Background
The prevalence of obesity and dyslipidemia is currently increasing, resulting in elevated morbidity of metabolic syndromes [
1]. Obesity is closely related to changes in the serum concentration of important lipid biomarkers, as well as to dyslipidemia and the development of complications such as arteriosclerosis [
1,
2]. Despite the large number of patients suffering from these conditions, the factors contributing to dyslipidemia and their effective treatment remain to be elucidated.
The cytokine, interleukin-18 (IL-18), was originally identified as an interferon (IFN)-γ-inducing proinflammatory factor. However, there is increasing evidence to support it has non-immunological effects on physiological functions [
3,
4]. IL-18 is produced as an inactive 24-kDa precursor and is processed by inflammasomes into an active 18-kDa mature form [
5‐
8]. Previous studies have reported that IL-18-deficient mice developed hyperphagia, obesity and insulin resistance [
9]. Moreover, these mice presented with dyslipidemia, non-alcoholic fatty liver diseases (NAFLD) and non-alcoholic steatohepatitis (NASH), and they were also related to kidney function [
10,
11]. In human studies, the serum concentration of IL-18 was significantly higher in patients with type 2 diabetes mellitus and with metabolic syndrome than in healthy controls [
12,
13].
Brown adipocytes (BAs) have been shown to be differentiated from myogenic factor 5 (
Myf5)-positive progenitors, and PR domain containing 16 (PRDM16), a transcription regulator of differentiation, induced differentiation into BAs and not skeletal muscle [
14]. PRDM16 acts as a transcription co-regulatory factor by activating peroxisome proliferator-activated receptor γ (PPARγ), which activates the PPARγ coactivator 1-α (
Pgc1α) promoter [
15,
16]. BAs and brown adipose tissue (BAT) express the unique molecule, uncoupling protein 1 (UCP1), which uncouples ATP production from oxidative phosphorylation, resulting in thermogenesis in mitochondria [
17‐
20]. BAT is indispensable for basal and inducible energy generation in mammals [
21,
22]. In addition, not only infant, but also adult humans have activated BAT for non-shivering thermogenesis. BAT is a very insulin-sensitive tissue, and chronic high-fat diet resulting in insulin resistance caused reduced glucose uptake by BAT in rodents, which led to diabetes [
21,
23]. Decreased BAT activation has been shown in obese humans, and BAT dysfunction has been associated with obesity and metabolic diseases [
24‐
26]. Therefore, BAT has an important role in maintaining human energy metabolism, especially the lipid balance.
A previous study has demonstrated that
IL-
18-knockout (
Il18−/−) mice have a remarkably increased body weight accompanied by high insulin resistance and dyslipidemia [
9,
10]. We have previously shown that IL-18 administration not only remarkably improved dyslipidemia, NAFLD and NASH in
Il18−/− mice, but it also inhibited body weight gain in
Il18+/+ mice [
10]. These results suggest there are certain relationships between IL-18 and glucose and lipid homeostasis. However, the pathophysiological mechanism and relationship between dyslipidemia and adipose tissue remain unclear.
Therefore, we investigated whether IL-18 promotes efficient use of energy through lipid metabolism in BAs and BAT. Furthermore, we examined whether IL-18 affects the differentiation and growth of BAs and BAT. To verify these hypotheses, we analyzed the role of IL-18 in the regulation of lipid concentration and cell growth in BAs and BAT as follows: (i) histopathological examination of BAs and BAT from Il18+/+ and Il18−/− mice; (ii) analysis of primary cell cultures was performed in vitro, and thermogenic and adipogenic genes were compared; (iii) we analyzed the molecular mechanisms affected during metabolic diseases; and (iv) we assessed the lipid-normalizing effect of recombinant IL-18 (rIL-18).
