Circadian peak dopaminergic activity response at the biological clock pacemaker (suprachiasmatic nucleus) area mediates the metabolic responsiveness to a high‐fat diet

Abstract Among vertebrate species of the major vertebrate classes in the wild, a seasonal rhythm of whole body fuel metabolism, oscillating from a lean to obese condition, is a common biological phenomenon. This annual cycle is driven in part by annual changes in the circadian dopaminergic signalling at the suprachiasmatic nuclei (SCN), with diminution of circadian peak dopaminergic activity at the SCN facilitating development of the seasonal obese insulin‐resistant condition. The present study investigated whether such an ancient circadian dopamine‐SCN activity system for expression of the seasonal obese, insulin‐resistant phenotype may be operative in animals made obese amd insulin resistant by high‐fat feeding and, if so, whether reinstatement of the circadian dopaminergic peak at the SCN would be sufficient to reverse the adverse metabolic impact of the high‐fat diet without any alteration of caloric intake. First, we identified the supramammillary nucleus as a novel site providing the majority of dopaminergic neuronal input to the SCN. We further identified dopamine D2 receptors within the peri‐SCN region as being functional in mediating SCN responsiveness to local dopamine. In lean, insulin‐sensitive rats, the peak in the circadian rhythm of dopamine release at the peri‐SCN coincided with the daily peak in SCN electrophysiological responsiveness to local dopamine administration. However, in rats made obese and insulin resistant by high‐fat diet (HFD) feeding, these coincident circadian peak activities were both markedly attenuated or abolished. Reinstatement of the circadian peak in dopamine level at the peri‐SCN by its appropriate circadian‐timed daily microinjection to this area (but not outside this circadian time‐interval) abrogated the obese, insulin‐resistant condition without altering the consumption of the HFD. These findings suggest that the circadian peak of dopaminergic activity at the peri‐SCN/SCN is a key modulator of metabolism and the responsiveness to adverse metabolic consequences of HFD consumption.


