Dietary Curcumin Intervention Targets Mouse White Adipose Tissue Inflammation and Brown Adipose Tissue UCP1 Expression

Materials

Escherichia coli LPS, peroxisome proliferator-activated receptor-gamma (PPARγ) antagonist GW9662 (M6191), and agonist rosiglitazone (R2408) were purchased from Sigma-Aldrich (L2880; Sigma-Aldrich Corp., St. Louis, Missouri). Murine IFNγ (AF-315-05) was purchased from Peprotech (Quebec, Canada). Human recombinant FGF21 was the product of Novoprotein Scientific Inc. (Summit, New Jersey). PPARα antagonist GW6471 (G5045) and agonist Wy14643 (C7081) were purchased from Santa Cruz Biotechnology (Santa Cruz, California). Curcumin was purchased from Sigma-Aldrich for cell culture experiments or from Organika Health Products (the 95% standardized curcumin extract; Richmond, British Columbia, Canada) for dietary intervention as we have documented previously (12).

Experimental animals

Male C57BL/6J mice (Charles River Laboratories, St. Laurent, Quebec, Canada) were housed under the conditions of a constant temperature (22°C, unless otherwise specified) and a 12-hour light/dark shift with free access to food and water.

For the experiments illustrated in Figure 1A, 45 mice were randomly divided into the following three groups: fed with the chow diet (low-fat diet [LFD], 3.6 kcal/g), HFD (Product# F3282 with 60% fat calories and 5.49 kcal/g; Bio-Serv, Flemington, New Jersey), or HFD plus 1% curcumin. Body weight and food intake were recorded weekly. Intraperitoneal glucose, insulin, and pyruvate tolerance tests were performed at indicated time points (Figure 1A) as we have previously described (15). Mice were sacrificed at the 18th week for blood and tissue collection.

For the experiments listed in Figure 5A, 24 mice were randomly divided into the following three groups: fed with the LFD, HFD, or HFD plus 1% curcumin. The cold challenge was performed after 2, 4, or 6 weeks after corresponding diet feeding. Mice were kept at room temperature prior to being put in a 4°C environmental chamber. Four hours later, rectal temperature recording was started by using an electronic thermometer equipped with a rectal probe (HH63K Stick Type Temperature Transducer and Thermometer; Omega, Quebec, Canada).

For metabolic cage analysis, mice fed HFD or HFD plus curcumin were housed individually in metabolic cages with sufficient food and water. After a 24-hour settlement period, data collection was started. Collected data for the 30-hour period were analyzed by using a lab animal monitoring system (Columbus Instruments Corp., Columbus, Ohio) as previously presented (16).

The animal experiments and protocols were approved by the University Health Network Animal Care Committee and performed in accordance with the guidelines of the Canadian Council of Animal Care.

Cell cultures and luciferase reporter analysis

The isolation of primary adipocytes from male Sprague Dawley rat (Charles River Laboratories) epididymal fat tissues has been previously described (17).

The mouse leukemic monocyte-macrophage Raw264.7 cell line was purchased from ATCC (Manassas, Virginia). M1-like macrophage differentiation was induced by LPS (100 ng/mL) and IFNγ (2.5 ng/mL).

The mouse brown adipocyte cell line (mBAC) was the gift of Shingo Kajimura (18). Cells were cultured with the growth medium (DMEM with 10% FBS). The differentiation medium contained the growth medium and the cocktail (5 µg/mL insulin, 1 nmol/L T3, 0.123 mmol/L indomethacin, 2 µg/mL dexamethasone, and 0.1 mmol/L IBMX).

For mouse peritoneal macrophage isolation, male C57BL/6J mice (8 weeks old) were injected intraperitoneally with 4% thioglycollate (1 mL per mouse). Mice were sacrificed 4 days later for peritoneal macrophage isolation.

The Ucp1-luciferase (LUC) reporter gene construct was described previously by the provider, in which the expression of the LUC reporter is driven by the 3.1-kilobase mouse Ucp1 gene 5' flanking region (19). The LUC assay was conducted as previously described (12).

