Humans and other animals must get adequate nutrition in order to survive. As a result, the body has several systems that work side by side to maintain a healthy body weight and ensure that enough food gets eaten to provide the energy that the body needs. Problems with these systems can contribute towards obesity and other eating disorders.

Certain types of cells in the brain play important roles in controlling weight and appetite, although the genes and cellular mechanisms that underlie these abilities are not well understood. When an animal is deprived of food, so-called AGRP neurons produce molecules that increase appetite and make it easier to gain weight. These neurons also go through structural changes and increase their electrical activity during weight loss. Another group of cells, called the POMC neurons, becomes less active when an animal is deprived of energy.

Using a technique called cell type-specific transcriptomics, Henry, Sugino et al. have now revealed that the expression of hundreds of genes in AGRP and POMC neurons changes depending on whether mice are well fed or food deprived. Food deprivation also affects more genes in AGRP neurons than has been seen in other types of brain cell, and the AGRP neurons are also more sensitive to a change in food intake than POMC neurons.

In the future, this gene expression data and knowledge of the pathways affected by the genes could help researchers to develop new treatments for obesity and other disorders that affect appetite. Henry, Sugino et al. then mapped how these changes in gene expression trigger molecular “pathways” in the neurons that alter how the cells work. These affect many parts of the cells, including ion channels, transcription factors, receptors, and secreted proteins. In addition, food deprivation activated pathways in AGRP neurons that protect the cells from damage and death caused by elevated neuron activity and also triggered signaling pathways that increase body weight.

In the future, this gene expression data and knowledge of the pathways affected by the genes could help researchers to develop new treatments for obesity and other disorders that affect appetite.

Neurons that express Agouti related protein (Agrp) and Proopiomelanocortin (Pomc) comprise two intermingled molecularly defined populations in the hypothalamic arcuate nucleus (ARC) that mediate whole-body energy homeostasis in conjunction with other cell types. AGRP and POMC neurons positively and negatively regulate body weight, respectively (Aponte et al., 2011; Krashes et al., 2011). AGRP neurons transduce circulating signals of energy deficit into increased output of the neuropeptides AGRP and Neuropeptide Y (NPY) as well as the neurotransmitter γ-aminobutyric acid (GABA), each of which contribute to increased appetite and body weight. Correspondingly, weight loss is accompanied by increased Agrp and Npy gene co-expression in AGRP neurons (Hahn et al., 1998), as well as increased electrical activity (Takahashi and Cone, 2005) and synaptic plasticity (Yang et al., 2011; Liu et al., 2012). In contrast, during energy deficit, POMC neurons decrease electrical activity due to inhibitory synaptic input from AGRP neurons (Takahashi and Cone, 2005; Atasoy et al., 2012), and Pomc neuropeptide gene expression is reduced (Schwartz et al., 1997). AGRP and POMC neurons are thus both associated with sensing and counteracting energy deficit states. Because these neurons play major reciprocal roles in energy homeostasis, investigations of the molecular response pathways for AGRP and POMC neurons to weight-loss are critical for identifying key control points associated with regulation of body weight.

AGRP and POMC neurons sense energy deficit, in part, through responding to the metabolic hormones ghrelin, leptin, and insulin. Signaling pathways downstream of the receptors for these hormones have been elucidated (Banks et al., 2000; Kitamura et al., 2006), but most of the other molecular processes involved in the cellular response to systemic metabolic challenge in AGRP and POMC neurons remain unexplored. In light of this, a transcriptome-wide view of gene expression changes can provide a foundation for investigating the neuronal cell biology of these energy homeostasis sensing neurons during a state of energy deficit.

The transcriptional response to food-deprivation has been reported previously using tissue samples from the entire hypothalamus (Guarnieri et al., 2012) or ARC (Li et al., 2005; Jovanovic et al., 2010), but these studies lacked cell type-specificity necessary to understand the molecular response properties of individual neural circuit nodes. Recent approaches employing immunoprecipitation of messenger RNA (mRNA) in molecularly defined and even projection-specific populations (Heiman et al., 2008; Dalal et al., 2013; Ekstrand et al., 2014; Allison et al., 2015) require large numbers of cells, and therefore have been challenging to perform for neurons with small population sizes, such as AGRP and POMC neurons. A transcriptional profile of AGRP neurons has been obtained previously from dissociated tissue in which fluorescently labeled AGRP neurons were sorted and pooled from ∼40 neonatal mice and compared to a similar number of neonatal AGRP neuron-specific Foxo1 knockout mice (Ren et al., 2012). In neonatal mice, cells are readily dissociated, but AGRP neurons are not necessary for early neonatal life and their axons are not developed (Bouret et al., 2004; Luquet et al., 2005), thus, the relevance of neonatal gene expression patterns to those in adult mice is uncertain. Moreover, comparing only one sample from two conditions prevents statistical analysis of differentially expressed genes (DEG). Recent technical improvements in cell sorting and transcriptional profiling methods have enabled the generation of high quality gene expression profiles from small numbers of fluorescently labeled neurons (typically 40–250 neurons) from single adult mouse brains (Sugino et al., 2006; Okaty et al., 2011). Importantly, this permits use of individual animals as replicates for comparing gene expression profiles under different conditions, which is the approach that we used here.

We performed RNA sequencing (RNA-Seq) using AGRP and POMC neurons from ad libitum fed young adult mice as well as from mice after 24-hr food deprivation. We confirmed a small number of previously reported changes in gene expression, and also identified hundreds of additional DEG. These changes in gene expression allowed identification of coordinated signaling pathways that are involved in the response to food deprivation, and we focus here on neuropeptides, G-protein coupled receptors (GPCRs), as well as pathways associated with neuron electrical activity, circadian regulation, and endoplasmic reticulum (ER)-stress signaling. This resource, which includes gene expression profiles for AGRP and POMC neurons as well as methods for validating and evaluating the functional significance of these changes, provides a foundation for in depth analysis of the molecular control points for neuron populations involved in energy homeostasis.

We obtained transcriptional profiles from AGRP and POMC neurons from male young adult mice (6.5–8 weeks old) under ad libitum fed (fed, AGRP, POMC: n = 5 mice each) or 24-hr food deprived (FD, AGRP: n = 6 mice, POMC: n = 5 mice) conditions during the middle of the light phase (percent weight change AGRP.fed: 0.4 ± 1.5%, POMC.fed: 1.0 ± 0.9%, AGRP.FD: −19.4 ± 1.8%, POMC.FD: −17.7 ± 1.6%; mean ± sd). AGRP and POMC neurons were dissociated and manually sorted from the hypothalamic ARC of Npy^hrGFP and Pomc^topazFP transgenic mice, respectively, where the neurons could be identified by fluorescent protein expression (Figure 1A). Nearly all (94.9%) ARC^NPY neurons are reported to express Agrp, which is selectively expressed in this brain area (Broberger et al., 1998), therefore we refer to these cells as AGRP neurons. Each pool of sorted fluorescent neurons (mean ± sd: 102 ± 40.5 neurons, range: 44–214 neurons) from a single brain was used for a separate sample from which RNA was extracted and then reverse transcribed, amplified, sequenced, and analyzed (Figure 1A, Figure 1—figure supplement 1A–C, and Supplementary file 1).

Figure 1 with 2 supplements see all

Download asset Open asset

Manual sorting of fluorescently labeled cells has been previously compared with other cell type-specific RNA isolation methods and shown to yield similar or improved RNA-Seq data quality (Okaty et al., 2011). In particular, manual-sorting can provide high purity samples (Okaty et al., 2011), which we initially assessed by comparing neuropeptide markers between AGRP and POMC neuron populations. The transcript Npy was enriched by 432-fold in AGRP neurons over POMC neurons, and Agrp was 252-fold enriched; conversely, Pomc was enriched by 351-fold in POMC neurons (Figure 1—figure supplement 1D). We also checked for non-neuronal contamination by looking at highly expressed marker genes for astrocytes, myelin oligodendrocytes, microglia, and endothelial cells. A prior effort to obtain transcriptomic profiles from manually sorted AGRP neurons (Ren et al., 2012) showed evidence of differential oligodendrocyte cell contamination in the two samples used for analysis (Figure 1—figure supplement 1E). The differential expression for these primarily non-neuronal markers in our AGRP and POMC neuron samples relative to purified samples of these non-neuronal cell types was at least 300-fold (range: 318- to 8904-fold), indicating a high level of sample purity (Figure 1B, Figure 1—figure supplement 1F).

AGRP and POMC neurons are functionally distinct cell types, and gene expression profiles can be used to measure their molecular differences (Figure 1C). Multidimensional scaling (MDS), which is a method for visualization of similarity in gene expression patterns, showed clear separation between AGRP and POMC neuron populations regardless of deprivation state, and AGRP neurons, but not POMC neurons, were also distinct after food deprivation (Figure 1D). The AGRP and POMC neuron cell types from fed mice showed substantial differences, with 694 DEG (Figure 1E and see ‘Materials and methods’ for criteria). Consistent with this difference between AGRP and POMC neurons, clear separation of the cell types by MDS was maintained after exclusion of Agrp, Npy, and Pomc (Figure 1D), the top 30-most DEG, or even all DEG (Figure 1—figure supplement 1G). Therefore, transcriptional profiles indicate that these canonical markers are not required to distinguish cellular identity of AGRP and POMC neurons.

Both AGRP and POMC neurons have been reported to show alterations of gene expression with food deprivation, but the global transcriptomic response of these two cell populations to an energy deficit state is not known. As expected, Agrp (+4.4-fold, q = 3.3e⁻⁶ [q-value is false discovery rate-corrected p-value]) and Npy (+3.3-fold, q = 1.5e⁻⁷) were upregulated and Pomc expression modestly decreased (−1.6-fold, q = 0.038) (Figure 1—figure supplement 1D). In addition, members of the Fos immediate-early-gene family were upregulated selectively in AGRP neurons from FD mice (Figure 1—figure supplement 2A), consistent with elevated AGRP neuron activity in energy deficit (Takahashi and Cone, 2005), and the selectivity of this response also confirmed that dissociation and sorting did not result in cellular activation, similar to a previous study (Okaty et al., 2011). Using our transcriptomic data, we found that gene expression changes in AGRP neurons were much more extensive than for POMC neurons in response to food-deprivation. AGRP neurons showed 826 DEG after food deprivation (51.8% upregulated), but only 47 genes (46.8% upregulated) for POMC neurons, of which 15 genes were in common between the two cell types (Figure 1C,E and Figure 1—figure supplement 2B). This surprising 17-fold difference in the number of DEG between the two cell types indicates that AGRP neurons are regulated by deprivation states much more strongly than POMC neurons.

