Relaxin/insulin-like family peptide receptor 4 (Rxfp4) expressing hypothalamic neurons modulate food intake and preference in mice

Objective Insulin-like peptide 5 (INSL5) signalling, through its cognate receptor relaxin/insulin-like family peptide receptor 4 (RXFP4), has been reported to be orexigenic, and the high fat diet (HFD) preference observed in wildtype mice is altered in Rxfp4 knock-out mice. In this study, we used a new Rxfp4-Cre mouse model to investigate the mechanisms underlying these observations. Methods We generated transgenic Rxfp4-Cre mice and investigated central expression of Rxfp4 by RT-qPCR, RNAscope and intraparenchymal infusion of INSL5. Rxfp4-expressing cells were chemogenetically manipulated in global Cre-reporter mice using designer receptors exclusively activated by designer drugs (DREADDs) or after stereotactic injection of a Cre-dependent AAV-DIO-Dq-DREADD targeting a population located in the ventromedial hypothalamus (RXFP4VMH). Food intake and feeding motivation were assessed in the presence and absence of a DREADD agonist. Rxfp4-expressing cells in the hypothalamus were characterised by single-cell RNA-sequencing (scRNAseq) and the connectivity of RXFP4VMH cells was investigated using viral tracing. Results Rxfp4-Cre mice displayed Cre-reporter expression in the hypothalamus. Active expression of Rxfp4 in the adult mouse brain was confirmed by RT-qPCR and RNAscope. Functional receptor expression was supported by cyclic AMP-responses to INSL5 application in ex vivo brain slices and increased HFD and highly palatable liquid meal (HPM), but not chow, intake after intra-VMH INSL5 infusion. scRNAseq of hypothalamic RXFP4 neurons defined a cluster expressing VMH markers, alongside known appetite-modulating neuropeptide receptors (Mc4r, Cckar and Nmur2). Viral tracing demonstrated RXFP4VMH neural projections to nuclei implicated in hedonic feeding behaviour. Whole body chemogenetic inhibition (Di-DREADD) of Rxfp4-expressing cells, mimicking physiological INSL5-RXFP4 Gi-signalling, increased intake of the HFD and HPM, but not chow, whilst activation (Dq-DREADD), either at whole body level or specifically within the VMH, reduced HFD and HPM intake and motivation to work for the HPM. Conclusion These findings identify RXFP4VMH neurons as regulators of food intake and preference, and hypothalamic RXFP4 signalling as a target for feeding behaviour manipulation.

, was blocked with a NPY2R antagonist [5]. Reports that Insl5 knockout (Insl5 À/À ) mice do not display an observable feeding phenotype and that, in some studies, pharmacological administration of INSL5 (both native and PEGylated forms) failed to increase food intake in lean and obese mice [7,8], have shed doubt on whether INSL5 plays a physiologically relevant role in the control of food intake. However, Rxfp4 knockout (Rxfp4 À/À ) mice exhibit shorter meal durations, particularly when fed a high fat diet (HFD), and lack the normal preference for HFD over standard chow diet observed in wildtype mice [4]. We therefore consider RXFP4 to be a potential target receptor for the manipulation of feeding behaviour. RXFP4 is a G ai/o -coupled receptor identified in 2003 through its homology to relaxin/insulin-like peptide receptor 3 (RXFP3) [9,10]. Binding of INSL5 activates downstream signalling pathways including phosphorylation of ERK1/2, Akt, p38MAPK and S6RP and reduces cytosolic cAMP levels through inhibition of adenylyl cyclase [2]. In addition to the selective ligand INSL5, relaxin-3 can also activate RXFP4, stimulating comparable signalling pathways [2,11]. Rxfp4 mRNA expression has been reported in the colon, kidney, heart, liver, testes and ovary of both mice and humans [1,10,11]. To identify and manipulate Rxfp4-expressing cells, we developed a new transgenic mouse model in which Cre-recombinase expression is driven by the Rxfp4 promoter (Rxfp4-Cre). Using this model, we found clear Credependent reporter expression within the central nervous system (CNS), including the ventromedial hypothalamus (VMH), a finding we confirmed via RT-qPCR and RNAScope. This contrasts with our previous study, in which we failed to detect Rxfp4 expression in the central nervous system by RT-qPCR, and in which intracerebroventricular administration of INSL5 failed to increase food intake significantly [4]. Revisiting the potential central role of RXFP4 in food intake regulation, we observed that infusion of INSL5 directly into the VMH induced a significant orexigenic response when mice were offered a HFD or a highly palatable liquid Ensure test meal (HPM), but not when offered a standard chow diet. We therefore used the Rxfp4-Cre mouse model to explore the role of Rxfp4-expressing cells in modulating food intake and preference.

