Introduction

Obesity and systemic insulin resistance are the primary causes of type 2 diabetes. Chronic low-grade inflammation is thought to account for the acquisition of insulin resistance.1 In both humans and rodents, monocyte chemoattractant protein-1 (MCP-1) secretion from hypertrophic adipocytes induces monocyte infiltration into adipose tissue, and this phenomenon causes the secretion of proinflammatory mediators from invasive macrophages and adipocytes in adipose tissue.2,3 Accordingly, the basis of systemic insulin resistance is thought to be the enhancement of proinflammatory cytokine production in adipose tissue.3 Moreover, it has been reported that the inhibition of the secretion of proinflammatory cytokines (for example, tumor necrosis factor-α (TNF-α) and MCP-1) or cytokines downstream in the signaling pathway and the inhibition of macrophage chemotaxis have the ability to alleviate insulin resistance.4, 5, 6, 7, 8

To find natural compounds that can suppress the secretion of MCP-1 in adipose tissue, we established a microplate-based high-throughput adipocyte and macrophage coculture assay system. Consequently, we screened a library of extracts and focused on Physalis pruinosa extract.

Recently, many physalins that are classified as secosteroids have been isolated from Physalis alkekengi, and these compounds have been reported to have anti-inflammatory9,10 and anticancer effects.11 However, the pharmacological activity of P. pruinosa or its active components has almost never been reported.

In contrast, some natural steroids and withanolides from plants or insects have been reported to improve impaired glucose tolerance and/or obesity.12, 13, 14, 15 In the present study, plant sterols with withanolide skeletons were isolated from P. pruinosa calyxes. Among the withanolides, moleculer mechanisms of its anticancer effect of Withaferin A (WA) are extensively well studied. The inhibition of the nuclear factor-κB (NF-κB) signaling pathway through the overphosphorylation of IκB kinase (IKK)-β16 and the upregulation of prostate apoptosis response-4 gene expression is considered as the key mechanism of WA's action.17 In addition, WA has anticancer effects resulting from direct interactions with cellular structural proteins such as annexin II18 and vimentin.19

In this study, the P. pruinosa calyx was found to improve glucose metabolism based on the anti-inflammatory effect that this plant extract exhibited in the macrophage and adipocyte coculture assay. We attempted to isolate the active compound from the P. pruinosa calyx and to elucidate its chemical structure. We also studied the effect of the isolated compound on the inflammatory response in adipose tissue and glucose tolerance in diabetic mice. We attempted to elucidate the intercellular signaling mechanisms of the anti-inflammatory effects of the active compound.

Materials and methods

Plant extract library

Available edible plants including vegetables, fruits, spices and herbs were purchased. The 1030 types of divided plant parts were freeze-dried and then crushed into powders with a food processor. The powders were extracted with methanol (that is, 1 g of each powder was soaked in 40 ml of methanol), the solid phase was removed by filtration, and the liquid phase was dried at 30 °C with an evaporator. The dried extract was dissolved in dimethylsulfoxide (DMSO) to a concentration of 100 mg ml−1 and stored at −80 °C.

TNF-α assay in macrophages

RAW264.7 cells (a mouse macrophage cell line; #TIB-71; ATCC, Rockville, MD, USA) were cultured overnight in RPMI 1640 (Invitrogen, Carlsbad, CA, USA) supplemented with 0.2% FBS. Plant extracts were added to the culture medium at concentrations of 3, 10, 30 and 100 μg ml−1, and then 400 μM BSA-conjugated palmitate was added to stimulate the cells. After 24 h, the medium was collected, and the TNF-α level was measured using an ELISA kit (Thermo Fisher Scientific, Tewksbury, MA, USA).