| INTRODUCTION
In the wild, vertebrate species, from teleosts to mammals, exhibit marked annual cycles of metabolism, oscillating between obese and lean conditions during particular seasons of the year. The seasonal obese condition is coupled with the hyperinsulinaemic, insulin-resistant state similar to the human metabolic syndrome. 1,2 This obese insulinresistant state imparts a survival advantage to the animal during an ensuing/existing season of low/no food (particularly glucose) availability. The available evidence suggests that, under such circumstances, insulin resistance facilitates increased endogenous glucose production to fuel central nervous system metabolism at the same time as the peripheral tissues increase their utilisation of stored fat. 1 year of such treatment. [7][8][9][10] The temporal phase relationships between the circadian peaks in dopaminergic and serotonergic neural activities at the SCN area differ with seasonal condition, even in animals held under the same photoperiod at the same time of year (seasonality) (eg, summer/autumn animals in the glucose tolerant condition on long daily photoperiods [>14 hours of light; termed photosensitive] vs summer/autumn animals in the glucose intolerant condition on the same long daily photoperiods [termed photorefractory]). 5 At the same time as seasonal animals transition from the insulin-sensitive, glucose tolerant state to the insulin-resistant, glucose intolerant state, when maintained on the same daily photoperiod, there is a marked two-thirds reduction in the circadian peak dopaminergic input activity at the area of the SCN. 5 Moreover, specific lesion of these SCN-area dopaminergic neurones in seasonally lean, insulin-sensitive animals results in the obese insulin-resistant state that cannot be explained by any change in food consumption. 6 It is suggested that the phase relationship between these two (dopaminergic and serotonergic) circadian neural oscillation input signals to the SCN regulates (circadian) output activities from the clock that modulate physiological status based upon their effect to synchronise the phase relationships of multiple peripheral circadian stimulus (eg, insulin) and response (eg, hepatic lipogenic responsiveness to insulin) rhythms. 1,2 Relevant to these findings are the independent observations of tyrosine-hydroxylase immunopositive (TH+) fibres observed in the SCN area (within the structure and around its perimeter) of perinatal rodents including Syrian hamsters, Siberian hamsters and rats that diminish in density (although they are still present) within the SCN but are still relatively prominent in the peri-SCN region of the adult. [11][12][13][14][15][16][17][18] The peri-SCN TH+ fibres were also dopamine β-hydroxylase negative and/or aromatic amino acid decarboxylase positive, suggesting a dopaminergic neuronal function. 11,12 The origin(s) of these TH+ fibres, however, have not been identified and could represent short local (inter)neurones, as well as projections from other anatomical sites, with this being one focus of the present study (see below).
We postulated that this ancient circadian control system for whole body regulation of fuel metabolism may modulate the sensitivity of the body to the obesity/insulin resistance-inducing effects of a high-fat diet (HFD) and, as such, play an important role in regulating the metabolic syndrome-inducing effects of the westernised diet of modern man. This clock system for the regulation of metabolism is sensitive to seasonal changes in nutrient quality (eg, changes in natural flora and fauna cycles) 1 and HFDs have been demonstrated to reduce striatal (mesolimbic) dopamine levels or dopamine receptor availability. [19][20][21][22][23][24][25][26] We therefore postulated that, in animals made obese and insulin resistant by HFD feeding, the circadian peak of dopaminergic input activity to the SCN area would be diminished and also that local SCN restoration of this circadian dopaminergic input activity would be sufficient to reverse the metabolic syndrome-inducing impact of the high fat diet upon metabolism. As such, investigations were undertaken aiming to 0.2 mm from midline and 9.3 mm below the skull). Ten days after FG microinjection, the rats were sacrificed under anesthesia by transcardiac perfusion with 4% paraformaldehyde. Brains were post-fixed in 4% paraformaldehyde overnight and cryoprotected in 30% sucrose/ phosphate-buffered saline (pH 7.4). Free floating coronal brain sections (30 μm) were cut from frozen brains on a cryostat and sequentially collected in a freezing solution containing 30% ethylene glycol, 25% glycerol and 0.05 m phosphate buffer saline (pH 7.4) and kept at −20°C until use. For the dual-labelling of fluorogold and TH, brain sections were first treated with 3% hydrogen peroxide for 10 minutes to quench the endogenous peroxidase activity followed by 1 hour of incubation in animal-free blocker (Vector Laboratories, Burlingame, CA, USA). Then, the sections were immunolabelled at 4°C overnight with a rabbit polyclonal antiserum for FG (dilution 1:3000; AB153; Millipore, Billerica, MA, USA) followed by 1 hour of incubation at room temperature with a goat anti-rabbit biotinylated secondary antiserum (dilution 1:500; Vector Laboratories). The FG immunoreactivity was amplified by an avidin-biotin complex (ABC) system and revealed as dark brown 3,3′-diaminobenzidine (DAB) punctates. Following thorough washing, the sections were re-blocked with animal-free blocker and then sequentially incubated with the mouse monoclonal antiserum for TH (dilution 1:1000; MAB5280; Millipore) at 4°C overnight and a horse anti-mouse biotinylated secondary antiserum (dilution 1:500; Vector Laboratories) for 1 hour at room temperature. Following amplification with the ABC system, the TH immunoreactivities were revealed as diffuse blue-grey staining in the cytoplasm (Vector SG peroxidase substrate kit, SK-4700; Vector Laboratories). The dual-labelled cells were examined under a light microscope (Nikon, Tokyo, Japan). Images were captured by a digital camera and processed using photoshop (Adobe Systems, San Jose, CA, USA).

| In vivo microdialysis
Animals (SD rats) utilised in microdialysis studies were anaesthetised with ketamine/xylazine (80/12 mg kg -1 body weight, i.p.) and mounted on a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA, USA) for implantation of microdialysis probes. In separate studies, a 30-gauge stainless steel guide cannula (CMA Microdialysis, Holliston, MA, USA) was permanently implanted aimed either at the top of the right SCN at coordinates 1.3 mm posterior to bregma, 0.25 right lateral to the midsagittal suture and 8.3 mm ventral to the dura, with the incisor bar set 3 mm below the interaural line or at the top of the ventromedial hypothalamus (VMH) at coordinates 2.6 mm posterior to bregma, 0.6 mm right lateral to the midsagittal suture and 9 mm ventral to the dura.
The guide cannula was anchored firmly to the skull with three stainless-steel screws and cemented in place with dental acrylic.
Animals were allowed 1 week to recover prior to the initiation of microdialysis experimentation. During microdialysis, each animal was placed in an acrylic bowl with free access to food and water. Microdialysis samples (0.12 μL min -1 flow rate) were collected into 300-μL vials (containing 2 μL of 0.1 n perchloric acid solution) at hourly intervals through an automated refrigerated fraction collector (modified CMA/170; CMA Microdialysis) over a 24-h period when animals were maintained on a 14 hour daily photoperiod and allowed free access to food and water.