Histology and immunohistochemistry

Mouse brown, epididymal (WAT), and inguinal (subcutaneous) adipose tissues were fixed in 4% paraformaldehyde followed by paraffin embedding. Tissue sections were stained with hematoxylin and eosin or immune-stained with the UCP1 antibody (15).

Processing of immune cells from epididymal adipose tissue and spleen

Epididymal adipose tissue immune cells were isolated as previously described (20). The cell culture medium was collected for the Luminex multiplex assay (Thermo Fisher Scientific, Waltham, Massachusetts), conducted by the Service Laboratories of University Health Network.

Flow cytometry

Immune cells were stained for 30 minutes with fluorophore-conjugated antibodies against CD206, CD11c, CD86, and CD80 by using recommended dilutions. Cells were identified by using an LSRFortessa cell analyzer (BD Biosciences, Mountain View, California). Data acquired on the flow cytometer were analyzed with FlowJo software (Tree Star; FlowJo LLC, Ashland, Oregon).

Western blotting and real-time reverse transcription polymerase chain reaction

Protein samples were subjected to Western blotting with antibodies against inducible nitric oxide synthase (iNOS) (13120S; Cell Signaling Technology Inc., Danvers, Massachusetts), UCP1 (Cell Signaling 14670S), or β-actin (Cell Signaling 3700S). Methods for RNA extraction, complementary DNA production, and real-time reverse transcription-polymerase chain reaction (RT-PCR) analyses have been described previously (15). RT-PCR primers are listed in Supporting Information Table S1.

Statistical analysis

All data were expressed as the mean ± SEM. Comparisons between groups were performed by using either a two-tailed t test or a one-way analysis of variance (ANOVA). P < 0.05 was considered significant.

Curcumin intervention improves metabolic profiles

We first assessed the effect of curcumin intervention in HFD-fed mice during the 17-week period, as illustrated in Figure 1A. The body weight difference between HFD- and LFD-fed mice started at the third week, while the attenuation effect of curcumin intervention on HFD-induced body weight gain started at the fifth week (Figure 1B). During the first 12-week period, total calorie intake for mice on an HFD was higher than for mice on an LFD, while no difference was found for HFD-fed mice with or without curcumin intervention (Supporting Information Figure S1a-S1b). Supporting Information Figure S1c shows the weekly calorie intake normalized by body weight. Results of glucose, pyruvate, and insulin tolerance tests among the three groups of mice performed at indicated time points (Figure 1A) are shown in Figure 1C-1F. These results, along with the attenuation of hyperglycemia, hyperinsulinemia, and hyperleptinemia (Figure 1G-1I) with curcumin intervention, allowed us to conclude that curcumin administration improved insulin signaling and metabolic profiles, which were impaired by HFD consumption.

We observed very recently that HFD consumption increased plasma Fgf21 levels, while curcumin intervention blocked the increase and improved hepatocyte Fgf21 sensitivity (12). We show here that HFD consumption also increased epididymal fat tissue Fgf21 levels, and the increase was attenuated by curcumin intervention (Figure 1J).

Curcumin intervention alters macrophage functional polarity

We then assessed the effect of curcumin intervention on adipose tissues. HFD-induced adipose tissue and liver weight gains were attenuated by curcumin intervention (Figure 2A-2B). We then compared epididymal fat tissue macrophage volumes and the compositions in HFD-fed mice in the presence and absence of curcumin intervention. LFD-fed mice were not included as there were very few infiltrated macrophages in those mice for conducting our assays. Figure 2C-2G shows that curcumin intervention not only reduced WAT macrophage volume, but also reduced the percentages of M1-like macrophages and increased the percentages of M2-like macrophages. Such effects, however, were not observed in the spleen tissue (Figure 2H). Table 1 shows that curcumin intervention altered the levels of a panel of cytokines secreted by infiltrated immune cells in WAT. The stimulation of the anti-inflammatory cytokine IL-10, but not IL-4 and IL-13, was also observed. Evidently, dietary curcumin intervention reduced the mean fluorescence intensity (MFI) in WAT CD86 on macrophages but not CD80 cells (Figure 2I-2J). CD86 is expressed in antigen-presenting cells. Reduced CD86 MFI suggests decreased activation of WAT immune cells (3). We therefore suggest that curcumin intervention not only reduces WAT macrophage infiltration, but it also alters WAT macrophage polarity.