MCP-1 assay using the adipocyte/macrophage coculture system

3T3-L1 cells (ATCC) were differentiated into mature adipocytes by using a conventional protocol.20 The medium was then removed, each plant extract was added to RAW264.7 cells, and the two cell lines were cocultured in DMEM supplemented with 0.5% BSA and 0.1% DMSO as a vehicle. After 24 h, the medium was collected, and the level of MCP-1 was determined using an ELISA kit (BD Biosciences-Pharmingen, San Jose, CA, USA). The viability of the cells was measured with the Cell Counting Kit-8 (Dojindo, Osaka, Japan) according to the manufacturer’s guidelines.

The purification of 4β-hydroxywithanolide E from P. pruinosa calyx extract

P. pruinosa calyx was obtained from the Akita prefecture in the northern part of Japan. Using the extraction method described above, 10.12 g of dried material was obtained from 340 g of P. pruinosa calyx. The dried material was extracted with methanol, and the extract was separated by column chromatography using Silica gel 60 N (Kanto Chemical, Tokyo, Japan). The fifth and sixth fractions exhibited anti-inflammatory activity, evaluated by using the MCP-1 assay. These fractions were separated again by Inertsil ODS-3 250 × 50 mm ID (GL Sciences, Torrance, CA, USA), and the purified active compound was analyzed using a total ion chromatogram collected with a photodiode array and LC/MS (Quattro Micro; Waters Corp., Milford, MA, USA). The compound was identified as 4β-hydroxywithanolide E (4βHWE; molecular weight 502.6) by NMR (AVANCE 400; Bruker BioSpin Corp., Billerica, MA, USA) and LC/MS. From 10.12 g dried P. pruinosa calyx methanol extract (PME), we obtained about 240 mg 4βHWE and these compounds were used to evaluate its effects in further experiments.

The effect of 4βHWE on macrophage and adipocyte cytotoxicity

4βHWE or the vehicle (0.5% BSA and 0.1% DMSO in DMEM) was added into mature 3T3-L1 or RAW264.7 cells. An equal volume of coculture medium was added simultaneously. After 24 h, the cell viability was measured as described above.

Immunoblot analysis

Mature 3T3-L1 cells were incubated with 4βHWE or withaferin A (WA; Alexis Lausanne, Switzerland) for 10 min. After incubation, these cells were stimulated with 10 ng ml−1 recombinant mouse TNF-α (R&D Systems, Minneapolis, MN, USA) or RAW264.7 cells for 5, 15 or 30 min. The cells were then collected with sample buffer, containing 10% glycerol, phosphatase inhibitor cocktail (Nacalai Tesque, Kyoto, Japan) and protease inhibitor cocktail (Nacalai Tesque) and stored in a freezer at −80 °C. Nucleotides were degraded with TurboNuclease (Accelagen, San Diego, CA, USA), and the lysate was loaded onto a 10.5% SDS-polyacrylamide gel, and electrophoresis was performed. The proteins were subsequently transferred to a nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA, USA) with a Mini Trans-Blot Cell (Bio-Rad Laboratories). Immunoblotting was carried out with the following primary antibodies: anti-phospho- transforming growth factor-β-activated kinase (TAK) 1 antibody (1:1000; Cell Signaling Technology, Beverly, MA, USA), anti-phospho-IKKα/β antibody (1:1000; Cell Signaling Technology), anti-IκB-α antibody (1:1000; Cell Signaling Technology) and anti-α-tubulin antibody (1:1000; Cell Signaling Technology).

To study the effect of 4βHWE on NF-κB translocation to the nucleus, nuclei were extracted from adipocytes that had been stimulated with 10 ng ml−1 recombinant mouse TNF-α for 30 min after a 10-min incubation with or without 4βHWE. The cells were washed with phosphate-buffered saline (PBS) and homogenized in lysis buffer with a Dounce homogenizer and centrifuged. Nuclear extraction buffer was added to the pellet, and the mixture was placed on ice for 45 min. The supernatant was obtained by centrifugation and stored at −80 °C. Each protein was loaded onto a 10.5% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. Immunoblotting was carried out with the NF-κB p50 antibody and Histone H1 antibody (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA). Histone H1 was used as a loading control.