| Evaluation of the impact of a HFD on circadian dopamine activities in the SCN and supramammillary nucleus (SuMN) of rats
Female Sprague-Dawley rats (14 weeks of age) were maintained on 14 hour daily photoperiods and allowed to feed and drink ad libitum.
Rats were randomly divided into two groups (N = 10 per group) that were fed either a RC diet or HFD to induce weight gain (25% more than RC fed rats) for 6 weeks. A glucose tolerance test (GTT) was then performed on all rats and then, after a 3-day rest, in vivo microdialysis was used to study daily extracellular profiles of dopamine metabolites in the SCN of rats. Microdialysis samples from the SCN of free-living rats held under a 14 hour daily photoperiod and allowed to feed and drink ad libitum during the sampling were collected every 1 hour continuously over a 24-hour period. Samples were assayed via HPLC for HVA and DOPAC. A sample of 5 μL from a total dialysate sample of

| Dual-label immunohistochemistry of c-Fos and TH
At the end of the above-described microdialysis experiment, HFD and RC fed SD rats were sacrificed under anesthesia by perfusion with 4% paraformaldehyde at either 5 or 16 hours after light onset (ZT5 [N HFD = 8; N RC = 8] or ZT16 [N HFD = 8; N RC = 8]), respectively, to quantify c-Fos and TH double-immunopositive neurones in the SuMN. The brains were similarly processed as described above (in section 2.2) for fluorogold tracing and the 30μm coronal sections were stored at −20°C until dual immunohistochemistry of c-Fos and TH was performed. Briefly, the brain sections containing the SuMN were first labelled with a rabbit polyclonal anti c-Fos antibody (dilution 1:20 000; PC38; Calbiochem Merck, Darmstadt, Germany) followed by a goat anti-rabbit biotinylated secondary antiserum (dilution 1:500; Vector Laboratories). c-Fos immunoreactivity was revealed as black Ni-DAB punctates in the nucleus. Subsequently, the brain sections were labelled with the mouse monoclonal antiserum for TH (dilution 1:1000; MAB5280; Millipore) followed by a horse anti-mouse biotinylated secondary antiserum (dilution 1:500, Vector Laboratories). TH immunoreactivity was revealed by DAB as brown staining in the cytoplasm. The c-FOS/TH dual-labelled cells were examined under a light microscope (Nikon). The images were captured with a digital camera and processed using photoshop without altering the ratios of signal from comparative sample regions. Dual immunopositive neurones for c-Fos and TH at the SuMN and the adjacent posterior hypothalamus (PH) region were identified and counted manually on sequential sections across the whole SuMN in a double-blind manner. The dual positive numbers obtained from sequential coronal sections of each brain in the same treatment group were used to generate a dual positive number per SuMN/PH test area. The between group difference in dual positive number per SuMN/PH area was then analysed by two-way analysis of variance (anova). As a control experiment, a parallel set of coronal brain sections at the level of SuMN was processed for dopamine β-hydroxylase (DBH) immunostaining and TH/DBH double-immunofluorescence labelling. The coronal brain sections at the level of SuMN were labelled with rabbit polyclonal anti-dopamine β-hydroxylase antibody (dilution 1:2000; ab209487) followed by biotinylated goat anti-rabbit IgG (dilution 1:500; Vector Laboratories).
The signals were amplified by the ABC complex and revealed by DAB chromogen. To further confirm that TH-immunopositive neurones detected in the SuMN are dopaminergic but not noradrenergic neurones, double-immunofluorescence labelling were sequentially performed on the same brain section using rabbit polyclonal anti-

| In vivo electrophysiology recordings
Female SD rats were anaesthetised with thiobutabarbital (120 mg kg -1 body weight, i.p.) and mounted in a stereotaxic apparatus. The core body temperature was maintained at 37°C with a heating pad. The skull was removed from the area overlying the right side SCN. A silver electrode was implanted at the coordinates: 1.3 mm posterior to bregma, 0.2 mm lateral and 9.2 mm ventral to the dura. The injection cannula was targeted to a region just exterior (0.25-0.4 mm lateral) to the SCN lateral edge (peri-SCN), whereas the electrode was placed within the SCN itself. After basal neuronal activity had stabilised, test chemicals at various doses were infused for 1 minute, with rest periods of 25 minutes between each subsequent increased dosage. Electrical signals were passed through an amplifier and surveyed using a Bio Amp ML136 (ADInstruments, Colorado Springs, CO, USA). Analyses of electrophysiological activities were conducted off line with the use of labchart, version 6 (ADInstruments) to isolate spike potentials from the background data. The dose-response curves were analysed via the Hill equation. All data are expressed as the mean ± SEM. Statistical analysis was performed using Student's t test and anova to determine the treatment difference in dose-response. A P < .05 value was considered statistically significant.

| Studies on SCN neurone electrophysiological responsiveness to dopamine and dopamine receptor modulators
Female SD rats (12 weeks old) were maintained under 14 hour daily photoperiods and allowed to feed ad libitum for at least 1 week before the initiation of experimentation. To study time of day dependent differences in neuronal responses to dopamine at the SCN, electrophysiology recordings were conducted at 14 hours after light onset (ZT14) (just at onset of darkness and the onset of locomotor activity in these nocturnal rodents) or at ZT5 (sleep time of day). After basal neuronal activity had stabilised, 70 mm glutamate (loaded cannula concentration) was injected at the peri-SCN at 14 nmol per 0.2 μL to evoke neuronal activity. 30 both with or without dopamine (5 mm) applied.