Hierarchical clustering of correlated gene expression patterns can reveal groups of genes that show similar responses to food-deprivation across different conditions and cell types. We identified many clusters of highly correlated genes based on their expression levels in AGRP and POMC neurons from fed and FD mice (Figure 1F). One small cluster of 15 genes included Foxo1, Agrp, and Npy, which are reported to be in the same pathway (Kitamura et al., 2006); moreover ghrelin receptor (Ghsr) and Acvr1c, both of which influence Npy expression (Sandoval-Guzman et al., 2012), were also highly correlated (Figure 1F). These data indicate that this resource can be a starting point for investigating patterns of gene regulation in AGRP and POMC neurons.

Our RNA-Seq data were also consistent with some results that have been obtained previously by transcriptional profiling of the hypothalamus without regard to cell type. For example, transcriptional profiling of whole-hypothalamus tissue samples indicated upregulation of the glucocorticoid-regulated gene Cdkn1a (p21) with food deprivation (Guarnieri et al., 2012; Tinkum et al., 2013). Here, we find that Cdkn1a is the second-most highly upregulated gene during food deprivation in AGRP neurons (+166-fold, q = 7.7e⁻⁷), but it is not significantly changed in POMC neurons (−1.5-fold, q = 0.79), illustrating the importance of cell type-specific neuronal profiling. We confirmed increased Cdkn1a expression in AGRP neurons from FD mice by double-label RNA single molecule fluorescence in situ hybridization (smFISH) (p = 1.5e⁻⁵³, Kolmogorov–Smirnov (ks) test, Figure 1G–I). Similarly, microarray profiling of the ARC has shown Asb4 was slightly downregulated (−1.25-fold) with food-deprivation (Li et al., 2005), but, with cell type-specific RNA-Seq we find that Asb4 is strongly down-regulated selectively in POMC neurons from FD mice (POMC: −5.8-fold, q = 2.2e⁻⁹; AGRP: +1.1-fold, q = 0.71, Figure 1C). Finally, Gpr17, which has been reported to be expressed in AGRP neurons based on transcriptional profiling of cells isolated using fluorescence activated cell sorting (FACS) from neonatal mice (Ren et al., 2012) was not detected in our samples from either AGRP or POMC neurons from adult mice, possibly indicating developmental differences in gene expression or differences arising from oligodendrocyte cell contamination of neonatal samples. These comparisons illustrate the advantages of using high purity sorted cells from individual adult mice to allow for statistical analysis of gene expression from intermingled cell types under different conditions.

ER stress pathways

Gene annotation enrichment analysis (Table 1) highlighted, selectively in AGRP neurons, a transcriptional response for genes associated with ER-stress: the unfolded protein response (UPR) and ER-associated degradation (ERAD) of misfolded proteins. This pathway has not been examined previously in AGRP neurons; instead, based on whole hypothalamus analysis, UPR has been primarily associated with overnutrition states and leptin resistance (Ozcan et al., 2009), as opposed to the energy-deficit condition examined here. ER-stress responses occur during high levels of protein translation where unfolded proteins elicit a program of downstream transcriptional responses to increase protein folding and processing in the ER, and this pathway showed the most prominent effect on gene expression that we identified in AGRP neurons. This pattern of differential gene expression was not observed in POMC neurons, indicating that it is not an artifact of the neuronal isolation procedure. Consequently we examined multiple aspects of ER-stress signaling by analysis of our RNA-Seq data, coupled with preliminary evaluation of ER-stress signaling using immunohistochemistry, a cell type-specific UPR pathway reporter, and smFISH.

ER-localized proteins were significantly overrepresented in the group of food-deprivation-regulated genes in AGRP neurons (p = 8.3e⁻¹⁶, hypergeometric-test), and 95% of these DEG were upregulated (Figure 2A). A key UPR marker, Hspa5, which encodes the canonical ER-localized unfolded protein-sensing chaperone BiP, was selectively upregulated in AGRP neurons during food deprivation (AGRP: +3.3-fold, q = 2.5e⁻¹⁰; POMC: −1.2-fold, q = 0.52). Correspondingly, BiP-immunoreactivity was significantly increased by food-deprivation in AGRP neurons (p = 2.5e⁻⁴², ks-test, Figure 2B,D), indicating UPR activation in AGRP neurons. In contrast, we also found that BiP-immunoreactivity was significantly reduced in POMC neurons during food-deprivation (p = 2.5e⁻¹⁹, ks-test, Figure 2C,D), suggestive of reduced protein translational-load in this population.

Figure 2

Download asset Open asset

UPR involves a program of gene expression that is regulated by Ern1/Ire1, which was selectively increased in AGRP neurons by food deprivation (+2.7-fold, q = 0.001). In response to the accumulation of unfolded proteins in the ER lumen, Ern1/Ire1 undergoes a conformational change, which activates endoribonuclease activity for the short-lived transcript Xbp1 and splices it into a more stable mRNA, Xbp1s. These RNA-Seq data reveal increased abundance for the Xbp1s spliced mRNA in AGRP neurons after food-deprivation (Xbp1s/Xbp1, fed: 7.4 ± 1.0%, FD: 15.6 ± 3.8%, p = 0.043, one-tailed t-test; Figure 2E), consistent with activation of Ern1/Ire1 endonuclease activity in AGRP neurons after 24-hr food deprivation.

Xbp1s is a transcription factor that regulates UPR-related gene expression. Two well-established Xbp1s-dependent gene targets (Lee et al., 2003) were selectively upregulated in AGRP neurons after food deprivation: Dnajc3/p58^IPK (+1.9-fold, q = 0.0007) and Dnajb9 (+2.2-fold, q = 0.0003). Many other genes regulated by Xbp1s were also increased in AGRP neurons during food-deprivation, including protein folding chaperones of the Hsp40 (Dnaj), Hsp70 (Hspa), and Hsp90 families, Calr and Canx, as well as protein disulfide isomerases (Pdia3-6), which aid in protein folding by catalyzing the formation of disulfide bonds (Figure 2A). Atm (−3.1-fold, q = 7e⁻⁵), a DNA damage sensing-enzyme that, when reduced leads to increased Xbp1 splicing (He et al., 2009), was downregulated with food deprivation. Taken together, Xbp1-splicing and the patterns of downstream gene expression indicate a previously unreported role for Ern1/Ire1 → Xbp1s signaling in the adaptive response of AGRP neurons to energy deficit.

We also investigated the involvement of two other UPR signaling pathways regulated by ER-bound transmembrane proteins, Eif2ak3 (also called PERK) and Atf6. The Eif2ak3/PERK (AGRP: −2.1-fold, q = 0.23) arm of the UPR pathway is the first to be engaged during ER stress and suppresses translation of most mRNA. We expected that prolonged translational arrest in AGRP neurons was unlikely because it is inconsistent with elevated neuropeptide production in AGRP neurons during energy deficit. Activation of Eif2ak3/PERK leads to mRNAs sequestration into messenger ribonucleoprotein particles that aggregate into stress granules containing the RNA binding protein, TDP43 (Colombrita et al., 2009). To assess stress granule formation, we used anti-TDP43 immunostaining. Elevated stress granule formation was not detected in AGRP neurons from FD mice (p = 0.76, ks-test, Figure 2F,G), which provides preliminary evidence that Eif2ak3/PERK-mediated translational arrest may not be engaged at this 24-hr food deprivation time-point. Moreover, transcripts for ER protein translocation (Srp68, Srp72, Sec61a1, Sec61b1, Sec63, Serp1/Ramp4) and Golgi trafficking (Sec14l1, Sec22b, Sec24d) were upregulated in AGRP neurons, possibly indicating increased protein translation and folding capacity during energy deficit and consistent with the requirement for increased peptidergic neurotransmission for AGRP neuron function.

An additional signaling arm of the UPR is mediated through the ER-bound protein Atf6 (AGRP: +2.0-fold, q = 0.0016). In response to ER stress, Atf6 is subject to proteolytic cleavage and releases an N-terminal transcription factor domain that translocates to the nucleus and initiates gene expression to increase protein folding. To examine the extent of Atf6 nuclear translocation in the response of AGRP neurons to food-deprivation, we developed a Cre recombinase-dependent N-terminal green fluorescent protein (GFP):Atf6 fusion (Samali et al., 2010) reporter of Atf6 cleavage for use in the brain (Figure 2H). After expression in AGRP neurons, we measured the nucleus:cytoplasm ratio of GFP, which was very low and was not significantly increased in FD vs ad libitum fed mice; whereas the ER-stress-inducer, tunicamycin, strongly increased nuclear localization of GFP in AGRP neurons (fed vs FD: p = 0.44, dimethyl sulfoxide (DMSO) vs tunicamycin: p = 2.9e⁻¹⁶, ks-test, Figure 2H–J). Low nuclear localization indicates minimal activation of the Atf6-signaling arm of UPR in AGRP neurons after 24-hr food deprivation. Consistent with this, a transcript selectively regulated by Atf6, Herpud1 (Lee et al., 2003), was not significantly (q = 0.22) elevated in AGRP neurons from FD mice. However, because many Atf6-regulated genes are also regulated by Xbp1s, a robust ER-stress response can be maintained without Atf6 signaling (Yamamoto et al., 2007).