To detect the primary antibodies, anti-rabbit IgG conjugated to horseradish peroxidase (1:10 000 GE Healthcare, Little Chalfont, UK) was used as a secondary antibody. Immunoblots were visualized by using an LAS-3000 imaging system (Fujifilm Corp., Tokyo, Japan) with the ECL Plus System (GE Healthcare) according to the manufacturer’s guidelines.

Luciferase assay for NF-κB transcriptional activity

CV-1 cells (African green monkey kidney cells) were transfected with the PathDetect NF-κB Cis-Reporting System (Stratagene, La Jolla, CA, USA) by the FuGENE6 Transfection Reagent (Roche, Mannheim, Germany). The transfected cells were incubated with 10 μM 4βHWE and 10 ng ml−1 TNF-α for 24 h, and the luciferase activity was detected with the ONE-Glo Luciferase Assay System (Promega, Madison, WI, USA) according to the manufacturer’s guidelines.

RNA extraction and real-time RT–PCR

Total RNA was extracted from cells or tissues and reverse transcribed into complementary DNA (cDNA). Real-time RT–PCR analysis was subsequently performed with the Thermal Cycler Dice Real Time System (Takara Bio, Shiga, Japan). Each reaction tube contained template, 0.5 μM each primer and SYBR Premix Ex Taq II (Takara Bio). The following real-time PCR conditions were used: 40 cycles of 94 °C for 5 s, 60 °C for 45 s and 95 °C for 30 s. The primer sequences are shown in Supplementary Table 1.

The results were analyzed using the 2-ΔΔCt method (ABI User Bulletin 2) and are presented as the ratio of each mRNA to the 18S rRNA to correct for variations in the quantities of the template DNA.

Animal experiment

The experimental protocol was reviewed and approved by the Animal Care Committee of Ajinomoto. Five-week-old male db/db and db/+ mice were purchased from the Charles River Laboratories (Yokohama, Japan) and were tamed for 1 week under the experimental conditions. The food and drinking water were provided ad libitum. The mice were divided into 4 groups: db/db mice administered with vehicle (0.5% carboxymethyl cellulose; Wako Pure Chemical Industries, Osaka, Japan), db/db mice administered with 10 mg kg−1 body weight of 4βHWE, db/db mice administered with 30 mg kg−1 body weight of 4βHWE and db/+ mice administered with vehicle. Mice were orally administered once per day for 2 or 4 weeks.

To quantify the mRNA expression levels of proinflammatory cytokines (TNF-α, MCP-1 and interleukin (IL)-6) and macrophage markers (F4/80 and CD68), the mice were killed and epididymal adipose tissue was collected after 2 weeks of treatment. A portion of the epididymal adipose tissue was fixed with neutral-buffered formalin for immunohistological analysis.

Glucose tolerance tests were performed after 3 weeks of treatment. In brief, mice were orally administered 0.75 g kg−1 body weight of glucose, and blood was collected before and 15, 30, 60 and 120 min after the administration. Blood glucose levels were measured with a glucose analyzer (ARKRAY, Kyoto, Japan). After 4 weeks of treatment, blood samples were collected, and the expression of mRNA in the adipose tissue was measured using the real-time RT–PCR method described above. The plasma MCP-1 and adiponectin concentrations were measured with an ELISA kit (BD Biosciences-Pharmingen and Otsuka Pharmaceutical, Tokushima, Japan, respectively). Plasma-free fatty acids were measured with the Wako NEFA C test kit (Wako Pure Chemical Industries).

Macrophages in the adipose tissue were detected by immunohistochemistry. In brief, tissues were embedded in paraffin and cut into 5-μm sections. An anti-mouse macrophage antibody (F4/80; BMA Biomedicals, Rheinstrasse, Switzerland) was used as the primary antibody (1:200), and immunoreactivity was detected with the avidin–biotin complex (VECTASTAIN ABC; Vector Laboratories, Burlingame, CA, USA). The tissue was subsequently counterstained with hematoxylin. The number of F4/80-positive cells was counted under a microscope in a 0.0625 mm2 field of view. More than 50 serial fields were examined, and the data were expressed as cells per mm2 of adipose tissue.