| Direct dopamine administration to peri-SCN
Surgery was performed on SHR rats at 16 weeks of age. Rats were  or vehicle at the onset of locomotor activity (ZT13), whereas two additional groups of rats received either the same dose of vehicle or dopamine infusion at the same peri-SCN area but at ZT19 as described above. Again, GTTs were performed after 2 weeks of such treatment.
At the end of the experiment, animals were killed and brains were collected and stored at −80°C for neurotransmitter analyses within the VMH and PVN.

| VMH and PVN NE analysis from animals treated for 2 weeks with dopamine or vehicle at the peri-SCN area at either ZT13 or ZT19
Frozen serial coronal brain sections were cut at a thickness of 300 μm on a cryostat maintained at −8°C. VMH and PVN tissues were punched out and placed in 40 μL of 2% trichloroacetic acid, sonicated and centrifuged. Supernatant was immediately analysed by HPLC with coulometric electrochemical detection (ESA) for NE and the NE metabolite MHPG content and quantified against a standard curve for each. The signal was analysed by EZChrom Elite data processing software (Agilent). Next, 10 μL of supernatant was injected into the system using a refrigerated autosampler (ESA 540). The results are expressed as pg per 10 μL of sample.

| GTT
GTT was performed 6 h after light onset. A 50% glucose solution was administered i.p. (3 g kg -1 body weight) and blood samples were and 120 min after glucose injection for plasma glucose and insulin analyses.

| Assay of metabolic parameters
Blood glucose concentrations were determined by a blood glucose monitor (OneTouch Ultra, LifeScan, Inc.; Milpitas, CA, USA). Plasma insulin and NE were assayed by an enzymeimmunoassay using commercially available assay kits utilising anti-rat serum and rat insulin and NE as standards (ALPCO Diagnostics, Salem, NH, USA). Liver tissue was homogenised in 5% NP-40, heated, centrifuged and supernatant assayed for triglyceride content using a Triglyceride Determination Kit (catalogue number TR0100; Sigma-Aldrich, St Louis, MO, USA). After drying at room temperature using an air blower, the slides were stored at −80°C in slide boxes with desiccant. Prior to using the tissue, the section mounted slides were gradually brought to room temperature. For dopamine D2 receptor binding determination, the slides were first equilibrated in Tris-ions assay buffer  TH-immunopositive cell bodies and processes revealed by brown DAB staining using antibody against TH were found in the SuMN at low magnification (E) and at high magnification (F). DBH-immunopositive terminals (but not cell bodies) revealed by brown DAB staining coupled with antibody against DBH were detected in the SuMN at low magnification (G) and at high magnification (H). Double-immunofluorescence staining of TH (I) and DBH (J) on the same brain section showed no co-localisation of TH-immunoreactivity (red) and DBH-immunoreactivity Primers:

| Statistical analysis
All data are expressed as the mean ± SEM. Statistical analysis were performed using an unpaired Student's t test for two group comparisons or one-way anova for more than two group comparisons, or twoway repeated measures anova for comparisons of treatment groups undergoing repeated measurements at different time points, as appropriate. When the overall anova result was statistically significant, a post-hoc Dunnett's test was carried out to highlight where these differences occur. A statistical value of P < .05 (2-tailed) was considered statistically significant.

