A total of 54 5-week-old BALB/c male mice (weight 19-20 g) were purchased from Japan SLC, Inc. (Hamamatsu, Japan) and housed under standardized conditions (temperature, 25±1˚C; humidity, 55±5%) in a room with a 12/12-h light/dark cycle. The mice were fed standard chow (cat. no. CE-2, CLEA Japan, Inc.) and water and acclimatized to the room for 1 week. All animal experiments were approved by Animal Experimental Committee of Tohoku Medical and Pharmaceutical University (Sendai, Japan (approval no. 23032-a) and performed in accordance with the ethical guidelines of the university. Euthanasia of mice was performed to adhere to the American Veterinary Medical Association Guidelines for the Euthanasia of Animals (15). The mice were euthanized by excessive inhalation of 5% isoflurane in a chamber; cardiac and respiratory arrest were confirmed and then the tissues were resected for analysis.

Lespedeza sp. were provided by Sendai Wild Plants Garden under the jurisdiction of Sendai City Office Construction Bureau Park Management Section. These plants were air-dried and extracted with 20-fold methanol volume of plant weight at room temperature for 7 days. The methanol extract was filtered and concentrated by a rotary evaporator under reduced pressure at 40˚C to obtain the dried extract. The yields of LB, LH, LM and LT-dried extracts were 94.35, 64.13, 79.17 and 51.72 mg/g, respectively.

PPARγ binding assay was conducted using the method described by Konno et al (16). Briefly, 0.2 µg per well CREB binding protein (CBP; BIOSS) was immobilized in the wells of a plastic 96-well plate at 37˚C for 1 h. After blocking with 5% skimmed milk at 37˚C for 1 h, PPARγ (PPARγ Human Recombinant, ProSpec-Tany TechnoGene Ltd.) was immobilized on the CBP in at 37˚C for 1 h. After the washing with wash buffer (0.1 M PBS containing 0.05% Tween 20), methanol extracts of Lespedeza sp. (the reaction concentrations: 0.038, 0.075, 0.15, 0.3 mg/ml) or pioglitazone (the reaction concentrations: 0.001, 0.01, 0.1, 1 µM) were added, respectively, and the plate was incubated for 1 h at 37˚C. 500-fold diluted PPARγ antibody (rabbit polyclonal, cat. No. AHP1269, Bio-Rad Laboratories, Inc.) was added to the wells after washing with the wash buffer and then incubated for 1 h at 37˚C. 40-fold diluted alkaline phosphatase-conjugated IgG antibody (goat anti-rabbit, catalog No. 170-6581, Bio-Rad Laboratories, Inc.) was added after the wells washing with the wash buffer and then incubated for 1 h at 37˚C. Following shaking with a microplate shaker at room temperature in the dark for 20 min, absorbance of the wells was measured at 405 nm with a spectrophotometer. The PPARγ binding activity was calculated using the following formula: PPARγ binding activity=absorbance of sample/absorbance of control. Pioglitazone (FUJIFILM Wako Pure Chemical Corporation) was used as a positive control.

DPPH radical scavenging assay was conducted using the method by Blois (17). Briefly, 100 mM (pH 5.5) acetate buffer, ethanol (95 vol%), 500 µM DPPH ethanol solution and sample solution (6.25, 125, 250, 500, 1,000, 2,000 µg/ml) or Trolox (0.25, 0.5, 1.0, 2.0 µM), respectively, were mixed in the same tube (1.00 and 1.25 ml and 250.00 and 50.00 µl, respectively). The mixture was incubated at room temperature for 30 min in the dark and absorbance was measured at 517 nm with a spectrophotometer. DPPH radical scavenging activity was calculated using the following formula: [(Absorbance of control-absorbance of sample)/absorbance of control] x100. Trolox (Tokyo Chemical Industry Co., Ltd.) was used as a positive control.