Statistical analyses

The data are presented as the mean±s.e. of the mean. For experiments with two factors, for example, concentration and drugs (Figures 2 and 3d–f) or time and treatment groups (Figure 5a), significant differences between groups and doses/time points were identified by a two-way ANOVA followed by Bonferroni’s multiple comparisons test. For experiments with one factor, for example, drug dose, significant differences were identified using a one-way analysis of variance (ANOVA) followed by Bonferroni’s multiple comparisons test.

Results

We screened 1030 plant extracts from the plant library using the adipocyte/macrophage coculture assay, and 10 plant extracts exhibited a strong inhibitory effect on MCP-1 production. These 10 plant extracts were derived from P. pruinosa, Rabdosia japonica, Coptis japonica Makino, Zingiber officinale Roscoe, Nuphar japonicum De Candolle, Ephedra sinica Stapf, Withania somnifera, Curcuma longa, Lycoris radiate, and Rosmarinus officinalis L. From the potent candidates, we focused on the extract of P. pruinosa because this extract had one of the strongest activities on MCP-1 production without inducing cytotoxicity.

The PME inhibited TNF-α production from macrophages that was induced by palmitate in a dose-dependent manner (Figure 1a). In addition, PME also inhibited MCP-1 production in the adipocyte/macrophage coculture system in a semi-dose-dependent manner without cytotoxicity in the range from 0 to 100 μg ml−1 (Figures 1b and c). These anti-inflammatory effects were exerted without any cytotoxicity.

Figure 1
figure 1

The effects of P. pruinosa methanol extract (PME) on the TNF-α production stimulated by palmitate (a), on the MCP-1 production induced by the RAW264.7 (macrophage) and 3T3-L1 (adipocyte) coculture (b) and on cytotoxicity in RAW264.7 and 3T3-L1 coculture (c). Values are presented as the mean±s.e.m. of three separate wells. * represents P<0.05 and ** represents P<0.01; these asterisks indicate statistically significant differences relative to the value obtained under stimulated conditions without PME, as identified through a one-way ANOVA with Bonferroni’s test.

Full size image

The active compound in the PME was identified as 4β-hydroxywithanolide E (4βHWE; Figure 2a) with a photodiode array and by NMR and LC/MS spectral analysis.

Figure 2
figure 2

The chemical structure of 4βHWE (a), the effect of withanolides on the production of MCP-1 induced by the adipocyte and macrophage coculture (b) and the effect of withanolides on the viability of 3T3-L1 adipocytes (c), RAW264.7 macrophages (d) and cocultured cells (e). Values are presented as the mean±s.e.m. of three well determinations. * represents P<0.001 and ** represents P<0.0001; these asterisks indicate statistically significant differences when relative to the treatment, as identified through a two-way ANOVA with Bonferroni’s test. NS, not significant.

Full size image

The inhibitory effects of the withanolides on MCP-1 secretion were evaluated using the adipocyte/macrophage coculture assay. The results are shown in Figure 2b, and the half maximal inhibitory concentrations of MCP-1 secretion were 1.59±0.28 μM for 4βHWE, 5.92±0.82 μM for WA, 10.48±1.8 μM for 3-methoxy-4βHWE and >30 μM for withanolide B. Neither withanolide S nor withaperuvin C, which are analogs of the withanolides purified from PME, showed any anti-inflammatory effects at concentrations up to 10 μM (data not shown).

The effects of 4βHWE and WA on cell viability were examined using 3T3-L1 adipocytes (Figure 2c), RAW264.7 macrophages (Figure 2d) and adipocyte/macrophage coculture (Figure 2e) with concentrations ranging from 0.1 to 30 μM. Neither 4βHWE nor WA affected the viability of adipocytes at concentrations between 0.1 and 10 μM. However, WA but not 4βHWE reduced the viability of adipocytes when added at concentrations of 30 μM (Figure 2c). In the coculture, the effects of these compounds on cell viability were similar to their effects on adipocytes alone (Figure 2e). However, both 4βHWE and WA decreased the viability of macrophages in a dose-dependent manner. The negative effect of 4βHWE on macrophage viability was significantly weaker than the negative effect of WA (Figure 2d).