RESULTS
To investigate a possible role of Rxfp4-expressing cells in feeding control, we generated a new bacterial artificial chromosome (BAC) transgenic mouse (Rxfp4-Cre) model in which Cre-recombinase is expressed under the control of the Rxfp4-promoter ( Figure 1A). By crossing Rxfp4-Cre mice with fluorescent protein reporter mice (e.g. Rosa26 fxSTOPfx-EYFP (RXFP4 EYFP )) ( Figure 1B) we observed Rxfp4 dependent expression in the colon ( Figure 1C), consistent with previous reports, but also detected reporter expression in the CNS (Suppl. Fig. 1) by GFP immunohistochemistry. Within the CNS we observed reporter expression in cells of the VMH, the accessory olfactory bulb, septofimbrial nucleus, retrosplenial cortex, and the mammillary body medial and lateral parts, as well as in a lower density of cells in the substantia innominata, cortical and central amygdala, periaqueductal grey, the lateral dorsal tegmental nucleus, nucleus of the tractus solitarius, area postrema and spinal trigeminal tract of the hindbrain (Suppl. Fig. 1). GFP-positive nerve fibres were also detected in the bed nucleus of the stria terminalis, anterodorsal thalamic nucleus, mammillothalamic tract, suprachiasmatic nucleus, paraventricular hypothalamus and ventral tegmental nucleus of Gudden (Suppl. Fig. 1). Closer examination of the VMH using RXFP4 GCaMP3 mice revealed transgene expression in the central and ventrolateral parts (VMHc and VMHvl, respectively) extending into the adjacent tuberal nucleus (TN) ( Figure 1D and E). The VMH is a central hub for the regulation of energy balance and integration of diverse nutritionally regulated hormonal and synaptic inputs, and the adjacent TN has also been implicated in feeding behaviour [12,13]. In the VMH, co-staining with the neuronal marker NeuN demonstrated GFP expression in mature Rxfp4 positive neurons ( Figure 1F). Active transcription of Rxfp4 mRNA in the adult mouse VMH was confirmed by RNAscope, at the low abundance expected for G protein-coupled receptors (Figure 1GeI), and RT-qPCR, with levels comparable to the distal colon, a known site of Rxfp4 expression ( Figure 1J). To test the functional activity of the RXFP4-INSL5 pathway in the Rxfp4expressing cells in the VMH (RXFP4 VMH ), we performed real time monitoring of cAMP levels in RXFP4 VMH cells using ex vivo CAMPER [14] imaging in hypothalamic slices from Rxfp4-Cre x CAMPER (RXFP4 CAMPER ) mice. We found that 15 out of 48 YFP-positive cells showed a clear reduction in the DFRET ratio (YFP/CFP), reflecting a reduction in cAMP levels, in response to INSL5 ( Figure 1K and L), whilst the remaining cells displayed no obvious change in FRET signal in response to INSL5 addition, within the limits of the assay sensitivity. To test the functional significance of the RXFP4 VMH cells in vivo, we treated wildtype ad libitum (AL) fed lean mice with an intra-VMH infusion of INSL5 after a 2 h fast. During the light phase, INSL5 had no effect on standard (std) chow intake ( Figure 1M) but in mice habituated to the appearance of a HFD or HPM for 1 h, treatment with INSL5 significantly increased HFD and HPM intake compared to the vehicle control treatment ( Figure 1N and O). As intra-VMH INSL5 significantly modulated feeding behaviour, we further characterised the Rxfp4-expressing cells in this region. Initially, we generated a single cell resolution transcriptomic profile of Rxfp4expressing cells in the hypothalamus. Fluorescent cell populations from the hypothalami of twelve RXFP4 EYFP mice were purified by fluorescence-activated cell sorting (FACS) and their transcriptomes analysed by single-cell RNA-sequencing (scRNAseq). After data normalisation, 350 cells passed quality control filtering. Graph-based clustering analysis revealed that the filtered cells separated into five populations ( Figure 2A). Cluster identities were assigned based on the expression patterns of canonical cell-type markers, identifying microglia (Tmem119, Siglech, P2ry12), neuronal cells (Snap25, Tubb3, Elavl2) and astrocytes (Aqp4, Cnn2, Fgfr3), together with smaller clusters of macrophages (Mrc1, Mgl2) and ependymocytes (Ccdc153, Hdc) ( Figure 2B). As IHC had suggested neuronal Rxfp4 expression ( Figure 1F) and hypothalamic neurons are known to modulate feeding behaviour, we analysed the neuronal cluster in more detail, identifying seven subclusters ( Figure 2C). Rxfp4-positive neurons expressed markers for both GABAergic (Slc32a1) and glutamatergic (Slc17a6) cells ( Figure 2D). Cluster 1 was enriched in markers previously associated with an estrogen receptor (Esr1)-positive VMHvl neuronal population [15e17], including preprotachykinin-1 (Tac-1), oxytocin receptor (Oxtr), cholecystokinin receptor A (Cckar), melanocortin 4 receptor (Mc4r) and neuromedin U receptor 2 (Nmur2) ( Figure 2D), suggesting crosstalk with known food regulatory networks. Receptors for other established feeding-neuromodulators, like glucagon-like peptide-1 receptor (Glp1r) and cholecystokinin receptor B (Cckbr), were preferentially expressed in cluster 6 ( Figure 2D). We subsequently aimed to establish the neuronal circuitry around RXFP4 VMH cells. Anterograde projections were mapped by stereotactically injecting Cre-dependent rAAV8-ChR2-mCherry into the VMH of RXFP4 GCaMP3 mice [18] (Figure 3A and B). Axonal transport of the ChR2-mCherry fusion protein revealed RXFP4 VMH projections to multiple regions including the bed nucleus of the stria terminalis (BNST), preoptic area (POA), anteroventral periventricular nucleus (AVPV), Original Article two-tailed t test). Animals adapted to the appearance of test meal over the course of two weeks prior to surgery and recovery, testing conducted between 12:00 and 15:00.  Figure 3C, Suppl. Fig. 3). Retrograde projections were assessed after AAV2-TVAeGFP-oG injection into the VMH of Rxfp4-Cre mice followed by Rab-DG-EnvA-mCherry [19] injection 21 days later ( Figure 3D and E). The retrograde monosynaptic transport of Rab-mCherry labelled inputs from several nuclei established in feeding regulation, including the ARC, DMH and PVH ( Figure 3F). The neuronal circuitry surrounding RXFP4 VMH cells is summarised in Figure 3G. Due to the previously reported altered feeding patterns and macronutrient preferences of Rxfp4 À/À mice [4], we further explored the physiology of the RXFP4 VMH population by ablating Rxfp4-expressing cells using rAAV-DTA injected into the VMH of RXFP4 GCaMP3 mice ( Figure 4A). Post-mortem histological analysis revealed a >90% reduction in RXFP4 VMH cells (identified by GFP immunohistochemistry) across the VMH ( Figure 4B). In the 8 weeks following the surgery, the RXFP4 VMHKO mice gained less body weight compared to the control rAAV-mCherry injected mice ( Figure 4C), when fed with a choice of