Methods
Animals
Il18−/− male mice were generated on the C57Bl/6 mouse background as previously described [
27]. Littermate C57Bl/6
Il18+/+ male mice were used as controls. Mice were housed in groups of 3–5 in polycarbonate cages placed in a colony room maintained at a constant temperature (22 ± 1 °C) and humidity (50–60%), under a 12-h light/dark cycle (lights on at 8 a.m) with free access to standard food (MF, Oriental Yeast Co., Ltd., Tokyo, Japan) and water. All mice were sacrificed at 10 a.m. Samples from
Il18+/+ and
Il18−/− mice for molecular, biochemical and histological analyses were taken at the same time points with n = 12–18, excluding in vitro and rIL-18 treatment experiments. Five to six and three mice per group were included in the short- and long-term rIL-18 treatment groups, respectively. The rIL-18 treatment details are described in the section, “
Short- and long-term treatment of mice with rIL-18”.
Animal experiments were conducted according to the “Guide for Care and Use of Laboratory Animals” published by the National Institutes of Health (NIH), and were approved by the Animal Care Committee of Hyogo College of Medicine (#28041 and #14-020).
Cell culture
The stromal vascular fraction containing adipose precursor cells was dissected from the interscapular brown fat pads of the Il18+/+ and Il18−/− male mice at 6 weeks of age. For cell culture, tissues from 5 to 9 animals were pooled and the equivalent of 1.5 animals was used for one culture. The tissues were minced and incubated in a buffer containing 2 mg/ml collagenase Type I (Wako Pure Chemical Industries, Ltd., Osaka, Japan) and 2% bovine serum albumin for 1 h at 37 °C in a shaking bath at 100 cycles/min. Digested tissues were filtered through a 100-μm cell strainer (BD Biosciences, Tokyo, Japan) and centrifuged at 1500 rpm (440×g) for 10 min. The cell pellet was treated with an ammonium–chloride–potassium (ACK) solution and centrifuged again. The cells were suspended in Dulbecco’s modified Eagle’s high glucose medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Nacalai Tesque, Inc., Kyoto, Japan), and seeded on 6-well plates after filtration through a 40-μm cell strainer (BD Biosciences). Then, cells were plated overnight, and attached cells were washed with phosphate-buffered saline (PBS) and grown for about 3 days in a humidified atmosphere of 5% CO2 and 20% oxygen until confluence. When the cells reached confluence, 0.5 mM 3-isobutyl-1-methylxanthine (IBMX) and 1 μM dexamethasone (Dex) were added to induce differentiation with or without rIL-18 (100 ng). Then, the cells were maintained in differentiation medium (10 μg/ml insulin, 50 nM triiodothyronine) for 7 days; the medium was changed every 2–3 days. The fully differentiated cells in the presence or absence of rIL-18 were harvested on day 7.
Histological analysis
In the in vitro study, fully differentiated BAs were fixed with 4% periodate-lysine-paraformaldehyde for 1 h and processed for histological staining with Oil Red O.
In the in vivo experiment, three to five mice per group were used for histopathological analysis. Mice were deeply anesthetized with isoflurane and perfused transcardially with periodate-lysine-paraformaldehyde fixative at 10 a.m. The fixed BATs were removed, cut into small pieces and immerged in the same fixative at 4 °C overnight. The specimens were processed for histological staining by embedding sections in paraffin for hematoxylin and eosin staining. Stained sections were mounted, and pathological diagnosis was determined in a blinded fashion by specialists. Stained cells and tissue sections were photographed using an optical microscope and CCD camera (AX-80 and DP-71, Olympus, Tokyo, Japan).
Molecular analysis
The protocols for sample collection, mRNA purification, microarray, Ingenuity
® Pathway Analysis (IPA; Ingenuity
® Systems, Redwood, CA, USA) and quantitative reverse transcription PCR (qRT-PCR) have been described previously [
10,
11,
28]. Mice were euthanized by decapitation at 10 a.m, and BATs were removed and immediately placed in liquid nitrogen, and then kept at − 80 °C until use.
Total RNA was purified from mouse BAT using a Sepasol-RNA I Super kit (Nacalai Tesque, Kyoto, Japan) according to the manufacturer’s instructions, and treated with 5 units of RNase-free DNase I at 37 °C for 30 min to remove genomic DNA contamination. After phenol/chloroform extraction and ethanol precipitation, total RNA was dissolved in deionized distilled water. RNA concentrations were determined by spectrophotometry.