To elucidate the anti-inflammatory mechanisms of 4βHWE, TNF-α was used to stimulate changes in the intracellular signaling of adipocytes. After 5 min of TNF-α stimulation, the phosphorylation levels of TAK-1 and IKKβ were increased, and the IκBα protein level was decreased in 3T3-L1 adipocytes (Figure 3a). In particular, TAK-1 phosphorylation increased and IκBα level decreased gradually until after 30 min of TNF-α stimulation. The addition of 10 μM 4βHWE to the medium inhibited the inflammatory changes in intracellular signaling in the adipocytes (Figure 3a).

Figure 3
figure 3

The effect of 4βHWE on the NF-κB signaling pathway after upregulation by TNF-α (10 ng ml−1) in 3T3-L1 adipocytes. Panel a shows the time course of intracellular signal transduction with or without 10 μM 4βHWE in adipocytes. Panel b shows intranuclear NF-κB in adipocytes after 30 min of TNF-α stimulation with or without 10 μM 4bHWE. Panel c shows the effect of 10 μM 4βHWE on NF-κB transcriptional activity. The effect of 5 μM 4βHWE on the mRNA expression of MCP-1 (d), TNF-α (e), and IL-6 (f) in adipocytes stimulated by TNF-α is shown. Values are presented as the mean±s.e.m. of three well determinations. In panel c, *** represents P<0.001 and indicates statistically significant differences when compared with the value obtained in the stimulated condition without 4βHWE, as identified through a one-way ANOVA with Bonferroni’s test. In panels d–f, ** represents P<0.01 and *** represents P<0.001; these asterisks indicate statistically significant differences relative to the treatment, as identified through a two-way ANOVA.

Full size image

Moreover, the level of NF-κB was increased in the nuclei of the adipocytes after 30 minutes of TNF-α stimulation. The addition of 10 μM 4βHWE to the medium inhibited the increase in intranuclear NF-κB induced by TNF-α (Figure 3b).

CV-1 cells that were transfected with a vector that contained five repeats of the NF-κB- binding site sequence upstream of the luciferase protein coding sequence were used to elucidate the effect of 4βHWE on the transcriptional activity of NF-κB. As shown in Figure 3c, 4βHWE significantly inhibited the upregulation of luciferase activity by TNF-α.

Next, the effect of 4βHWE on the mRNA expression of inflammatory cytokines induced by TNF-α was examined in adipocytes. MCP-1 mRNA expression increased approximately 30–40 times after 2, 4 and 6 h of TNF-α stimulation, and 4βHWE significantly inhibited this upregulation of MCP-1 mRNA expression (Figure 3d). The TNF-α mRNA level increased approximately 3000-fold after a 2-h stimulation, and this elevation was attenuated by 4βHWE (Figure 3e). The IL-6 mRNA level also increased twofold after a 2-h stimulation, and 4βHWE attenuated this increase in expression (Figure 3f). These suppressions were similar to the effects of WA after a 2-h stimulation (Figures 3d–f).

To evaluate the effects of 4βHWE on the inflammatory response and glucose tolerance in diabetic mice, 4βHWE (10 or 30 mg kg−1 body weight) was orally administered daily to db/db mice for 4 weeks.

A group treated with 10 mg kg−1 of 4βHWE exhibited a significant increase in body weight compared with the vehicle group; however, at a higher dose (30 mg kg−1 of 4βHWE), there were no significant differences (Supplementary Table 2). The weight of subcutaneous, epididymal and mesenteric adipose tissues were not significantly different among the groups. There were no significant differences in the number of white blood cells (WBC), glutamate oxaloacetate transferase (GOT), glutamate pyruvate transaminase (GPT) and creatinine concentrations among the groups (Supplementary Table 3).

