All animal procedures were approved by the Animal Welfare Ethical Review Body at the University of Bradford and performed according to the Animals (Scientific Procedures) Act, 1986. Animal experiments were licensed by the United Kingdom Home Office (project license number: P0D6AA50D).

Male Sprague Dawley rats were obtained from Envigo (Oxon, United Kingdom). Initially, rats were acclimatized for 7 days under 12 h light:12 h dark photoperiod in groups of four with ad libitum access to water and standard chow diet (2018 Teklad global 18% protein rodent diet, Envigo). After acclimatization, rats received a high-fat diet (45% fat by kcal, TD.06415, Envigo) or a nutrient-matched chow diet (10%fat by kcal, TD.06416, Envigo) depending on the experimental design (Figure 1). Rats were housed in standard rat cages (Type III rat caging, 2017 cm² floor area, Arrowmight, Hereford, United Kingdom) with soft woodchip bedding (EC06 chips, Datesand, Manchester, United Kingdom) and a red plastic tunnel and shredded paper (Sizzle Nest, Datesand) for enrichment. All environmental conditions were kept constant with a temperature of 21 ± 10°C, humidity of 50 ± 5% and average light intensity of 150 lux. Body weights of individual rats and food intake of the cage was measured regularly at the beginning of the light phase as described in detail in (Mclean et al., 2021). Health checks were carried out twice daily and no welfare-related issues were observed. Animals were assigned random numbers before killing to allow all analyses to be conducted blind to the groupings.

Experiment 1

Antagonism of CMKLR1 with 2-(α naphtyl) ethyltrimethylammonium iodide (α-NETA)

Intracerebroventricular (i.c.v) cannulation and i.c.v. injections were carried out as previously described (Helfer et al., 2013). Briefly, after 4 weeks of high-fat feeding, rats (10–11 weeks of age, weight range 267.8–332.3 g) received 2.5 ml of 0.9% saline for circulatory support and 5.0% w/v Carprieve (Norbrook^® Labaratories, Newry, Northern Ireland). Rats were then anesthetized by inhalation (2.5–3% isoflurane) and were placed in a Kopf stereotaxic frame (David Kopf, New York, NY). EMLA Cream 5% (Lidocaine 25mg, Prolocaine 25 mg/g; Astrazeneca, Cambridge, United Kingdom) was applied to the periosteum, before a permanent 22-gauge stainless steel guide cannula (Plastics One Inc., Roanoke, VA, United States) was implanted into the third ventricle of the hypothalamus on the mid-sagittal line, 0.5 mm posterior to the bregma line and 6.5 mm below the outer surface of the scull. Cannulae were fixed with dental cement (AgnTho’s, Lindingoe, Sweden) to three stainless steel screws (AgnTho’s) inserted into the cranium. After post-operative recovery, rats were single housed. To allow injections under conscious conditions, rats were handled daily pre- and post-surgery and habituated to the injections process. Compounds were injected in a volume of 5 µl over 1 min using a 28-gauge stainless steel injection (Plastics One) projecting 1 mm below the tip of the cannula. Correct i.c.v cannula placement was confirmed by a positive dipsogenic response to angiotensin II (100 ng in 5 µl of saline; Sigma-Alrich, Poole, United Kingdom). One week after angiotensin II injection, rats were randomly divided into control group and α-NETA injection group and received either 5 µl of saline or 9.2 ng α-NETA (Sigma Aldrich) in 5 µl of saline over 1 min, respectively. I.c.v. injections were performed in the early light phase at 3–4 h after lights on (10:00–11:00; ZT3 - 4) with food intake and body weight measured shortly before and 1, 2, 4 and 24 h post injection. One rat was removed from data analysis as the cannula was blocked during the injection process, thus final animal numbers were: control group n = 8 and α-NETA injection group n = 6. Rats were killed 24 h after the i.c.v. injection at ZT3–4 by decapitation following isoflurane inhalation (Figure 1A). Guide cannulae were carefully removed and brains were immediately dissected, and the hypothalamus was frozen on dry ice and stored at −80^оC.

