Nucleus of the Solitary Tract Serotonin 5-HT2C Receptors Modulate Food Intake.

To meet the challenge to human health posed by obesity, a better understanding of the regulation of feeding is essential. Medications targeting 5-hydroxytryptamine (5-HT; serotonin) 2C receptors (htr2c; 5-HT2CR) improve obesity. Here we probed the functional significance of 5-HT2CRs specifically within the brainstem nucleus of the solitary tract (5-HT2CRNTS) in feeding behavior. Selective activation of 5-HT2CRNTS decreased feeding and was sufficient to mediate acute food intake reductions elicited by the 5-HT2CR agonist obesity medication lorcaserin. Similar to pro-opiomelanocortin neurons expressed within the hypothalamic arcuate nucleus (POMCARC), a subset of POMCNTS neurons co-expressed 5-HT2CRs and were activated by 5-HT2CR agonists. Knockdown of POMCNTS prevented the acute appetite-suppressive effect of lorcaserin, whereas POMCARC knockdown prevented the full anorectic effect. These data identify 5-HT2CRNTS as a sufficient subpopulation of 5-HT2CRs in reducing food intake when activated and reveal that 5-HT2CR agonist obesity medications require POMC within the NTS and ARC to reduce food intake.

Introduction

As demonstrated 40 years ago, brain 5-hydroxytryptamine (5-HT; serotonin) is a significant regulator of appetite and body weight (Breisch et al., 1976). Targeting this neurochemical machinery, medications increasing 5-HT bioavailability were developed to treat obesity. However, increasing 5-HT activity at peripheral 5-HT receptors (5-HTRs) contributed to side effects that led to the withdrawal of medications such as d-fenfluramine in the 1990s and sibutramine in the 2000s (Burke and Heisler, 2015). Efforts to delineate 5-HTRs mediating 5-HT's therapeutic effects on food intake revealed the 2C receptor subtype (Htr2cr, 5-HT2CR) as the principal mediator (Tecott et al., 1995, Williams et al., 2011). The clinical significance of this discovery is underscored by a new obesity medication, the 5-HT2CR agonist, lorcaserin, which was recently launched in the United States (Aronne et al., 2014, Smith et al., 2010).

5-HT2CRs are widely distributed within the brain where they act to modulate a diverse array of behaviors and physiological processes (Burke and Heisler, 2015, Julius et al., 1988). Of the neurochemically defined cells expressing 5-HT2CRs, it is a subpopulation of neurons within the arcuate nucleus of the hypothalamus (ARC) that co-express the precursor polypeptide pro-opiomelanocortin (POMC) that have been proposed to be a principal mediator of 5-HT2CR's effects on metabolic functions (Berglund et al., 2013, Burke et al., 2017, Doslikova et al., 2013, Heisler et al., 2002, Xu et al., 2010). However, anorectic doses of 5-HT2CR agonists have been shown to increase markers of neuronal activity in caudal brainstem regions, including the nucleus of the solitary tract (NTS) (Lam et al., 2009, Stark et al., 2006). Although the NTS is a brainstem structure known to be a key regulatory center for the integration of food related signals of both peripheral and central origin (Fan et al., 2004, Grill and Hayes, 2009, Zhan et al., 2013), no direct functional assessment of 5-HT2CRs in this brain region has been undertaken. Here we probed the functional significance of 5-HT2CRs within the NTS, their contribution to the therapeutic effect of 5-HT2CR agonist obesity medications, and the mechanism underpinning their appetite-reducing effect via exclusive action within the NTS.

Results

Selective 5-HT2CRNTS Neuron Activation Decreases Food Intake

To determine whether 5-HT2CR-expressing NTS (5-HT2CRNTS) neurons play a role in appetite control, we first employed the chemogenetic designer receptors exclusively activated by designer drugs (DREADDs; Armbruster et al., 2007) approach to specifically express Gq-coupled (hM3Dq) or Gi-coupled (hM4Di) designer receptors in 5-HT2CRNTS neurons. To target 5-HT2CRNTS neurons, we stereotactically delivered Cre-dependent adeno-associated viruses (AAVs) encoding designer receptor hM3Dq or hM4Di into the NTS of 5-HTR mice (Burke et al., 2016, Burke et al., 2017, Marcinkiewcz et al., 2016) to generate 5-HTR-hM3Dq (Figures 1A and 1B) and 5-HTR-hM4Di cohorts (Figures S1A and S1B). Cre recombinase activity in 5-HTR mice efficiently recombined the DREADD-mCherry allele in the NTS, as seen by the expression of the mCherry reporter protein in transduced cells (Figure 1C). A degree of spread was observed in adjacent regions following amplification of mCherry signal with immunohistochemistry (Figure 1C). Following treatment with designer drug clozapine-N-oxide (CNO; 1 mg/kg, intraperitoneally [i.p.]), c-fos immunoreactivity (FOS-IR; a surrogate marker of neuronal activation) was observed within the NTS of 5-HTR-hM3Dq mice. In 5-HTRNTS-hM3Dq mice that had been fasted overnight to decrease baseline FOS-IR, quantification analysis revealed that approximately 76.4% ± 3.7% of mCherry-positive NTS cells expressed FOS-IR following CNO treatment whereas less than 1% expressed FOS-IR following saline treatment (Figure 1C). Consistent with these in vivo data, in ex vivo NTS slices, bath application of CNO (10 μM) increased the firing frequency of 5-HTRNTS-hM3Dq-expressing neurons (n = 5/5; Figure 1D). These data indicate that CNO allows remote chemical control of 5-HT2CRNTS neuronal activity in 5-HTR-hM3Dq-expressing mice. We next examined the effect of CNO on feeding behavior. CNO (1 mg/kg, i.p.) significantly reduced the first 5 hr of ad libitum dark cycle chow intake compared with saline treatment in 5-HTR-hM3Dq mice (Figure 1E).

Figure 1

Activation of 5-HT2CRNTS-Expressing Neurons Reduces Food Intake

(A) Schematic of the strategy used to selectively activate 5-HT2CRNTS neurons; stereotactic injection of Cre-dependent DREADD vectors into NTS of 5-HTR mice.

(B) 5-HTR construct and Cre-mediated recombination of DREADD allele.

(C) Expression of the DREADD-fused mCherry reporter protein within the NTS using immunohistochemistry for mCherry (mCherry-IR; red). In vivo, CNO (1 mg/kg, intraperitoneally [i.p.]) increased FOS-IR (green) expression in the majority of 5-HTRNTS-hM3Dq-expressing neurons (red, bottom panels), whereas saline treatment did not (top panels).

(D) Ex vivo, bath application of CNO (10 μM) increased the firing frequency of 5-HTRNTS-hM3Dq-expressing neurons.

(E) CNO (1 mg/kg, i.p.) elicited a significant reduction in ad libitum dark cycle food intake compared with saline treatment in 5-HTRNTS-hM3Dq mice (n = 6–7; 3 hr: t12 = 3.57, p = 0.004; 5 hr: t12 = 3.379, p = 0.005; 24 hr: t12 = 1.876, p = 0.0874). Data are presented as mean ± SEM. ∗∗p < 0.01.

NTS, nucleus of the solitary tract; AP, area postrema; CC, central canal. Scale bar, 200 μm. See also Figure S1.

Despite the CNO (10 μM)-induced reduction in membrane potential observed in ex vivo NTS slices from 5-HTR-hM4Di mice (n = 5/5; Figure S1C), CNO (1 mg/kg, i.p.) did not affect ad libitum food intake in 5-HTR-hM4Di mice compared with saline (Figure S1D). The effect of CNO on feeding behavior in 5-HTRNTS-hM3Dq but not 5-HTRNTS-hM4Di mice illustrates that CNO is not altering food intake via action at endogenous receptors. Rather, the designer drug CNO is reducing food intake via action at the designer 5-HTR-hM3Dq receptors. These data reveal that activation of 5-HT2CRNTS neurons significantly suppresses feeding and identifies a novel population of NTS cells that may be targeted for food intake reduction.

