Resolvin RvD2 reduces hypothalamic inflammation and rescues mice from diet-induced obesity

All of the reagents for SDS-polyacrylamide gel electrophoresis and immunoblotting were from Bio-Rad (Richmond, CA, USA). HEPES, phenylmethylsulfonyl fluoride, aprotinin, dithiothreitol, Triton X-100, Tween 20, glycerol and BSA (fraction V) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). The antibodies against GPR18 (sc79503), NPY (sc133080) and Iba1 (sc28530) were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The reagents for chemiluminescence protein labelling in immunoblots were purchased from Amersham (Aylesbury, UK). FITC-conjugated anti-rabbit (sc2012), FITC-conjugated anti-goat (sc2024), Cy3-conjugated goat anti-mouse (ab6946), Cy3-conjugated donkey anti-goat (ab6949), rhodamine-conjugated anti-rabbit (sc2091) and rhodamine-conjugated anti-goat (sc2094) antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The lipid mediator resolvin D2 (sc-351847A) was from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Reagents for the real-time PCR analysis were from Invitrogen (Carlsbad, CA, USA) and Applied Biosystems (Foster City, CA, USA). TaqMan primers for PLA2 (Mm00448161_m1), 15-LOX (Mm00507789_m1), 5-LOX (Mm01182747_m1), GPR18 (Mm01224541_s1), TNFα (Mm00443258_m1), IL1β (Mm00434228_m1), IL6 (Mm00446190_m1), IL10 (Mm01288386_m1), PGC1α (Mm00447183_m1), UCP1 (Mm01244861_m1) and glyceraldehyde-3-phosphate dehydrogenase (GAPD) (#4352339E) were obtained from Applied Biosystems.

Male Swiss mice originally imported from Jackson Laboratory and currently bred at the University of Campinas Breeding Center were used in the study. The investigation was conducted in accordance with the principles and procedures described by the National Institutes of Health Guidelines for the Care and Use of Experimental Animals and was previously approved by the University of Campinas Ethical Committee (ID 2011/2341-1). The animals were maintained at 21 ± 3 °C, on a 12-h artificial light/dark cycle and housed in individual cages. By 5 weeks old, the mice were assigned in three groups, with the same body weight mean: standard rodent chow diet (CT), high-fat diet (HF; 60% of energy value from fat, Table 1) and high-fat diet supplemented with omega 3 (HFS; HF supplemented with 20% omega 3, Table 1) by the time specified in the protocol.

For evaluation and characterisation of biosynthetic pathways of RvD2, generated enzymatically from DHA, mice were fed for 16 weeks on either the chow diet or the HF. By the end of this period, the hypothalamus was removed and RNA extracts were employed in the real-time PCR analysis. In another set of experiments, mice were assigned to the HF or chow diet for 16 weeks and another group of mice was assigned to the HF for 8 weeks following 8 weeks on a HF supplemented with 20% omega 3 (HFS). Food intake and body mass were measured during this period. At the 15th week, the animals were subjected to the intraperitoneal glucose tolerance test. Subsequently, the hypothalamus was removed and employed in the MALDI-MSI analysis (as described below) and real-time PCR analysis for the identification and characterisation of RvD2 in this experimental model. In order to determine the impact of direct intracerebroventricular (icv) injection of DHA on the RvD2 system in the hypothalamus, mice were fed a HF for 4 weeks and were subsequently stereotaxically instrumented in a Stoelting stereotaxic apparatus to receive a cannula placed in the lateral hypothalamic ventricle, using the following stereotaxic coordinates: anteroposterior 0.34 mm, lateral 1.0 mm and dorsoventral 2.2 mm. The correct position of the cannula was tested 5 days after surgery by evaluation of the thirst response elicited by intracerebroventricular angiotensin II (10⁶ M). After 1 week, icv-cannulated mice were treated once a day for 4 days with 2 μL of saline or 2 μL of DHA (5, 10 or 20 ng). In another experiment, mice were fed a HF for 8 weeks and were subsequently stereotaxically instrumented, as previously described in this section. After 1 week, the icv-cannulated mice were treated once a day for 11 days with 2 μL of saline or 2 μL of RvD2 (3 or 50 ng). The doses of RvD2 were defined based on a previous study [9]. Food intake and body mass were measured during the treatment period. On the seventh and tenth days of treatment, the animals were subjected to an intracerebroventricular leptin tolerance test and intraperitoneal glucose tolerance test, respectively. In each group, some mice were randomly selected for indirect calorimetry and spontaneous physical activity measurements. At the end of the experiment, the hypothalamus and brown adipose tissue were collected for real-time PCR analysis. Outlines of the different protocols are depicted in the figures that show the respective experiments.