In vitro curcumin treatment attenuates Raw264.7 macrophage M1-like cell differentiation

It has been reported previously that curcumin treatment affects the polarization of the murine Raw264.7 macrophages (21). We assessed the direct effect of curcumin treatment in vitro in this monocyte-macrophage cell model with the dosages of curcumin that generated no appreciable inhibition on cell viability (Supporting Information Figure S2). Indeed, LPS/IFNγ treatment in this cell model induced the dendritic morphological change, while curcumin treatment reversed the effect of LPS/IFNγ (Supporting Information Figure S3). Curcumin treatment also dose-dependently attenuated the stimulatory effect of LPS/IFNγ on iNOS protein levels (Figure 3A). Figure 3B-3D shows that LPS/IFNγ-induced expression of iNOS, MCP-1, and Il-6 messenger RNA (mRNA) (differentiation markers of the M1-like subtype) can be dose-dependently inhibited by curcumin treatment. The stimulation of the differentiation markers of the M2-like subtype (KFL4, PPARγ, and ARG1) as well as the anti-inflammation cytokine gene IL-4, however, was observed only when the dosage of curcumin reached 20 μM (Figure 3E-3H). Such a high dosage is unlikely reached in the in vivo settings. When mouse peritoneal macrophages were treated with 0.25 to 2 μM curcumin for 10 hours, we also did not observe the stimulation of M2-like subtype differentiation marker expression (Supporting Information Figure S4).

Adipocytes can produce cytokines that may affect macrophage activity and differentiation (22, 23). We performed the in vitro assay, asking whether mature adipocyte-secreted cytokines affect the differentiation of Raw264.7 cells. Figure 3I shows that rat adipocyte conditioned cell culture medium also increased expression of the M1-like subtype markers in the Raw264.7 cell model. When rat adipocytes were treated with curcumin for 10 hours, the expression of a battery of proinflammatory cytokine genes was reduced, while the expression of the anti-inflammatory cytokine IL-4 was increased (Figure 3J-3K). We therefore suggest that in addition to the regulation of macrophage differentiation directly, curcumin can also inhibit the expression of proinflammatory cytokines produced by adipocytes, indirectly repressing M1 macrophage differentiation.

We reported very recently that in vitro curcumin treatment increased Fgf21 expression in mouse hepatocytes (12). Here, we show that curcumin also stimulated Fgf21 mRNA expression in rat primary adipocytes (Supporting Information Figure S5a). In vitro treatment with human recombinant FGF21, however, did not reduce the expression of the differentiation markers of the M1-like subtype in Raw264.7 cells (Supporting Information Figure S5b-S5e).

Dietary curcumin intervention increases energy expenditure and the thermogenic capacity

Mice fed with an HFD without or with curcumin intervention at the 16th week were subjected to metabolic cage analyses. Curcumin intervention was shown to increase CO₂ production, O₂ consumption, and energy expenditure (Figure 4A-4C), regardless of whether the data were expressed per animal or per body weight of the animals. During the metabolic cage analysis period, there was a trend of increased on respiratory exchange ratio (RER) in mice that received curcumin intervention (Figure 4D-4E). When we separated the data into the dark and the light cycles, a significant increase in RER was found in mice with curcumin intervention in the dark cycle but not in the light cycle (Figure 4F-4G). Increased RER usually suggests that the mice utilize relatively more carbohydrates as the source of energy (24, 25). How this can be found in curcumin intervention treated mice with improved insulin signaling, associated with increased energy expenditure, is an open question for further investigations.