standard chow and HFD in parallel. Five weeks post-surgery the mice were studied in metabolic cages. RXFP4 VMHKO mice had a lower respiratory exchange ratio (RER, Figure 4D), whilst energy expenditure (in kcal/h [ Figure 4E] and using body weight as a covariate [Suppl. Fig. 4d]) and ambulatory activity ( Figure 4F) were unaffected by ablation. Average 24 h food intake was significantly reduced, measured by ANCOVA using body weight as a covariate (see Suppl. Fig. 4f), a consequence of reduced HFD intake ( Figure 4G). Meal duration was unaffected by treatment, however the interval between meals ( Figure 4I) was significantly increased. These data demonstrate the importance of this neuronal population in governing long-term feeding behaviour. There was no observable difference between male and female animals (see suppl. Fig. 4aei). Given that RXFP4 VHMKO mice ate less HFD, yet central Insl5 infusion, which we would predict to inhibit Rxfp4-expressing neurons, increased HFD intake, we investigated the effects of acute activation or inhibition of Rxfp4-expressing cells. Initially we used a whole-body hM4Di DREADD Cre-reporter (RXFP4 wbÀDi ) to mimic the established RXFP4signalling via pertussis-toxin sensitive Gi pathways [2] ( Figure 5A). During the light phase, activation of Di in Rxfp4-expressing cells using the DREADD agonist clozapine-N-oxide (CNO) had no measurable effect on food intake in AL fed male and female mice ( Figure 5B, as indicated). However, when animals were habituated to the appearance of a HFD or HPM for 1 h during the light phase, CNO application resulted in increased food intake ( Figure 5C and D). These results were consistent with the response to infusion of INSL5 into the VMH. To investigate this further, we gave male mice housed in metabolic cages the choice between standard chow and HFD. CNO injection during the light phase significantly increased HFD but not chow intake in RXFP4 wbÀDi mice ( Figure 5E). This effect was transient ( Figure 5F), consistent with the pharmacokinetics of CNO [20], and occurred without any significant differences in ambulatory activity or energy expenditure compared to the vehicle cross-over control (Suppl. Fig. 5a and b). We next investigated the effects of whole-body hM3Dq Cre-reporter activation in Rxfp4-expressing cells (RXFP4 wbÀDq ) ( Figure 6A). After a 2 h brief fast, activation of Dq in Rxfp4-expressing cells at the onset of the dark phase had no measurable effect on the food intake of male and female mice offered only standard chow ( Figure 6B, as indicated). However, when RXFP4 wbÀDq animals were habituated to the appearance of HFD or a HPM for 1 h at the onset of the dark phase, activation of Dq-expressing cells with CNO resulted in a marked reduction in HFD and HPM intake ( Figure 6C and D). These results were consolidated in AL fed male animals tested in metabolic cages with parallel access to standard chow and HFD. RXFP4 wbÀDq activation in AL fed animals at the onset of the dark phase had no effect on standard chow intake, but significantly and transiently reduced HFD consumption ( Figure 6E and F). RXFP4 wbÀDq activation also attenuated the increase in energy expenditure associated with the onset of the dark phase, however, there was no effect on ambulatory activity (Suppl. Fig. 5c and d).
To probe whether Rxfp4-expressing cells play a role in the motivational aspects of feeding, we calorically restricted male RXFP4 wbÀDq animals to 95% body weight and placed them in operant chambers. Mice were tested with a fixed ratio (FR) schedule, requiring 5 nose pokes to release a food reward (liquid Ensure, HPM), or a progressive ratio (PR) schedule requiring increasing number of nose pokes for each subsequently earned reward (in this case, þ4, i.e. 1, 5, 9, 13, etc). RXFP4 wbÀDq mice treated with CNO completed fewer attempts under FR to earn individual Ensure rewards ( Figure 6G). Under a PR schedule, they exhibited a reduced breakpoint, i.e. CNO-treated mice stopped working for the HPM-reward at lower ratios than when receiving vehicle treatment ( Figure 6H). In an effort related choice (ERC) paradigm, where animals had the choice of working for a HPM (FR8, liquid Ensure) or consuming freely available standard chow, CNO treatment reduced HPM consumption ( Figure 6I). However, animals consumed similar amounts of standard chow and displayed otherwise normal behaviour (supplementary video 1), suggesting that activation of Rxfp4-expressing cells reduced motivation for the HPM rather than inducing generalised malaise.
(Supplementary video related to this article can be found at https://doi. org/10.1016/j.molmet.2022.101604) To assess whether the Rxfp4-expressing population in the VMH is involved in the feeding phenotype observed in RXFP4 wbÀDq mice, the effect of acute chemogenetic manipulation of RXFP4 VMH neuron activity on food intake was investigated. Male and female RXFP4 GCaMP3 mice received bilateral VMH injections of Cre-dependent hM3Dq-expressing rAAVs (rAAV-hSyn-DIO-hM3D(G)q-mCherry) designed to preferentially target neurons, to produce RXFP4 VMHDq mice ( Figure 7A). Targeting efficiency was subsequently determined by immunohistochemistry ( Figure 7B), or in a subset of mice by fluorescent microscopy in live slices (see below). All mice demonstrated robust transduction that was limited to the target region. To confirm the functional activation of these RXFP4 VMHDq neurons, we generated ex vivo brain slices from a subset of RXFP4 VMHDq mice ( Figure 7C). Calcium imaging demonstrated that ex vivo treatment with CNO activated mCherry/GCaMP3 positive RXFP4 VMHDq , stimulating an increase in the frequency of calcium oscillations and an increase in GCaMP3 fluorescence in Rxfp4expressing somas ( Figure 7C and D). The effect of chemogenetic activation of this cell population on food intake in vivo was studied in a crossover design. In chow fed male and female mice, and in line with RXFP4 wbÀDq animals, CNO treatment of RXFP4 VMHDq mice had no effect on standard chow intake at the onset of the dark phase ( Figure 7E, as indicated). When animals were habituated to the appearance of a HFD or HPM, CNO resulted in a significant reduction in food intake at the onset of the dark phase ( Figure 7F and G). When offered HFD and chow diet in parallel in metabolic cages, CNO significantly reduced intake of the HFD whilst intake of standard chow was not altered, resulting in an overall reduced caloric intake ( Figure 7H) in male and female mice (Suppl. Fig. 5e and f). As seen with RXFP4 wbÀDq animals, this was a transient effect was no longer apparent in the 2nd hour post CNO administration ( Figure 7I). CNO had no effect on energy expenditure or ambulatory activity in these RXFP4 VMHDq animals (Suppl. Fig. 5g and h).