For microarray analysis, expression profiling was performed using SurePrint G3 Mouse GE 8 × 60 K Microarray G4852A (Agilent Technologies Inc., Santa Clara, CA, USA). Twelve microarray analyses as a one-color experiment were performed in biological triplicates. Each gene expression profile was compared between Il18+/+ and Il18−/− mice at 6 and 12 weeks of age. Total RNA (200 ng) was reverse transcribed into double-stranded cDNA using AffinityScript multiple temperature reverse transcriptase (Agilent Technologies Inc.), and amplified. The resulting cDNAs were used for in vitro transcription by T7-polymerase and labeled with cyanine-3-labeled cytosine triphosphate (Perkin Elmer, Wellesley, MA, USA) using a Low Input Quick-Amp Labeling Kit (Agilent Technologies Inc.). After the labeled cDNAs were fragmented, each cDNA sample was hybridized to SurePrint G3 Mouse GE 8 × 60 K Microarray (Agilent Design # 028005). After washing, the slides were scanned with an Agilent Microarray Scanner (G2505C). Feature extraction software (version 10.5.1.1; Agilent Technologies Inc.) was used to convert images into gene expression data. For microarray data analysis, raw data were imported into Subio platform version 1.18 (Subio Inc., Kagoshima, Japan), and raw intensity data were normalized to the 75th percentile intensity of probes above the background levels (gIsWellAbove = 1). The BAT genes in Il18+/+ and Il18−/− mice were defined to show signal ratios of > 2.0-fold and < 0.5-fold. Details of the microarray analysis and results can also be found in the Gene Expression Omnibus (GEO; Accession No. GSE64308).
IPA software was used for microarray analysis for the interpretation of gene expression data. The IPA network explorer was used to detect relevant interactions among the
Il18+/+ and
Il18−/− genes, and identify the shortest direct paths between genes [
10,
28].
For qRT-PCR, BAT samples at 6 and 12 weeks of age were obtained from the same animals used for the microarray analysis. Total RNA (10 ng/reaction) was used in the RNA-direct SYBR Green Real-Time PCR Master Mix: One-step qPCR kit (Toyobo Co. Ltd., Tokyo, Japan). Samples were run in duplicate reactions in 96-well plates. Median threshold cycle values were used to calculate the fold changes (FC) between the samples of the groups. The FC values were normalized to glyceraldehyde 3-phosphate dehydrogenase (
Gapdh) or β-actin (
Actb) levels. The following temperature profile was used according to the manufacturer’s instructions: 30 s at 90 °C and 20 min at 61 °C for reverse transcription, followed by 45 cycles of 98 °C for 1 s, 67 °C for 15 s and 74 °C for 35 s. The primer sequences for qRT-PCR are shown in Additional file
1.
Western blotting
Mouse BAs/BATs were minced in lysis buffer (1% Nonidat P-40, 20 mM Tris–HCl, pH 8.0, 150 mM NaCl, 10% glycerol) containing protease inhibitor cocktail (Complete™, Roche, Mannheim, Germany), and then homogenized on ice using a sonicator (Sonifier II, Branson, MO, USA). Each lysate was centrifuged at 13,000×
g for 3 min and the supernatant was collected. The protein concentration of each specimen was determined with a Bradford protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). Samples were denatured in Laemmli sample buffer for 5 min at 95 °C, electrophoresed in a 12.5% sodium dodecyl sulfate polyacrylamide gel, and transferred onto polyvinylidene difluoride membranes (Hybond-P, Amersham Bioscience, Little Chalfont, UK). The antibodies used were monoclonal rabbit anti-β-actin (ACTB; Cat. No. 5125S; Cell Signaling Technology, Inc., Danvers, MA, USA), adipose triglyceride lipase (ATGL; Cat. No. 2439, Cell Signaling), fatty acid binding protein (FABP4; Cat. No. 3544, Cell Signaling), fibroblast growth factor 21 (FGF21; Cat. No. LS-B5864, LifeSpan Biosciences, Inc. Seattle, WA, USA), IL-18 (IL18; Cat. No. D046-3, Medical & Biological Laboratories Co., Ltd., Tokyo, Japan), monoclonal rabbit anti-GAPDH (Cat. No. 3683S; Cell Signaling Technology, Inc.), PGC1α (Cat. No. Ab3242, Merck