Oxidative stress pathways associated with elevated protein production are also increased in AGRP neurons during food deprivation. The transcription factor Nfe2l2 (also called Nrf2) regulates oxidative stress and is normally rapidly degraded in the cytoplasm, in part through association with Keap1 (Kobayashi et al., 2004), which shows reduced expression with food-deprivation (−2.2-fold, q = 0.05). Nfe2l2/Nrf2 upregulates expression of the key glutathione biosynthetic enzymes Gclc (+2.3-fold, q = 0.0002) and Gclm (+1.8-fold, q = 0.002), which are elevated in AGRP neurons with food-deprivation. Nfe2l2/Nrf2 increases other transcripts associated with oxidative stress, such as Mt1 (+34-fold, q = 0.004), Mt2 (79-fold, q = 0.004), and Srxn1 (+17-fold, q = 1.2e⁻⁶), which are among the most strongly upregulated transcripts in AGRP neurons during food-deprivation. Other oxidative stress transcripts were also upregulated, such as peroxiredoxin 2, Prdx2 (+1.9-fold, p = 0.0004), the transcription factor Hif1a (+1.8-fold, p = 0.005) and the oxidative regulatory enzyme prolyl hydroxylase (P4hb, +2.7-fold, p = 8e⁻⁷). Together, this group of cell type-selectively upregulated genes is consistent with an adaptive response to increased oxidative stress in AGRP neurons during energy deficit.

UPR also typically leads to ERAD of unfolded proteins. E2 ubiquitin conjugating enzyme subunits (Ube2j1, Ube2g2, Ube2q1, Ube2z, Ube4b) and the E3 ligase Park2 were upregulated in AGRP neurons during food-deprivation (Figure 2A). Os9, a lectin that senses misfolded proteins, and Edem3 and Stt3b (Figure 2A), which are enzymes that mark unfolded proteins for ERAD (Sato et al., 2012), were also increased. In addition, Vcp, a component for retrotranslocation of polyubiquitinylated unfolded proteins for degradation, was elevated in AGRP neurons during energy deficit. Collectively, this pattern of gene expression indicates engagement of ERAD during food-deprivation.

UPR is protective to cells for short periods of elevated ER-stress, but prolonged activation can result in apoptosis. The pro-apoptotic Ddit3 (Chop) is typically upregulated during UPR, however Ddit3/Chop was unchanged in AGRP neurons after 24-hr food deprivation (Figure 2K). Instead, a variety of anti-apoptotic transcripts Bcl2l (Bcl-xl), Manf, Mcl1, and the caspase inhibitor Cflar were upregulated (Figure 2K). Mcl1 upregulation is consistent with a previously established pathway Creb3l2 → Atf5 → Mcl1 (Izumi et al., 2012), each member of which is upregulated in food deprivation (Figure 2K). Irak2 (−4.7-fold, q = 3.7e⁻⁵) was strongly suppressed, consistent with reports that its reduction is protective against apoptosis (Benosman et al., 2013). In addition, the cellular stress-induced transcript Atf3 (+15.5-fold, q = 0.005) also promotes neuron survival (Francis et al., 2004) and is increased with food deprivation. This pattern of gene expression provides preliminary evidence of a potential role for anti-apoptotic pathways in AGRP neurons during activation in response to energy deficit.

Taken together, cell type-specific transcriptional profiling, immunohistochemistry, and a cell type-specific UPR reporter construct indicate engagement of ER-stress pathways selectively in AGRP neurons during energy deficit. Increased neuron activity and elevated neuropeptide production associated with food-deprivation is expected to increase the translational-load in AGRP but not POMC neurons, and UPR may serve to cope with elevated neuropeptide and synaptic output. This is due, in part, to Xbp1s signaling and also results in induction of gene expression associated with oxidative stress responses, ERAD, and protection of cells from apoptotic pathways that might otherwise be associated with prolonged ER-stress (Figure 2L).

Regulation of circadian genes by food-deprivation

Expression of genes associated with circadian regulation was strongly altered by food-deprivation. We examined the expression levels of 19 core circadian reference genes (Yan et al., 2008; Rey et al., 2011) and found that after food-deprivation 9/19 of these genes were downregulated in AGRP neurons (Bhlhe40 [Dec1], Bhlhe41 [Dec2], Nr1d1 [Rev-erbα], Nr1d2 [Rev-erbβ], Dbp, Hlf, Tef, Per2, and Per3), and only 3/19 were upregulated (Rorb, Nfil3, Per1) (Figure 3A). For POMC neurons, only one transcript was significantly changed (Nr1d2/Rev-erbβ) (Figure 3A). The nine circadian genes downregulated in AGRP neurons during food-deprivation are established targets of two important circadian transcription factors, Clock and Arntl (also called, Bmal), which heterodimerize to regulate a large number of downstream genes through E-box transcriptional response elements (Rey et al., 2011). We analyzed a collection of E-box regulated genes that have been previously determined by Arntl (Bmal) chromatin binding, bioinformatic analysis, and transcriptional profiling (Rey et al., 2011). E-box-containing genes were over-represented among differentially expressed transcripts in AGRP neurons but not in POMC neurons (p = 0.00075 and 0.31, respectively; hypergeometric-test). For the E-box-containing genes that were significantly differentially expressed more than twofold in AGRP neurons after food-deprivation, 90% (18/20) were downregulated (Figure 3B). Conversely, in POMC neurons, Nr1d2/Rev-erbβ was the only E-box-containing gene that was significantly differentially expressed more than twofold, and it showed increased expression (Figure 3B). This indicates that differentially expressed E-box-containing genes during energy deficit are selectively reduced in AGRP neurons.

Figure 3 with 1 supplement see all

Download asset Open asset

Many genes regulated by E-box transcriptional response elements show reduced expression in Arntl^−/− mice (Rey et al., 2011), but Clock and Arntl expression levels in AGRP neurons were not significantly altered by food-deprivation. Immunohistochemistry for Arntl protein expression in AGRP neurons from fed and FD mice also showed similar levels (fed: 115 ± 4 a.u., n = 158 neurons; FD: 127 ± 4 a.u., n = 195 neurons; p = 0.058, rank sum test, Figure 3—figure supplement 1). An alternative pathway that has been shown to regulate E-box genes is the transcriptional splicing factor Sfpq (Psf) (Duong et al., 2011), which shows selectively increased expression in AGRP neurons after food-deprivation (AGRP: +2.0-fold, q = 0.00064; POMC: 1.0-fold, q = 1.0). However, detailed examination of pathways that regulate circadian E-box containing genes in AGRP neurons during energy deficit states is required.

Ion channels

AGRP and POMC neuron electrical properties are determined by expression of distinct groups of ion channels. Despite the importance of neuron electrical activity for influencing appetite and body weight, only a few ion channel subunits have been determined in these cell types and little is known about their regulation by energy deficit state. Previous work has shown transient receptor potential C (TRPC) ion channel subunit expression in POMC neurons (Qiu et al., 2010), which is confirmed by our RNA-Seq data, and we find that an overlapping set of TRPC channels (Trpc1, Trpc3-Trpc7) is also expressed in AGRP neurons (Figure 4A). Moreover, Kcnq3 is downregulated in AGRP neurons with food-deprivation (Figure 4A), as previously reported (Roepke et al., 2011). Thus, these RNA-Seq data are consistent with established changes in ion channel expression.

Figure 4 with 1 supplement see all

Download asset Open asset

This transcriptional profiling resource also identifies other potentially important ion channel genes that cell type-selectively change expression with energetic state and have not been previously investigated in AGRP and POMC neurons (Figure 4A–C). One of the most striking changes in ion channel gene expression in AGRP neurons during food-deprivation was a sharp reduction of Kcnn3 (also called Sk3, −5.4-fold, q = 0.0006), which is a small conductance calcium-activated potassium channel (SK) that attenuates action potential firing rate during elevated activity. Other SK channels were not appreciably expressed in AGRP neurons (Figure 4A). Electrophysiological characterization of AGRP neurons showed the presence of SK-mediated tail currents, confirmed by blockade with the highly selective SK channel antagonist, apamin (Figure 4D). Notably, the apamin-sensitive SK conductance was absent in FD mice (Figure 4G), consistent with reduced Kcnn3 expression. Apamin blockade of SK channels in AGRP neurons resulted in higher firing rates (Figure 4E,F), which often elicited bursts and plateau potentials (Figure 4—figure supplement 1A). Similarly, AGRP neurons from FD mice showed a comparable increase in excitability, but in this state the neurons were not sensitive to apamin, consistent with greatly reduced SK-channel expression (Figure 4H,I and Figure 4—figure supplement 1B). Thus, by starting from cell type-specific RNA-Seq analysis, we found that Kcnn3 regulates firing rate as well as burst firing in AGRP neurons, and its reduction with food-deprivation plays an important role in increasing the excitability of AGRP neurons.

In addition to the examples highlighted here, this resource shows many other ion channels and regulatory subunits that are altered by energy deficit in AGRP neurons, including ionotropic glutamate receptors, GABA receptors, the ionotropic ATP receptor P2rx4, sodium channels, calcium channels, additional potassium channels and others (Figure 4A–C). This resource provides a list of ion channels expressed in AGRP and POMC neurons, and it is a foundation for a concrete understanding of the electrical properties of AGRP and POMC neurons in basal and energy deprived states.

GPCRs

GPCRs are critical molecular control points for neuronal function. AGRP and POMC neurons express multiple GPCRs, many of which respond to circulating hormones, neuropeptides, or neurotransmitters. RNA-Seq provides a cell type-specific taxonomy of GPCRs expressed under different conditions. For each cell type, many GPCRs were detected at expression levels greater than 20 transcripts per million (TPM) under either fed or FD conditions (AGRP: 59, POMC: 61, AGRP or POMC: 80) (Figure 5A), which is more than previous estimates of GPCRs for hypothalamic regulation of energy homeostasis (Schioth, 2006). Also, several GPCRs show high differential expression between AGRP and POMC neurons: 8 GPCRs were >10-fold differentially expressed in AGRP neurons and 13 GPCRs were >10-fold overrepresented in POMC neurons (Figure 5A). Therefore, differential GPCR expression can separately regulate the function of AGRP and POMC neurons.