Selective 5-HT2CRNTS/DMV Activation Is Sufficient to Promote 5-HT2CR Agonist Hypophagia

We next evaluated whether an obesity medication currently in human use employs the subset of 5-HT2CRNTS to promote a reduction in feeding behavior. To exclusively activate 5-HT2CRNTS with clinical and preclinical obesity medications, we restricted the expression of the endogenous receptor to the caudal aspects of the dorsal vagal complex (the NTS and dorsal motor nucleus of the vagus [DMV]; 5-HTR). We achieved this using a reversible 5-HTR null line, in which expression of Htr2cr is prevented by a loxP-flanked transcriptional blocker (Xu et al., 2008). The expression of Htr2cr was site-specifically reactivated via stereotactic delivery of an AAV-expressing Cre recombinase (AAV-Cre) into the NTS/DMV, which removes the transcriptional blocker (Figures 2A and 2B), allowing 5-HT2CRs expression only in the NTS/DMV (Figures 2C and 2D). To confirm the functional re-expression of Htr2c, we made ex vivo whole-cell electrophysiological recordings from neurons within NTS slices in 5-HTR (wild-type), 5-HTR (knockout), and mice with 5-HT2CRs restored exclusively within the NTS/DMV (5-HTR). As expected, NTS neurons from 5-HTR mice failed to respond to lorcaserin, while similar proportions of NTS neurons from 5-HTR and 5-HTR mice were excited by lorcaserin (Figure 2E). These data indicate that AAV-Cre restores functional 5-HT2CR expression in 5-HTR mice.

Figure 2

5-HT2CRNTS Is Sufficient to Mediate the Appetite-Suppressive Effect of 5-HT2CR Agonists

(A) Schematic of the strategy used to selectively restore 5-HT2CRs within the NTS of loxTB5-HTR null mice through stereotactic injection of Cre-dependent DREADD vectors into NTS.

(B) 5-HTR allele expression is disrupted by a loxP-flanked transcriptional blocker (loxTB) inserted between exons 3 and 4 of the htr2c gene in loxTB5-HTR mice (5-HTR). Stereotactic injection of AAV vector expressing Cre recombinase and mCherry reporter driven by the hSyn promoter removes the TB and allows 5-HT2CR expression only in the NTS (5-HTR).

(C) Representative image of Cre-mCherry expression within the rostrocaudal aspect of the NTS/DMV.

(D) Cre-mediated recombination of the mutant allele and TB removal specifically within the NTS/DMV in 5-HTR mice (wild-type band 437 bp, mutant 310 bp).

(E) Representative current-clamp traces and proportion of neurons responding to bath application of lorcaserin in ex vivo NTS slices prepared from 5-HTR (n = 7), 5-HTR (n = 12), and 5-HTR (n = 8) mice.

(F) Sidak's post hoc comparisons revealed that lorcaserin and WAY161,503 (7 mg/kg, i.p.) reduced ad libitum food intake at specific time points within the first 3 hr of the dark cycle (tick marks on x axis represent 20-min intervals) in 5-HTR (lorcaserin, n = 5; time: F6,24 = 55.19, p < 0.0001; treatment: F1,4 = 16.81, p = 0.0149; interaction: F6,24 = 8.991, p < 0.0001. WAY161,503, n = 6; time: F6,30 = 53.37, p < 0.0001; treatment: F1,5 = 61.55, p = 0.0005; interaction: F6,30 = 26.16, p < 0.0001) and 5-HTR mice (lorcaserin, n = 5; time: F6,24 = 63.53, p < 0.0001; treatment: F1,4 = 23.78, p = 0.0082; interaction: F6,24 = 4.099, p = 0.0057. WAY161,503, n = 6; time: F6,30 = 51.89, p < 0.0001; treatment: F1,5 = 4.028, p = 0.1010; interaction: F6,30 = 1.878, p = 0.1175), but not 5-HTR mice. Data are presented as mean ± SEM. Sidak's post hoc comparisons ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

NTS, nucleus of the solitary tract; AP, area postrema; CC, central canal; DMV, dorsal motor nucleus of the vagus. See also Figure S2.

Consistent with previous reports, body weight was significantly higher in adult 5-HTR compared with 5-HTR mice (Nonogaki et al., 2003, Xu et al., 2008); an effect that failed to normalize following post-developmental re-expression of the 5-HT2CR exclusively within the NTS/DMV (Figure S2A). Further analysis of energy balance revealed no difference in daily, nocturnal, or diurnal energy expenditure between genotypes (Figures S2B and S2C). We were unable to detect the predicted increase in ad libitum chow intake (Figure S2D) or chow intake following an overnight fast in 5-HTR mice (Figure S2E), which may be related to the previously reported variability in the magnitude of hyperphagia on a chow diet with age in 5-HTR mice (Wade et al., 2008). These results indicate that restoration of 5-HT2CRs only within the NTS/DMV is not sufficient to correct the increased body weight phenotype of chow-fed 5-HTR mice.

To test the significance of 5-HTR in a therapeutic context, we treated 5-HTR, 5-HTR, and 5-HTR mice with 5-HT2CR agonists lorcaserin and WAY161,503. In line with previous research (Burke et al., 2014, Fletcher et al., 2009), the obesity medication lorcaserin and the preclinical compound WAY161,503 significantly reduced ad libitum food intake in 5-HTR mice compared with saline treatment (Figure 2F). This effect was absent in 5-HTR mice but predominantly restored in 5-HTR mice (Figure 2F). Thus, consistent with chemogenetic activation of 5-HT2CRNTS cells, pharmacological activation of this discrete subset of cells has a significant impact on feeding behavior. These data also reveal that 5-HT2CRs expressed within the NTS/DMV are sufficient to mediate acute therapeutic effects of a medication currently in clinical use to treat obesity, thereby highlighting a specific subset of 5-HT2CRs that may be exploited for future obesity medication development.

5-HT2CR Agonists Directly Activate POMCNTS Cells

We next interrogated the neurochemical phenotype of 5-HT2CRNTS cells to probe the mechanism through which 5-HT2CRNTS-mediated effects on appetite are achieved. Though POMC neurons within the ARC have been proposed to be a primary mediator of 5-HT2CR's effects on metabolic functions (Berglund et al., 2013, Burke et al., 2014, Burke et al., 2017, Doslikova et al., 2013, Heisler et al., 2002, Xu et al., 2010), a smaller and less well characterized population of POMC expressing neurons reside within the caudal aspect of the NTS (POMCNTS), a region receiving inputs from vagal afferent fibers that transmit meal-related information from the gastrointestinal system (Appleyard et al., 2005, Fan et al., 2004, Joseph et al., 1983, Zhan et al., 2013). We thus hypothesized that POMCNTS is a functional component of 5-HT2CRNTS-mediated hypophagia.