Millipore, MA, USA), Ser563-phosphorylated hormone-sensitive lipase (pHSL563; Cat. No. 4139, Cell Signaling), Ser565-phosphorylated HSL (pHSL565; Cat. No. 4137, Cell Signaling), PPARγ (Cat. No. 2430, Cell Signaling), PRDM16 (Cat. No. Ab106410, Abcam plc, Cambridge, UK) and UCP1 (Cat. No. Ab10983, Abcam plc). Membranes were blocked with 1% bovine serum albumin in PBS containing 0.1% Triton X-100 (TPBS), incubated with primary antibodies at 4 °C overnight, followed by incubation with horseradish peroxidase-conjugated secondary antibodies (#NA9340V and #RPN1025, GE Healthcare, Buckinghamshire HP7 9NA, UK). Washing with TPBS was performed after each treatment. Antibody reactions were captured using the photo-image analyzer, LAS-4010 (Fuji Photo Film Co., Ltd., Tokyo, Japan). The density of specific protein bands was measured with ImageJ (
http://rsbweb.nih.gov/ij/, version 1.6), and the results obtained were normalized to β-actin levels. The mean of measured bands in the controls was set to one. We also assessed positive controls.
Short- and long-term treatment of mice with rIL-18
To determine the response to rIL-18 treatment, mice were administered 2 µg of rIL-18 dissolved in saline containing heat-inactivated normal mouse serum (0.5%). They were injected twice a week via the caudal vein for 2 weeks (short-term study) from 10 weeks of age, and for 12 weeks (long-term study) from 12 or 37 weeks of age, as previously reported [
10]. For control experiments, saline was injected using the same procedure. Five to six and three mice per group were included in the short- and long-term treatment groups, respectively.
Statistical analysis
All results are expressed as mean ± SD. Sigmaplot™ (version 11.0 Systat Software, Inc., San Jose, CA, USA) was used for all statistical analyses. Body weight, serum measurement, qRT-PCR and western blotting were analyzed by the Student t-test or Mann–Whitney U-test. Analysis of rIL-18 administration was examined by two-way analysis of variance (ANOVA). Correlations between microarray and qRT-PCR were examined by Spearman’s rank correlations test. Differences were considered statistically significant when p < 0.05. All analyses were performed at least twice to confirm the results.
Discussion
In the present study we revealed the following novel pathophysiological features of IL-18 functions: (1) BAs extracted from Il18−/− mice were more differentiated than those from Il18+/+ mice; (2) this phenomenon was induced by upregulation of the differentiation inducer, PRDM16, and resulted in not only upregulated PRDM16-dependent genes such as Pparg and Pgc1a, but also in upregulated PRDM16-independent genes such as Fgf21; (3) BAT in Il18−/− mice developed severe adiposity during growth compared with that in Il18+/+ mice; (4) several molecules related to adiposity were affected by IL-18 deficiency; and (5) some of these abnormal conditions were recovered by rIL-18 administration, which also affected some of the identified molecules.
Regarding BA differentiation, in
Myf5-positive progenitors, PRDM16 led not only to skeletal muscle differentiation, but also to BA differentiation of brown preadipocytes. Furthermore, PPARγ was indispensable for BA differentiation [
14,
30,
31]. PGC1α, a coactivator of PPARγ, is a crucial regulator of mitochondrial biogenesis and oxidative metabolism, and it activates thermogenic functions [
32]. This heat production depends on UCP1 activity [
33]. Moreover, PRDM16 increases the transcriptional activity of PGC1α and PPARγ [
34]. FGF21, a member of the FGF superfamily, is mainly released from the liver and has recently emerged as a novel regulator of glucose and lipid metabolism [
35,
36]. FGF21 has been shown to promote thermogenic activity in neonatal BAT and in isolated BAs [
37]. Recently, a study has demonstrated that FGF21 was secreted from BAT and that BAT has an endocrine role during thermogenic activation [
38].
According to our results, BAs from
Il18−/− mice were more differentiated than those from
Il18+/+ mice (Fig.
