POMC expression. 17 In contrast, conditional deletion of the Notch cofactor Rbpjk in NKX2-1 cells causes a decrease in the progenitor cell count and an increase in the differentiated POMC cell count, resulting in the increase of POMC expression. 17 CK1a, encoded by Csnk1a1, comes from the serine/threonine-protein kinase family. [18][19][20] CK1a regulates multiple signaling pathways, including the circadian rhythm, 21 autophagy, 22,23 apoptosis, 24,25 and developmental differentiation. 26 CK1a functions in the maintenance of epidermal progenitors 27 as well as in spermatogonia-committed progenitors. 28 However, the function of CK1a in neural progenitor cells is unknown, although its expression has been reported in the hypothalamus. 26 Therefore, we hypothesized that CK1a may control POMC neuron function and may have important implications for systemic energy metabolism. To test this hypothesis, CK1a function in murine POMC neurons was assessed and whether the loss of CK1a expression affects POMC neuron development was investigated.
Here, we report that CK1a regulates POMC expression via NICD-POMC, which is essential for the regulation of POMC neuron development.

RESULTS
CK1a and pro-opiomelanocortin are co-expressed in developing and adult arcuate nucleus To examine whether CK1a plays a role in hypothalamic development, Csnk1a1 mRNA and CK1a protein expressions of the mouse hypothalamus were measured using real-time PCR and western blotting, respectively. In the hypothalamus of E15.5 male and female mice, we found that Csnk1a1 mRNA and CK1a protein expression levels were at their highest values ( Figures 1A-1D), suggesting that CK1a is involved in the development of embryonic hypothalami. Csnk1a1 mRNA and CK1a protein levels then decreased from E15.5 to postnatal day (P) 60 ( Figures 1A-1D). We next evaluated CK1a expression in the hypothalamus by immunofluorescence. CK1a and POMC were co-expressed in ARC as early as at E12.5 (93.4%) and E15.5 (95.6%) ( Figure 1E) and continued to be expressed in adulthood (96.3%; Figures 1F and S1B, Table S1), suggesting that CK1a may affect POMC expression in mouse hypothalamic development.