To visualize anatomical expression, we utilized PomcDsRED reporter mice (Hentges et al., 2009) and fluorescent in situ hybridization to label endogenous 5-HTR mRNA. Although more than one-third (38.2% ± 1.3%) of POMCNTS cells express 5-HTR mRNA, POMCNTS neurons represent only a small (approximately 4%) subset of the total NTS 5-HTR mRNA-expressing population (Figure 3A). Next, to evaluate the impact of 5-HT2CR activation on POMCNTS neuronal activity, we performed whole-cell electrophysiological recordings in PomcDsRED NTS slices. Consistent with the anatomical co-expression profile, 45.5% (10/22) of PomcDsRED cells were responsive to 5-HT (10 μM), the endogenous 5-HT2CR agonist (Figures 3B and 3C). A similar proportion of responders was observed following application of 5-HT2CR agonists lorcaserin (20 μM) and WAY161,503 (20 μM), with 47.1% (8/17) and 50% (11/22) of PomcDsRED cells responding, respectively (Figures 3B and 3C). All responses were characterized by depolarization and commencement/enhancement of action potential discharge. These effects were independent of action potential-dependent synaptic transmission, as depolarizations endured in the presence of 500 nM tetrodotoxin (TTX) (5-HT, 13.2 ± 3.8 mV, n = 7; lorcaserin, 11.7 ± 3.6 mV, n = 7; WAY161,503, 12.8 ± 2.7 mV, n = 8; Figure 3C), indicating that 5-HT and 5-HT2CR agonists directly activate POMCNTS cells via a post-synaptic mechanism. 5-HT and 5-HT2CR agonists induced a similar degree of depolarization and activated a similar percentage of POMCNTS cells (Figure 3D). These data reveal that 5-HT2CRs are anatomically positioned to increase the activity of a subset of POMCNTS cells.

Figure 3

5-HT and 5-HT2CR Agonists Directly Activate POMCNTS Neurons

(A) Representative image illustrating 38% of DsRED-positive POMCNTS neurons (red) express 5-HTR mRNA (green; co-expressed, yellow) in PomcDsRED mice. Scale bar, 200 μm; inset scale bar, 50 μm.

(B) Representative current-clamp recordings of POMCNTS neurons in slices from PomcDsRED mice following (i) 10 μM 5-HT, (ii) 20 μM lorcaserin, (iii) and 20 μM WAY161,503.

(C) (i) 5-HT (n = 7; F2,18 = 12.9, p = 0.001), (ii) lorcaserin (n = 7; F2,18 = 8.2, p = 0.006), and (iii) WAY161,503 (n = 8; F2,21 = 10.3, p = 0.002) elicit a significant and reversible change in membrane potential (recorded in 500 nM TTX and from an initial membrane potential of −65 mV, a value achieved through the injection of DC current). Thick lines represent mean ± SEM; dashed lines, raw data. ∗∗∗p < 0.001.

(D) Summary of membrane potential changes induced by 5-HT, lorcaserin, and WAY161,503. Boxes represent 25th, 50th, and 75th percentile with superimposed mean ± SEM and maximum and minimum values.

NTS, nucleus of the solitary tract; AP, area postrema; CC, central canal.

To examine the post-synaptic ionic mechanisms underpinning the excitatory effects of 5-HT and lorcaserin, we recorded POMCNTS neurons in a voltage clamp (VC) and in the presence of TTX. At a holding potential of −60 mV, application of 5-HT (10 μM) induced a reversible inward current of −23.1 ± 2.5 pA (n = 10; Figure 4A). The current-voltage relationship of I5-HT was examined using VC ramps, which drove the membrane potential from −120 mV to 0 mV at a rate of 45.7 mV s−1 (Figure 4B; n = 10). The digital subtraction of the ramp in control from the ramp at the peak of the response revealed I5-HT to be outwardly rectifying and to reverse at −27.0 ± 3.6 mV (Figure 4C; n = 10). Likewise, lorcaserin (20 μM) induced an inward current of −15.5 ± 2.5 pA (n = 12; Figure 4D) that was shown to be outwardly rectifying and to reverse at −24.6 ± 2.9 mV (Figures 4E and 4F). These data illustrate that 5-HT and lorcaserin excite POMCNTS neurons via the activation of a post-synaptic, mixed cationic current. This mechanism is consistent with that observed for 5-HT2CR activation of POMCARC neurons (Sohn et al., 2011).

Figure 4

5-HT and 5-HT2CR Agonists Activate POMCNTS Neurons Via Post-synaptic, Mixed Cationic Current

(A) Representative voltage-clamp recording of a POMCNTS neuron from PomcDsRED NTS slices in the presence of TTX (500 nM). Application of 5-HT (10 μM) results in an inward current (scale bar, 20 s). To the right, sharing its y axis with the raw trace, are Gaussian fits of averaged (solid lines) holding current frequency distributions in control (green), 5-HT (red), and wash (black). Raw data used to produce averages shown as dashed lines (n = 10, F2,27 = 18.1, p = 1.02 × 10−5; post hoc Tukey test, ∗∗∗p < 0.001).

(B) Averaged voltage-clamp ramps (n = 10) acquired in control (green) and at the peak of response (red). Line denotes mean, shading the SEM.

(C) 5-HT-induced current (I5-HT) obtained by the digital subtraction of the traces displayed in (B). Note the reversal at −27.0 ± 3.6 mV (n = 10). Line denotes mean, shading the SEM.

(D) Representative voltage-clamp recording of a POMCNTS neuron in the presence of TTX. Application of lorcaserin (20 μM) results in an inward current (scale bar, 20 s). To the right, sharing its y axis with the raw trace, are Gaussian fits of averaged (solid lines) holding current frequency distributions in control (green) and lorcaserin (blue). Raw data used to produce averages shown as dashed lines (n = 13, t12 = 6.2, p = 4.5 × 10−5, t-test paired, ∗∗∗p < 0.001).

(E) Averaged voltage-clamp ramps (n = 12) acquired in control (green) and at the peak of response (blue). Line denotes mean, shading the SEM.

(F) Lorcaserin-induced current (ILorcaserin) obtained by the digital subtraction of the traces displayed in (E). Note the reversal at −24.6 ± 2.9 mV (n = 12). Line denotes mean, shading the SEM.

Brain POMC Is Necessary to Mediate 5-HT2CR Agonist Feeding Effects

Thus far, we have established that the 5-HT2CRs are anatomically positioned to influence the activity of approximately 40% of POMCNTS neurons and previous reports indicate that 5-HT2CRs are co-expressed with approximately 40% of POMCARC neurons (Heisler et al., 2002, Lam et al., 2008). Earlier transgenic approaches have revealed that the subset of 5-HT2CRs that are necessary and sufficient for 5-HT2CR agonists to reduce food intake are co-expressed with POMC (Berglund et al., 2013, Xu et al., 2008). These approaches manipulate the expression of the receptor, not POMC production. We aimed to determine the function of POMC in lorcaserin's anorectic effect. As a first step toward this goal, we employed a mouse model with full POMCARC knockout and approximately half POMCNTS knockdown (Pomc) (Bumaschny et al., 2012). Pomc and wild-type (Pomc) littermates were treated with saline or lorcaserin (7 mg/kg, i.p.) at the onset of the dark cycle and ad libitum home cage chow intake was measured 1, 3, and 6 hr later. As expected, Pomc mice treated with lorcaserin significantly reduced food intake compared with saline-treated littermates at each time point (Figure 5A). However, lorcaserin did not have a significant effect on food intake at any of these time points in Pomc mice compared with saline-treated littermates (Figure 5B). These results were replicated in further cohorts of Pomc and Pomc littermates using a within-subjects experimental design (data not shown). These data reveal that brain POMC is required for lorcaserin to promote its effects on food intake.

Figure 5

Brain POMC Is a Necessary Neurochemical Mediating 5-HT2CR Agonist Feeding Effects

(A) Male and female wild-type mice (Pomc) treated with lorcaserin (7 mg/kg, i.p.) significantly reduced 1-, 3-, and 6-hr dark cycle food intake compared with saline-treated Pomc littermates (n = 11 per group; 1 hr: t20 = 2.884, p = 0.0092; 3 hr: t20 = 2.916, p = 0.0085; 6 hr: t20 = 3.373, p = 0.0030).

