Interleukin‐6 in the central amygdala is bioactive and co‐localised with glucagon‐like peptide‐1 receptor

Abstract Neuronal circuits involving the central amygdala (CeA) are gaining prominence as important centres for regulation of metabolic functions. As a part of the subcortical food motivation circuitry, CeA is associated with food motivation and hunger. We have previously shown that interleukin (IL)‐6 can act as a downstream mediator of the metabolic effects of glucagon‐like peptide‐1 (GLP‐1) receptor (R) stimulation in the brain, although the sites of these effects are largely unknown. In the present study, we used the newly generated and validated RedIL6 reporter mouse strain to investigate the presence of IL‐6 in the CeA, as well as possible interactions between IL‐6 and GLP‐1 in this nucleus. IL‐6 was present in the CeA, mostly in cells in the medial and lateral parts of this structure, and a majority of IL‐6‐containing cells also co‐expressed GLP‐1R. Triple staining showed GLP‐1 containing fibres co‐staining with synaptophysin close to or overlapping with IL‐6 containing cells. GLP‐1R stimulation enhanced IL‐6 mRNA levels. IL‐6 receptor‐alpha (IL‐6Rα) was found to a large part in neuronal CeA cells. Using electrophysiology, we determined that cells with neuronal properties in the CeA could be rapidly stimulated by IL‐6 administration in vitro. Moreover, microinjections of IL‐6 into the CeA could slightly reduce food intake in vivo in overnight fasted rats. In conclusion, IL‐6 containing cells in the CeA express GLP‐1R, are close to GLP‐1‐containing synapses, and demonstrate increased IL‐6 mRNA in response to GLP‐1R agonist treatment. IL‐6, in turn, exerts biological effects in the CeA, possibly via IL‐6Rα present in this nucleus.


| INTRODUC TI ON
Interleukin-6 (IL-6) is a cytokine with a large variety of functions in disease and health. 1 Released together with other classic pro-inflammatory cytokines, such as tumour necrosis factor (TNF)-α, IL-6 appears to have deleterious effects on metabolism. [2][3][4] By contrast, when IL-6 is released together with metabolically beneficial hormones such as leptin, glucagon-like peptide-1 (GLP-1) and amylin, IL-6 appears to have positive effects on metabolism. [5][6][7][8][9] IL-6 is not only produced in adipose tissue, in skeletal muscle during exercise and in immune cells, but also is abundant in the central nervous system (CNS). 1,5,[10][11][12] The membrane bound IL-6 receptor, IL-6 receptor-alpha (IL-6Rα) is expressed in a major portion of fundamental energy-regulatory hypothalamic nuclei in the healthy mouse brain, mainly in neurones, thus forming an anatomical basis for the local effects of IL-6. [13][14][15][16] IL-6 −/− mice develop mature-onset obesity at around 6 months of age. 17 Intracerebroventricular administration (but not peripheral administration) of IL-6 decreases body fat in rodents, suggesting that the anti-obesity effect of IL-6 is exerted at the level of the CNS. 18,19 Central IL-6 has been shown to play a role in the anti-obesity effects of GLP-1 analogues, such as exendin-4 (Ex-4), mainly via reduced food intake. 20,21 In the hypothalamus and hindbrain, IL-6 appears to act as a downstream mediator for the effects of GLP-1R-stimulation, 21 as well as amylin action. 5 Thus, IL-6 may act as a regulator of anorexia and body weight in healthy animals. In addition, the cachexia observed in various diseases appears to depend partially on IL-6 release. 2,22,23 In mouse cancer models, blocking of IL-6 synthesis has been shown to attenuate cachexia progression. 24 GLP-1, a post-translational proglucagon product, is a neuropeptide produced peripherally in ileal L-cells of the intestine, as well as centrally in preproglucagon (PPG) neurones of the nucleus of the solitary tract (NTS). 25 GLP-1 production causes an increase of insulin release from pancreatic β-cells and a decrease of glucagon release from pancreatic α-cells. GLP-1 has a very short half-life in serum because it is rapidly degraded by the enzyme dipeptidyl-peptidase 4. However, long acting GLP-1 analogues used to treat obesity and type 2 diabetes, such as Ex-4, are less easily degraded. 26,27 GLP-1 and the ability of its analogues to increase insulin secretion and control blood glucose levels, as well as their anti-obesity effects, have been widely explored. 20,[28][29][30] However, the potential for IL-6 to act as a downstream signalling peptide of GLP-1 has not yet been fully explored. It has been observed that GLP-1 analogues decrease food intake as well as body weight. Previous studies have reported that these anorexic effects are exerted at the level of the CNS 27,31,32 and it has been suggested that GLP-1 may act on POMC/CART expressing neurones in the arcuate nucleus of the hypothalamus. 33 Moreover, peripheral GLP-1 and its analogues have been reported to cross the blood-brain barrier, making the central GLP-1 receptor (GLP-1R) a possible pharmaceutical target. 34,35 GLP-1Rs are highly expressed in energy-balance-regulating areas, such as the hypothalamus, hindbrain and amygdala. 36 It is likely that the principal source of ligand to these receptors is the GLP-1-producing-neurones of the NTS that produce PPG, a precursor for e g GLP-1. Furthermore, it has been shown that PPG-projections reach, among other nuclei, the central amygdala (CeA). 28 The CeA is involved in fear-, stress-and drug-related responses. 37 Aside from these functions, the CeA was implied to be important in regulation of food intake and energy expenditure. 38 Bilateral lesions of the amygdala in rats induce hyperphagia and weight gain, further supporting its role as an anorexigenic nucleus. 39 Studies of the amygdala have shown that PPG-projections lie in proximity to the same neurones, mainly in the capsular part of the CeA. Moreover, cells in both the CeA and medial amygdala display GLP-1R immunoreactivity. 20,40 Previous studies have shown that GLP-1R stimulation increases neural activity in the CeA, which leads to a decrease in food consumption in rats, 38 although the regulating mechanism remains unclarified. Thus, there are plenty of data in the literature suggesting that GLP-1 exerts an anorectic effect in CeA. 38 In the present study, we investigated the effects exerted on CeA by IL-6, another peptide with well-known anorectic and fat mass suppressing effects in the CNS according to most (but not all) studies. 17,19,21,41 Moreover, we investigated the possible interactions between IL-6 and GLP-1, two peptides shown to interact in the brain in regulation of metabolic functions. 21  For intra-amygdala injections and electrophysiology, 7-8-weekold male Sprague-Dawley rats were used.