To further verify the stimulatory effect of curcumin intervention on thermogenesis, we conducted another set of mouse experiments with a relatively short-term curcumin intervention, illustrated in Figure 5A. Mice were housed at room temperature (∼22^oC) before being placed in a 4^oC cold chamber. Two, four, and six weeks after the corresponding diet consumption, mice were subjected to cold exposure. Four hours after the mice were in the cold chamber, we started rectal temperature recording. As shown in Figure 5B-5D, curcumin intervention increased the thermogenic capacity during the cold exposure process, starting as early as 2 weeks after curcumin intervention. We noticed that the significant temperature differences appeared mostly during the dark periods. The differences, however, were not observed for these three groups of mice in the absence of the cold challenge (Supporting Information Figure S6).

Curcumin intervention increases BAT UCP1 expression

An improved ability of the mice to handle a cold challenge could be due to the increase of shivering thermogenesis, the increase of BAT-mediated nonshivering thermogenesis, or both (26-28). As curcumin intervention was shown to increase energy expenditure in our metabolic cage analysis, and curcumin gavage was shown to stimulate the inguinal adipocyte “browning” that was reported by another team (14), we asked whether long-term (17 weeks) curcumin intervention affects UCP1 expression and subcutaneous (inguinal) WAT browning. We did not see changes in UCP1 expression in subcutaneous adipose tissues after curcumin intervention and could not detect their UCP1 expression by Western blotting (Supporting Information Figure S7a-S7b). In addition, mice that received the 6-week curcumin intervention showed elevated BAT but not inguinal UCP1 expression (Supporting Information Figure S7c-S7d). We did find the “browning-like clusters” within the inguinal adipose tissue in mice that received curcumin intervention, but the numbers were few. Nevertheless, the 17-week dietary curcumin intervention reduced the sizes of adipocytes in both inguinal and epididymal adipose tissues (Supporting Information Figure S7e-S7h).

Curcumin intervention appears to reduce BAT adipocyte size, which is associated with increased UCP1 mRNA and protein expression (Figure 6A-6D). In BAT, increased UCP1 and Prdm16 (which encodes the protein PR domain containing 16) expression in response to long-term curcumin intervention was associated with increased PPARα but not PPARγ or PPARγ coactivator (PGC-1α) (Figure 6C), while in rat mature adipocytes, in vitro curcumin treatment increased UCP1 mRNA levels, which were associated with PPARα and PGC-1α elevation (Figure 6E). As PPARs are known transcriptional activators of UCP1 (29-31), the above observation indicates that curcumin may directly stimulate UCP1 transcription via PPAR activation. To examine this, we transfected the UCP-1/LUC fusion gene plasmid (19) into the 293T cells, followed by curcumin treatment in the absence and presence of PPARα or PPARγ antagonist. Figure 6F shows that curcumin or the PPARα agonist WY14643 treatment increased UCP1 promoter activity, while the PPARα antagonist GW6471 inhibited its activity. The stimulatory effect of curcumin, however, was not blocked by GW6471 pretreatment. Figure 6G shows that PPARγ agonist rosiglitazone also stimulated UCP1 promoter activity, while the stimulatory effect of curcumin was not blocked by the PPARγ antagonist GW9662. These observations suggest that curcumin can stimulate UCP1 expression via PPARα and PPARγ, while the existence of PPARα- and PPARγ-independent mechanisms cannot be excluded.

Finally, we tested the effect of curcumin on UCP1 expression in a mBAC after its differentiation for 3 days (Supporting Information Figure S8a-S8b). As shown in Figure 6H, 10-hour curcumin treatment increased UCP1, PPARα, PPARγ, and PGC-1α mRNA levels in differentiated mBAC. Supporting Information Figure S8c shows that curcumin-stimulated UCP1 mRNA expression in mBAC cells could not be completely blocked by PPARα or PPARγ inhibition.