Figure 5 with 1 supplement see all

Download asset Open asset

GPCR expression was significantly overrepresented in genes that were differentially expressed after food deprivation in AGRP and POMC neurons (AGRP: p = 0.0002, POMC: p = 3.3e⁻⁵, hypergeometric-test). For AGRP neurons, 13 genes were >twofold upregulated and 11 genes were >twofold downregulated by food deprivation (Figure 5B). All significantly upregulated GPCRs were either G_q-protein coupled (e.g., Ghsr, Nmur2, Gpr83, Htr2a, Nmbr, Hcrtr2) or G_s-protein coupled receptors, several of which show high ligand-independent basal G_s-protein activity (Gpr3, Gpr6, Gpr64). The most strongly downregulated genes were G_i-protein coupled receptors (Npy2r, Hrh1, Hrh3, Htr1a). Conversely, POMC neurons showed the opposite pattern with six differentially regulated GPCR transcripts where all upregulated genes were G_i-protein coupled (4/6) and the downregulated genes were G_q- or G_s-protein coupled (2/6). For example, two GPCRs were reciprocally regulated in AGRP and POMC neurons with food-deprivation, most strikingly G_i-protein-coupled Hrh3 (AGRP: −9.5-fold, q = 4.7e⁻⁸; POMC: +5.3-fold, q = 0.003; Figure 5—figure supplement 1). Taken together, these changes reveal an opposite response pattern during food deprivation for GPCR expression levels in AGRP and POMC neurons, highlighting the importance of cell type-specific transcriptional profiling for assessing gene expression regulation in the brain. Because G_q-protein signaling increases AGRP and POMC neuron activity and G_i-protein signaling reduces activity in these cell types (Krashes et al., 2011; Atasoy et al., 2012), the pattern of GPCR expression is consistent with elevated AGRP neuron activity and reduced POMC neuron activity during food deprivation.

In addition to G_q- and G_i-protein coupled signaling, elevated G_s-protein coupled signaling was prominent in AGRP neurons, especially upregulation of Gpr3, Gpr6, and Gpr64, which are constitutively active G_s-coupled receptors (Uhlenbrock et al., 2002). Moreover, the G_s subunit Gnas was upregulated selectively in AGRP neurons (AGRP: +1.8-fold, q = 0.0002, POMC: −1.1-fold, q = 0.73). However, the physiological consequences of G_s-protein coupled signaling in AGRP neurons are not well understood. We extended our RNA-Seq observations by examining Gpr6 expression using quantitative smFISH, which showed a significant increase in the number of Gpr6 transcripts in AGRP neurons after food deprivation (p = 4.8e⁻¹⁰, ks-test, Figure 5C–E). To test the functional consequence of elevated Gpr6 expression in AGRP neurons, we transduced Agrp^Cre mice with a Cre-dependent virus co-expressing Gpr6 and a fluorescent protein (AGRP^Gpr6 mice; Figure 5F–H). AGRP^Gpr6 mice showed significantly elevated body weight compared to mice expressing a fluorescent protein alone (Figure 5I). These experiments reveal a potential role for Gpr6 and G_s-coupled signaling in AGRP neurons for positive regulation of body weight.

Secreted peptides

Neuropeptide expression is typically used to discriminate AGRP and POMC neurons. Transcriptional profiling shows that these two cell populations are distinguished by several additional secreted proteins (Figure 6A). Neuropeptides were also among the most highly regulated genes in response to food-deprivation in AGRP neurons (Figure 6B). Several neuropeptide transcripts with increased expression in our dataset have been previously shown to increase appetite, for example, Agrp, Npy, Vgf, Pdyn. We also observed increased expression of peptides that are associated with reduced appetite, based on pharmacological experiments: Nmb (+8.0-fold, q = 6.0e⁻⁵), Nts (+107-fold, q = 0.0001), Nucb2 (+2.3-fold, q = 0.0002). However, these neuropeptides may have a local signaling role that promotes appetite. For example, Nmbr (+8.8-fold, q = 0.0008) and Ntsr1 (+41-fold, q = 0.005) were also strongly upregulated in AGRP neurons, are G_q-coupled receptors, and Nmb has been previously demonstrated to increase AGRP neuron electrical activity (van den Pol et al., 2009).

Figure 6

Download asset Open asset

Other genes were upregulated by energy deficit that have not been previously associated with appetite regulation, for example Ccl17 (+21.6-fold, q = 0.0012) and Thbs1 (+12.4-fold, q = 0.0002). We further investigated the chemokine CCL17, which was highly upregulated in AGRP neurons. CCL17 is a 103 aa member of the CC chemokine group with no previously reported functional characterization in the central nervous system. Assessment of Ccl17 expression using double smFISH in AGRP neurons confirmed food deprivation-induced upregulation (p = 1.4e⁻²⁰, ks-test, Figure 6C–E). Microinjection of recombinant CCL17 (500 ng) into the lateral cerebral ventricle elicited a small increase of food intake during the light period (1-hr: p = 0.02, 2-hr: p > 0.05; rank sum-test, Figure 6F). More strikingly, chronic viral overexpression of Ccl17 and GFP in adult mice by selective expression in AGRP neurons (AGRP^Ccl17 mice) resulted in a progressive elevation of weight gain compared to mice expressing a fluorescent protein alone (Figure 6J), providing preliminary evidence that CCL17 may play a role in regulating body weight.

Conversely, we considered the possibility that genes coding for secreted proteins with anorexigenic properties were selectively repressed in AGRP neurons after food-deprivation. For example, Fgf1, which has been shown to reduce appetite (Sasaki et al., 1991), was strongly downregulated (−4.1-fold, q = 0.0004). To explore this idea, we tested a number of secreted proteins that were downregulated in AGRP neurons after food deprivation and previously had not been examined for regulation of food intake. Based on these criteria, we selected four peptides for further analysis: Pleiotrophin (Ptn, −2.3-fold, q = 0.005), a heparin-binding cytokine, which inhibits receptor protein tyrosine phosphatase β/ζ; autotaxin (Enpp2, −2.2-fold, q = 0.00013), a secreted enzyme that converts lysophosphatidylcholine into the lipid second messenger lysophosphatidic acid; cerebellin 4 (Cbln4, −7.6-fold, q = 2.9e⁻⁵), a transneuronal regulator of synaptic function; and bone morphogenic protein 3 (Bmp3, −17.7-fold, q = 2.0e⁻⁵), a member of the TGFβ superfamily (Figure 6B). Based on smFISH, we confirmed that Cbln4 and Bmp3 were downregulated in AGRP neurons after food-deprivation (Cbln4: p = 6.8e⁻¹⁷, Bmp3: p = e −12, ks-test; Figure 7A–F). To test whether these proteins influence food intake, we delivered them by intracerebroventricular injection. All four proteins significantly reduced food intake over a 24-hr period (Figure 7G), consistent with reduction of their expression levels during food deprivation. Therefore, cell type-specific RNA-Seq is also effective for identifying new secreted proteins that regulate appetite.

Figure 7

Download asset Open asset

The molecular processes governing the function of AGRP and POMC neurons have been a central focus for understanding energy homeostasis and for identifying new approaches to influence appetite and body weight in humans. This cell type-specific RNA-Seq resource from AGRP and POMC neurons reveals an extensive program of gene expression changes selectively in AGRP neurons during energy deficit associated with increased protein translation and folding, circadian gene expression, increased neuronal activity and synaptic release of neurotransmitter and neuropeptides, and alterations in secreted protein and cell surface receptor expression. POMC neurons show a much smaller number of changes, and some of these are associated with reduced activity, as would be expected from suppression of POMC neuron function under energy deficit conditions. Therefore, AGRP neurons are much more sensitive to energy deficit states than POMC neurons.

Differential gene expression in AGRP neurons after food deprivation is also considerably greater than has been observed in other cell types after various perturbations of brain state. Miller et al. (2011) compared gene expression in fast-spiking GABAergic interneurons from 48-hr muscimol-treated and saline-treated sides of motor cortex, which led to dramatic alteration in firing properties of these neurons but only 13 DEG (q < 0.05). We observed >160-fold more DEG in AGRP neurons after food-deprivation using the analysis criteria as Miller et al. (2011). Another strong neuronal perturbation, genetic ablation of the important transcriptional repressor Mecp2, only resulted in 10–50% of the number of DEG in various cell types (Sugino et al., 2014) compared to AGRP neurons from FD mice using the same analysis criteria. Finally, RNA-Seq analysis of motor neurons in an amyotrophic lateral sclerosis mouse model showed only 62 genes differentially expressed (q < 0.05) relative to wildtype motor neurons (Bandyopadhyay et al., 2013), which is 34-fold fewer differentially expressed transcripts than from AGRP neurons under food deprivation using the same analysis criteria. The high number of DEG in our dataset is influenced by our study's greater statistical power, use of RNA-Seq (instead of microarrays), as well as high purity and more narrowly defined cell types. However, the relatively small number of DEG measured in POMC under similar experimental conditions indicates that the magnitude of changes in gene expression in AGRP neurons is also based on a selectively tuned response in these neurons to the energy deficit physiological state. Therefore, our results show that neuronal gene expression in vivo can have much greater dynamic changes from alteration of genetic, physiological, or behavioral state than previously reported.

Many DEG were consistent with ER-stress responses associated with the transition to elevated protein translation in AGRP neurons during periods of increased neuropeptide release. Indeed, secreted proteins are among the most abundant transcripts in AGRP neurons and their abundance increases further during energy deficit. We observed increased splicing of the UPR regulator Xbp1s, as well as multiple protein folding chaperones, transcripts encoding protein degradation machinery, and oxidative stress signaling molecules during food deprivation. This adaptive cellular response to food deprivation in AGRP neurons is different to what has been found using whole hypothalamus analysis, where UPR signaling has been primarily associated with overnutrition states (Ozcan et al., 2009). Interestingly, overexpression of Xbp1s in POMC neurons has been shown to facilitate the function of those neurons (Williams et al., 2014), and a similar effect may be operating in AGRP neurons. In addition, we found that anti-apoptotic pathways were also prominently increased in AGRP neurons during food-deprivation, which is also indicative of considerable cellular stress and the importance of protecting against cell death for this critical energy homeostasis neuron population.