Experiment 2

shRNA ARC-CMKLR1 knockdown

To block chemerin signaling in the ARC, a recombinant mammalian shRNA knockdown lentiviral vector was designed to inhibit the expression of rat CMKLR1 protein. CMKLR1 shRNA (sequence: 5′- GGA​AAG​CCA​TGT​GCA​AGA​TTA -3′) or control shRNA (sequence: 5′- CCT​AAG​GTT​AAG​TCG​CCC​TCG -3′) was inserted into the lentivirus vector pLV [shRNA]-EGFP:T2A:Puro-U6 under the control of the U6 promoter. A dual marker, EGFP fused to puromycin via T2A linker was inserted into the vector which is expressed under the control of the Human phosphoglycerate kinase 1 (hPGK1) promoter. The lentiviruses were named pLV-EGFP:T2A:Puro-U6>rCmklr1 [shRNA#4] and pLV-EGFP/Puro-U6>Scramble_shRNA. pLV-EGFP:T2A:Puro-U6>rCmklr1 [shRNA] and control vectors were constructed and packaged by VectorBuilder Inc. (Chicago, United States). The viral prep was concentrated (>10⁹ TU/mL) and ultra-purified (in HBSS) suitable for in vivo injection by VectorBuilder Inc. The efficiency to reduce expression of CMKLR1 was tested in vitro using rat primary cortical neurons cultured from E18 Sprague Dawley embryos before in vivo injections.

At 10–11 weeks of age (weight range 254.6–347.4 g), rats were randomly divided into groups and general surgical procedures were followed as described above. 3 μl of lentivirus vector expressing shRNA targeted against CMKLR1 (ARC-CMKLR1-KD, n = 8) or 3 μl of scrambled virus (control, n = 8) were bilaterally injected into the ARC using a micro syringe: 3.6 mm posterior to the bregma line, ±0.30 mm lateral to the mid-sagittal line, 9.8 mm below the outer surface of the skull. Stereotactic coordinates were calculated using the rat brain atlas of Paxinos and Watson (Paxinos and Watson 2005). After surgery, rats remained on high-fat diet for the first 17 days, but due to delivery problems of the high-fat diet, rats were then moved to nutrient-matched chow diet (10%fat by kcal, TD.06416, Envigo). Body weight and food intake were measured daily. Rats were killed 28 days after surgery, trunk blood was collected for serum and plasma and tissues (hypothalamus, white adipose tissue (WAT), brown adipose tissue (BAT)) were dissected and immediately frozen on dry ice and stored at −80^оC. Glucose levels were measured in non-fasted rats from trunk blood after killing using a blood glucose monitoring system (OneTouch^® Select Plus, LifeScan UK&Ireland, United Kingdom) (Figure 1B).

Experiment 3

Behavioral testing high-fat vs. chow

18 male Sprague Dawley rats were obtained from Envigo (Oxon, United Kingdom) at 4–5 weeks old. After acclimatization rats (weight range 171.9–220.9 g) were randomly assigned to two groups of eight rats. One group was provided with ad libitum access to water and high-fat diet and the other group received chow diet. Rats were housed in groups of 4–5/cage to avoid effects of isolation on behavior. Rats were tested in the Novel object recognition (NOR) task after 14 and 28 days of diet intervention. After 29 days, rats were killed by terminal anesthesia using isoflurane followed by decapitation at ZT3 (Figure 1C).

Hypothalamic total RNA was isolated using the PureLink RNA mini kit (Invitrogen, United Kingdom) or PARIS™ kit (Thermo Fisher Scientific, United Kingdom) with on-column DNAse treatment. The hypothalamus was cut to approximately 60–80 mg per sample, transferred to cell disruption buffer and homogenized with 1.5 mm homogenization beads (Triple-Pure™, Molecular Biology Grade Zirconium Beads) using the BeadBug microtube homogenizer. After homogenization, the supernatant was transferred to Eppendorf tubes and RNA extraction was carried out according to the manufacturer’s protocols. The RNA was quantified using a Nanodrop™ spectrophotometer (Nanodrop Lite, Thermo Fisher Scientific, United Kingdom). 500 ng of RNA was used to synthesize the cDNA using iScriptTM cDNA synthesis kit (Bio-Rad, United Kingdom). qPCR was performed on an Eco™ Real-Time PCR system (Illumina Inc., CA, United States) to amplify gene expression levels using following parameters: 40 cycles of 2 min at 50°C, 5 min at 95°C, 10 s at 95°C, and 30 s at 60°C. mRNA levels were normalized to D-Box and fold changes were calculated using the 2^−ΔΔCt method. The sequences of each primer set were designed using NCBI gene sequences (Supplementary Table S1). Standard curves were performed for each primer set to confirm qPCR efficiency.