| Animals
Animals had free access to water and standard chow pellets (Tekland Global, Harlan, The Netherlands) and were kept under a 12:12 hour light/dark cycle (lights on 6.00 am) at 24-26°C and 50%-60% relative humidity, with food available ad lib., unless otherwise specified. The local ethics committee for animal care at the University of Gothenburg approved all of the animal procedures.

| Tissue preparation for immunohistochemistry
Mice were deeply anaesthetised and perfused transcardially with

| Confocal microscopy and cell counting
Images of the stained sections were obtained using either a confo-  TA B L E 1 List of antibodies used in the present study clearly above background, as measured by fluorescence intensity measurements. The emission spectrum of the secondary fluorescent antibody is well known. By adjusting the beam splitter of the confocal microscope, the signal of the fluorophore was maximised at the same time as minimising background fluorescence. Cells were only considered labelled if their nucleus was in the focal plane. Four mice per treatment group and experiment were used for immunohistochemistry. Cell counting was performed from one brain slice from each of the four animals. Representative confocal micrographs from these animals were used to construct images. Student's t test was used to determine statistical significance of the differences in co-localisation.

| Electrophysiology
Rats were anaesthetised using isoflurane inhalation. The brain was removed rapidly and immersed in ice-cold sodium-free solution. 22 Acute 300-μm-thick coronal slices containing the CeA were pre-
Following two quick washes in phosphate-buffered saline, brain slices were dehydrated in 50% (5 minutes To test the effects of IL-6, rats were injected with IL-6 (0.2 or 1 μg per 0.5 μL) or vehicle (aCSF). Rats were food-restricted to 50% of their normal intake overnight. Chow intake was measured 1, 3 and 6 hours after injection. Each treatment was counterbalanced, where each condition was separated by a minimum of 2 days.

| Statistical analysis
Graphs were constructed using prism, version 8 (GraphPad Software Inc., San Diego, CA, USA). Statistical analyses were conducted using Students t test. P < 0.05 was considered statistically significant. Using the novel RedIL6 reporter mouse strain, we found a pattern of fluorescence similar to that with the anti-IL-6 antibody staining ( Figure 3A,B).

| GLP-1 is present in synapses close to IL-6 fluorescent cells
PPG-reporter (Venus) mice were co-stained with IL-6 and synaptophysin antibodies to determine whether PPG-synapses were present near IL-6 cells in the CeA (Figure 4). We found examples of cells where triple co-localisation was present, suggesting that some PPG-fibres formed synapses with IL-6-immunoreactive cells ( Figure 4D-F).

| Central Ex-4 administration increases IL-6 mRNA expression in the CeA
Administration of Ex-4 (3 mg kg -1 ) to the lateral ventricle of mice resulted in a statistically significant 50% increase in IL-6 mRNA-expression after 90 minutes compared to vehicle (aCSF) controls (P = 0.04) ( Figure 3C).