We also found reduced expression of E-box-regulated circadian genes during energy deficit in AGRP neurons. AGRP neurons contribute to increased locomotor activity during scheduled feeding (Tan et al., 2014), which is associated with the food-entrained circadian rhythm. Moreover, E-box genes are expressed primarily during the light period (Rey et al., 2011), when mice normally eat little, and suppression of these genes is similar to the expression level of these genes at night, a time when mice consume the most food. The cause and downstream consequences of this ‘night-like’ pattern of circadian gene expression are an important area for further investigation into the circadian control of AGRP neurons, but transcriptional repression involving Sfpq (Psf) (Duong et al., 2011) is a candidate molecule for this pathway.

GPCR gene expression was also substantially altered by food-deprivation in both AGRP and POMC neurons. For AGRP neurons, this was largely associated with G_q- and G_s-coupled GPCR upregulation as well as reduced G_i-coupled GPCRs. This is in line with activation of AGRP neurons by G_q-protein-coupled signaling pathways through the G_q-coupled DREADD hM3Dq (Krashes et al., 2011). In addition, overexpression of the constitutively active G_s-protein-coupled Gpr6 in AGRP neurons led to a significant increase in body weight. A similar role may extend to other constitutively active G_s-protein-coupled receptors that are upregulated by energy deficit, such as Gpr3 and Gpr64, and this effect is likely enhanced by the concomitant fall in G_i-protein coupled receptor expression. Interestingly, we noted that Hrh1 and Hrh3 are sharply reduced in AGRP neurons during energy deficit, indicating reduced responsiveness to histamine. Because GPCRs are of considerable interest as targets for drug development, this resource of AGRP and POMC neuron-expressed GPCRs provides multiple possibilities for modulating the function of these key energy homeostasis-regulating neuron populations.

Neuropeptide genes were strongly differentially expressed in AGRP neurons. Ccl17 has not been functionally investigated in the brain, but it was highly increased by food-deprivation as confirmed by RNA-Seq and smFISH, and overexpression in AGRP neurons increased body weight, consistent with a role in energy homeostasis. Notably, several neuropeptide genes that were strongly downregulated reduced appetite when injected in the lateral cerebral ventricle. For example, the neuropeptide CBLN4 strongly reduced food intake, and Cbln4 has been implicated in regulating inhibitory synapse function and is thought to play a neuroprotective role (Chacon et al., 2015). In addition, PTN is an inhibitor of receptor tyrosine phosphatase β/ζ, and its reduction might reduce cytokine signaling through tyrosine kinase receptors, which include leptin and insulin receptors.

Another important application of cell type-specific RNA-Seq data is to investigate the electrical properties of neurons. A list of expressed ion channels and how they are altered with food-deprivation provides a molecular framework for electrical activity in these neurons. Ex vivo and in vivo recordings of AGRP neuron activity have shown elevated activity and indicated the importance of burst firing in AGRP neurons during energy deficit (Betley et al., 2015). Here, we show that an important contributor to AGRP neuron excitability and burst firing in FD mice is reduced expression of the SK3 channel transcript Kcnn3. In fed mice, this calcium-activated potassium channel reduces electrical activity during elevated firing, and we find that SK-channel blockade in AGRP neurons promotes burst firing in brain slices.

Taken together, the data described in this resource provide a framework for extensive analysis of energy homeostasis circuits as well as other cell type-specific circuit nodes in the brain. Transcriptomic profiling was performed only on male mice, and future work should examine gene expression in female mice. However, these data allow insights into the cell biology of energy deficit-responsive neurons and highlight the remarkable specificity of gene expression responses of AGRP and POMC neurons. Moreover, in light of the intensive focus on new therapies for obesity as well as undereating disorders (Gautron et al., 2015), this detailed resource of molecular components in key energy homeostasis neurons will be valuable to researchers aiming to devise small molecule and peptide modulators of these circuits. In addition, together with the rapid increase in known genetic variants associated with obesity (van der Klaauw and Farooqi, 2015), cell type-specific RNA-Seq data from these and numerous other populations essential for energy homeostasis may strengthen understanding of the cell type basis of obesity found in the human population.

All experimental protocols were conducted according to U.S. National Institutes of Health guidelines for animal research and approved by the Institutional Animal Care and Use Committee at Janelia Research Campus (protocol 13-92). Experiments conducted in the UK were licensed (PPL 70/7652) under the UK Animals (Scientific Procedures) Act of 1986 following local ethical approval. All surgery was performed under isoflurance anesthesia to minimize suffering.

1. 1. Sugino K
   2. Henry T
   3. Sternson S

   (2015) Cell type transcriptomics of hypothalamic energy-sensing neuron responses to fasting

   Publicly available at the NCBI Gene Expression Omnibus (Accession no: GSE68177).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE68177

1. 1. Accili D
   2. Seigneur EM
   3. Panicker LM
   4. Ren H
   5. Chen C
   6. Simonds WF

   (2013) FoxO1 target Gpr17 activates AgRP neurons to regulate food intake

   Publicly available at the NCBI Gene Expression Omnibus (Accession no: GSE45858).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE45858
2. 1. Zhang Y
   2. Chen K
   3. Sloan SA
   4. Scholze AR
   5. Caneda C
   6. Ruderisch N
   7. Deng S
   8. Daneman R
   9. Barres BA
   10. Wu JQ

   (2014) An RNA-Seq transcriptome and splicing database of neurons, glia, and vascular cells of the cerebral cortex

   Publicly available at the NCBI Gene Expression Omnibus (Accession no: GSE52564).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE52564
3. 1. Cahoy JD
   2. Emery B
   3. Xing Y
   4. Kaushal A

   (2007) A Transcriptome Database for Astrocytes, Neurons, and Oligodendrocytes

   Publicly available at the NCBI Gene Expression Omnibus (Accession no: GSE9566).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE9566

1. 1. Accili D
   2. Seigneur EM
   3. Panicker LM
   4. Ren H
   5. Chen C
   6. Simonds WF

   (2013) FoxO1 target Gpr17 activates AgRP neurons to regulate food intake

   FoxO1 target Gpr17 activates AgRP neurons to regulate food intake, NCBI Gene Expression Omnibus, GSE45858, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE45858.

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE45858

   • Google Scholar
2. 1. Allison MB
   2. Patterson CM
   3. Krashes MJ
   4. Lowell BB
   5. Myers MG Jr
   6. Olson DP

   (2015) TRAP-seq defines markers for novel populations of hypothalamic and brainstem LepRb neurons

   Molecular Metabolism 4:299–309.

   https://doi.org/10.1016/j.molmet.2015.01.012

   • Google Scholar
3. 1. Anders S
   2. Pyl PT
   3. Huber W

   (2015) HTSeq–a Python framework to work with high-throughput sequencing data

   Bioinformatics 31:166–169.

   https://doi.org/10.1093/bioinformatics/btu638

   • Google Scholar
4. 1. Aponte Y
   2. Atasoy D
   3. Sternson SM

   (2011) AGRP neurons are sufficient to orchestrate feeding behavior rapidly and without training

   Nature Neuroscience 14:351–355.

   https://doi.org/10.1038/nn.2739

   • Google Scholar
5. 1. Atasoy D
   2. Aponte Y
   3. Su HH
   4. Sternson SM

   (2008) A FLEX switch targets Channelrhodopsin-2 to multiple cell types for imaging and long-range circuit mapping

   The Journal of Neuroscience 28:7025–7030.

   https://doi.org/10.1523/JNEUROSCI.1954-08.2008

   • Google Scholar
6. 1. Atasoy D
   2. Betley JN
   3. Su HH
   4. Sternson SM

   (2012) Deconstruction of a neural circuit for hunger

   Nature 488:172–177.

   https://doi.org/10.1038/nature11270

   • Google Scholar
7. 1. Bandyopadhyay U
   2. Cotney J
   3. Nagy M
   4. Oh S
   5. Leng J
   6. Mahajan M
   7. Mane S
   8. Fenton WA
   9. Noonan JP
   10. Horwich AL

   (2013) RNA-Seq profiling of spinal cord motor neurons from a presymptomatic SOD1 ALS mouse

   PLOS ONE 8:e53575.

   https://doi.org/10.1371/journal.pone.0053575

   • Google Scholar
8. 1. Banks AS
   2. Davis SM
   3. Bates SH
   4. Myers MG Jr

   (2000) Activation of downstream signals by the long form of the leptin receptor

   The Journal of Biological Chemistry 275:14563–14572.

   https://doi.org/10.1074/jbc.275.19.14563

   • Google Scholar
9. 1. Bartos JA
   2. Ulrich JD
   3. Li H
   4. Beazely MA
   5. Chen Y
   6. Macdonald JF
   7. Hell JW

   (2010) Postsynaptic clustering and activation of Pyk2 by PSD-95

   The Journal of Neuroscience 30:449–463.

   https://doi.org/10.1523/JNEUROSCI.4992-08.2010

   • Google Scholar
10. 1. Baskin DG
    2. Breininger JF
    3. Schwartz MW

    (1999) Leptin receptor mRNA identifies a subpopulation of neuropeptide Y neurons activated by fasting in rat hypothalamus

    Diabetes 48:828–833.

    https://doi.org/10.2337/diabetes.48.4.828

    • Google Scholar
11. 1. Benosman S
    2. Ravanan P
    3. Correa RG
    4. Hou YC
    5. Yu M
    6. Gulen MF
    7. Li X
    8. Thomas J
    9. Cuddy M
    10. Matsuzawa Y
    11. Sano R
    12. Diaz P
    13. Matsuzawa S
    14. Reed JC

    (2013) Interleukin-1 receptor-associated kinase-2 (IRAK2) is a critical mediator of endoplasmic reticulum (ER) stress signaling

    PLOS ONE 8:e64256.

    https://doi.org/10.1371/journal.pone.0064256

    • Google Scholar
12. 1. Betley JN
    2. Xu S
    3. Cao ZF
    4. Gong R
    5. Magnus CJ
    6. Yu Y
    7. Sternson SM