After 2 weeks of 4βHWE administration, F4/80-positive cells, which were used as indicator of macrophages, infiltrated into the epididymal adipose tissue of db/db mice; the level of macrophage infiltration was reduced by 4βHWE (Figures 4a–d). Figure 4e shows the effect of 4βHWE on the number of macrophages in adipose tissue; 4βHWE reduced the infiltration of macrophages in a dose-dependent manner. In addition, the mRNA expression levels of macrophage-specific cellular membrane proteins (F4/80 and CD68) in adipose tissue were also reduced by 4βHWE (Figures 4f and g). The mRNA expression of the obesity-induced proinflammatory cytokines (MCP-1, TNF-α and IL-6) was simultaneously suppressed by 4βHWE in a dose-dependent manner (Figures 4h–j).

Figure 4
figure 4

The effect of the oral administration of 4βHWE (10 or 30 mg kg−1 body weight per day) for 2 weeks on the inflammatory responses in the epididymal adipose tissue of db/db mice. Representative examples of tissue sections in which the macrophages were stained with an anti-F4/80 antibody and the nucleus was counterstained with hematoxylin are shown in panels (a–d). The effect of 4βHWE on the number of F4/80-positive cells (e) and on the mRNA expression of macrophage markers (F4/80 (f) and CD68 (g)) and inflammatory cytokines (TNF-α (h), MCP-1 (i) and IL-6 (j)) in adipose tissue are shown. The scale bar indicates 50 μm in panels a–d, and the arrowhead indicates macrophages. The values are presented as the mean±s.e.m. of 5 and 8 individual determinations in panels e and f–j, respectively. * represents P<0.05 and ** represents P<0.01; the asterisks indicate statistically significant decreases relative to vehicle-treated db/db mice, as identified by a one-way ANOVA with Bonferroni’s test.

Full size image

After 3 weeks of 4βHWE administration, the mice were subjected to a glucose tolerance test. db/db Mice exhibited increased blood glucose levels, and this elevation was attenuated by treatment with 30 mg kg−1 of 4βHWE for 3 weeks (Figures 5a and b). In addition, the administration of 30 mg kg−1 of 4βHWE for 4 weeks significantly decreased the blood glucose levels in db/db mice relative to the levels in db/db mice that were treated with the vehicle alone (Figure 5c). The plasma insulin concentrations did not change at this time point (Figure 5d).

Figure 5
figure 5

The effect of the oral administration of 4βHWE (10 or 30 mg kg−1 body weight per day) for 3 weeks on glucose availability (a), area under the curve (AUC) of glucose concentration (b) in db/db mice. The effect of oral administration of 4βHWE for 4 weeks on basal plasma glucose concentration (c), and plasma insulin concentration (d) in db/db mice. The values are presented as the mean±s.e.m. of 8–10 individual determinations. In panel a, * represents P<0.05, which indicates significant decreases relative to vehicle-treated db/db mice, as identified through a two-way ANOVA with Bonferroni’s test. In panels b and c, * represents P<0.05 and *** represents P<0.001; the asterisks indicate statistically significant decreases relative to vehicle-treated db/db mice, as identified through a one-way ANOVA with Bonferroni’s test.

Full size image

The treatment of 30 mg kg−1 of 4βHWE for 4 weeks significantly increased plasma adiponectin and decreased MCP-1 levels compared with the vehicle groups. (Figures 6a and b). Also, the concentration of plasma-free fatty acids decreased significantly (Figure 6c).

Figure 6
figure 6

The effect of the oral administration of 4βHWE (10 or 30 mg kg−1 body weight per day) for 4 weeks on the plasma concentrations of adiponectin (a), MCP-1 (b) and free fatty acids (c) in db/db mice. The values are presented as the mean±s.e.m. of 10 individual determinations. * represents P<0.05 and ** represents P<0.01; the asterisks indicate statistically significant differences relative to vehicle-treated db/db mice, as identified through a one-way ANOVA with Bonferroni’s test.