1a). The expression of PRDM16, PGC1α, PPARγ and UCP1 in
Il18−/− mice was significantly higher than that in
Il18+/+ mice (Fig.
1b, c). Additionally, the relative expression of FGF21, which is a thermogenic activator in BAT, in
Il18−/− mice was significantly higher than that in
Il18+/+ mice (Fig.
1b, c). These results suggest that IL-18 deficiency directly induces transcription regulators, resulting in promotion of BA differentiation, thermogenesis activation and increased energy production.
Similar to the differentiation and thermogenic regulators, FABP4 expression in BAs extracted from
Il18−/− mice was increased compared with that in BAs from
Il18+/+ mice (Fig.
1b, c). FABP4 is expressed in adipose tissue, and is related to obesity and diabetes mellitus [
39]. Moreover, high expression of FABP4 in adipose tissue has been suggested to increase insulin resistance [
40,
41]. FABP4 is induced by PPARγ, and these two molecules interact with each other and have opposite effects on energy metabolism [
42,
43]. Furthermore, in vivo, the amount of food and energy intake of
Il18−/− mice increased compared with the controls, which led to obesity, increased insulin resistance and diabetes mellitus [
9]. We have also previously found that
Il18−/− mice developed dyslipidemia, NAFLD and NASH [
10]. Our results suggest that higher expression of FABP4 in BAs is one of the key factors in these metabolic disorders.
ATGL and HSL in adipose tissue are key enzymes involved in intracellular degradation of triacylglycerol. Overexpression of ATGL in adipose tissue increased thermogenesis, resulting in higher energy expenditure and resistance to obesity [
44]. Moreover, ablation of ATGL in BAT led to transformation of BAT into white adipose tissue (WAT)-like tissues and deceased the expression of BAT-related molecules, resulting in suppressed thermogenesis [
45]. HSL is activated by phosphorylation at Ser563, leading to lipolysis [
46,
47]. However, HSL phosphorylation at Ser565 suppresses HSL phosphorylation at Ser563 and inhibits HSL activity [
47,
48]. In our study, the expression of ATGL, pHSL563 and pHSL565 in cells from
Il18−/− mice was significantly increased compared with that in cells from
Il18+/+ mice (Fig.
1d). These results suggest that lipolysis in BAs from
Il18−/− mice was more robustly promoted than in BAs from
Il18+/+ mice through HSL activation.
Further examination of other molecules related to BA differentiation and thermogenesis such as
Elovl3 and
Adrb3 was performed by qRT-PCR.
Elovl3 is a specific marker of BAT, CD36 is indispensable for thermogenesis under low temperatures, and thermogenesis is promoted by
Adrb3 signaling [
49].
Dio2 is a prohormone that converts tyrosine into triiodothyronine (T3) [
50]. Signal transducer and activator of transcription 3 (STAT3) is phosphorylated by Leptin [
51]. The expression of all of these molecules, except for
Dio2, was increased under IL-18 deficiency, whereas
Dio2 was decreased (Fig.
1e). These results suggest that differentiation and thermogenesis in BAs from
Il18−/− mice were accelerated though decreased expression of
Dio2, which may stem from a negative feedback due to T3-induced differentiation.
In vivo, while the amount of food and energy intake of
Il18−/− mice was higher compared with the controls, resulting in obesity, insulin resistance and diabetes mellitus, the metabolic rate of
Il18−/− mice was similar to that of the controls [
9]. In addition, our previous study has revealed that
Il18−/− mice showed dyslipidemia, resulting in NAFLD and steatohepatitis [
10]. Especially, hypertriglyceridemia occurred at 6 weeks of age, which was the same age at which we extracted adipose precursor cells for primary cell culture. Our results showed that in
Il18−/− mice, BAT stored more lipids with age (Fig.
2a). Moreover, the gene expression of
Prdm16,
Ucp1,
Pgc1a,
Fgf21 and
Dio2 at 6 and 12 weeks of age in
Il18−/− mice was similar to the in vitro results (Figs.
1b–e,
2b, c). Thus, the in vivo results were similar to the in vitro results, suggesting that the histological and molecular differences were due to negative feedback of dyslipidemia.