(B) In contrast, male and female mice with full POMCARC knockout and partial POMCNTS knockdown (Pomc) exhibited a blunted response to lorcaserin treatment as compared with Pomc littermates treated with saline (n = 6–7 per group; 1 hr: t11 = 1.3204, p = 0.216; 3 hr: t11 = 0.4098, p = 0.6898; 6 hr: t11 = 0.9486, p = 0.3632).

Data are presented as mean ± SEM. ∗∗p < 0.01.

POMCARC Is Required for the Full Effect of 5-HT2CR Agonist on Food Intake

We next sought to determine the functional significance of the discrete subset of POMCNTS versus POMCARC in 5-HT2CR agonist-induced reduction of food intake. We first targeted Pomc in the hypothalamus, since this population of Pomc expressing neurons is larger, better characterized, and projects widely within the hypothalamus, allowing efficient tissue sampling and recovery of peptides that localize at neuron terminals (e.g., POMC-derived peptide α-melanocyte-stimulating hormone [α-MSH] [D'Agostino and Diano, 2010]). To knock down POMCARC, we employed the clustered, regularly interspaced, short palindromic repeats (CRISPR) and associated endonuclease 9 (Cas9) gene-editing technique (Doudna and Charpentier, 2014, Hsu et al., 2014). Specifically targeted by synthetic guide RNA (sgRNA) sequences, Cas9-induced double-stranded breaks typically result in frameshifting insertion/deletions (indels) owing to the action of non-homologous end-joining (NHEJ) DNA repair mechanisms, leading to disruption of the encoded protein. Following an in vitro screening of selected sgRNAs targeting the third exon of the Pomc gene, Cas9-induced indels at the predicted positions were detected using a T7 endonuclease I assay (Figure 6A). Two active sgRNAs were cloned into an AAV vector that also expressed eGFP to enable site-specific brain delivery (Swiech et al., 2015).

Figure 6

PomcARC Is Required for 5-HT2CR Agonist Hypophagia

(A) Representative screening of sgRNAs targeting the exon 3 of the Pomc gene in Neuro2A cells and indels detection using a T7 endonuclease I assay. (A, left) Sites of potential indels in exon 3 are indicated by red or green bars, and the presence of indels results in the fragments indicated by red or green asterisks, respectively, following endonuclease treatment. (A, right) Bands corresponding to these fragments are indicated with red and green asterisks on the gel.

(B) Schematic of bilateral mediobasal hypothalamic stereotactic delivery of AAVs expressing Pomc-targeting CRISPR/Cas9 and EGFP (PomcARC-CRISPR). Representative image of GFP within the arcuate nucleus of the hypothalamus (ARC) third ventricle (3V).

(C) Hypothalamic content of Pomc peptide product α-melanocyte-stimulating hormone (α-MSH) 6 weeks after AAV injection is significantly reduced as measured by quantitative fluorescent EIA assay in PomcARC-CRISPR mice compared with controls injected with AAV-expressing sgRNA-eGFP alone (PomcARC-WT), illustrating that in vivo CRISPR targeting knocked down Pomc in the ARC (p = 0.0059; Mann-Whitney test).

(D) PAGE analysis identified heteroduplexes (pink arrowheads) from exon 3 of Pomc amplified from the mediobasal hypothalamus of PomcARC-CRISPR mice.

(E) PomcARC-CRISPR mice exhibit 24 hr of hyperphagia for home cage chow following food restriction compared with PomcARC-WT mice (n = 7–8; F1,13 = 11.27, p = 0.005).

(F) Likewise, PomcARC-CRISPR mice show greater motivation to work for palatable food compared with PomcARC-WT mice as assessed by the breakpoint of lever presses for chocolate pellets using a progressive-ratio schedule (n = 7–8; t12 = 2.463, p = 0.029).

(G) (Left) Lorcaserin (7 mg/kg, i.p.) reduced acute food intake within the first 3 hr (tick marks on x axis represent 20-min intervals) of the dark cycle in PomcARC-WT mice (n = 7; time: F9,54 = 38.05, p < 0.0001; treatment: F1,6 = 39.94, p = 0.0007; interaction: F9,54 = 13.91, p < 0.0001) and (right) as analyzed as total 3- and 6-hr cumulative intake (time: F1,6 = 169.7, p < 0.0001; treatment: F1,6 = 11.22, p = 0.0154; interaction: F1,6 = 0.8351, p = 0.3960).

(H) (Left) Lorcaserin (7 mg/kg, i.p.) also reduced acute food intake within the first 3 hr in PomcARC-CRISPR mice (n = 10; time: F9,81 = 168.1, p < 0.0001; treatment: F1,9 = 16.46, p = 0.0029; interaction: F9,81 = 2.108, p = 0.0381). (Right) However, lorcaserin was not effective in reducing cumulative food intake compared with saline treatment in PomcARC-CRISPR mice when analyzed at 3 or 6 hr food intake with Sidak's post hoc comparisons (n = 10; time: F1,9 = 130.8, p < 0.0001; treatment: F1,9 = 110.44, p = 0.0103; interaction: F1,9 = 0.01045, p = 0.9208).

Data are presented as mean ± SEM. Sidak's post hoc comparisons ∗p < 0.05, ∗∗p < 0.01.

AAVs expressing Pomc-targeting Cas9 and sgRNA-eGFP (PomcARC-CRISPR) or AAVs expressing sgRNA-eGFP alone (PomcARC-WT) were then stereotactically injected into the ARC (Figure 6B). Quantitative analysis of hypothalamic α-MSH content revealed PomcARC-CRISPR mice had an approximately 60% reduction in α-MSH (Figure 6C). To control for the occurrence of CRISPR/Cas9-induced indels, we implemented a method allowing ex vivo detection of DNA heteroduplexes. This is a widely used proxy of indel occurrence in vitro. Genomic DNA from PomcARC-CRISPR and PomcARC-WT mice was extracted and the third exon of the Pomc gene was amplified by PCR. Subsequent PAGE analysis revealed that DNA heteroduplexes were exclusively in samples obtained from the mediobasal hypothalamus in PomcARC-CRISPR mice. DNA heteroduplexes were not found in the hypothalamus of PomcARC-WT mice (Figure 6D). Behaviorally, PomcARC-CRISPR mice displayed phenotypes consistent with reduced POMCARC function, including hyperphagia following food deprivation (Figure 6E) and greater operant responding for palatable food (Figure 6F). Next, we assessed the role of POMCARC in lorcaserin's effects on food intake. As expected, lorcaserin significantly reduced ad libitum dark cycle food intake compared with saline treatment in PomcARC-WT mice when analyzed over the first 3 hr of the dark cycle (Figure 6G, left) and as cumulative 3 and 6 hr of food intake (Figure 6G, right). In contrast, PomcARC-CRISPR mice only significantly responded to the acute anorectic effect of lorcaserin (Figure 6H, left). Analysis of 3- and 6-hr cumulative intake revealed that PomcARC-CRISPR mice exhibited an attenuated response to lorcaserin. Specifically, lorcaserin did not significantly alter food intake compared with saline treatment at either time point in PomcARC-CRISPR mice (Figure 6H, right). One interpretation of these data is that POMCARC is not a principal mediator of lorcaserin's acute effects on food intake; rather, it is necessary for the longer-term effect. However, it is possible that the acute response to lorcaserin in PomcARC-CRISPR mice is due to incomplete knockdown of POMCARC and that the subset remaining is sufficient for lorcaserin to reduce feeding at this earlier time point. Nevertheless, these data indicate that lorcaserin requires functional Pomc within the ARC to promote its full effect on food intake.