For the WAT and BAT gene expression analysis, RNA was extracted using Qiagen lipid tissue kit (Qiagen, Valencia, CA, United States), treated with DNase I (Qiagen) and cDNA synthesized according to kit instructions (Promega, Southampton, United Kingdom). qPCR was carried out on 56.25 ng cDNA on a QuantStudio™ 7 Flex Real-Time PCR System (Applied Biosystems) with PowerTrack™ SYBR Green Master Mix (Applied Biosystems). Relative expression was calculated by 2^-ΔΔCT method, using β-Actin (Actb) as the housekeeping gene and expressed relative to the scrambled control virus group.

18 male rats were tested in the NOR task as previously described (Mclean et al., 2021). Rats were habituated to the test box for 20 min on 3 days. On the day of testing, rats were placed in the NOR chamber (52 cm wide × 40 cm high × 52 cm long) for 3 min. In the acquisition trial, rats were permitted to explore two identical objects for 3 min, they were then taken out of the NOR chamber and returned to their home cage for 1 min intertrial interval. During the intertrial interval, both objects were removed and the objects and the NOR chamber were cleaned with 70% EtOH. Rats were then exposed to a familiar object and a novel object for 3 min during the retention trial. The location of the novel object was randomly assigned using a Gellerman schedule. Acquisition and retention trials were recorded using a video recording system (Home Guard CCTV Home Security Kit). At the end of the experiments, the exploration time (sec) spent at each object in each trial was analyzed by an experimenter blind to the groups using two stopwatches. Rats that failed to explore both objects in one or both trials of the task were excluded from the analysis.

Group sizes were calculated based on gene expression analysis in surgical prepared rats from previous studies (Helfer et al., 2016). A sample size of seven to eight animals per group was found to provide the appropriate power to identify significant differences (α = 0.05). Body weight, food intake and energy intake data were analyzed by two-way repeated measures ANOVA (treatment group × time interaction). Area under the curve (AUC) of energy intake was calculated by setting the baseline as the mean of the first timepoint data for both groups and compared by two-tailed unpaired Student’s t-test. Gene expression was analyzed by two-tailed unpaired Student’s t-test or Mann-Whitney test as appropriate (MW: WAT Adipoq, Dio2, Cebpb, Cmklr1, Rarres2. BAT: Rarres2). The NOR data are expressed as mean exploration time ±SD. Data passed normality (Shapiro-Wilk) and student’s two-tailed paired t-test was performed to compare time spent exploring the familiar versus the novel object. Differences were considered statistically significant if p < 0.05. Data are presented as mean ± SD, n refers to the number of animals.

To evaluate whether the small molecule CMKLR1 antagonist α-NETA (Graham et al., 2014) has the potential to regulate body weight and food intake in vivo, 5 µM α-NETA was administered i.c.v. into high-fat fed Sprague Dawley rats as a bolus injection. Acute injection of α-NETA caused a marked reduction in body weight (t₍₁₂₎ = 2.526, p = 0.0266; Figure 2A). The reduction in body weight was accompanied by a decrease in food intake after 24hrs, however this was not significant (Figure 2B). Next, we tested whether α-NETA selectively targets Cmklr1 and Gpr1 in the hypothalamus. Interestingly, we found that both Cmklr1 and Gpr1 were significantly upregulated in α-NETA-injected rats compared to saline-injected rats (t₍₁₁₎ = 2.230, p = 0.0476, t₍₁₁₎ = 2.813, p = 0.0169, respectively; Figure 2C). Given these results, we decided to block CMKLR1 using a genetic manipulation approach.