| Neuroanatomical and neurophysiological studies identify dopaminergic neuronal projections from the SuMN to dopamine D2 receptor site regions within the peri-SCN/SCN area
Our previous studies suggested a role for circadian responsiveness to systemic l-DOPA administration in the regulation of seasonal metabolism. [7][8][9][10] Other studies centred on clock region neurophysiology identified a potential cause-effect relationship between a diminution of the circadian peak in dopamine release at the general peri-SCN region and seasonal glucose intolerance. 5,6 However, the identification of specific dopaminergic neurones projecting to a peri-SCN/SCN neural circuit that in turn regulates peripheral metabolism has not been established. Therefore, our initial studies into a potential role of peri-SCN/SCN circadian dopaminergic regulation of HFD-induced insulin resistance focused on identifying the anatomy of dopaminergic F I G U R E 2 (A-E) Dopamine receptor binding and mRNA present in peri-suprachiasmatic nuclei (SCN)/SCN area. (A) Autoradiography using radioligands selective for dopamine receptor D2 (I 125 -iodosulpride) in brain sections at the level of the SCN (bregma -1.30 mm). [I 125 ]iodosulpride revealed low density D2 dopamine receptor binding sites within SCN (blue cycle) and higher (moderate) density binding in peri-SCN (red semicycle). Insert: higher magnification of peri-SCN/SCN. (B) The binding specificity of [I 125 ]-iodosulpride (0.5 nm, K d = 1.6 nm) to D2 dopamine receptors was confirmed by the displacement of the [I 125 ]-iodosulpride binding sites with a saturation concentration of dopamine D2 receptor antagonist haloperidol (10 μm). (C) Autoradiograph ligand binding study with [I 125 ]-SCH23982 revealed a moderate density of dopamine D1 receptor binding sites within the SCN and low binding density in the peri-SCN (red semicycle). Dopamine D1 receptors are defined by [I 125 ]-SCH23982 binding sites (0.1 nm, K d = 0.12 nm) in the presence of 5-HT 2A/2C antagonists (ketanserin, 50 nm and mianserin, 100 nm). Insert: high magnification of peri-SCN/SCN. (D) The binding specificity to dopamine D1 receptors was confirmed by the displacement of the [I 125 ]-SCH23982 binding sites with a saturation concentration of dopamine D1 receptor antagonist R-(+)-SCH23390 (10 μm). (E) Dopamine D1 and D2 receptor mRNA at the medial preoptic area (mPOA), periSCN/SCN and SCN regions of the hypothalamus quantified by a quantitative reverse transcriptase-polymerase chain reaction (PCR). Dopamine D1 and D2 receptor mRNA actual transcript number per mm 3 of tissue at the mPOA, peri-SCN/SCN and SCN areas were each quantified by generation of standard curves with a Bio-Rad PrimePCR template (assay ID qRnoCEP0027016 for dopamine D1 receptor and qRnoCIP0023714 for dopamine D2 receptor) as standard. Such transcripts for dopamine D2 receptor were much lower than that at the striatum (15 million copies per mm 3 of tissue). Relative concentrations of Dopamine D2 and D1 receptor mRNA among these brain regions were not altered when normalised to GAPDH mRNA (GAPDH quantified with Bio-Rad assay qRnoCIP0050838). Reduction of dopamine D1 receptor mRNA transcript density in peri-SCN/SCN vs SCN area reflects dilution of SCN transcript with peri-SCN tissue of much reduced D1 mRNA content. Results are the mean ± SEM of tissue samples from 5 animals. (F, G) Neurophysiological dopamine communication from the supramammillary nucleus (SuMN) to SCN. Acute intra-SuMN AMPA administration increases the extracellular dopamine metabolites 3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) at the SCN. Extracellular profiles of HVA (F) and DOPAC (G) in microdialysate samples from the SCN of freely-moving rats that received either acute intra-SuMN AMPA (•) or vehicle (○). Data are expressed as percentage changes from the baseline (mean ± SEM, n = 6 per group). Two-way anova with repeated measures on HVA (F) indicates a time effect (F 9,90 = 2.026, P < .05) and also a time and treatment interaction effect (F 9,90 = 3.368, P <0.005). SCN DOPAC (G) is increased in AMPA treated vs vehicle groups (F 1,10 = 5.387, P < .05). There is also a time effect (F 9,90 = 2.509, P < .05) and a time and treatment interaction effect (F 9,90 = 2.065, P < .05) neurones within the peri-SCN region, thereby aiming to investigate a role for their circadian neurophysiology in the regulation of SCN activity and peripheral metabolism.
Neuroanatomical studies utilised nanolitre injections of the retrograde tracer fluorogold (which labels primary projecting neurone cell bodies when injected at the neuronal terminal region) at the SCN and its perimeter followed by double-immunohistochemical staining with fluorogold and TH (the rate-limiting enzyme in dopamine synthesis) antibodies to trace the origin of primary dopaminergic neurones that project to the peri-SCN (the anatomical area defined as F I G U R E 3 The high-fat diet (HFD) feeding-induced obese and insulin-resistant condition is accompanied by a concurrent abolishment of the circadian peak in suprachiasmatic nuclei (SCN) dopaminergic activity and of the coincident daily peak in dopaminergic neurone activity in supramammillary nucleus (SuMN) neurones. Animals were fed HFD for 9 weeks and, after 32% of gain in body weight, analyses of SCN area dopamine release and SuMN dopamine activity at different circadian time points were performed. (A) Body weight of regular chow (RC, white bar) and HFD (black bar) fed rats (*P < .0001, HFD fed vs RC fed group) (Student's t test). Plasma glucose (B) and insulin (C) during a glucose tolerance test (*P < .05, difference between the two groups at same time) (anova with repeated measures followed by post-hoc t test). The area under the glucose and insulin tolerance curve in the HFD fed group increased by 23% and 57% respectively, compared to the RC fed group (P < .05, Student's t test). HFD feeding induces insulin resistance (reduces Belfiore and Matsuda insulin sensitivity indices by 50% [D] or 34% [E], respectively, *P < .005 [Student's t test]). (F,G) Daily profiles of homovanillic acid (HVA) and 3,4-dihydroxyphenylacetic acid (DOPAC), respectively in 5-μL microdialysate samples from the SCN of freely-moving rats fed either HFD (•) or RC (○) (n = 8 per group). The horizontal bar indicates light and dark phases of the daily photoperiod. Two-way anova with repeated measures on HVA indicates a time of day effect (F 21,294 = 4.3, P < .0001). There is also a time and group interaction effect (F 21,294 = 3.2, P < .0001), which indicates a circadian difference of SCN dopamine activity between the HFD and RC fed groups. Two-way anova with repeated measures on DOPAC (G) reveals a time of day effect (F 21,294 = 5.488, P < .0001) and a time and group interaction effect (F 21,294 = 2.578, P < .0002). All data are expressed as the mean ± SEM (n = 10 per group). (H) The daily peak dopamine neuronal activities at the SuMN and adjacent posterior hypothalamus (PH) were reduced by HFD feeding. The brains from HFD fed obese rats and RC fed lean rats on LD 14:10 h photoperiods were collected during the day (ZT4 [Zeitgeber time] hours after light on set) and night (ZT16), respectively. The activated dopamine neurones were identified as double immune-positive neurones using antibodies against tyrosine hydroxylase (TH) (a rate-limiting enzyme for dopamine synthesis) and c-Fos (neuronal activation marker). The number of activated dopamine neurones at the SuMN/adjacent posterior hypothalamus (determined as number per total sampled areas) in the brains from RC lean rats was 46% higher at ZT16 than at ZT4 (two-way anova analysis: *P < .05; n = 8 or 9) and this circadian peak was abolished in the brains from HFD-fed obese rats (*P < .05). Insert: Number of double positive neurons at each sampled area within the SuMN/PH for animals within each group (mean ± SEM).