DISCUSSION
The Rxfp4-Cre mouse model generated in this study has enabled the identification of central Rxpf4-expressing cells, and aided the transcriptomic, functional and anatomical characterisation of a hypothalamic Rxfp4-expressing neuronal population. We show that Rxfp4 is expressed in key feeding centres of the brain, including neurons of the VMH. Ex vivo CAMPER imaging revealed that INSL5 reduces cAMP levels in Rxfp4-expressing VMH neurons, and central administration of INSL5 into the VMH increased HFD and HPM intake in lean AL fed mice Original Article in vivo. Mimicking the native physiology of INSL5-RXFP4 via Di recapitulated these results, whereas chemogenetic cell activation via Dq suppressed HFD and HPM intake. Selective chemogenetic activation of Rxfp4-expressing cells in the VMH alone (targeted via Cre-dependent rAAV-Dq) also suppressed HFD and HPM intake, reflecting their position in brain circuits implicated in homeostatic and hedonic regulation of food intake [21]. Ablation of this population further demonstrated the importance of these neurones in governing long-term feeding behaviour. These data identify hypothalamic RXFP4 signalling as a key regulator of food intake and preference. Rxfp4 expression has been difficult to localise due to low mRNA expression levels and the lack of suitable verified antibodies. Rxfp4  Original Article 10 expression was previously reported in the colon [4] and in enteroendocrine tumor cell lines [22,23]. In contrast to our previous report [4], we detected Rxfp4 mRNA in the hypothalamus ( Figure 1G,H,J) and found substantial Rxfp4-dependent reporter expression in multiple brain regions, with distinct Rxfp4-expressing cell populations in the accessory olfactory bulb, RSC, VMHvl, and mammillary body (Suppl. Fig. 1). While it could be argued that this reflects lineage tracing from Rxfp4-positive precursor cells, the detection of Rxfp4 mRNA by RT-qPCR and RNAscope, the activation of Cre-dependent rAAV-constructs when stereotactically injected into the adult VMH, the cAMP responsiveness of Rxfp4-positive cells to locally-perfused INSL5 in slice preparations, and the effect of stereotactically injected INSL5 on feeding behaviour, confirm active Rxfp4-promoter activity and consolidate active Rxfp4 expression and function in the adult mouse brain (Figs. 1G, H and 7B). To characterise the transcriptomic profile of hypothalamic Rxfp4expressing cells we performed scRNAseq. Although some cells will have been lost and some genes may have exhibited altered expression during the cell dissociation and sorting process, the results allowed us to cluster Rxfp4-expressing cells into several subpopulations each characterised by a profile of cell-specific marker genes (Figure 2A and B). Rxfp4 was identified in microglia and ependymocytes that potentially constitute the blood brain barrier, in addition to astrocytes, suggesting that INSL5 may additionally exert effects on non-neuronal cells [24]. Clustering of the Rxfp4-expressing neuronal population indicated a predominance of glutamatergic neurons with few GABAergic clusters ( Figure 2D). Cluster 6 is of note given the expression of Glp1r, Cckbr, Sstr1 and Sstr2, suggesting an overlap with other known appetite-modulating gut peptide receptors. Cocaine-and amphetamine-regulated transcript (Cartpt), expressed in cluster 3, and cannabinoid receptor 1 (Cnr1), expressed in clusters 1, 4 and 5, (Figure 2D, Suppl. Fig. 3) have also been implicated in energy homeostasis [25,26]. Cluster 1 displayed markers of a VMHvl Esr1 population (Esr1, Pgr, Tac1, Cckar, Rprm and Oxtr) previously associated with food intake and energy expenditure [12,15,16]. However, chemogenetic activation of RXFP4 VMHDq neurons did not result in increased energy expenditure or ambulatory activity, contrasting with the previously described Esr1 VMHDq phenotype [16]. The role of RXFP4 VMH cells in feeding regulation is further suggested by the co-expression of the neuropeptide receptors Mc4r and Nmur2. MC4R activation has been linked to suppressed food intake through regulation of Bdnf expression in the VMH [27] e Bdnf was co-expressed in cluster 1 neurons in our dataset ( Figure 2D). Acute administration of NMUR2 agonists have been shown to decrease feeding, with one agonist being somewhat selective to HFD intake regulation [28], mirroring RXFP4 VMHDq cell activation (Figure 7). Mapping of the retrograde inputs to and anterograde projections from the Rxfp4-expressing cells in the VMH revealed a distinct neural circuitry surrounding this hypothalamic population. While we aimed to target the VMH specifically during stereotactic injections, disparities between current mouse brain atlases make it difficult to distinguish the VMH from the adjacent TN, which has also been implicated in feeding behaviour [13,29]. It is therefore possible that some Rxfp4-expressing cells in the TN were transfected by the viral vectors and acted as starter cells in these tracing experiments. Monosynaptic inputs to RXFP4 VMH cells were labelled predominantly from brain regions involved in homeostatic regulation of food intake such as the ARC, PVH and LHA ( Figure 3F). Within the ARC, two populations of neurons have been intensely studied with regards to feeding regulation: proopiomelanocortin (POMC)/cocaine-and amphetamine-regulated transcript (CART)-expressing neurons inhibit food intake while Agouti-Related Peptide (AgRP)/neuropeptide Y (NPY)-expressing neurons stimulate food intake. These neurons integrate nutritional and hormonal signals from the periphery and send projections to multiple brain regions including the VMH [30]. Although the VMH is not thought to be the main target of arcuate POMC projections [31], the co-expression of Mc4r in the RXFP4 VMH neurons suggests that POMC neurons may be part of the ARC innervation of RXFP4 VMH neurons. RXFP4 VMH cells predominantly project onto regions associated with reward and motivation-related behaviours ( Figure 3C, Suppl. Fig. 3) such as the BNST, POA, CeA, paraventricular thalamic