Full size image

Discussion

In this study, we screened 1030 plant extracts to identify potent compounds with the ability to inhibit the diabetes progression. From those, P. pruinosa calyx extract was identified to possess strong inhibitory effect on MCP-1 production induced in adipocyte/macrophage coculture assay. We then attempted to isolate the possible active compound from P. pruinosa calyx and identified it as 4βHWE, which is a member of the withanolide family of compounds. 4βHWE had a stronger anti-inflammatory effect and lower toxicity than any other withanolides and that its anti-inflammatory effect was dependent on the attenuation of NF-κB signaling. I n vivo experiments showed that the oral administration of 4βHWE to obese db/db mice inhibited the inflammatory responses in adipose tissue and improved impaired glucose tolerance.

First, we verified that PME attenuated palmitate-induced production of TNF-α in macrophages and MCP-1 in the macrophage/adipocyte coculture system (Figures 1a and b). Cocultures of 3T3-L1 adipocytes and RAW264.7 macrophages are used as in vitro models of inflammatory changes in obese animals.21,22 These results indicate that PME contains anti-inflammatory compounds that inhibit macrophage activation and infiltration into adipose tissue caused by adipocyte hypertrophy. Physalis extracts have been reported to regulate the estrus cycle,23 the activity of glucose-6-P dehydrogenase in hepatocytes,24 induce cancer cell apotosis,25,26 histopathological changes27 and inflammatory changes.28, 29, 30 In addition, the anti-proliferative effect on HTLV-1-infected T-cell lines has been reported only with the extract of P. pruinosa.31 Therefore, we attempted to isolate and identify anti-inflammatory compounds present in PME and identified 4βHWE. 4βHWE is a withanolide, a family that has more than 60 types. Most of the known withanolides have been isolated from Tubocapsicum anomalum or Withania somnifera Dunal not from the genus physalis, and these compounds have been reported to possess variety of physiological activities in vivo and/or in vitro, including anticancer,32,33 anti-inflammatory16,34 and neuroprotective 35,36 activities. 4βHWE was originally found in Physalis peruviana L.37 and has been intensively studied as an inhibitor of cancer cell proliferation in vitro.38,39 The present study is the first report of the isolation of 4βHWE from P. pruinosa, and our experiments suggested a novel physiological function of 4βHWE as a compound that has anti-inflammatory effects that involve an increase in glucose tolerance.

4βHWE exhibited the strongest inhibition of MCP-1 production, with the least cytotoxicity than the other four withanolides that were isolated during the isolation of the active compound (Figures 2b–e). Yen et al.38 reported that 2, 10 and 20 μM 4βHWE significantly induced DNA damage in a dose-dependent manner in a human lung cancer cell line (H1299). However, in our experiments, 30 μM 4βHWE was not toxic to a mature adipocyte cell line (3T3-L1; Figures 1c and 2c), whereas it showed toxic effects at 3 μM in a macrophage cell line (RAW264.7; Figure 2d). These results suggest that the toxic effects of 4βHWE might differ according to cell type. To elucidate the mechanisms of the anti-inflammatory effects of 4βHWE, NF-κB signal transduction was evaluated in TNF-α-stimulated 3T3-L1 cells. 4βHWE significantly inhibited TNF-α-induced proinflammatory responses, including the induction of TAK-1, IKKβ phosphorylation and the degradation of IκBα (Figure 3a). Moreover, 4βHWE suppressed the translocation of NF-κB into the nucleus and thus inhibited the transcriptional activity of NF-κB (Figures 3b and c). The mRNA expression of proinflammatory cytokines such as MCP-1, TNF-α and IL-6 was decreased by 4βHWE (Figures 3d–f). Generally, mRNA expression of these cytokines is upregulated by the activation of NF-κB and the acceleration of the inflammatory response.40, 41, 42 Thus, the anti-inflammatory effects of 4βHWE could possibly be caused by the inhibition of NF-κB signaling. Although WA has been reported to inhibit the IKKβ activation through the induction of IKKβ overphosphorylation,16 4βHWE inhibited the IKKβ activation through the suppression of IKKβ phosphorylation in adipocytes (Figure 3a). The diverse regulation of IKKβ phosphorylation observed in this study might be caused by the differences between the chemical structures of 4βHWE and WA. In the calculated stable structural form of 4βHWE, the flat surface of the lactone ring is skewed toward the flat surfaces of the A–D rings (PubChem Compound Identifier (CID) 73621). However, in the calculated stable structural form of WA, both the flat surface of the lactone ring and the surfaces of the A–D rings lie in the same plane (CID 265237). Thus, these differences in the stable chemical structure might be the cause of the distinct mechanisms of action of these two compounds. Taken together, these results led us to speculate that the anti-inflammatory mechanisms or molecular targets of 4βHWE and WA are different.