| The circadian peak in SuMN-SCN dopamine release is abolished by HFD feeding
Based upon results of several previous studies demonstrating the import of (i) a circadian rhythm of responsiveness to systemically administered l-DOPA 7-10 and (ii) peri-SCN dopamine activity in the modulation of seasonal metabolism, 5,6 we postulated that these SuMN-SCN dopaminergic neurones may represent a circadian dopaminergic circuit operative in the modulation of SCN function and its regulation of peripheral metabolism. We further postulated that this circuit may represent not only a clock mechanism regulating seasonal metabolism, but also a biological target for the fattening/insulin resistance-inducing effects of a HFD. Therefore, our next series of studies investigated the potential existence of a daily rhythm of SuMN dopaminergic neuronal activity and a daily rhythm of dopamine release from peri-SCN neurones among lean animals and those made obese/insulin resistant by HFD feeding. Compared to RC fed animals, HFD fed animals had increased body weight (P < .0001, Student's t test) ( Figure 3A), as well as increased plasma glucose ( Figure 3B) and insulin ( Figure 3C) during a GTT (difference between the two groups at same time P < .05, anova with repeated measures followed by a t test). The area under the glucose and insulin GTT curve in the HFD fed group increased by 23% and 57%, respectively, compared to the RC fed group (P < .05, Student's t test). HFD feeding reduced the insulin sensitivity (ie, Belfiore and Matsuda insulin sensitivity indices by 50% [ Figure 3D] and 34%, respectively [ Figure  from RC lean rats was 46% higher at ZT16 than at ZT4 (two-way anova analysis: P < .05; n = 8 or 9 per group) and this daily peak is abolished in the brains from HFD fed obese rats ( Figure 3H), thus corroborating the microdialysis study results reported above.

| The circadian peak in electrophysiological responsiveness to dopamine at the SCN coincides with the circadian peak in dopamine release at the SCN in lean insulin-sensitive rats and is attenuated by HFD feeding
To gain insight into a potential neurophysiological role for the circa-  was reduced 50% relative to RC fed controls (two-way anova with repeated measures: F 6,48 = 54.3, P < .0001) ( Figure 4D).
Consequently, in healthy, non-obese animals, the daily peak of dopamine release at the peri-SCN region is "in-phase" with the daily peak in SCN responsiveness to peri-SCN dopamine, whereas, among animals made obese/insulin resistant by HFD feeding, these coincident daily peaks in peri-SCN dopamine release and SCN re-