nucleus (PVT) and ventral tegmental area (VTA) [32e34], potentially underlying our finding that chemogenetic activation of Rxfp4-expressing cells in RXFP4 wbÀDq mice reduced an animal's drive to seek out and work for a highly palatable food reward (Figure 6GeI). Taken together, these data suggest that Rxfp4-expressing cells may influence motivation and reward-related behaviour via regulation of central reward signalling pathways. RXFP4 VMH cells also send projections to the PAG and PBN, two integration sites responsible for relaying sensory information between the forebrain and hindbrain and coordinating behaviour in response to various stimuli including metabolic, gustatory and nociceptive inputs [35e37]. This RXFP4 VMH projection map broadly recapitulates previously identified efferent projections from the VMH [38e40] and SST-expressing cells in the TN [13,38], as well as affarent projections to the VMH [41,42]. Interestingly, all retrograde-labelled input regions also received projections from RXFP4 VMH cells suggesting a high level of bidirectional connectivity within the RXFP4 VMH signalling network ( Figure 3G). Similar bidirectional connectivity has been shown for an Esr1þve VMHvl population [38]. The RXFP4 VMH neural network established in the present study suggests these cells may integrate metabolic and nutritional cues either directly or via other hypothalamic regions and regulate the reward system to influence ingestive behaviours. The low number of RXFP4 VMH input regions compared to projection regions suggests these cells may comprise an early node in this network. Inhibition of Rxfp4-expressing cells via direct intra-VMH infusion of INSL5 or whole body Di acutely increased intake of both a HFD and a HPM in the home cage and when animals were offered a choice of standard laboratory chow and HFD in metabolic cages, without altering energy expenditure and activity in both male and female mice (Figs. 1N, O, 5, Suppl. Fig. 5a and b). This is consistent with our previous demonstration that INSL5 administration increased food intake [4,5]. Activation of Di receptors should, at least in part, mimic the Ga i/o -coupling of RXFP4 [2]. By contrast, activation of Rxfp4expressing cells with Dq produced a robust suppression of HFD and HPM intake ( Figure 6). Targeting the RXFP4 VMH population with an rAAV-Dq reporter recapitulated the findings with the whole body-Dq approach (Figure 7). Whilst any of the identified Rxfp4-expression sites could participate in the feeding phenotype of the global DREADD reporter mice, the VMH-specific rAAV-Dq reporter phenotype indicates that this hypothalamic population either underlies or at least contributes to the observed anorexigenic effects in the RXFP4 wbÀDq model. Our data suggest that activation of Rxfp4-expressing cells in the VMH suppresses the consumption and drive to work for calorie dense HFD and HPM. The importance of the VMH in the regulation of feeding and metabolism has been disputed (reviewed in [43]). Initial studies suggested the VMH might be a "satiety centre", as VMH-lesioned rats, particularly females, over-consumed a HFD when fed AL [44], despite being seemingly less willing to work for food on a fixed ratio leverpressing paradigm [45]. The observed hyperphagia has subsequently been linked to additional damage to adjacent hypothalamic structures (see King 2006 [43] for discussion), while the reduced motivation was not observed when rats were trained pre-operatively, suggesting that the VMH lesion may have altered "trainability" rather than feeding motivation [43,46]. Lesioning studies, whilst informative, lack cellular precision and damage neural connections to other feeding centres within the brain. More recent work employing immunohistochemistry, RNA sequencing, chemogenetics and neuronal projection mapping, has demonstrated that the VMH consists of anatomical subdivisions made up of distinct cell populations [15,47,48]. Functional studies suggest neurons in the central and dorsomedial VMH regulate feeding, energy expenditure and glucose homeostasis [29,49], while the VMHvl is more frequently implicated in the control of social and sexual behaviours [48,50,51]. However, several studies have demonstrated the involvement of VMHvl neurons in energy expenditure and feeding behaviour [12,52]. Increases in physical activity have been observed following chemogenetic activation of NK2 homeobox transcription factor 1 (Nkx2-1)-expressing neurons in the VMHvl of female rats, while knockout of Nkx2-1 in the VMHvl leads to decreased physical activity and thermogenesis [53]. Furthermore, chemogenetic activation of Esr1-expressing VMHvl neurons was found to stimulate physical activity and thermogenesis in both sexes [16]. Esr1 signalling in the VMHvl was previously demonstrated to influence food intake, energy expenditure and glucose tolerance and VMHvlrestricted knockdown of Esr1 resulted in increased food intake, decreased physical activity and thermogenesis, reduced glucose tolerance and obesity in female rats [12]. Activation of RXFP4 VMHDq neurons reduced HFD and HPM intake but had no effect on chow intake, energy expenditure or ambulatory activity in both male and female mice (Figure 7, Suppl. Fig. 5eeh). This, alongside the transcriptomic profile and neural circuitry of RXFP4 VMH neurons, suggests that these neurons comprise a distinct VMH population modulating food intake and preference based on the rewarding aspects of food rather than the homeostatic food intake or energy expenditure responses observed during chemogenetic manipulation of other VMH populations. We recognise several limitations to this study. First, the exact classification of Rxfp4-expressing brain regions is difficult given the disparities between current brain atlases. Whilst we aimed to target Rxfp4-expressing cells in the VMH, the inexact nature of stereotactic injections may have resulted in additional targeting of neurons in adjacent LHA and TN regions, which may have contributed to the food intake phenotype of RXFP4 VMHDq mice. The