Previously, we showed that i.c.v. administration of chemerin into rats altered food intake and body weight (Helfer et al., 2016). To demonstrate that this is achieved via the chemerin-CMKLR1 axis, we manipulated chemerin signaling in the arcuate nucleus of the hypothalamus. We designed a short-hairpin-RNA (shRNA) lentivirus construct targeting expression of CMKLR1 and tested for efficiency to reduce expression of this receptor RNA in vitro in rat primary cortical neurons. The shRNA construct, or a control construct was injected bilaterally into the arcuate nucleus of adult Sprague Dawley rats on high-fat diet. After surgery, these rats were maintained on high-fat diet for 2 weeks and then switched to chow diet for a further 2 weeks. Weight gain was inhibited in ARC-CMKLR1-KD rats (time x treatment, F_(22,308) = 2.300, p = 0.001) and this became apparent following the change in diet (Figure 3A). While there was no difference detected in food intake (Figure 3B), a small but statistically significant difference in energy intake between the ARC-CMKLR1-KD and scrambled virus injected rats was detected (time x treatment, F_(7,98) = 2.727, p = 0.0125; Figure 3C). This was confirmed when analyzing the area under the curve for energy intake (t₍₁₄₎ = 2.288, p = 0.0382; Figure 3D). At 29 days following surgery, no significant differences were found in unfasted glucose levels (Figure 3E).

Next, we tested if the chemerin-CMKLR1 axis was involved in appetite regulating neuropeptides and inflammatory markers. Initially, hypothalamic neuropeptide gene expression of Agrp, Npy, Pomc and Cart was measured in rats 29 days after knockdown of CMKLR1 in the ARC and showed no difference in gene expression (Figure 4A). Given that chemerin is a potential regulator of inflammation, we then tested gene expression levels of Il-6, Tnf-α and Nfκb. While there was no difference of levels Il-6 and Nfκb mRNAs, Tnf-α mRNA was significantly reduced in ARC-CMKLR1-KD rats compared to scrambled virus injected rats (t₍₉₎ = 2.428, p = 0.0381; Figure 4B), suggesting that knockdown of CMKLR1 protects from high-fat induced neuroinflammation. Bdnf, a neurotrophic factor that protects the hypothalamus from high-fat diet induced inflammation, remained unchanged (Figure 4C).

Chemerin is secreted from and has several known actions in white and brown adipose tissue. Therefore, to determine whether central lentiviral knockdown of CMKLR1 leads to altered chemerin expression, signaling or adipose tissue physiology, qPCR was carried out on key mRNA markers. Chemerin and its receptor (Rarres2, Cmklr1, Gpr1) expression was not significantly altered in white or brown adipose tissue (Figures 5A,C). In white adipose tissue, markers of adipogenesis (Cebpa, Cebpb, Pparg, Fabp4), lipid and glucose metabolism (Fasn, Lipe, Plin1, Glut4), and markers of browning (Ucp1, Pgc1a) remained unaltered in the arcuate nucleus-CMKLR1-knockdown rats (Figure 5B). Similarly, markers of adipogenesis (Cebpa, Cebpb), mitochondrial biogenesis (Pparg) and non-shivering thermogenesis activation (Ucp1, Dio2) were unaltered in arcuate nucleus-CMKLR1-knockdown rats (Figure 5D).

We then tested whether knockdown of CMKLR1 in the arcuate nuceleus had an impact on Novel object recognition (NOR). Initially, we tested rats on chow and high-fat diet for 2 and 4 weeks (Figure 1C, Experiment 3). One rat was excluded from the analysis as it failed to explore both objects in one trial of the task. After 2 weeks, there was no significant difference spent exploring the two identical objects during the acquisition trial in the chow and high-fat groups (Figure 6A). In the retention trial, both groups explored the novel object significantly more than the familiar object (chow: t₍₈₎ = 4.221, p = 0.0029; high-fat: t₍₇₎ = 03.958, p = 0.0055). However, when re-tested after 4 weeks, only the chow diet group explored the novel object significantly more than the familiar object (t(8) = 2.445, p = 0.0403; Figure 6D). These data show that 4 weeks of high-fat feeding is sufficient to induce an impairment in memory in our rat model. Next, to determine if central knockdown of CMKLR1 protects against high-fat diet inducing short term memory impairment, we conducted NOR test in the arcuate nucleus-CMKLR1-knockdown rats and control rats. Four rats were excluded from the control group and two rats were excluded from the arcuate nucleus-CMKLR1-knockdown group because they failed to explore both objects in one of the trials. As expected, at 2 weeks following surgery for control rats the high-fat diet impaired the ability of rats to differentiate between novel and familiar objects, whereas arcuate nucleus-CMKLR1-knock down rats showed increased exploration of the novel object, however, this was not significant (Figure 6F). 4 weeks following surgery, none of the groups showed a preference (Figure 6H).