    (2015) Neurons for hunger and thirst transmit a negative-valence teaching signal

    Nature 521:180–185.

    https://doi.org/10.1038/nature14416

    • Google Scholar
13. 1. Boda B
    2. Alberi S
    3. Nikonenko I
    4. Node-Langlois R
    5. Jourdain P
    6. Moosmayer M
    7. Parisi-Jourdain L
    8. Muller D

    (2004) The mental retardation protein PAK3 contributes to synapse formation and plasticity in hippocampus

    The Journal of Neuroscience 24:10816–10825.

    https://doi.org/10.1523/JNEUROSCI.2931-04.2004

    • Google Scholar
14. 1. Bouret SG
    2. Draper SJ
    3. Simerly RB

    (2004) Trophic action of leptin on hypothalamic neurons that regulate feeding

    Science 304:108–110.

    https://doi.org/10.1126/science.1095004

    • Google Scholar
15. 1. Broberger C
    2. Johansen J
    3. Johansson C
    4. Schalling M
    5. Hokfelt T

    (1998) The neuropeptide Y/agouti gene-related protein (AGRP) brain circuitry in normal, anorectic, and monosodium glutamate-treated mice

    Proceedings of the National Academy of Sciences of USA 95:15043–15048.

    https://doi.org/10.1073/pnas.95.25.15043

    • Google Scholar
16. 1. Caenepeel S
    2. Charydczak G
    3. Sudarsanam S
    4. Hunter T
    5. Manning G

    (2004) The mouse kinome: discovery and comparative genomics of all mouse protein kinases

    Proceedings of the National Academy of Sciences of USA 101:11707–11712.

    https://doi.org/10.1073/pnas.0306880101

    • Google Scholar
17. 1. Cahoy JD
    2. Emery B
    3. Kaushal A
    4. Foo LC
    5. Zamanian JL
    6. Christopherson KS
    7. Xing Y
    8. Lubischer JL
    9. Krieg PA
    10. Krupenko SA
    11. Thompson WJ
    12. Barres BA

    (2008) A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function

    The Journal of Neuroscience 28:264–278.

    https://doi.org/10.1523/JNEUROSCI.4178-07.2008

    • Google Scholar
18. 1. Cahoy JD
    2. Emery B
    3. Xing Y
    4. Kaushal A

    (2007) A Transcriptome Database for Astrocytes, Neurons, and Oligodendrocytes

    A Transcriptome Database for Astrocytes, Neurons, and Oligodendrocytes, NCBI Gene Expression Omnibus, GSE9566, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE9566.

    http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE9566

    • Google Scholar
19. 1. Cao P
    2. Maximov A
    3. Sudhof TC

    (2011) Activity-dependent IGF-1 exocytosis is controlled by the Ca(2+)-sensor synaptotagmin-10

    Cell 145:300–311.

    https://doi.org/10.1016/j.cell.2011.03.034

    • Google Scholar
20. 1. Chacon PJ
    2. del Marco A
    3. Arevalo A
    4. Dominguez-Gimenez P
    5. Garcia-Segura LM
    6. Rodriguez-Tebar A

    (2015) Cerebellin 4, a synaptic protein, enhances inhibitory activity and resistance of neurons to amyloid-beta toxicity

    Neurobiology of Aging 36:1057–1071.

    https://doi.org/10.1016/j.neurobiolaging.2014.11.006

    • Google Scholar
21. 1. Chen Y
    2. Zhang Y
    3. Yin Y
    4. Gao G
    5. Li S
    6. Jiang Y
    7. Gu X
    8. Luo J

    (2005) SPD–a web-based secreted protein database

    Nucleic Acids Research 33:D169–D173.

    https://doi.org/10.1093/nar/gki093

    • Google Scholar
22. 1. Colombrita C
    2. Zennaro E
    3. Fallini C
    4. Weber M
    5. Sommacal A
    6. Buratti E
    7. Silani V
    8. Ratti A

    (2009) TDP-43 is recruited to stress granules in conditions of oxidative insult

    Journal of Neurochemistry 111:1051–1061.

    https://doi.org/10.1111/j.1471-4159.2009.06383.x

    • Google Scholar
23. 1. Dalal J
    2. Roh JH
    3. Maloney SE
    4. Akuffo A
    5. Shah S
    6. Yuan H
    7. Wamsley B
    8. Jones WB
    9. de Guzman Strong C
    10. Gray PA
    11. Holtzman DM
    12. Heintz N
    13. Dougherty JD

    (2013) Translational profiling of hypocretin neurons identifies candidate molecules for sleep regulation

    Genes & Development 27:565–578.

    https://doi.org/10.1101/gad.207654.112

    • Google Scholar
24. 1. Dobin A
    2. Davis CA
    3. Schlesinger F
    4. Drenkow J
    5. Zaleski C
    6. Jha S
    7. Batut P
    8. Chaisson M
    9. Gingeras TR

    (2013) STAR: ultrafast universal RNA-seq aligner

    Bioinformatics 29:15–21.

    https://doi.org/10.1093/bioinformatics/bts635

    • Google Scholar
25. 1. Duong HA
    2. Robles MS
    3. Knutti D
    4. Weitz CJ

    (2011) A molecular mechanism for circadian clock negative feedback

    Science 332:1436–1439.

    https://doi.org/10.1126/science.1196766

    • Google Scholar
26. 1. Ekstrand MI
    2. Nectow AR
    3. Knight ZA
    4. Latcha KN
    5. Pomeranz LE
    6. Friedman JM

    (2014) Molecular profiling of neurons based on connectivity

    Cell 157:1230–1242.

    https://doi.org/10.1016/j.cell.2014.03.059

    • Google Scholar
27. 1. Francis JS
    2. Dragunow M
    3. During MJ

    (2004) Over expression of ATF-3 protects rat hippocampal neurons from in vivo injection of kainic acid

    Molecular Brain Research 124:199–203.

    https://doi.org/10.1016/j.molbrainres.2003.10.027

    • Google Scholar
28. 1. Gautron L
    2. Elmquist JoelK
    3. Williams KevinW

    (2015) Neural control of energy balance: translating circuits to therapies

    Cell 161:133–145.

    https://doi.org/10.1016/j.cell.2015.02.023

    • Google Scholar
29. 1. Guarnieri DJ
    2. Brayton CE
    3. Richards SM
    4. Maldonado-Aviles J
    5. Trinko JR
    6. Nelson J
    7. Taylor JR
    8. Gourley SL
    9. DiLeone RJ

    (2012) Gene profiling reveals a role for stress hormones in the molecular and behavioral response to food restriction

    Biological Psychiatry 71:358–365.

    https://doi.org/10.1016/j.biopsych.2011.06.028

    • Google Scholar
30. 1. Hahn T
    2. Breininger J
    3. Baskin D
    4. Schwartz M

    (1998) Coexpression of Agrp and NPY in fasting-activated hypothalamic neurons

    Nature Neuroscience 1:271–272.

    https://doi.org/10.1038/1082

    • Google Scholar
31. 1. Hart SN
    2. Therneau TM
    3. Zhang Y
    4. Poland GA
    5. Kocher JP

    (2013) Calculating sample size estimates for RNA sequencing data

    Journal of Computational Biology 20:970–978.

    https://doi.org/10.1089/cmb.2012.0283

    • Google Scholar
32. 1. He L
    2. Kim SO
    3. Kwon O
    4. Jeong SJ
    5. Kim MS
    6. Lee HG
    7. Osada H
    8. Jung M
    9. Ahn JS
    10. Kim BY

    (2009) ATM blocks tunicamycin-induced endoplasmic reticulum stress

    FEBS Letters 583:903–908.

    https://doi.org/10.1016/j.febslet.2009.02.002

    • Google Scholar
33. 1. Heiman M
    2. Schaefer A
    3. Gong S
    4. Peterson JD
    5. Day M
    6. Ramsey KE
    7. Suarez-Farinas M
    8. Schwarz C
    9. Stephan DA
    10. Surmeier DJ
    11. Greengard P
    12. Heintz N

    (2008) A translational profiling approach for the molecular characterization of CNS cell types

    Cell 135:738–748.

    https://doi.org/10.1016/j.cell.2008.10.028

    • Google Scholar
34. 1. Hempel CM
    2. Sugino K
    3. Nelson SB

    (2007) A manual method for the purification of fluorescently labeled neurons from the mammalian brain

    Nature Protocols 2:2924–2929.

    https://doi.org/10.1038/nprot.2007.416

    • Google Scholar
35. 1. Izumi S
    2. Saito A
    3. Kanemoto S
    4. Kawasaki N
    5. Asada R
    6. Iwamoto H
    7. Oki M
    8. Miyagi H
    9. Ochi M
    10. Imaizumi K

    (2012) The endoplasmic reticulum stress transducer BBF2H7 suppresses apoptosis by activating the ATF5-MCL1 pathway in growth plate cartilage

    The Journal of Biological Chemistry 287:36190–36200.

    https://doi.org/10.1074/jbc.M112.373746

    • Google Scholar
36. 1. Jovanovic Z
    2. Tung YC
    3. Lam BY
    4. O'Rahilly S
    5. Yeo GS

    (2010) Identification of the global transcriptomic response of the hypothalamic arcuate nucleus to fasting and leptin

    Journal of Neuroendocrinology 22:915–925.

    https://doi.org/10.1111/j.1365-2826.2010.02026.x

    • Google Scholar
37. 1. Kalashnikova E
    2. Lorca RA
    3. Kaur I
    4. Barisone GA
    5. Li B
    6. Ishimaru T
    7. Trimmer JS
    8. Mohapatra DP
    9. Diaz E

    (2010) SynDIG1: an activity-regulated, AMPA- receptor-interacting transmembrane protein that regulates excitatory synapse development

    Neuron 65:80–93.