| Restoration of the circadian peak of dopamine at the SCN of insulin-resistant/glucose intolerant animals held on a HFD attenuates the insulin resistance/glucose intolerance
We next hypothesised that the circadian peak in peri-SCN dopa-   Figure 6). Repeated measures anova on NE indicated a group effect (F 1,13 = 14.01, P < .005), as did repeated measures anova on MHPG (F 1,13 = 5.334, P < .05). Repeated measures anova on 5-HIAA also revealed a group effect (F 1,13 = 7.630, P < .05). Following 2 weeks of infusion of dopamine or vehicle into the peri-SCN/SCN region of obese/ insulin-resistant rats at either ZT13 or ZT19 and subsequent to death, brain punches of frozen sections of the hypothalamus were obtained for the analysis of VMH and PVN NE and its metabolites within these study groups. It was found that daily dopamine infusion at the peri-SCN area at ZT13 (the daily peak in dopaminergic activity at this site) but not at ZT19 significantly reduced NE activity (measured as total NE content and NE × NE metabolite product in both the VMH [by 46% and 47%, respectively; P < .05] and the PVN [by 33% and 43%, respectively; P < .05]; unpaired Student's t test) (Figure 7).

| DISCUSSION
These studies demonstrate the functionality of a unique, potent and previously unrecognised role for the circadian organisation of dopa- Subsequent to our initial postulate that temporal interactions of circadian neuroendocrine input activity rhythms to the biological clock modulate its output signals that regulate peripheral metabolism, 1,2,9,10,61,62 a multitude of studies have provided evidence supporting such a regulatory role for the SCN in the modulation of peripheral fuel metabolism. [63][64][65][66] First, complete destruction of the SCN was shown to lead to insulin resistance, glucose intolerance and weight gain, clearly identifying a functional SCN as necessary to maintain normal fuel metabolism. 67 However, the results from more detailed investigations of the role of the SCN in the regulation of peripheral fuel metabolism ascribe major roles for specific temporal interactions of specific circadian input signals to the SCN clock system with respect to directing its regulation of metabolism. The SCN sends direct and indirect projections to many hypothalamic centres that modulate peripheral metabolism, including the VMH, PVN, dorsal medial hypothalamus, lateral hypothalamus and arcuate nuclei, and also to behavioural/feeding centres in the mesolimbic system. 41,42,68,69 With regard to the present study, the construct of this postulate is that the circadian dopaminergic message at the peri-SCN/SCN area is integrated within the SCN clock system with other local environmental information to modulate SCN output signalling to other downstream metabolic regulatory sites including but not limited to noradrenergic activity at the VMH and PVN activity, each of which in turn regulates peripheral metabolism, in part by adjusting phase relations of numerous metabolic circadian stimulus and response rhythms in target tissues (eg, circadian rhythm of plasma insulin stimulus interacting with a circadian rhythm of hepatic lipogenic responsiveness to insulin). [1][2][3]61 The present studies suggest that a contributing mediator of this  [75][76][77][78] The SCN also sends strong projections directly to the PVN 41,79-82 a neural centre with strong regulatory control over peripheral metabolism. [83][84] In seasonally obese/insulin-resistant animals, the diminution of the circadian peak of dopaminergic activity at the peri-SCN is coupled with substantial increases in NE release at the VMH. 5,35 Exogenous infusion of noradrenaline into the VMH of seasonally lean/insulinsensitive animals induces the obese/insulin-resistant condition within just a few weeks without any alteration in the food consumption of a low-fat diet. 36 Similarly, elevations of VMH noradrenergic activity have been consistently documented in a wide variety of other animal models of insulin resistance, including ob/ob mice, db/db mice, A y /J mice, offspring of malnourished mother rats and offspring of insulintreated mother rats. 2 Moreover, infusion of NE into the VMH of lean/insulin-sensitive inbred laboratory rats held on a low-fat diet to raise the extracellular NE levels induces a rapid and sustained (chronic) simultaneous increase in sympathetic nervous system (SNS) tone, plasma NE, blood pressure, insulin, glucagon and leptin levels, as well as adipocyte lipogenic activity and adipocyte responsiveness to the lipolytic effects of noradrenergic stimulation (resulting in increased free fatty acid [FFA] release that potentiates insulin resistance in liver and muscle 37 ). This VMH NE effect ultimately leads to the obese, insulin-resistant, glucose intolerant, leptin resistant and hypertensive state without altering food consumption. 36,37,85 Increased VMH NE activity also results in a loss of appropriate fuel (FFA and glucose) sensing by these VMH neurones, such that, instead of responding to increased meal-time FFA and glucose by sending neuroendocrine signals to increase peripheral insulin sensitivity as would normally occur, the NE-overstimulated VMH now sends neuroendocrine signals to the peripheral tissues that counter insulin action in liver and muscle. 2,86 At the PVN, increased F I G U R E 7 Dopamine administration at the peri-suprachiasmatic nuclei (peri-SCN) at Zeitgeber time (ZT)13 in high-fat diet (HFD) fed rats to restore the normal circadian peak of dopamine at this site but not when administered outside this circadian peak window (at ZT19) reduced noradrenergic turnover/activity (3-    It is clear that it is the circadian rhythm of peri-SCN dopaminergic activity and not merely the absolute level of such activity that is critical in manifesting the attenuation of the insulin-resistant/glucose intolerant state because the addition of dopamine to this site at the time of its normal circadian peak in insulin-sensitive/glucose tolerant animals (but not at a time outside this daily) interval manifests this physiological response. It is equally instructive that the impact of the HFD feeding to reduce peri-SCN/SCN dopamine activity was restricted to the circadian peak dopaminergic activity period of the day at this site and showed no such effect during the long low trough activity period of the day (Figure 3), implicating an impact on a circadian coupling/ expression mechanism and not merely comprising a biochemical inhibitory phenomenon.