The intraperitoneal glucose tolerance test (ipGTT) and intracerebroventricular leptin tolerance test (icvLTT) were performed on food-deprived (6 h) non-anaesthetised mice. Blood glucose levels were measured using an OptiumTM mini (Abbott Diabetes Care, Alameda, CA, USA) handheld glucometer with appropriate test strips. For ipGTT, a solution of 20% glucose (2.0 g/kg body weight) was administered into the peritoneal cavity. Blood samples were collected from the tail vein at 30, 60, 90 and 120 min for determination of glucose concentrations. The area under the curve (AUC) was calculated using these values. For icvLTT, food intake was measured 2, 4, 6 and 12 h following icv injection of leptin (10⁻⁶ M). These values were used for determining leptin sensitivity.

Total RNA was extracted using a commercially available acid-phenol reagent Trizol (Invitrogen Corp.). RNA concentration, purity and integrity were confirmed spectrophotometrically using a Nanodrop (ND-1000; Nanodrop Technologies, Wilmington, DE). The first-strand cDNA was synthesised using SuperScript III reverse transcriptase and random hexamer primers, as described in the manufacturer’s protocol (Invitrogen Corp.) The quantitative PCR was run to determine the expression of TNFα, IL1β, IL6, IL10, PLA2, 15-LOX, 5-LOX and GPR18 in the hypothalamus of mice and to determine the expression of PGC1α and UCP1 in the brown adipose tissue (BAT) of mice using primers supplied with commercially available assays from Applied Biosystems. The endogenous gene was GAPD (glyceraldehyde-3-phosphate dehydrogenase (Applied Biosystems)). A real-time PCR analysis of gene expression was carried out in an ABI Prism 7500 sequence detection system (Applied Biosystems). The optimal concentration of complementary DNA and primers and the maximum efficiency of amplification were obtained through a 5-point, two-fold dilution curve analysis for each gene. Amplification was performed in a 20-μL final volume containing 40–50 ng of reverse-transcribed RNA according to the manufacturer’s recommendations using the TaqMan PCR master mix. Real-time data were analysed using the Sequence Detector System 1.7 (Applied Biosystems). Results were expressed as the relative transcript amount, as previously optimised [25].

For the histological analysis, hypothalamic tissue samples were frozen sectioned and processed routinely for immunofluorescence staining. Coronal sections of the hypothalamus (5 μm) were double-labelled with anti-GPR18 antibodies and specific primary antibodies against markers related to the different cell types, including NPY, POMC and Iba1. Thereafter, the sections were incubated with specific FITC or rhodamine-conjugated IgG secondary antibodies. Immunofluorescence imaging was performed to evaluate the distribution of GPR18 in the mouse’s hypothalamus.