    https://doi.org/10.1016/j.neuron.2009.12.021

    • Google Scholar
38. 1. Kitahara H
    2. Masumoto J
    3. Parker AL
    4. Maruta F
    5. Kubo N
    6. Shimizu A
    7. Akita N
    8. Miwa S
    9. Kobayashi N
    10. Nakayama J
    11. Miyagawa S

    (2011) COP35, a cholangiocarcinoma-binding oligopeptide, interacts with the clathrin heavy chain accompanied by GRP78

    Molecular Cancer Research 9:688–701.

    https://doi.org/10.1158/1541-7786.MCR-10-0470

    • Google Scholar
39. 1. Kitamura T
    2. Feng Y
    3. Kitamura YI
    4. Chua SC Jr
    5. Xu AW
    6. Barsh GS
    7. Rossetti L
    8. Accili D

    (2006) Forkhead protein FoxO1 mediates Agrp-dependent effects of leptin on food intake

    Nature Medicine 12:534–540.

    https://doi.org/10.1038/nm1392

    • Google Scholar
40. 1. Kobayashi A
    2. Kang MI
    3. Okawa H
    4. Ohtsuji M
    5. Zenke Y
    6. Chiba T
    7. Igarashi K
    8. Yamamoto M

    (2004) Oxidative stress sensor Keap1 functions as an adaptor for Cul3-based E3 ligase to regulate proteasomal degradation of Nrf2

    Molecular and Cellular Biology 24:7130–7139.

    https://doi.org/10.1128/MCB.24.16.7130-7139.2004

    • Google Scholar
41. 1. Krashes MJ
    2. Koda S
    3. Ye C
    4. Rogan SC
    5. Adams AC
    6. Cusher DS
    7. Maratos-Flier E
    8. Roth BL
    9. Lowell BB

    (2011) Rapid, reversible activation of AgRP neurons drives feeding behavior in mice

    The Journal of Clinical Investigation 121:1424–1428.

    https://doi.org/10.1172/JCI46229

    • Google Scholar
42. 1. Langmead B
    2. Salzberg SL

    (2012) Fast gapped-read alignment with Bowtie 2

    Nature Methods 9:357–359.

    https://doi.org/10.1038/nmeth.1923

    • Google Scholar
43. 1. Law CW
    2. Chen Y
    3. Shi W
    4. Smyth GK

    (2014) voom: Precision weights unlock linear model analysis tools for RNA-seq read counts

    Genome Biology 15:R29.

    https://doi.org/10.1186/gb-2014-15-2-r29

    • Google Scholar
44. 1. Lee AH
    2. Iwakoshi NN
    3. Glimcher LH

    (2003) XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response

    Molecular and Cellular Biology 23:7448–7459.

    https://doi.org/10.1128/MCB.23.21.7448-7459.2003

    • Google Scholar
45. 1. Lee S
    2. Lee TA
    3. Lee E
    4. Kang S
    5. Park A
    6. Kim SW
    7. Park HJ
    8. Yoon JH
    9. Ha SJ
    10. Park T
    11. Lee JS
    12. Cheon JH
    13. Park B

    (2015) Identification of a subnuclear body involved in sequence-specific cytokine RNA processing

    Nature Communications 6:5791.

    https://doi.org/10.1038/ncomms6791

    • Google Scholar
46. 1. LeSauter J
    2. Lambert CM
    3. Robotham MR
    4. Model Z
    5. Silver R
    6. Weaver DR

    (2012) Antibodies for assessing circadian clock proteins in the rodent suprachiasmatic nucleus

    PLOS ONE 7:e35938.

    https://doi.org/10.1371/journal.pone.0035938

    • Google Scholar
47. 1. Li JY
    2. Kuick R
    3. Thompson RC
    4. Misek DE
    5. Lai YM
    6. Liu YQ
    7. Chai BX
    8. Hanash SM
    9. Gantz I

    (2005) Arcuate nucleus transcriptome profiling identifies ankyrin repeat and suppressor of cytokine signalling box-containing protein 4 as a gene regulated by fasting in central nervous system feeding circuits

    Journal of Neuroendocrinology 17:394–404.

    https://doi.org/10.1111/j.1365-2826.2005.01317.x

    • Google Scholar
48. 1. Liu T
    2. Kong D
    3. Shah BP
    4. Ye C
    5. Koda S
    6. Saunders A
    7. Ding JB
    8. Yang Z
    9. Sabatini BL
    10. Lowell BB

    (2012) Fasting activation of AgRP neurons requires NMDA receptors and involves spinogenesis and increased excitatory tone

    Neuron 73:511–522.

    https://doi.org/10.1016/j.neuron.2011.11.027

    • Google Scholar
49. 1. Lovero KL
    2. Blankenship SM
    3. Shi Y
    4. Nicoll RA

    (2013) SynDIG1 promotes excitatory synaptogenesis independent of AMPA receptor trafficking and biophysical regulation

    PLOS ONE 8:e66171.

    https://doi.org/10.1371/journal.pone.0066171

    • Google Scholar
50. 1. Luquet S
    2. Perez FA
    3. Hnasko TS
    4. Palmiter RD

    (2005) NPY/AgRP neurons are essential for feeding in adult mice but can be ablated in neonates

    Science 310:683–685.

    https://doi.org/10.1126/science.1115524

    • Google Scholar
51. 1. Maddalo D
    2. Neeb A
    3. Jehle K
    4. Schmitz K
    5. Muhle-Goll C
    6. Shatkina L
    7. Walther TV
    8. Bruchmann A
    9. Gopal SM
    10. Wenzel W
    11. Ulrich AS
    12. Cato AC

    (2012) A peptidic unconjugated GRP78/BiP ligand modulates the unfolded protein response and induces prostate cancer cell death

    PLOS ONE 7:e45690.

    https://doi.org/10.1371/journal.pone.0045690

    • Google Scholar
52. 1. Miller MN
    2. Okaty BW
    3. Kato S
    4. Nelson SB

    (2011) Activity-dependent changes in the firing properties of neocortical fast-spiking interneurons in the absence of large changes in gene expression

    Developmental Neurobiology 71:62–70.

    https://doi.org/10.1002/dneu.20811

    • Google Scholar
53. 1. Okaty BW
    2. Sugino K
    3. Nelson SB

    (2011) A quantitative comparison of cell-type-specific microarray gene expression profiling methods in the mouse brain

    PLOS ONE 6:e16493.

    https://doi.org/10.1371/journal.pone.0016493

    • Google Scholar
54. 1. Ozcan L
    2. Ergin AS
    3. Lu A
    4. Chung J
    5. Sarkar S
    6. Nie D
    7. Myers MG Jr
    8. Ozcan U

    (2009) Endoplasmic reticulum stress plays a central role in development of leptin resistance

    Cell Metabolism 9:35–51.

    https://doi.org/10.1016/j.cmet.2008.12.004

    • Google Scholar
55. 1. Pawson AJ
    2. Sharman JL
    3. Benson HE
    4. Faccenda E
    5. Alexander SP
    6. Buneman OP
    7. Davenport AP
    8. McGrath JC
    9. Peters JA
    10. Southan C
    11. Spedding M
    12. Yu W
    13. Harmar AJ
    14. Nc I

    (2014) The IUPHAR/BPS Guide to PHARMACOLOGY: an expert-driven knowledgebase of drug targets and their ligands

    Nucleic Acids Research 42:D1098–D1106.

    https://doi.org/10.1093/nar/gkt1143

    • Google Scholar
56. 1. Qiu J
    2. Fang Y
    3. Ronnekleiv OK
    4. Kelly MJ

    (2010) Leptin excites proopiomelanocortin neurons via activation of TRPC channels

    The Journal of Neuroscience 30:1560–1565.

    https://doi.org/10.1523/JNEUROSCI.4816-09.2010

    • Google Scholar
57. 1. Ren H
    2. Orozco IJ
    3. Su Y
    4. Suyama S
    5. Gutierrez-Juarez R
    6. Horvath TL
    7. Wardlaw SL
    8. Plum L
    9. Arancio O
    10. Accili D

    (2012) FoxO1 target Gpr17 activates AgRP neurons to regulate food intake

    Cell 149:1314–1326.

    https://doi.org/10.1016/j.cell.2012.04.032

    • Google Scholar
58. 1. Rey G
    2. Cesbron F
    3. Rougemont J
    4. Reinke H
    5. Brunner M
    6. Naef F

    (2011) Genome-wide and phase-specific DNA-binding rhythms of BMAL1 control circadian output functions in mouse liver

    PLOS Biology 9:e1000595.

    https://doi.org/10.1371/journal.pbio.1000595

    • Google Scholar
59. 1. Ritchie ME
    2. Phipson B
    3. Wu D
    4. Hu Y
    5. Law CW
    6. Shi W
    7. Smyth GK

    (2015) limma powers differential expression analyses for RNA-sequencing and microarray studies

    Nucleic Acids Research 43:e47.

    https://doi.org/10.1093/nar/gkv007

    • Google Scholar
60. 1. Robinson MD
    2. Oshlack A

    (2010) A scaling normalization method for differential expression analysis of RNA-seq data

    Genome Biology 11:R25.

    https://doi.org/10.1186/gb-2010-11-3-r25

    • Google Scholar
61. 1. Roepke TA
    2. Qiu J
    3. Smith AW
    4. Ronnekleiv OK
    5. Kelly MJ

    (2011) Fasting and 17beta-estradiol differentially modulate the M-current in neuropeptide Y neurons

    The Journal of Neuroscience 31:11825–11835.

    https://doi.org/10.1523/JNEUROSCI.1395-11.2011

    • Google Scholar
62. 1. Sacco F
    2. Perfetto L
    3. Castagnoli L
    4. Cesareni G

    (2012) The human phosphatase interactome: an intricate family portrait

    FEBS Letters 586:2732–2739.

    https://doi.org/10.1016/j.febslet.2012.05.008

    • Google Scholar
63. 1. Samali A
    2. Fitzgerald U
    3. Deegan S
    4. Gupta S

    (2010) Methods for monitoring endoplasmic reticulum stress and the unfolded protein response

    International Journal of Cell Biology 2010:830307.