POMCNTS Is Required for the Acute Effect of 5-HT2CR Agonists on Food Intake

We next examined the functional significance of POMCNTS in the therapeutic effects of 5-HT2CR agonist obesity medications. We created two cohorts of mice by stereotactically delivering AAVs expressing Pomc-targeting Cas9 and sgRNA-eGFP (PomcNTS-CRISPR) or AAVs expressing sgRNA-eGFP alone (PomcNTS-WT) to the NTS (Figures 7A and 7B). Since α-MSH expression from NTS explants is not at a high enough level for reliable measurement, we adopted the method used above for PomcARC-CRISPR validation, allowing ex vivo detection of DNA heteroduplexes to control for the occurrence of CRISPR/Cas9-induced indels. Genomic DNA from PomcNTS-CRISPR and PomcNTS-WT mice was extracted and the third exon of the Pomc gene was amplified by PCR. Subsequent PAGE analysis revealed that DNA heteroduplexes were exclusively in samples obtained from the NTS/DMV and not an internal control region in PomcNTS-CRISPR mice (Figures 7C and 7D). As expected, DNA heteroduplexes were not found in either the NTS/DMV or an internal control region of PomcNTS-WT mice (Figure 7C). These data suggest that Pomc has been successfully targeted in PomcNTS-CRISPR mice, although this method does not provide quantification of the degree of knockdown. As expected, lorcaserin and WAY161,503 significantly reduced ad libitum dark cycle food intake in PomcNTS-WT mice over 6 hr (Figures 7E and 7G). However, following administration of 5-HT2CR agonists, Pomc mice exhibited an attenuated anorectic response to both lorcaserin and WAY161,503 during the first 3 hr of the dark cycle (Figure 7F). This attenuated effect to lorcaserin and WAY161,503 did not persist for the full 6 hr in PomcNTS-CRISPR mice. Lorcaserin significantly reduced 6-hr cumulative intake in PomcNTS-CRISPR mice compared with saline treatment (Figure 7H). Taken together with the Pomc and PomcARC-CRISPR results, these data suggest that POMCNTS is required for the acute effect of lorcaserin on food intake, whereas POMCARC is required for the longer-term effect on feeding. These findings are generally consistent with previous reports using chemogenetic and optogenetic approaches to examine POMC neuron function, which suggest that POMCARC and POMCNTS neurons affect feeding on different time scales (Aponte et al., 2011, Koch et al., 2015, Zhan et al., 2013).

Figure 7

PomcNTS Is Required for 5-HT2CR Agonist Hypophagia

(A) AAV-mediated strategy to express in the nucleus of the solitary tract (NTS) Pomc-targeting sgRNAs and Cas9. Schematic of bilateral NTS stereotactic delivery of AAVs expressing Pomc-targeting CRISPR/Cas9 and EGFP (PomcNTS-CRISPR).

(B) Representative image of eGFP expression within the dorsal vagal complex (DVC).

(C) PAGE analysis identified heteroduplexes (pink arrowheads) from exon 3 of Pomc amplified from the DVC of PomcNTS-CRISPR mice.

(D) Representative image of tissue samples taken from the NTS and broader dorsal vagal complex (DVC) and an internal control region (Int Ctrl).

(E and F) 5-HT2CR agonists lorcaserin and WAY161,503 (7 mg/kg, i.p.) reduced acute food intake in control PomcNTS-WT mice (E) within the first 3 hr (tick marks on x axis represent 20-min intervals) of the dark cycle (lorcaserin, n = 8; time: F9,63 = 102.2, p < 0.0001; treatment: F1,7 = 18.55, p = 0.0035; interaction: F9,63 = 6.245, p < 0.001. WAY161,503, n = 8; time: F9,63 = 54.95, p < 0.0001; treatment: F1,7 = 24.28, p = 0.0017; interaction: F9,63 = 3.369, p = 0.0020), but not in PomcNTS-CRISPR mice (F) (lorcaserin, n = 7; time: F9,54 = 38.39, p < 0.0001; treatment: F1,6 = 0.1580, p = 0.7048; interaction: F9,54 = 0.2533, p = 0.9841. WAY161,503, n = 7; time: F9,54 = 62.69, p < 0.0001; treatment: F1,6 = 2.444, p = 0.1690; interaction: F9,54 = 0.7558, p = 0.6567).

(G) Likewise, when analyzed as total cumulative 3- and 6-hr dark cycle intake, lorcaserin and WAY161,503 significantly reduced food intake in control PomcNTS-WT mice (lorcaserin, n = 8; time: F1,7 = 96.71, p < 0.0001; treatment: F1,7 = 70.40, p < 0.0001; interaction: F1,7 = 0.6355, p = 0.4515. WAY161,503, n = 8; time: F1,7 = 62.88, p < 0.0001; treatment: F1,7 = 17.53, p = 0.0041; interaction: F1,7 = 4.238, p = 0.0785).

(H) POMCNTS knockdown in PomcNTS-CRISPR mice prevented the acute 3-hr anorectic effect of 5-HT2CR agonists, but was not sufficient to prevent the full effect 6 hr post treatment as analyzed with Sidak's post hoc comparisons (lorcaserin, n = 7; time: F1,6 = 92.76, p < 0.0001; treatment: F1,6 = 2.823, p = 0.1439; interaction: F1,6 = 5.309, p = 0.0608. WAY161,503, n = 7; time: F1,6 = 66.68, p = 0.0002; treatment: F1,6 = 4.152, p = 0.00877; interaction: F1,6 = 4.005, p = 0.0923).

Data are presented as mean ± SEM. Sidak's post hoc comparisons ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Discussion

Obesity negatively affects human health on a global scale. If we are to meet this challenge and design new medications, it is paramount that we gain a more complete understanding of the mechanisms regulating food intake and body weight. The 5-HT2CR agonist lorcaserin is a current obesity medication; however, neither the specific subset of 5-HT2CRs coordinating the therapeutic effect nor the neurochemical mediator targeted by these receptors has been fully defined. Earlier research genetically manipulating the receptor revealed that the subset of 5-HT2CRs co-expressed with brain POMC are both sufficient (Xu et al., 2010) and necessary (Berglund et al., 2013) to modulate the anorectic effect of preclinical 5-HT2CR agonists. Given that 5-HT2CR agonists increase the activity of POMCARC neurons (Burke and Heisler, 2015, Doslikova et al., 2013, Heisler et al., 2002), it was presumed that these neurons were the mediator of this effect. However, in light of recent findings that POMCARC neurons suppress appetite on a time scale of hours (Aponte et al., 2011, Koch et al., 2015, Zhan et al., 2013), we considered whether 5-HT2CRs outside the ARC are necessary for 5-HT2CR's acute effects on feeding. Here we identify the function of a previously uncharacterized subset of 5-HT2CR expressing neurons, localized within the NTS, and reveal that the activation of this subpopulation has a significant and rapid impact on feeding behavior, and furthermore is sufficient to drive lorcaserin's acute reduction in food intake. In addition, we report that a subset of POMCNTS neurons express 5-HT2CRs and that 5-HT and 5-HT2CR agonists directly activate POMCNTS neurons via a post-synaptic, mixed cationic current. We also clarify that brain POMC is necessary for lorcaserin to influence feeding. Although POMC is not abundantly expressed within the NTS, we report that POMC in this region is required for 5-HT2CR agonist obesity medications to acutely reduce food intake. These data thereby reveal a subpopulation of 5-HT2CRs, which when pharmacologically activated significantly reduce food intake, identify that the mechanistic underpinnings of this acute effect is via POMCNTS, and demonstrate that this discrete 5-HT2CRNTS subpopulation is therapeutically relevant to obesity medications targeting the 5-HT2CRs.