Obtained brain tissue sections were set on MALDI-appropriate stainless steel plates (GMS-Thermo, CA, USA) and then coated with a 10 mg/mL (50% methanol to acetonitrile) solution of alpha-cyano-4-hydroxycinnamic acid (CHCA) matrix (Sigma Aldrich, PA, USA). For an even distribution, a customised system with a commercial airbrush was utilised to spray the matrix. The MALDI-LTQ-XL instrument (Thermo Fisher, San Jose, CA, USA) with a tissue imaging feature was utilised in the mass spectrometry data acquisition. Operating conditions were set as follows: 4 μJ laser power, 50 μm raster step size, standardised sample size of 600 × 600 μm focused on the coronal sections of the hypothalamus, three laser shots per step and 30 eV for the helium collision-induced dissociation in fragmentation reactions (MS/MS). A survey scan was performed in the mass-to-charge ratio (m/z) range of 50 to 500. Samples were analysed in the negative ion mode. Compound structures were proposed using MS/MS data supported by software calculations with Mass Frontier (v. 6.0, Thermo Scientific, CA, USA), as well as previous data from the literature [25]. Imaging data were run in replicates for all described animal conditions (chow, HF and HDFω3) and were processed using ImageQuest software (Thermo Scientific, San Jose, CA, USA). Relative quantification was performed using ImageJ (National Institutes of Health, USA–Open Source) on images in grey scale. Since the analysed areas were the same size (in pixel numbers) for all the replicates, ImageJ was able to assign a non-dimensional value for each sample image. This result is based on the intensity of each pixel and can be compared among all samples to determine the relative levels of the desired molecule.

Oxygen consumption/carbon dioxide production and spontaneous physical activity were measured in the fed animals through a computer-controlled, open-circuit calorimeter system (LE405 gas analyser; Panlab-Harvard Apparatus). Mice were singly housed in clear respiratory chambers and room air was passed through chambers at a flow rate of ten times the body weight of each animal. The air-flow within each chamber was monitored using a sensor (Air Supply and Switching; Panlab-Harvard Apparatus). Gas sensors were calibrated prior to the onset of the experiments using primary gas standards containing known concentrations of O₂, CO₂ and N₂ (liquid air). The analyses were performed over a 24-h period. Outdoor air reference values were sampled after every four measurements. Sample air was sequentially passed through O₂ and CO₂ sensors for the determination of O₂ and CO₂ content, from which the measures of oxygen consumption (VO₂) and carbon dioxide production (VCO₂) were estimated. VO₂ and VCO₂ were calculated by Metabolism version 2.2 software based on the Withers equation and are expressed in millilitres per hour per gram. The respiratory quotient was calculated as VCO₂/VO₂. Energy expenditure was estimated as VO₂/body mass (grams) [26].

All results are reported as mean ± SEM. Differences between the treatment groups were evaluated using an unpaired Student t test or a one-way analysis of variance (ANOVA). When the ANOVA indicated significance, a Tukey-Kramer post hoc test was performed (GraphPad Software, San Diego, CA). p < 0.05 was accepted as being statistically significant.

GPR18 was recently defined as the receptor for RvD2 [27]. We employed immunofluorescence staining to determine the expression and distribution of GPR18 in the hypothalamus of lean mice. Figure 1 shows that some, but not all, of the NPY (Fig. 1a) and POMC neurons (Fig. 1b) express GPR18. However, we could not detect any Iba1-positive cells (microglia) expressing GPR18 (Fig. 1c).

High consumption of dietary fats is one of the most important environmental factors leading to obesity [28]. Here, mice were fed either chow or a high-fat diet (HF) that was rich in saturated fats. Then, we evaluated the hypothalamic expression of transcripts encoding for proteins involved in the synthesis and action of RvD2. In Fig. 2a, there is a schematic representation of the main enzymes involved in the synthesis of RvD2. The receptor for RvD2, GPR18 is also depicted in the scheme (Fig. 2a). PLA2 and 15-LOX, which are involved in the initial steps of RvD2 synthesis, are inhibited early after HF introduction and then undergo a significant increase at middle and late phase obesity (Fig. 2b, c). Conversely, 5-LOX, the enzyme involved in the final step of RvD2 synthesis, undergoes an early increase and then returns to levels similar to control by middle and late phase obesity (Fig. 2d). GPR18 is also regulated by dietary fats, undergoing an early reduction, then increasing at middle obesity (8 weeks) and finally reducing again at late obesity (16 weeks) (Fig. 2e).

To measure hypothalamic RvD2, we employed a MALDI method, which was adapted from a mass spectrometry method, as previously described [29]. As depicted in Fig. 2f, g, RvD2, which corresponds to m/z 375, is significantly reduced in the hypothalamus of mice fed for 16 weeks on HF.