fact that only a subset of RXFP4 CAMPER neurons responded to INSL5 in acute slices likely reflects technical limitations of this preparation and the difficulty in recording reductions in cAMP downstream of Gi. The similarities in feeding outcomes of intraparenchymal INSL5 in wildtype mice and CNO application in RXFP4 VMHDq mice together with RNAscope confirmation of Rxfp4-Cre labelled VMH neurons makes the alternative explanation of substantial additional off target Cre-expression unlikely, but we cannot fully rule this out. It is also uncertain whether INSL5 is the endogenous ligand acting on Rxfp4-expressing cells in the brain. We have previously been unable to identify Insl5-expressing cells in the mouse brain [5] and there is no evidence that INSL5 can cross the blood brain barrier. In addition, relaxin-3 also activates RXFP4 [2] and is expressed in the mouse brain [54], hence it is possible that relaxin-3 is involved in central RXFP4 action. Intracerebroventricular injection of relaxin-3 stimulates food intake in rats [55e57], but given that Rxfp4 is a pseudogene in rats, this orexigenic action is likely due to activation of RXFP3 [54,58]. Relaxin-3 has not been demonstrated to be orexigenic in mice, but both relaxin-3 and Rxfp3 knockout mice demonstrate dark phase hypoactivity (assessed via voluntary home-cage wheel running) [59,60], potentially reflecting reduced motivational drive. Whilst an RXFP3-selective antagonist reduced motivated food seeking behaviour and palatable food intake in wild-type but not Rxfp3 knockout mice, icv injection of a synthetic RXFP3-agonist, which cross reacts with RXFP4, failed to stimulate chow or palatable food intake in wild type mice [61e63]. However, our own previous attempts to increase food intake by injecting INSL5 icv also failed to produce significant results [4] contrasting with the results after intraparenchymal injection presented here. The source of a central RXFP4 ligand in the mouse CNS remains obscure and should be addressed in future research. The viral tracing techniques used to decipher the RXFP4 VMH neuronal network also have some limitations. In the retrograde tracing experiments, it was difficult to detect rAAV-GFP immunoreactive cells, making it hard to confirm the exact starter cells in this experiment. Furthermore, while the ChR2-mCherry construct is preferentially targeted to axon terminals, it is possible that some of the mCherrylabelled fibres in adjacent regions, such as the ARC, are in fact dendrites [64] which may underlie our inability to detect retrograde-only labelled regions. Nevertheless, we have been able to identify distinct regions that project onto and receive projections from Rxfp4expressing cells in the VMH that connect these cells to known feedingrelated neural networks. The data generated following ablation of RXFP4 VMH neurons are somewhat unexpected given native INSL5-RXFP4 physiology, the INSL5 effects on cAMP in RXFP4 CAMPER cells in ex vivo brain slice imaging and the DREADD-based feeding phenotypes, which are all consistent with the model that RXFP4dependent inhibition of RXFP4 VMH neurons results in increased HFD and HPF intake. The reduced HFD and HPM intake after ablation of RXFP4 VMH neurons and the reduced weight gain of HFD fed RXFP4 VMHKO animals likely reflects loss of the plethora of neurotransmitters and receptors of these neurons, exemplified by the scRNA-seq data, and the chronic disruption to the neural circuits surrounding these neurons, possibly also affecting other behaviours, given that the VMHvl also modulates social behaviour and aggression [15] even though we did not observe any obvious behavioural abnormalities. Finally, most of the functional data included in this manuscript reflect ablation, activation and inhibition of Rxfp4-expressing neurons rather than of the receptor itself. Nevertheless, the phenotypes observed highlight that these neurons are involved in regulating food intake and that manipulation of these neurons, via RXFP4, could affect long-term feeding behaviour.
In summary, we have characterised a previously unrecognised population of ventromedial hypothalamic cells that express Rxfp4 in mice, identified projections in homeostatic and hedonic feeding centres in the CNS, and demonstrated that acute manipulation of these cells modulates HFD/HPM intake without affecting chow intake or energy expenditure. Together, these findings suggest Rxfp4-expressing neurons in the VMH are key regulators of food preference and represent a target for the modulation of feeding behaviour.

Animals
All experiments were performed under the UK Home Office project licences 70/7824 and PE50F6065 in accordance with the UK Animals (Scientific Procedures) Act, 1986 and approved by the University of Cambridge Animal Welfare and Ethical Review Body. All mice were from a C57BL/6 background and were group-housed and maintained in individual ventilated cages with standard bedding and enrichment in a temperature and humidity controlled room on a 12 h light:dark cycle (lights on 7:00) with ad libitum (AL) access to food (Scientific Animal Food Engineering) and water unless otherwise stated. Groups were randomised by body weight and the researcher was blinded to treatment.

Mouse models
To express Cre recombinase under the control of the Rxfp4 promoter, we replaced the sequence between the start codon and the stop codon in the single coding exon of Rxfp4 in the murine-based BAC RP24-72I4 (Children's Hospital Oakland Research Institute) with iCre [65] sequence using Red/ET recombination technology (GeneBridges) ( Figure 1A). The resulting BAC was used to create BAC-transgenic mice e of four initial founders, two passed the transgene to their offspring; both resulting lines showed similar Cre-reporter expression and one line was used throughout this manuscript. Several Crereporter transgenes, in which expression is only activated after removal of a fxSTOPfx cassette, were used, resulting in expression of EYFP [66], hM3Dq, hM4Di [67], or GCaMP3 [68], respectively.