Limitations of Study

It is possible that the results obtained are affected by incomplete knockdown of Pomc in the ARC and NTS. Although altering POMCNTS expression significantly attenuated the acute effect of lorcaserin on food intake, at baseline PomcNTS-CRISPR mice displayed food intake comparable with PomcNTS-WT mice during ad libitum feeding (Figures 7E and 7F; and data not shown). These data are consistent with a previous report of diphtheria toxin POMCNTS cell ablation in a Pomc mouse line, which did not produce an effect on ad libitum food intake (Zhan et al., 2013). However, we cannot exclude the possibility that this absence of baseline phenotype PomcNTS-CRISPR mice is due to incomplete knockdown of PomcNTS. Further studies are required to fully interrogate the physiological role of POMCNTS. Nevertheless, our data reveal that the relatively small source of POMC specifically derived within the NTS is necessary for the obesity medication lorcaserin and the preclinical compound WAY161,503 to promote their full effects on feeding behavior.

STAR★Methods

Key Resources Table

Contact for Reagent and Resource Sharing

Further information and requests for reagents may be directed to the Lead Contact Lora Heisler (lora.heisler@abdn.ac.uk).

Experimental Model and Subject Details

Mice

5-HTR-ires (Burke et al., 2016, Marcinkiewcz et al., 2016), loxtb5-HTR (B6.129-Htr2ctm1Jke/J; The Jackson Laboratory, Bar Harbor, ME USA), PomcDsRED (Hentges et al., 2009) and Pomc (Bumaschny et al., 2012) mice were housed with ad libitum food and water access (unless otherwise stated) in a light- (12 h on/12 h off) and temperature-controlled (21.5°C to 22.5°C) environment. All procedures were performed in accordance with the U.K. Animals (Scientific Procedures) Act 1986 and local ethical approvals.

Method Details

Stereotaxic Surgery and Viral Vectors

Viral constructs were packaged in AAV serotype-8 and were delivered to the NTS via bilateral stereotaxic injections, as previously described (D'Agostino et al., 2016).

AAV vectors

The DREADD viruses used have been described previously: AAV8-hSyn-DIO-hM3Dq-mCherry, AAV8-hSyn-DIO-hM4Di-mCherry and AAV8-hSyn-mCherry-Cre were packaged in AAV serotype-8 at a titer of 1.3 x 1013 vg/ml (University North Carolina Vector Core, Chapel Hill, NC, USA) (Krashes et al., 2011, D'Agostino et al., 2016). AAV/DJ-CMV-spCas9 was obtained from Vector Biolabs (Cat No. 7120; Vector Biolabs, Malvern, PA, USA). High titer pAAV-U6sgRNA(SapI)_hSyn-GFP-KASH-bGH (PX552) were packaged in AAV serotype-8 at a titer of 4.1 x 1012 vg/ml (Unitat de Producció de Vectors, Universitat Autonoma Barcelona, Spain).

Stereotaxic Surgery

NTS delivery of AAVs was achieved through a stereotaxic procedure. Briefly, 5–8 week old male mice were anesthetized with a mixture of ketamine and xylazine dissolved in saline (100 and 10 mg/kg, respectively injected in a volume of 10 ml/kg, i.p.). Mice were placed in a stereotaxic frame (Kopf Instruments, Tujunga, CA USA) using ear bars with the head angled at 45°. Under magnification, an incision was made at the level of the cisterna magna and neck muscles carefully retracted. A 33G needle was used for dura incision. The obex served as reference point for injection. Bilateral injections were performed using a glass micropipette (diameter 20–40 μm). NTS coordinates used were A/P, -0.2; M/L, ±0.2; D/V, -0.2 from the obex. Virus was delivered under air pressure using a PLI-100A Pico-Injector (Harvard Apparatus, Cambridge, UK). 150 nl of virus per side was delivered with multiple microinjections over 4–5 minutes. The pipette remained in place for a minimum of 3 minutes after injection. Following all surgery, animals were closely monitored, kept warm and appropriate postsurgical care was taken. Animals were also administered an analgesic (5 mg/kg carprofen, s.c.) daily for 2 days post-operatively. Prior to experimentation, mice were acclimated to handling and i.p. injections.

Mice were given a minimum of 14 days of surgical recovery before experimentation. For chemogenetic experiments, mCherry was visualised by immunohistochemistry. Mice in which the expression of reporter fluorescent protein was bilaterally detected in the caudal aspect of the DVC were included in the analysis, whereas mice without reporter fluorescent expression or expression outside the DVC were excluded from the analysis. Caudal DVC was defined by the presence of the area postrema in the same section of brain tissue.

DNA Constructs

For SpCas9 target selection and generation of single guide RNA (sgRNA), the 20-nt target sequences were selected to precede a 5’-NGG protospacer-adjacent motif (PAM) sequence. To minimize off-targeting effects, the CRISPR design tool was used (http://crispr.mit.edu/). The 20-nt target and complement sequences were synthesized to include BbsI overhangs. Oligos were annealed, phosphorylated and ligated into the pX330-U6-Chimeric_BB-CBh-hSpCas9 (PX330; Addgene plasmid #42230) plasmid after BbsI digestion.

pAAV-U6sgRNA(SapI)_hSyn-GFP-KASH-bGH (PX552; Addgene plasmid # 60958) was used for cloning sgRNAs into AAV the backbone and generate viral particles. The plasmid was first digested with SapI to create sticky ends for ligation. The 20-nt target and complement sequences were synthesized to include SapI overhangs and ligated. A second sgRNA was then inserted into this vector. The second sgRNA was first PCR-amplified from PX330. Primers were designed to include 15bp homology arms to the PspOMI-lineraised PX552 into the PCR clone sgRNA guide. The PCR product was column-purified and then cloned into the PspOMI-lineraised PX552 vector via homologous recombination-assisted cloning (In-Fusion HD cloning Kit®, Cat No. 638910; Clonotech Laboratories). Primers sequences: AGACTGCAGAGGGCCGAGGGCCTATTTCCCATGATTCCT, forward; CTCATACGCAGGGCCCTAAAACAAAAAAGCACCGACTCGG, reverse. All obtained constructs were verified by sequencing. All plasmids were amplified using One Shot® Stbl3™ Chemically Competent E. coli (Cat No. C737303; Invitrogen) and column-purified (Cat No. 270104; Qiagen).

Pomc targeting sgRNAs (5’ to 3’): sgRNA 1- GGTGGGCAAGAAACGGCGCC(cgg); sgRNA 2- GTGACCCATGACGTACTTCCG(ggg).

Cell Line Culture and Transfection

Mouse Neuro-2a (N2a) cells were grown in DMEM containing 10% FBS. Cells were maintained at 37°C in 5% CO2 atmosphere. Transfection were performed using FuGENE ® HD Transfection Reagent (Cat No E2311; Promega) according to the manufacturer’s protocol. Briefly, 1.5x104 cells per well were plated into a 12 well plate the day before transfection. 1 ug of PX330 plasmid was mixed with Fugene HD in Optimem medium (ratio 1:3). Cells were incubated for 48 hours. Genomic DNA was extracted from transfected cells using PureLink® Genomic DNA Mini Kit (Cat No K182001, Invitrogen) following the manufacturer’s instructions.