    https://doi.org/10.1155/2010/830307

    • Google Scholar
64. 1. Sandoval-Guzman T
    2. Gongrich C
    3. Moliner A
    4. Guo T
    5. Wu H
    6. Broberger C
    7. Ibanez CF

    (2012) Neuroendocrine control of female reproductive function by the activin receptor ALK7

    FASEB Journal 26:4966–4976.

    https://doi.org/10.1096/fj.11-199059

    • Google Scholar
65. 1. Sasaki K
    2. Oomura Y
    3. Suzuki K
    4. Muto T
    5. Hanai K
    6. Tooyama I
    7. Kimura H
    8. Yanaihara N

    (1991) Effects of fibroblast growth factors and platelet-derived growth factor on food intake in rats

    Brain Research Bulletin 27:327–332.

    https://doi.org/10.1016/0361-9230(91)90120-9

    • Google Scholar
66. 1. Sato T
    2. Sako Y
    3. Sho M
    4. Momohara M
    5. Suico MA
    6. Shuto T
    7. Nishitoh H
    8. Okiyoneda T
    9. Kokame K
    10. Kaneko M
    11. Taura M
    12. Miyata M
    13. Chosa K
    14. Koga T
    15. Morino-Koga S
    16. Wada I
    17. Kai H

    (2012) STT3B-dependent posttranslational N-glycosylation as a surveillance system for secretory protein

    Molecular Cell 47:99–110.

    https://doi.org/10.1016/j.molcel.2012.04.015

    • Google Scholar
67. 1. Schioth HB

    (2006) G protein-coupled receptors in regulation of body weight

    CNS & Neurological Disorders Drug Targets 5:241–249.

    https://doi.org/10.2174/187152706777452263

    • Google Scholar
68. 1. Schwartz MW
    2. Seeley RJ
    3. Woods SC
    4. Weigle DS
    5. Campfield LA
    6. Burn P
    7. Baskin DG

    (1997) Leptin increases hypothalamic pro-opiomelanocortin mRNA expression in the rostral arcuate nucleus

    Diabetes 46:2119–2123.

    https://doi.org/10.2337/diab.46.12.2119

    • Google Scholar
69. 1. Seeburg DP
    2. Feliu-Mojer M
    3. Gaiottino J
    4. Pak DT
    5. Sheng M

    (2008) Critical role of CDK5 and Polo-like kinase 2 in homeostatic synaptic plasticity during elevated activity

    Neuron 58:571–583.

    https://doi.org/10.1016/j.neuron.2008.03.021

    • Google Scholar
70. 1. Sugino K
    2. Hempel CM
    3. Miller MN
    4. Hattox AM
    5. Shapiro P
    6. Wu C
    7. Huang ZJ
    8. Nelson SB

    (2006) Molecular taxonomy of major neuronal classes in the adult mouse forebrain

    Nature Neuroscience 9:99–107.

    https://doi.org/10.1038/nn1618

    • Google Scholar
71. 1. Sugino K
    2. Hempel CM
    3. Okaty BW
    4. Arnson HA
    5. Kato S
    6. Dani VS
    7. Nelson SB

    (2014) Cell-type-specific repression by methyl-CpG-binding protein 2 is biased toward long genes

    The Journal of Neuroscience 34:12877–12883.

    https://doi.org/10.1523/JNEUROSCI.2674-14.2014

    • Google Scholar
72. 1. Takahashi KA
    2. Cone RD

    (2005) Fasting induces a large, leptin-dependent increase in the intrinsic action potential frequency of orexigenic arcuate nucleus neuropeptide Y/Agouti-related protein neurons

    Endocrinology 146:1043–1047.

    https://doi.org/10.1210/en.2004-1397

    • Google Scholar
73. 1. Tan K
    2. Knight ZA
    3. Friedman JM

    (2014) Ablation of AgRP neurons impairs adaption to restricted feeding

    Molecular Metabolism 3:694–704.

    https://doi.org/10.1016/j.molmet.2014.07.002

    • Google Scholar
74. 1. Tinkum KL
    2. White LS
    3. Marpegan L
    4. Herzog E
    5. Piwnica-Worms D
    6. Piwnica-Worms H

    (2013) Forkhead box O1 (FOXO1) protein, but not p53, contributes to robust induction of p21 expression in fasted mice

    The Journal of Biological Chemistry 288:27999–28008.

    https://doi.org/10.1074/jbc.M113.494328

    • Google Scholar
75. 1. Uhlenbrock K
    2. Gassenhuber H
    3. Kostenis E

    (2002) Sphingosine 1-phosphate is a ligand of the human gpr3, gpr6 and gpr12 family of constitutively active G protein-coupled receptors

    Cellular Signalling 14:941–953.

    https://doi.org/10.1016/S0898-6568(02)00041-4

    • Google Scholar
76. 1. van den Pol AN
    2. Yao Y
    3. Fu LY
    4. Foo K
    5. Huang H
    6. Coppari R
    7. Lowell BB
    8. Broberger C

    (2009) Neuromedin B and gastrin-releasing peptide excite arcuate nucleus neuropeptide Y neurons in a novel transgenic mouse expressing strong Renilla green fluorescent protein in NPY neurons

    The Journal of Neuroscience 29:4622–4639.

    https://doi.org/10.1523/JNEUROSCI.3249-08.2009

    • Google Scholar
77. 1. van der Klaauw AgathaA
    2. Farooqi IS

    (2015) The hunger genes: pathways to obesity

    Cell 161:119–132.

    https://doi.org/10.1016/j.cell.2015.03.008

    • Google Scholar
78. 1. Williams KW
    2. Liu T
    3. Kong X
    4. Fukuda M
    5. Deng Y
    6. Berglund ED
    7. Deng Z
    8. Gao Y
    9. Liu T
    10. Sohn JW
    11. Jia L
    12. Fujikawa T
    13. Kohno D
    14. Scott MM
    15. Lee S
    16. Lee CE
    17. Sun K
    18. Chang Y
    19. Scherer PE
    20. Elmquist JK

    (2014) Xbp1s in Pomc neurons connects ER stress with energy balance and glucose homeostasis

    Cell Metabolism 20:471–482.

    https://doi.org/10.1016/j.cmet.2014.06.002

    • Google Scholar
79. 1. Xu T
    2. Bajjalieh SM

    (2001) SV2 modulates the size of the readily releasable pool of secretory vesicles

    Nature Cell Biology 3:691–698.

    https://doi.org/10.1038/35087000

    • Google Scholar
80. 1. Yamamoto K
    2. Sato T
    3. Matsui T
    4. Sato M
    5. Okada T
    6. Yoshida H
    7. Harada A
    8. Mori K

    (2007) Transcriptional induction of mammalian ER quality control proteins is mediated by single or combined action of ATF6alpha and XBP1

    Developmental Cell 13:365–376.

    https://doi.org/10.1016/j.devcel.2007.07.018

    • Google Scholar
81. 1. Yan J
    2. Wang H
    3. Liu Y
    4. Shao C

    (2008) Analysis of gene regulatory networks in the mammalian circadian rhythm

    PLOS Computational Biology 4:e1000193.

    https://doi.org/10.1371/journal.pcbi.1000193

    • Google Scholar
82. 1. Yang Y
    2. Atasoy D
    3. Su HH
    4. Sternson SM

    (2011) Hunger states switch a flip-flop memory circuit via a synaptic AMPK-dependent positive feedback loop

    Cell 146:992–1003.

    https://doi.org/10.1016/j.cell.2011.07.039

    • Google Scholar
83. 1. Zeisel A
    2. Munoz-Manchado AB
    3. Codeluppi S
    4. Lonnerberg P
    5. La Manno G
    6. Jureus A
    7. Marques S
    8. Munguba H
    9. He L
    10. Betsholtz C
    11. Rolny C
    12. Castelo-Branco G
    13. Hjerling-Leffler J
    14. Linnarsson S

    (2015) Brain structure. Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq

    Science 347:1138–1142.

    https://doi.org/10.1126/science.aaa1934

    • Google Scholar
84. 1. Zhang Y
    2. Chen K
    3. Sloan SA
    4. Bennett ML
    5. Scholze AR
    6. O'Keeffe S
    7. Phatnani HP
    8. Guarnieri P
    9. Caneda C
    10. Ruderisch N
    11. Deng S
    12. Liddelow SA
    13. Zhang C
    14. Daneman R
    15. Maniatis T
    16. Barres BA
    17. Wu JQ

    (2014a) An RNA-sequencing transcriptome and splicing database of glia, neurons, and vascular cells of the cerebral cortex

    The Journal of Neuroscience 34:11929–11947.

    https://doi.org/10.1523/JNEUROSCI.1860-14.2014

    • Google Scholar
85. 1. Zhang Y
    2. Chen K
    3. Sloan SA
    4. Scholze AR
    5. Caneda C
    6. Ruderisch N
    7. Deng S
    8. Daneman R
    9. Barres BA
    10. Wu JQ

    (2014) An RNA-Seq transcriptome and splicing database of neurons, glia, and vascular cells of the cerebral cortex

    An RNA-Seq transcriptome and splicing database of neurons, glia, and vascular cells of the cerebral cortex, NCBI Gene Expression Omnibus, GSE52564, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE52564.

    http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE52564

    • Google Scholar

Author details

1. Fredrick E Henry

   Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, United States

   Contribution

   FEH, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Ken Sugino

   Competing interests

   The authors declare that no competing interests exist.

2. Ken Sugino

   Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, United States

   Contribution

   KS, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Fredrick E Henry

   Competing interests

   The authors declare that no competing interests exist.

3. Adam Tozer

   Division of Neurobiology, Medical Research Council Laboratory of Molecular Biology, Cambridge, United Kingdom

   Contribution

   AT, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

4. Tiago Branco

   Division of Neurobiology, Medical Research Council Laboratory of Molecular Biology, Cambridge, United Kingdom

   Contribution

   TB, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

5. Scott M Sternson

   Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, United States

   Contribution

   SMS, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   sternsons@janelia.hhmi.org

   Competing interests

   The authors declare that no competing interests exist.

• 9,432

  views

• 2,575

  downloads

• 191

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.