Important unanswered questions of the present study include: (i)
what is the genesis of these circadian dopamine stimulus and response rhythms at the clock (ie, does the circadian rhythm of SuMN dopamine release at the SCN derive from entrainment by the SCN itself and/or from other sources such as the gut-brain axis?) and (ii) how does HFD feeding reduce the concurrent circadian peaks in SuMN-SCN tyrosine hydroxylase activity and dopamine release, as well as SCN responsiveness to peri-SCN dopamine stimulation? However, the finding of reduced dopaminergic activity is generally consistent with a multitude of studies indicating a reduction of dopamine and/or dopamine receptor levels in other brain areas, particularly the striatal-mesolimbic system, following chronic HFD feeding, although no specific investigation of such feeding on circadian aspects of dopaminergic activity at these brain sites has been investigated. [19][20][21][22][23][24][25][26]101,102 Interestingly, striatal reduction of dopamine levels may partly be the result of reduced gut synthesis of the diet-derived satiety factor, oleoylethanolamine, following HFD feeding, the gastric presentation of which is known to inhibit dopamine efflux in the striatum via vagal inputs. 103 Although attenuation of dopamine function within the mesolimbic system has been associated with a reduction in appropriate reward signalling and consequent overfeeding as a compensatory response to chronic reduced dopaminergic signalling, the presently described circadian dopamine-SCN clock system for regulation of glucose metabolism and body fat appears to be quite a different aspect of CNS dopamine regulation of metabolism in that its influences to attenuate the obese/ insulin-resistant state do not require and are independent of a reduction of feeding in animals held on a HFD. The present findings suggest that this circadian dopamine-SCN clock regulatory pathway may be an important modulator of sensitivity to the metabolic effects of a HFD (ie, inducing a HFD sensitive vs HFD resistant phenotype) dependent upon the circadian nature of the dopamine input message to the SCN clock system.
Although this circadian dopamine input signalling system to peri-SCN/SCN neurones to regulate clock functions controlling peripheral fuel metabolism is a unique finding, it mirrors the similar circadian dopamine regulation of striatal clock gene expression. 104,105 In the striatum and several other areas of the brain, circadian dopamine-dopamine receptor interactions regulate the circadian expression of cellular clock genes that in turn modulate (i) particular functionalities of the neurone, as well as (ii) the dopamine-dopamine receptor circadian interaction (ie, feedback). [106][107][108][109] The circadian dopamine-dopamine receptor interactions at the peri-SCN/SCN may well function to regulate circadian dopamine-dopamine receptor interactions governing clock gene expression in the striatum in that the SCN has been observed to modulate striatal circadian neuronal activities. 68,110 This circuit may have major implications for linking visceral metabolism and behaviour as discussed below.
Neuroanatomical studies indicate that the major source of the peri-SCN dopamine derives from the SuMN. The SuMN is a hypothalamic nucleus with connections to widespread regions of the hippocampus, forebrain areas, raphe nuclei and limbic system areas. 38 Its role has thus far been defined as an integration centre for cognitive and emotional aspects of behaviour, including reward functions of the nucleus accumbens. 38,111,112 The SuMN contributes greatly to theta rhythm generation in the hippocampus 38  dopamine release at the nucleus accumbens leading to overfeeding to achieve reward is concurrent with a diminution of its circadian peak dopamine release at the SCN that facilitates insulin resistance, then these events would synergise to potentiate obesity. Indeed, in preliminary studies, we have observed that attenuation of overall SuMN activity (with GABA agonist, glutamate antagonist cocktail) resulted in overfeeding, obesity and insulin resistance/glucose intolerance within a couple of weeks. 113 In conclusion, the present study has identified a previously un-