Csnk1a1 deletion causes obesity
Pomc-specific CK1a knockout mice were generated by crossing Csnk1a1 Flox/Flox mice with POMC cre mice. To evaluate deletion efficacy, ROSA mT/mG reporter mice were crossed with POMC cre and PKO (Csnk1a1 fl/fl ; POMC cre ) mice for cell lineage tracing, to label cre-mediated neuronal protrusions and recombination cells permanently and genetically, and CK1a immunofluorescence staining was performed. CK1a-positive and GFP-positive cell numbers significantly decreased by 87.7% in the ARC of adult Csnk1a1 fl/fl ;POMC cre ;ROSA mT/mG mice relative to POMC cre ;ROSA mT/mG mice (Figures 2A and 2B), although CK1a was still expressed in some remaining POMC neurons in PKO mice (Figures S2A and S2B). These data indicate that POMC cre deleted CK1a in POMC neurons successfully.
As POMC cre mice could also recombine in the pituitary, pituitary function was examined. The pituitary gene Pomc, thyroid-stimulating hormone beta-chain (Tshb), and growth hormone (Gh) mRNA expressions were unchanged ( Figure S3A). Consistently, immunofluorescence showed that the number of POMC-, GH-, and TSHB-positive cells were unaltered in control and PKO mouse pituitaries ( Figure S3B). These results indicate that CK1a deletion did not cause pituitary defects.
To investigate whether CK1a deletion influences energy balance, we monitored the weights of Csnk1a1 +/+ , Csnk1a1 fl/+ ;POMC cre , and PKO mice under a chow diet (CD). PKO mice were heavier starting at 6 weeks of age (Figures 2C and 2D) and had higher fat mass, whereas their lean mass component was unchanged (Figures 2E and 2F), as seen by an increase in epididymal white adipose tissue (eWAT), inguinal adipose tissue (iWAT), and brown adipose tissue (BAT; Figures S4A-S4D). The higher body weight of PKO mice was also associated with increased food intake and decreased energy expenditure ( Figures 2G, S5A, and S5B). Female PKO mice exhibited similar phenotypes; therefore, we performed subsequent studies on male mice.
Obesity is associated with changes in glucose metabolism, leptin resistance, and hepatic steatosis. The 12-week-old PKO mice exhibited impaired glucose and insulin tolerance ( Figures S6A-S6D). Furthermore, serum leptin levels were higher in PKO mice than in controls ( Figure S7A). To examine whether Csnk1a1  iScience Article knockout affected leptin sensitivity, we injected mice with 1.5 mg/kg leptin intraperitoneally 29 twice a day for three days and then measured the change in body weight gain and food intake. Leptin had little impact on body weight gain ( Figure S7B) and cumulative food intake in PKO mice ( Figure S7C) relative to that in controls. Consistently, western blotting results showed that the increase in phosphorylated signal transducer and activator of transcription 3 (p-STAT3, Tyr705) protein levels, which represents leptin signaling activity, was significantly reduced in PKO mice (Figures S7D and S7E CK1a loss in pro-opiomelanocortin neurons leads to a significant decrease in the numbers of pro-opiomelanocortin-positive cells, resulting in reduced a-melanocyte-stimulating hormone expression To investigate the mechanisms underlying metabolic dysregulation in PKO mice, we examined whether the absence of CK1a in POMC neurons leads to altered neurodevelopment. Immunofluorescence was performed to count POMC + neurons in the PKO ARC, which at E12.5, E15.5, and P60 were 73.5, 26.9, and 17.6% of those observed in controls, respectively (Figures 3A and 3B,and  We next examined the expression of genes encoding hypothalamic hormones. We did not observe significant changes in melanin-concentrating hormone (Mch), thyrotropin-releasing hormone (Trh), corticotropin-releasing hormone (Crh), or orexin (Hcrt) mRNA expression ( Figure S9C). POMC generates a-MSH to PVN through processing enzymes to regulate energy homeostasis. Thus, we examined a-MSH peptide expression in the PVN and a-MSH content of hypothalamic extracts using immunohistochemistry and ELISA from 12-week-old control and PKO mice. The a-MSH content decreased by 63.5% ( Figure 3G), and a-MSH staining decreased significantly in PKO mice neuronal projections (Figures 3H and 3I). However, the decrease in a-MSH did not result from processing enzyme expression changes, as carboxypeptidase E (Cpe), a-amidating monooxygenase (Pam), prolylcarboxypeptidase (Prcp), prohormone convertase 2 (Pc2), and prohormone convertase 1 (Pc1/3) mRNA levels were unaltered ( Figure 3J).
To investigate whether the decrease in the a-MSH content led to obesity, recombinant a-MSH was used to improve PKO mice obesity. Daily a-MSH injections markedly reduced weight gain and food intake in PKO mice (Figures 3K and 3L). Together, the results indicate that CK1a conditional deletion results in the marked reduction of POMC + neurons, and that the decrease in these a-MSH content causes the obese PKO mice phenotype.
CK1a is essential for pro-opiomelanocortin expression in mouse arcuate nucleus As previously reported in multiple neuronal systems, 30,31 cell death or proliferation may lead to a decrease in POMC + neurons during embryonic development. To test this hypothesis, we assessed apoptosis using a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. Compared with those of controls, no significant differences were examined in TUNEL + POMC + neuron numbers in PKO hypothalamus ARCs at E15.5 (Figures 4A and 4C) and P60 ( Figures 4B and 4D). Furthermore, we analyzed the expression of the apoptosis-related genes Casp3 and Bax/Bcl2 in the hypothalami of control and PKO mice. We observed no significant differences at E15.5 ( Figure 4E) and P60 ( Figure 4F). Consistently, western blotting  iScience Article results showed no difference in c-Caspase3 expression between control and PKO hypothalami at E15.5 (Figures S10A and S10C) and P60 (Figures S10B and S10D). These results indicate that CK1a deletion had no influence on apoptosis.
Next, the proliferation of POMC + neurons was examined by POMC and EdU dual staining in the E12.5 hypothalamus before the massive reduction in their number. The proportions of EdU + POMC + cells were not different between control and PKO mice (Figures 4G and 4H). To further determine how CK1a affected POMC neurons, ROSA mT/mG reporter mice were crossed with control and PKO mice for cell lineage tracing, to label cre-mediated neuronal protrusions and recombination cells permanently and genetically. The neuronal-fiber density and GFP neuron numbers were not different between control and PKO mice (Figures S11A-S11C), which was consistent with the lack of apoptotic differences found at E15.5 and P60 ( Figures 4A-4D). In short, neither neuronal architecture nor cellular survival was affected by CK1a deletion in Pomc-expressing cells.