| IL-6Rα is present in NeuN-immunoreactive cells in the central amygdala and IL-6 stimulates the firing rate of neurones in the capsular part of the CeA
IL-6Rα antibody was co-stained with the neurone marker NeuN. We found co-localisation between the two antibodies in the CeA, mainly in  Figure 5A,B). Cell counting showed that approximately 50% of IL-6Rα-immunoreactive cells were also NeuN-immunoreactive ( Figure 5). Conversely, approximately 50% of the NeuN-immunoreactive cells were also IL-6Rα-immunoreactive ( Figure 5C).
Given that we found IL-6 and IL-6Rα in the CeA, we next set out to investigate whether IL-6 could exert effects in CeA (ie, if the IL-6Rα found there were functional). Accordingly, IL-6 was added to neurones of the CeA of acute rat brain slices. Action current firing was then recorded by loose-patch method. A control period of 5 minutes at the start of the recordings showed that most of these cells fired spontaneously in bursts. The burst frequency was 0.15 ± 0.10 Hz, the number of spikes in a burst was 12 ± 6 and the intraburst frequency of the action current spikes was 6.6 ± 2.34 Hz.
Application of a single bolus of IL-6 (1 nmol L -1 ) rapidly increased the firing rate up to 356 ± 48% of the baseline frequency (2.09 ± 0.43 Hz; 12 neurones/12 brain slices/five rats) ( Figure 5E) and this significant (P < 0.05, Student's t test) elevation was mostly because the decrease in the time between bursts resulted in a high-frequency tonic firing.

| IL-6 administration to CeA slightly reduces 1hour chow intake in overnight-restricted rats
Administration of 0.2 or 1 μg ( Figure 6) of IL-6 to the CeA in rats slightly decreased 1-hour chow intake ( Figure 6A).
By contrast to the 1-hour time point, there was no effect on 3hour ( Figure 6B) or 6-hour ( Figure 6C) chow intake after IL-6 administration. The reduction of food intake was seen in animals that had been fasted overnight, whereas there was no effect in ad lib. fed animals (not shown). Administration of an IL-6/CGRP cocktail did not potentiate the suppression of food intake seen in IL-6 treated and food-restricted rats (not shown).

| Potential interactions between GLP-1 and IL-6 in the CeA
Formerly, the amygdala has been seen mainly as a nucleus important for fear-and stress conditioning. 37 However, recently, it has also been implicated as important for the regulation of energy balance, including In the present study, we used RedIL6, an IL-6 reporter mouse that was described recently. 42 Briefly, the RedI6 reporter mouse was containing fibres are likely to originate in the NTS, which is a major location of GLP-1 producing neuronal cell bodies. 25

| The presence of functional IL-6Rα in the CeA
IL-6Rα immunoreactivity was present in over half of the CeA cells with NeuN (a neuronal marker) immunoreactivity. This is consistent with the previously reported co-localisation of these two antibodies in both the hypothalamus and hindbrain. [13][14][15][16]55 We next determined whether these receptors were functional. Using electrophysiology, we found that these receptors indeed responded to IL-6 treatment with an increased firing rate and that this effect could be attenuated by an IL-6 neutralising antibody.
Further support for functional IL-6Rα in CeA came from the finding that administration of IL-6 to the CeA partially inhibited feeding.
The latter finding is in line with studies by Timper et al 6 reporting a similar effect after i.c.v. administration. However, further studies are needed to fully determine whether or not IL-6 is an important player in amygdala energy balance regulation.

| Summary
Taken together, our data indicate that IL-6 is present in approximately 40% of GLP-1R cells in the CeA and also that PPG-fibres appear to synapse with IL-6-immunoreactive cells in this nucleus.
Moreover, i.c.v. administration of Ex-4 induced an increase in IL-6 mRNA in this nucleus. Approximately 50% of the neurones in the CeA expressed the ligand binding receptor IL-6Rα. The administration of exogenous IL-6 to CeA brain slices caused electrophysiological stimulation of IL-6Rα, an effect that is quite rapid and therefore likely to be exerted directly. Moreover, treatment with IL-6 to CeA of overnight fasted mice reduced food intake, further supporting a biological role for IL-6Rα in the CeA, although further studies are needed to fully clarify this role. Several brain areas involved in regulation of energy balance appear to contain IL-6 ligand (present study, as well as Fredrik, A,Jansson, JO, unpublished results) and IL-6Rα, [13][14][15][16]56 which, in some cases, have been shown to be functional, as indicated by their responsiveness to IL-6 treatment. 13,55 Deprivation of endogenous IL-6 by different means results in metabolic effects, including an increased fat mass. 10,17,21,57 Moreover, the results obtained in several other experimental paradigms are accordance with the effects of IL-6 in the CNS on energy balance. [58][59][60] Taken together, these findings are in line with the hypothesis that IL-6 could act as a neuropeptide regulating energy balance in different parts of the brain.

ACK N OWLED G EM ENTS
We thank the Centre for Cellular Imaging at the University of