Intra-VMH infusion
Mice were stereotactically implanted with a guide cannula (Plastics One) positioned 1 mm above the VMH (A/P: À1.7 mm, D/V: À4.5 mm, M/L: AE0.75 mm from bregma). Bevelled stainless steel injectors (33 gauge, Plastics One) extending 1 mm from the tip of the guide were used for infusions in free-moving mice. Animals was given 1 week to recover. Mice were given AL access to standard (std) chow.

Food intake
Food intake studies were performed in a cross-over design, on agematched groups, a minimum of 72 h apart. Mice were administered vehicle or INSL5 (5 mg/animal, Phoenix Biotech, 035-40 A) at 12:00-15:00 following a 2 h fast. Food was weighed 1 h post-injection. For experiments assessing the effect of global RXFP4 Di and Dq activation, animals were singly housed prior to the experiment. Mice were administered 1 mg/kg CNO (Sigma, C0832) ip or an equivalent volume of vehicle containing a matched concentration of DMSO (Sigma, D2650). For light phase activation, animals were injected with vehicle or CNO at 11:00 (AE30 min) following a 2 h fast. Food was weighed at 1 h post-injection. For dark phase activation, animals were injected with vehicle or CNO at 19:00 at the onset of the dark phase following a 2 h fast. Food was weighed at 1 h post-injection. In trials with a high fat diet (HFD, Open Source Diets, D12451) and a highly palatable meal (HPM, liquid Ensure, Abbott Laboratories, 353-3601), mice were habituated to the appearance of the test meal for 1 h (5 days per week) for two weeks prior to testing.

Operant chambers
Twelve male RXFP4 wbÀDq mice (weighed 3 times weekly) were food restricted to maintain 95% body weight for two weeks prior to training and testing in standard mouse Bussey-Saksida touchscreen chambers (Campden Instruments Ltd, Loughborough, UK). Training and testing procedures were conducted as previously described [69]. Briefly, mice were trained to touch a screen for a reward (HPM, 20 mL) under a fixed ratio (FR) schedule for 2 weeks, progressing from FR1 to FR5 (training deemed successful when the animal earnt 30 rewards within 1 h), followed by testing. Mice then progressed to progressive ratio (PR, increment þ4 i.e. 1, 5, 9, 13, etc.), where the breakpoint was defined as the last reward earned before 5 min elapsed without operant response. Following testing of the breakpoint, mice progressed to the effort related choice schedule (ERC) -mice were trained on FR8, with the addition of standard chow to the operant arena. Once animals successfully earned 30 rewards within 60 min, testing was undertaken. The 60-minute training and testing sessions took place at the same time each day (between 10:00 and 13:00).

METABOLIC CAGES
Animals were acclimated to metabolic cages prior to study and data collection. Oxygen consumption and carbon dioxide production were determined using an indirect calorimetry system (Promethion, Sable Systems, Las Vegas, NV). The system consisted of 8 metabolic cages (similar to home cages), equipped with water bottles and food hoppers connected to load cells for continuous monitoring, in a temperature and humidity-controlled cabinet. The respiratory exchange ratio (RER) and energy expenditure (via the Weir equation) were calculated, whilst ambulatory activity was determined simultaneously. Raw data were processed using ExpeData (Sable Systems). Animals were exposed to standard chow and a HFD during metabolic assessment as indicated.
Imaging was performed using a Zeiss Axioskop2 FS Plus microscope. Slices were immobilised in a glass bottom chamber using a slice anchor (Warner instruments) and imaged using a 40Â water immersion objective. For CAMPER imaging, RXFP4 CAMPER neurons of the VMH were excited using a LED light source (Dual Optoled, Cairn Research) at 435 for 50 ms every 3 s and CFP and YFP emission were simultaneously recorded after passing through an optosplit (Cairn; 465e505 and 525e600 nm, respectively). For Ca 2þ imaging, RXFP4 GCaMP3 neurons were excited at 470 nm for 75 ms every 2 s and emission from 500 to 550 nm was captured. Images were captured using a charged-coupled camera (Evolve 512 Delta, Photometrics), acquisitions were saved and data logged using Metamorph software (Molecular Devices). Slices were continuously perfused with room temperature standard ACSF at a flow rate of 1 mL per minute. All drugs were added directly in standard ACSF. CNO (15 mM, C0832), IBMX (100 mM, I7018) and forskolin (10 mM, F6886) were purchased from Sigma. INSL5 (100 nM, 035-40 A) was purchased from Phoenix Biotech.
Regions of interest (ROIs) and an area determined for background fluorescence were outlined for each image and the mean pixel intensity extracted for each ROI. The mean intensity of the background was subtracted from each ROI. For CAMPER imaging, the fluorescence energy transfer ratio (DFRET) of the mTurquoise FRET donor (CFP) and Venus FRET acceptor (YFP) was calculated and is presented on a reverse axis. INSL5 responses were measured by calculating the percentage change in DFRET between the maximum IMBX response and the maximum IBMX þ INSL5 response, both compared to the maximum IBMX þ Forskolin response. Changes of more than 20% of the IBMX þ Forskolin responses were considered responses. For Ca 2þ imaging, fluorescence intensity of the GCaMP3 is presented as DF/F 0 with F 0 being the mean intensity for 5 min before stimulus and DF the fluorescence intensity F minus F 0 . Responses were observed on raw and DF/F 0 traces and quantified by calculating areas under the curve (AUCs) over 5 min before the stimulus for the ACSF condition and over 5 min after the stimulus for the treated conditions. Cells were considered activated by stimuli if the AUC increased by 25% after treatment compared to the AUC before treatment.