Inhibition of notch signaling restores the phenotype of PKO mice
As CK1a regulates biological processes through the Wnt/b-catenin pathway, we hypothesized that it regulates POMC expression through Wnt/b-catenin signaling. To test this hypothesis, we detected active b-catenin signaling using western blotting. No significant differences were observed in E15.5 ( Figures 5A and 5C) and P60 (Figures 5B and 5D) PKO hypothalami compared with controls. Our data indicate that the Wnt/b-catenin pathway is unlikely to be the primary pathway regulating POMC expression.
Notch signaling negatively regulates POMC expression. Thus, we determined whether Notch signaling was activated in the hypothalamus of PKO mice using western blotting. NICD expression in E15.5 ( Figures 5E and 5G) and P60 (Figures 5F and 5H) PKO hypothalami was increased 2.15-and 2.18-fold as compared to that in control hypothalami. If the regulation of Notch signaling by CK1a greatly influenced POMC expression, we predicted that such regulation may rescue PKO mouse phenotypes. To confirm this, the Notch signaling pathway inhibitor N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester (DAPT) was injected into pregnant dams intraperitoneally from E12.5 until E15.5 and   5M) and from 11 G 2 to 29 G 4 in P60 PKO hypothalami ( Figures S13D and S13E). In summary, CK1a regulates the development of murine POMC neurons via NICD-POMC signaling.

DISCUSSION
Our previous work suggested that CK1a was involved in the development of murine testis 28 and ovaries 32 ; however, its role in neural development and control of energy homeostasis remains elusive. We found that  iScience Article CK1a deletion in POMC neurons led to food intake, metabolic defects, and increased body weight, which were associated with a decrease in POMC-positive cell counts. Our data show that CK1a is required to maintain normal levels of POMC expression in ARC neurons mediated by NICD-POMC signaling during embryonic life.
Neuron numbers are regulated by their migration, proliferation, death, and differentiation. 33 In the murine hypothalamic ARC, Pomc mRNA was first detected at E10.5. Recent studies confirm that POMC progenitor cells differentiate to terminal phenotypes at E15.5 using genetic lineage tracing. As early as E15.5, we found a large reduction in the number of POMC-positive cells in PKO mice, suggesting the impairment of neurogenesis. Compared with the controls, the distribution of POMC neurons in adult PKO mice ARCs was similar, suggesting that POMC cell migration was not affected. Alternatively, the large reduction in the number of POMC neurons may be caused by increased cell death, which would be consistent with the results of previous reports, demonstrating that CK1a knockout leads to cell apoptosis in the intestinal epithelium. 34 Cell lineage tracing proved that these neurons survived in PKO mice. Consistently, we found no significant difference in ARC apoptotic signals between PKO and control mice. Most Cre recombinant neurons in PKO mice lacked POMC immunoreactivity, suggesting that CK1a ablation in POMC progenitor populations disrupted their ability to express POMC. Our results show that CK1a is essential for POMC expression in ARC neurons.
We found that CK1a deletion did not affect POMC expression via Wnt/b-catenin signaling. As a Wnt/b-catenin signaling negative regulator, 35 CK1a deletion causes the activation of b-catenin 18 in the intestinal epithelium, 34 keratinocytes, 36 and spermatogonia. 28 Recent reports showed that knockdown of b-catenin by CaMKIIa-iCre decreased the numbers of POMC neurons, 37 which is not consistent with our results. However, we found that Wnt/b-catenin signaling was not affected in POMC neurons. Thus, the Wnt/b-catenin pathway may not regulate POMC expression.
Finally, we showed that CK1a regulates POMC expression via NICD-POMC signaling. Corroborating the results of earlier reports on the inhibitory effect of Notch on hypothalamic neuronal development, 38 we detected a significant increase in NICD protein expression in CK1a knockout mice, along with a significant decrease in POMC neuron counts. In addition, in-vivo DAPT treatment restored POMC expression. These results provide strong evidence that CK1a regulates POMC expression through the Notch signaling pathway.
Collectively, our findings confirm a mechanism in the regulation of POMC expression that involves CK1a and NICD/Pomc signaling. CK1a conditional deletion in mouse POMC neurons caused a decrease in their numbers and an increase in NICD protein levels, which subsequently resulted in a decrease in a-MSH peptide expression, eventually leading to obesity in PKO mice. Our data confirm a regulatory mechanism of POMC expression during embryogenesis that is critical to our understanding of the pathogenesis of obesity.