Detection of Insertions/Deletions (Indels)

The third exon of the mouse Pomc gene was PCR amplified from genomic DNA using the following primers: GCTTGCATCCGGGCTTGC, forward; GACTTTATTTACGCAGTTTTTATTGAAGATCAGAGC, reverse. This PCR reaction generates a single PCR product of the expected 726bp size. PCR product was sequence verified. To detect sgRNA/spCas9-induced indels, genomic DNA from cells was amplified using the same primers and this 726bp product corresponding to third exon of the mouse Pomc gene was used for the endonuclease I assay (Surveyor® Mutation Detection Kit; Cat No. 706020; Integrated DNA Technologies) following the manufacturer’s instructions. Briefly, PCR amplicons were heat-denaturated and re-annealed using a thermocycler to create DNA heteroduplexes. DNA heteroduplexes were incubated with endonuclease I enzyme, which cleaves DNA at the site of a base mismatch with high specificity, for 1 hour at 42°C. Genomic modifications were verified using gel agarose electrophoresis.

Feeding Studies

Dark Cycle Food Intake

Mice were injected with vehicle or drug 30 minutes prior to the onset of the dark cycle and food was removed. At the onset of dark, food was returned and manually (home cage) or automatically (Phenomaster chambers, TSE, Bad Homburg, Germany) weighed using methods previously described (D'Agostino et al., 2016).

Post-fast Refeeding

Food was removed at the onset of the dark cycle and returned 2 hours after the onset of the light cycle the following day. Food intake was measured manually (home cage) or automatically (TSE Phenomaster chambers) using methods previously described (D'Agostino et al., 2016).

Progressive Ratio

Mice were trained in operant chambers housed in sound-attenuating boxes and controlled by a PC Med-PC-IV programming language (MED Associates, Inc., Fairfax, VT USA). Chambers (21.6 cm long × 17.8 cm wide × 12.7 cm high) have a retractable lever, a pellet receptacle and a 3 watt house-light on the opposite wall (Cat No. MED-307W-D1, MED Associates, Inc.). Under food restriction, mice were trained to press the lever for 20 mg chocolate pellet reinforcers (TestDiet, St. Louis, MO USA) under a fixed ratio (FR) 1 schedule for 3 days followed by FR 5 for a further 3 days. Then they were fed ad libitum and underwent daily training under a PR schedule based on an exponential progression derived from the formula (5 × e0.2n)−5, rounded to the nearest integer, where n is the position in the ratio sequence (Richardson and Roberts, 1996). 50 minute sessions took place at the same time each day during the light phase (between 08:00 and 13:00 hours) for 1 week. The breakpoint or alternatively the highest completed ratio (Olarte-Sánchez et al., 2012) was defined as the last ratio completed before 5 minutes elapsed without any responding.

Immunohistochemistry (IHC) and Fluorescent In Situ Hybridization Histochemistry (FISH)

IHC. Tissue was processed for GFP, mCherry, dsRed and/or c-Fos (FOS-IR) IHC as previously described (D’Agostino et al., 2016). Briefly, mice were transcardially perfused with phosphate buffered saline (PBS) followed by 10% neutral buffered formalin (Sigma-Aldrich). Brains were extracted, post-fixed in 10% neutral buffered formalin at 4°C, cryoprotected in 20% sucrose at 4°C and then sectioned coronally on a freezing sliding microtome at 25 μm. Tissue was processed for chicken anti-GFP (1:1000; AbCam, ab13970), rabbit anti-dsRED (1:1000; Rockland, 600-401-379), goat anti-mCherry/RFP (1:1000; Sicgen, AB0040-200) or anti-c-FOS (1:5000, rabbit, Calbiochem, PC38) primary antibodies and a biotinylated donkey anti-rabbit (1:500, Jackson ImmunoResearch Laboratories, Inc.) or Alexa Fluor (1:500, Life Technologies) secondary antibodies using standard protocols previously described (Lam et al., 2009, Heisler et al., 2006). Tissue was then mounted on slides, cover slipped and the NTS visualized using an Axioskop II microscope (Carl Zeiss, Oberkochen, Germany) and Adobe Photoshop CS5 software. Images of single-label immuoreactivity (IR) for GFP or mCherry were used to visualize and analyze NTS injection sites.

Dual-IHC

Dual-IHC was performed to visualize co-expression of mCherry-IR and FOS-IR in 5-HTRNTS-hM3Dq mice. Specifically, mice were fasted overnight and then injected with designer drug CNO or saline. Brains were extracted 90 minutes later and dual-label mCherry and FOS-IR performed using method outlined above. Quantitative analysis of single and dual-labelled cells was performed manually using an Axioskop II microscope and Adobe Photoshop CS5 software.

IHC and FISH

IHC and FISH were performed to visualize co-expression of POMC and 5-HT2CR within the NTS in POMC mice. To label 5-HTR expressing neurons, a RNA expression vector (pBluescript SK-) containing the 3-kilobase (kb) coding region of 5-HT2CR cDNA was used to generate single-stranded RNA probes (Julius et al., 1988). Briefly, in vitro transcription was performed using Digoxigenin (DIG)-RNA Labelling Mix (Roche, Mannheim, Germany). Sections were treated with 1% Sodium Borohydride solution and 0.25% Acetic Anhydride in Triethanolamine (TEA) solution. Sections were then incubated in a hybridization buffer containing a DIG-UTP-labelled 5-HT2CR riboprobe overnight at 55°C. Sections were rinsed in a standard sodium citrate/50% formamide solution, rinsed in an RNase (0.02 mg/ml RNase A) solution, incubated in a 3% H2O2 solution for 30 minutes and blocked in a 2% sheep serum (Sigma-Aldrich, Saint Louis, MO USA). Sections were then incubated with an anti-DIG antibody (1:5000, Roche, Mannheim, Germany), treated with TSA PLUS Biotin Kit (PerkinElmer Inc., Waltham, MA USA) and revealed with streptavidin conjugated Alexa Fluor® 568 (1:2000, Life Technologies, Carlsbad, CA USA). To label POMC, sections were blocked again and incubated with rabbit anti-dsRED (1:1000; Rockland, Limerick, PA USA; 600-401-379) overnight at 4°C using the IHC protocol described above. The sections were incubated with Alexa Fluor® 568 donkey anti-rabbit secondary antibody (1:500, Invitrogen™ Life Technologies, Carlsbad, CA USA) for 1 hour. Tissue was then mounted on slides, cover slipped and single- and dual-labelling assessed in the NTS using an Axioskop II microscope and Adobe Photoshop CS5 software. Quantitative analysis of single and dual-labelled dsRed-IR and 5-HTR mRNA cells was conducted manually.

Enzyme Immunoassay (EIA) for Alpha-Melanocyte-Stimulating Hormone (Alpha-MSH)

Mice were anesthetized with sodium pentobarbital (Euthatal) and decapitated. The brain was rapidly removed, the hypothalamus was dissected using razor blades and mouse acrylic brain matrices (Stoelting Co.) and frozen in dry ice. Hypothalamic lysates were obtained using a probe sonicator. Lysis consisted of an acid-ethanol solution obtained by combining 1 part of concentrated HCl with 7 parts pure ethanol. 0.1ml of lysis was used per 10 mg of tissue (weight of each hypothalamic explant was about 30 mg). 10 μl of crude homogenate were combined with 50 μl of fresh lysis buffer and centrifuged 3500 g for 30 min. Supernatants were collected and dried using a vacuum centrifuge in low protein binding tubes (Eppendorf® LoBind, Sigma-Aldrich, Cat No. Z666505). Dry fractions were solubilised in 150 μl of assay buffer and alpha-MSH content analysed using a fluorescent EIA ultra-sensitive assay (Phoenix Pharmaceutical Inc.; Cat No FEK-043-01) following the manufacturer’s instructions.