Anti-inflammatory compounds, including salicylates,43 IKKβ inhibitors,44 curcumin45 and natural steroids or triterpenoids,12, 13, 14, 15 have been reported to improve impaired glucose tolerance in mouse obesity models, such as high-fat-diet-induced obese mice, db/db and ob/ob mice. In these studies, the inflammatory responses in adipose tissue were alleviated when the secretion of adipokines (for example, TNF-α, MCP-1 and FFA) decreased and adiponectin level increased. As a result, systemic insulin sensitivity might be increased and/or glucose tolerance was improved. Thus, we attempted to examine the effects of 4βHWE on the inflammatory responses in the adipose tissue of an obese mouse model. The oral administration of 4βHWE (30 mg kg−1) to db/db mice for 2 weeks decreased the number of macrophages, the mRNA expression of macrophage-specific membrane antigens in abdominal adipose tissue (Figures 4a–g) and the expression of proinflammatory cytokines (Figures 4h–j). The inhibition of the expression of proinflammatory cytokines by 4βHWE might be dependent on the suppression of macrophage recruitment in obese animals’ adipose tissue. This inhibition of macrophage infiltration effect might be due to inhibition of obesity-induced inflammatory cytokine production from adipocyte as seen in in vitro experiments. Moreover, administration of 30 mg kg−1 4βHWE for 3 or 4 weeks suggested improvement of impaired glucose tolerance (Figures 5a and b), which involved an increase in plasma adiponectin levels and decrease in the plasma MCP-1 and free fatty acid levels in db/db mice (Figures 6a–c). These effects also may be dependent on the suppression of the inflammatory responses in the adipose tissue.

Recently, the gut microbiome has been recognized as a key regulator of metabolic health in humans.46,47 Because it is not clear whether 4βHWE altered the microbiome in this study, further studies on the effects of 4βHWE on the microbiome are required. We investigated basal glucose and insulin concentrations after a 4-week 4βHWE administration (Figures 5c and d). The plasma glucose concentrations decreased in a dose-dependent manner; however, the insulin concentrations did not change significantly. These results suggest that rather than acting on insulin secretion, 4βHWE affects insulin sensitivity. To confirm the effects on insulin sensitivity, an insulin tolerance test was performed. 4βHWE slightly increased glucose clearance, but the effect was not significant (data not shown). Although the mechanism has not been completely elucidated, long-term 4βHWE administration might produce qualitative changes in adipose tissue rather than directly influencing glucose uptake and cellular signaling. 4βHWE inhibited the inflammatory response in adipose tissue and may have the potential to improve impaired glucose tolerance in db/db mice. 4βHWE is expected to prevent the development of glucose metabolism impairment in obese individuals.

In conclusion, this study showed that 4βHWE could have greater beneficial anti-inflammatory effects than other withanolides, which are known to have a wide range of physiological activities. 4βHWE may be useful as a natural anti-inflammatory compound for the prevention of type 2 diabetes and metabolic syndrome.