Limitations of the study
Our present study suggests that CK1a plays crucial roles in the development of POMC neurons. There are a few limitations to our study. First, CK1a conditional knockout results in the decrease of POMC neurons during embryonic development, however, the function of CK1a in mature POMC neurons after birth requires iScience Article further investigation using TAM-inducible CK1a conditional knockout mice. Furthermore, we have shown that CK1a regulates POMC expression via NICD-POMC signaling, however, the link between CK1a and NICD is unknown. This awaits further molecular genetic investigation.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

DECLARATION OF INTERESTS
The authors declare no competing interests.

Materials availability
This study did not generate new unique reagents and all materials in this study are commercially available.
Data and code availability d This paper does not report original code.
d The data reported in this paper will be shared by the lead contact upon request.
d The additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
The plug day was considered as embryonic day E0.5. On E12.5 and E15.5, female pregnant mice were sacrificed to collect embryos for immunohistochemistry, real-time PCR, and western blotting. The birth day was considered as postnatal day 1. E15.5 to P60 mice were performed for CK1a expression in hypothalamic development. 5-week-old and 12-week-old male mice were performed for glucose tolerance test and insulin tolerance test. 12-week-old male mice were performed for indirect calorimetry, body composition analysis, leptin function and sensitivity, a-MSH treatment and histology. Unless above noted, male P60 mice were used for other experiments.

Physiological measures
Mice were housed individually to measure food consumption. Body weight and food intake were calculated at 8:00 a.m. daily for 5 days to calculate mean 5-day intakes. For the glucose tolerance test, mice were injected with 1.5 g/kg glucose after overnight fasting. At 0, 15, 30, 60, 90, and 120 min post-injection, tail blood glucose concentration (mg/dl) was measured using a handheld glucometer (One Touch). For the insulin tolerance test, mice were injected with 0.5 U/kg insulin after having fasted for 4 h beginning at 9:00 a.m. At 0, 30, 60, 90, and 120 min post-injection, blood glucose levels were measured from tail blood as described above. To measure body composition (lean and fat mass), mice were assayed using EchoMRI.

Plasma leptin and hypothalamic a-MSH level assessment
Plasma leptin and hypothalamic a-MSH protein contents were analyzed using ELISA (Ruixin Biotechnology) according to the manufacturer's instructions. Each mouse serum sample was prepared for leptin ELISA and each mouse hypothalamic tissue homogenate was prepared for a-MSH ELISA. Firstly, standard and sample holes were set up, and different concentrations of 50 mL standard products were added into the standard holes. Blank and sample holes were arranged. We added 40 mL of sample dilution and another 10 mL of ll OPEN ACCESS iScience Article and NPY + cells were manually counted using the ImageJ software for the quantitative analysis of cell number. Only cells with a DAPI nucleus were counted in our analysis.
For the quantitative analysis of fiber density, the relative density of a-MSH + fibers and POMC + fibers in PVN were analyzed using the ImageJ software. To separate the labeled fibers from the background, ImageJ was used to compensate for the difference in fluorescence intensity, and each image plane was binarized. Then, the combined intensity of each image was calculated and summed.
For DAPT Intracerebroventricular (i.c.v.) injection and treatments, DAPT was dissolved in 0.9 NaCl with 5% DMSO to prepare 8.3 mg/mL concentration. I.c.v. injection was carried out on P60 male control (Csnk1a1 fl/+ ;POMC cre ) and PKO mice. DAPT solution was stereotactically injected into the third ventricle immediately. The stereotactic injections into the third ventricle were performed at midline 0 mm, G 0.8 mm posterior from bregma, and À4.5 mm depth from the bregma. Control (Csnk1a1 fl/+ ;POMC cre ) and PKO mice were injected with 2 ml of either vehicle (aCSF) or 0.03 mg/kg DAPT. Then, 6 h after injection, the hypothalamus was dissected for western blotting or perfused for immunohistochemistry studies.

QUANTIFICATION AND STATISTICAL ANALYSIS
Data are expressed as means G SEM. All analyses of real-time PCR and imaging data were performed with at least three independent replicates. Prism 7 (GraphPad Software) was used for all analyses. When two independent groups were compared, a two-tailed Student's t-test was performed. When multiple independent groups were compared, two-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test was performed. Statistical significance was set at P < 0.05.

ll
OPEN ACCESS