The mice were initially fed the HF diet for 8 weeks and then were randomly selected for either continuing on the current HF or transferring to a HF on which lard was substituted by a PUFA-rich oil (HFS) (Table 1); lean controls (CT) were maintained on chow throughout the experiment (Fig. 3a). PUFA substitution resulted in an increased amount of RvD2 in the hypothalamus (Fig. 3b). This was accompanied by a reduction of diet-induced expression of the pro-inflammatory cytokine transcripts TNFα (Fig. 3c) and IL1β (Fig. 3d) in the hypothalamus. In addition, the PUFA-rich diet reduced body mass gain without affecting caloric intake (Fig. 3e, f). The consumption of a PUFA-rich diet was accompanied by improved glucose tolerance, as determined via a glucose tolerance test (Fig. 3g) and by increased sensitivity to insulin, as determined by the constant of glucose decay (kITT) (Fig. 3h) during an insulin tolerance test.

Next, we tested the hypothesis that injecting DHA, the substrate for RvD2 synthesis, directly in the hypothalamus could modulate the proteins involved in its synthesis. For that, mice fed the HF diet for 4 weeks were subjected to an icv cannulation and treated for 4 days, either with saline or with different amounts of DHA (Fig. 4a). As depicted in Fig. 4b–e, DHA exerted a positive effect on the expressions of PLA2, 15-LOX, 5-LOX and GPR18. In the case of 15-LOX, 5-LOX and GPR18, this was a dose-dependent effect. Because most proteins involved in the synthesis/signalling of DHA are rapidly modulated by diet (as depicted in Fig. 2b–e), we evaluated the effect of icv-injected DHA in mice fed the HF for 3 days. Additional file 1: Figure S1 shows that DHA produced increases of 15-LOX, 5-LOX and GPR18.

The consumption of dietary saturated fats induces hypothalamic inflammation and increases body mass [15‐19]. When mice were icv-treated with DHA, there was an increased expression of two anti-inflammatory cytokines in the hypothalamus—IL6 (Fig. 4f) and IL10 (Fig. 4g)—and no modification of the expression of TNFα and IL1β (not shown). This was accompanied by reduced caloric intake (Fig. 4h) and reduced body mass (Fig. 4i). DHA was also capable of inducing the increased hypothalamic expression of IL6 and IL10 in mice fed the HF for 3 days (Additional file 1: Figure S1).

Obese mice were submitted to icv cannulation and treated with RvD2 icv for 11 days (Fig. 5a). Either 3.0 or 50 ng RvD2 was sufficient to increase the hypothalamic expression of GPR18 (Fig. 5b). RvD2 icv was not capable of modifying caloric intake (Fig. 5c); nevertheless, 3.0 ng icv was sufficient to reduce body mass gain (Fig. 5d), while both 3.0 and 50 ng produced a significant reduction of visceral fat (Fig. 5e). The 3.0 ng, but not the 50 ng, dose improved glucose tolerance, as determined by the area under the glucose curve during a GTT (Fig. 5f). The 3.0 ng, but not the 50 ng, dose led to an increased expression of the anti-inflammatory cytokines IL6 (Fig. 5g) and IL10 (Fig. 5h) and no modification of the expression of TNFα and IL1β (not shown) in the hypothalamus. In addition, 3.0 ng RvD2-injected icv increased the expressions of UCP1 (Fig. 5i) and PGC1α (Fig. 5j) in the BAT. This was accompanied by the increased whole body consumption of O₂ (Fig. 5k) and reduced respiratory quotient (Fig. 5m). Resistance to leptin is a hallmark of diet-induced obesity [30]. In the final part of this study, we tested the hypothesis that RvD2 could revert hypothalamic leptin resistance in obese mice. For that, obese mice were initially treated icv with saline or RvD2 and then treated with either saline or leptin. Spontaneous intake of diet was measured over a period of 12 h. As depicted in Fig. 5n, RvD2 was sufficient to significantly reduce food intake in response to leptin.