Immunohistochemistry
Colonic tissues were fixed in 4% paraformaldehyde (PFA; Alfa Aeser, J61899), dehydrated in 15% and 30% sucrose, and frozen in OCT embedding media (CellPath, Newtown, U.K.). Cryostat-cut sections (8e10 mm) were mounted directly onto poly-L-lysine-covered glass slides (VWR, Leuven, Belgium) by the Institute of Metabolic Science Histology Core. Slides were incubated for 1 h in blocking solution containing 10% goat or donkey serum. Slides were stained overnight at 4 C with primary antisera (table 1) in PBS/0.05% Triton X-100/10% serum. Slides were washed with blocking solution and incubated with appropriate secondary antisera (donkey or goat Alexa FluorÒ 488, 546, 555, 633 or 647; Invitrogen) diluted 1:400 and Hoechst diluted 1:1500 for 1 h. Control sections were stained with secondary antisera alone. Sections were mounted with Prolong Gold (Life Technologies) or hydromount (National Diagnostics, Atlanta, Georgia, USA) prior to confocal microscopy (Leica TCS SP8 X, Wetzlar, Germany). Quantification of cell number was performed using Leica Application Suite X and Image J. Brain tissue was collected from perfusion fixed mice as previously described [70]. Animals were anaesthetised with dolethal sodium pentobarbital solution at 125 mg/kg ip (in saline) and transcardially perfused with heparinised 0.1 M phosphate buffered saline (1xPBS) followed by 4% PFA in PBS. Brains were extracted and post-fixed in 4% PFA for 24 h at 4 C then transferred to 30% sucrose solution at 4 C for 48 h. Brains were sectioned coronally from olfactory bulb to the spinomedullary junction at 25 mm using a freezing microtome and stored in cryoprotectant medium. For diaminobenzidine (DAB) staining, antigen retrieval was used for all experiments prior to antibody incubation. Sections were incubated in 10 mM sodium citrate (Alfa Aesar, A12274) at 80 C for 10 min then washed in PBS. Sections were incubated in 0.5% hydrogen peroxide (Sigma, H1009) in milliQ water for 15 min then washed in PBS. Sections were blocked with 5% donkey serum in 0.3% Tween20 (VWR, 437082Q) in PBS (PBST) for 1 h at room temperature, then incubated with GFP antiserum (1:4000; ab5450, Abcam) in blocking solution overnight at 4 C. After washing in 0.1% PBST, sections were incubated with biotinylated anti-goat IgG (1:400; AP180B, Millipore) in 0.3% PBST for 1.5 h at room temperature, followed by a 1 h incubation with streptavidin conjugated to horseradish peroxidase (Vectastain Elite ABC kit, Vector Laboratories, PK-6100) and developed by DAB substrate (Abcam, ab94665). Sections were washed in PBS, prior to dehydration in ethanol and xylene, then mounting/coverslipping with Pertex mounting medium (Pioneer Research Chemicals Ltd., PRC/R/750). For immunofluorescent staining, slices were washed in PBS, prior to blocking for 1 h in 5% donkey serum then incubation with primary antisera (Table 1) in blocking solution overnight at 4 C. Slices were washed in and incubated with the appropriate secondary antisera (Alexa FluorÒ 488 or 555; Invitrogen) diluted 1:500 for 2 h at room temperature. Following washing, mounted sections were coverslipped on superfrost slides using Vectashield (Vector Laboratories, H-1400-10). Slides were imaged using an Axio Scan. Z1 slide scanner (Zeiss) and confocal microscope (Leica TCS SP8 X, Wetzlar, Germany) with a 20Â or 40Â objective as indicated. Images were analysed in Halo Image Analysis Software (Indica Labs) and ImageJ. 5.3. Tissue extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR) Animals were culled by cervical dislocation. Brain and colon tissues were extracted and frozen immediately on dry ice. For pancreatic islet extraction, pancreases were injected with collagenase V (0.5 mg/mL) and digested at 37 C. Islets were hand-picked into HBSS containing 0.1% wt/vol. fatty acid-free BSA. Each pancreas yielded approximately 150e300 islets which were pooled. RNA from whole tissue segments was extracted using TRIzol Reagent (Sigma, T9424) according to manufacturer's instructions. RT-qPCR was performed and analysed as described in [71] using probes Rxfp4: Mm00731536_s1 and b-actin: Mm02619580_g1 (Applied Biosystems). Mean, SEM and statistics were obtained for the DCT data and converted to relative expression levels (2DCT) for presentation only. 5.4. Dissociation, fluorescence-activated cell sorting (FACS) and single cell RNA sequencing For hypothalamic samples, single cell suspensions were prepared and pooled from twelve female Rxfp4-Cre x EYFP mice (biological replicates) as previously described [70]. Briefly, mice were sacrificed by  To prepare the sections, animals were anaesthetized with sodium pentobarbital solution (50 mg/kg in saline) and transcardially perfused with PBS followed by 4% PFA in PBS. Brains were extracted and post-fixed in 4% PFA for 24 h before being transferred to 25% sucrose for 24 h at 4 C. Brains were embedded in OCT compound, frozen in a Novec 7000 (Sigma)/dry ice slurry and stored at À80 C. 16 mm cryosections containing the hypothalamus were prepared on a Leica CM1950 cryostat (Wetzlar, Germany) at À12 C and stored at À20 C until required. Slides were thawed at room temperature for 10 min before baking at 60 C for 45 min. The sections were then post-fixed in pre-chilled 4% PFA for 15 min at 4 C, washed in 3 changes of PBS for 5 min each before dehydration through 50%, 70%, 100% and 100% ethanol for 5 min each. The slides were air-dried for 5 min before loading onto a Bond Rx instrument (Leica Biosystems). Slides were prepared using the frozen slide delay prior to pre-treatments using Epitope Retrieval