To examine the molecular mechanism of lipid accumulation, comprehensive analysis was performed by microarray. Similar to our previous liver analyses, the function of ‘Quantity of adipose tissue’-related genes, which was common between 6 and 12 weeks of age, was extracted automatically, and
Apoc3,
Insig1 and
Vdr were detected (Fig.
3).
Apoc3 is closely associated with high density lipoprotein (HDL), modulates its function and induces BAT metabolic activation [
52]. INSIG1 affects cholesterol biosynthesis [
53]. In addition, increased expression of
Insig1 in the liver and in WAT may reduce cholesterol biosynthesis, resulting in obesity and dyslipidemia [
54]. In our study,
Apoc3 and
Insig1 expression in
Il18−/− mice at 6 and 12 weeks of age was significantly increased compared with that in
Il18+/+ mice (Fig.
3b). Thus, our results suggest that increased
Apoc3 expression may be induced by excessive lipid accumulation in BAT, and that increased
Insig1 was one of the mediators promoting fat accumulation.
As for
Vdr, energy expenditure is increased in
VDR-knockout mice, however, it is decreased in VDR-overexpressing mice [
55,
56]. At 6 weeks of age,
Vdr expression in
Il18−/− mice was increased, but at 12 weeks it significantly decreased (Fig.
3b). Therefore, our results suggest that at 12 weeks of age, decreased
Vdr was a response to excessive lipid accumulation in BAT of
Il18−/− mice.
Although protein expression in BAs from
Il18−/− mice, which was significantly different such as UCP1 (Fig.
1), was not affected by rIL-18 administration (Additional file
6), in
Il18−/− mice, dyslipidemia was improved by short-term (2 weeks) rIL-18 administration [
10]. Contrary to our expectation, the expression of PRDM16, FGF21 and PGC1α in BAT was similar. However, the increased expression of UCP1 in
Il18−/− mice was decreased by rIL-18 administration and there were no significant differences between the
Il18−/− and
Il18+/+ mice (Fig.
4a, b). Regarding the other molecules,
Adrb3 and
Elovl3 were decreased and
Vdr was increased after short-term rIL-18 treatment (Fig.
4c, d). Our results suggest that there was no direct effect of IL-18 on BAs, and lipid normalization by rIL-18 may result in decreased expression of
Adrb3,
Elovl3,
Vdr and UCP1.
We have previously revealed that long-term intravenous administration of rIL-18 inhibited the body weight gain of
Il18+/+ mice and prevented the onset of NASH in
Il18−/− mice [
10]. Therefore, long-term exposure to excessive IL-18 expression may affect lipid homeostasis. Even though there was no effect on body weight in
Il18−/− mice, the expression of PGC1α and UCP1 in
Il18+/+ mice after long-term rIL-18 treatment was significantly lower compared with that in mice treated with saline (Fig.
5a, b) [
10]. Moreover, the size of cells containing lipids in BAT from
Il18−/− mice treated with rIL-18 from 37 to 49 weeks of age was much smaller than that of cells from
Il18+/+ mice (Figs.
2a and
5c). Therefore, similar to previously shown liver results [
10], IL-18 was indispensable for normal lipid metabolism not only in the liver, but also in BAT, and may contribute to the development of novel treatment options for lipid dyshomeostasis.
Limitations of the present study include a limited number of mice that received long-term (12 weeks) rIL-18 administration because of animal ethical restrictions enforced by the animal ethical committee in our college, which may be insufficient for obtaining statistical significance. Even though the direct effect on BAs under IL-18-deficient conditions in vitro was clarified, there were inconsistent results between the in vitro and in vivo experiments. Moreover, this study only focused on BAT. However, we are currently investigating the basic relationship between IL-18 and physiological homeostasis in other organs [
10,
11]. Furthermore, additional analysis is required, for example, analysis of the metabolic function of IL-18 in BAT under cold exposure should be performed.
Authors’ contributions
KY, KM, TH, KI and HM conducted the study. KY, SM, SK-O, KM, KI, NG, YE, DO and HN performed the experiments. KY, KN, YW, HY and HM analyzed the data. KY and HM prepared the tables and figures and wrote the manuscript. All authors read and approved the final manuscript.