Electrophysiology and Chemogenetic Validation Analysis

For electrophysiological experiments, male and female POMC (n=44), male loxtb5-HTR (n=9), male 5-HTRNTS-hM3Dq (n=2) or male 5-HTRNTS-hM4Di (n=2) mice aged 2-6 months were anesthetized with sodium pentobarbital (Euthatal) and decapitated. The brain was rapidly removed and placed in cold (0–4°C), oxygenated (95%O2/5%CO2) ‘slicing’ solution containing (in mM) sucrose (214), KCl (2.5), NaH2PO4 (1.2), NaHCO3 (26), MgSO4 (4), CaCl2 (0.1), D-glucose (10). The brain was fixed to a vibrating microtome (Campden Instruments, Loughborough, UK) and 200-μm thick coronal sections of the brainstem containing the NTS were prepared. Slices were immediately transferred to a ‘recording’ solution containing (in mM) NaCl (127), KCl (2.5), NaH2PO4 (1.2), NaHCO3 (26), MgCl2 (1.3), CaCl2 (2.4) and D-glucose (10) in a continuously oxygenated holding chamber at 35°C for a period of 25 minutes. Slices were remained in the recording solution at room temperature for a minimum of 1 hour before recording. For whole-cell recordings, slices were transferred to a submerged chamber and a Slicescope upright microscope (Scientifica, Uckfield, UK) was used for infrared - differential interference contrast and fluorescence visualization of cells. During recording, slices were continuously perfused at a rate of ca. 2 ml/minute with oxygenated recording solution at room temperature. All pharmacological compounds were bath applied. Whole cell current-clamp recordings were performed with pipettes (3–7 MΩ when filled with intracellular solution) made from borosilicate glass capillaries (World Precision Instruments, Aston, UK) pulled on a Zeitz DMZ micropipette puller (Zeitz Instruments GmBH, Martinsried, Germany). The intracellular recording solution contained (in mM) K-gluconate (140), KCl (10), HEPES (10), EGTA (1), Na2ATP (2), pH 7.3 (with KOH). Recordings were performed using a Multiclamp 700B amplifier and pClamp10 software (Molecular Devices, Sunnyvale, CA USA). Liquid junction potential was 16.4 mV and not compensated. The recorded current was sampled at 10 kHz and filtered at 2 kHz unless otherwise stated.

Drugs

Lorcaserin (LGM Pharma, Erlanger KY USA), WAY161,503 (Tocris Bioscience, Abingdon UK) and clozapine-n-oxide (CNO; Tocris Bioscience) were prepared in double distilled H2O, diluted in sterile phosphate-buffered saline (PBS) and administered in vivo i.p. in a volume of 10 ml/kg. 5-HT (Sigma-Aldrich), lorcaserin, WAY161,503 and CNO were dissolved in artificial cerebrospinal fluid (aCSF) for electrophysiology experiments.

Quantification and Statistical Analysis

Data were analyzed with t-test, Mann Whitney U test, One-way, Two-way or Repeated Measures ANOVA followed by Tukey’s or Sidak’s post hoc tests, where appropriate. For all analyses, significance was assigned at p<0.05. Data are presented as mean±SEM. All statistical analyses were completed with GraphPad Prism 7.0 (GraphPad Software, San Diego, CA USA).

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Chicken anti-GFP	AbCam	Cat# ab13970; RRID: AB_300798
Rabbit anti-dsRED	Rockland	Cat# 600-401-379; RRID: RRID: AB_2209751
Rabbit anti-c-FOS	Calbiochem	Cat# PC38; RRID:AB_2106755
Goat mCherry (RFP)	Sicgen	Cat# AB0040-200; RRID:AB_2333092
Donkey polyclonal anti-chicken alexa 488	Jackson ImmunoResearch	Cat# 703-545-155; RRID: AB_2340375
Goat anti-Rabbit IgG Secondary Antibody, Alexa Fluor 594	Life Technologies	Cat# A-11012, RRID:AB_141359
Donkey anti-Goat IgG Secondary Antibody, Alexa Fluor 594	Life Technologies	Cat# A-11057; RRID:AB_142581
Biotin-SP-AffiniPure F(ab')2 Fragment Donkey Anti-Rabbit IgG	Jackson ImmunoResearch	Cat# 711-066-152, RRID:AB_2340594
Streptavidin, Alexa Fluor® 568 conjugate antibody	Life Technologies	Cat# S-11226, RRID:AB_2315774
Bacterial and Virus Strains
One Shot® Stbl3™ Chemically Competent E. coli	Invitrogen	Cat# C737303
AAV8-hSyn-DIO-hM3Dq-mCherry	University North Carolina Vector Core	N/A
AAV8-hSyn-DIO-hM4Di-mCherry	University North Carolina Vector Core	N/A
AAV8-hSyn-mCherry-Cre	University North Carolina Vector Core	N/A
AAV-U6sgRNA(SapI)_hSyn-GFP-KASH-bGH	Unitat de Producció de Vectors, Universitat Autonoma Barcelona, Spain	N/A
AAV/DJ-CMV-spCas9	Vector Biolabs	Cat# 7120
Chemicals, Peptides, and Recombinant Proteins
Clozapine-N-oxide (CNO)	Tocris Bioscience	Cat. No. 4936; CAS: 34233-69-7
Lorcaserin Hydrochloride	LGM Pharma	Cat# 846589-98-8Lot# 20130115
WAY 161503 hydrochloride	Tocris Bioscience	Cat. No. 1801; CAS: 276695-22-8
Critical Commercial Assays
In-Fusion HD cloning Kit	Clonotech Laboratories	Cat No. 638910
Plasmid plus Maxi kit	QIAGEN	Cat No. 12943
QIAquick Gel Extraction Kit	QIAGEN	Cat No. 28704
FuGENE HD Transfection Reagent	Promega	Cat No E2311
PureLink Genomic DNA Mini Kit	Invitrogen	Cat No K182001
Surveyor Mutation Detection Kit	Integrated DNA Technologies	Cat No. 706020
MSH, alpha EIA kit	Phoenix Pharmaceutical Inc.	Cat No FEK-043-01
Experimental Models: Cell Lines
Neuro-2a (N2a) cells	ATCC	CCL-131
Experimental Models: Organisms/Strains
Mouse: 5-HT2CR-iresCre	Burke et al., 2016, Marcinkiewcz et al., 2016	N/A
Mouse: POMCdsRed	Hentges et al., 2009	N/A
Mouse: POMCNEO	Bumaschny et al., 2012	N/A
Mouse: loxtb5-HT2CR (B6.129-Htr2ctm1Jke/J)	Jackson Laboratory	JAX: 015821
Mouse: C57BL/6J	Jackson Laboratory	JAX: 000664
Oligonucleotides
Genomic DNA Pomc PCR primer F:GCTTGCATCCGGGCTTGC	This paper	N/A
genomic DNA Pomc PCR primer R:GACTTTATTTACGCAGTTTTTATTGAAGATCAGAGC	This paper	N/A
Pomc gRNA1:GGTGGGCAAGAAACGGCGCC	This paper	N/A
Pomc gRNA2:GTGACCCATGACGTACTTCCG	This paper	N/A
Recombinant DNA
pX330-U6-Chimeric_BB-CBh-hSpCas9 (PX330)	Addgene	Addgene Plasmid #42230
pAAV-U6sgRNA(SapI)_hSyn-GFP-KASH-bGH (PX552)	Addgene	Addgene Plasmid #60958
cDNA HTR2C cDNA	Julius et al., 1988	N/A
Software and Algorithms
pClamp10 software	Molecular Devices
CRISPR Design	Zhang Lab, MIT	http://crispr.mit.edu/
PhenoMaster	TSE	https://www.tse-systems.com/product-details/phenomaster
MATLAB	Mathworks	https://www.mathworks.com/
Prism6	GraphPad	https://www.graphpad.com/scientific-software/prism/