The spleen of male BALB/c mice was extirpated under sterile conditions following euthanasia via inhalation of isoflurane. The spleen was minced by scissors and ground by slide glass in RPMI-1640 medium (FUJIFILM Wako Pure Chemical Corporation) and filtered through a 40-µM mesh. The cell suspension was hemolyzed using 1X RBC Lysis Buffer, pluriSelect Life Science) to obtain splenocytes. The splenocytes were suspended in RPMI-1640 containing 0.05 mM 2-mercaptoethanol (FUJIFILM Wako Pure Chemical Corporation), 10% heat-inactivated FBS (Thermo Fisher Scientific, Inc.) and Antibiotic-Antimycotic Mixed Stock Solution (100X; Nacalai Tesque, Inc.).

A total of 5x10⁴ splenocytes and 4 µg concanavalin A (derived from Canavalia esiformis; Merck KGaA) were added to the individual wells of a 96-well plate and incubated at 37˚C in a humidified atmosphere containing 5% CO₂ for 48 h. The medium was exchanged for RPMI-1640 medium containing LH and incubated at 37˚C in a humidified atmosphere containing 5% CO₂ for 48 h. The splenocyte proliferative activity was determined by adding Cell Count Reagent SF (Nacalai Tesque, Inc.) and incubating for 4 h at 37˚C in a humidified atmosphere containing 5% CO₂ and measuring the absorbance at 450 nm with a spectrophotometer.

A total of 5x10⁶ splenocytes and 4 µg concanavalin A were added to the individual wells of a 24-well plate. The plate was incubated at 37˚C in a humidified atmosphere containing 5% CO₂ for 48 h. The medium was exchanged for RPMI-1640 medium containing LH and incubated at 37˚C in a humidified atmosphere containing 5% CO₂ for 48 h. The medium in the individual wells was collected and stored at -80˚C until measurement of the IL-2 concentrations in the medium using an ELISA kit (Mouse IL-2 Quantikine ELISA kit, cat. no. DY402-05, R&D Systems, Inc.) according to the manufacturer's instructions.

Medium-low-glucose DMEM (Merck KGaA) containing 10% FBS, isobutyl-methylxanthine (IBMX), dexamethasone (DEX) and insulin (all FUJIFILM Wako Pure Chemical Corporation) and Antibiotic-Antimycotic Mixed Stock Solution were used for cell culture at 37˚C in a humidified atmosphere containing 5% CO₂.

3T3-L1 cells, a fibroblast isolated from the embryo of a mouse, were purchased from the Japanese Collection of Research Bioresources Cell Bank (National Institutes of Biomedical Innovation, Health, and Nutrition, Ibaraki, Japan). 3T3-L1 cell cultivation and differentiation were performed using a previously described method (18). 3T3-L1 cells at the 7th passage were cultured at 37˚C with 5% CO₂ in DMEM supplemented with 10% FBS and antibiotic/antimycotic. To individual wells of a 24-well plate, 3x10⁴ cells/well were seeded and the plate was incubated for 48 h in maintenance medium (10% FBS and 1% antibiotic/antimycotic in DMEM) at 37˚C in a humidified atmosphere containing 5% CO₂. After reaching 90% confluence, the medium was changed to differentiation induction medium (0.5 mM IBMX and 1.0 µM DEX in the maintenance medium) and incubated at 37˚C for 48 h in a humidified atmosphere containing 5% CO₂. The medium was replaced with differentiation medium (1.7 µM insulin in the maintenance medium) and incubated at 37˚C for 48 h in a humidified atmosphere containing 5% CO₂.

3T3-L1-derived adipocytes were used to determine the lipid accumulation and adiponectin secretion values. After the 3T3-L1 cell differentiation, the medium was replaced with maintenance medium containing samples (0.00001, 0.0001, 0.001, 0.01 mg/ml) or pioglitazone (0.28, 2.8, 28, 280 µM), respectively, and incubated for 96 h (medium was changed every 48 h) at 37˚C in a humidified atmosphere containing 5% CO₂. Medium in the individual wells was collected and stored at -80˚C until use for the determination of the adiponectin level using the ELISA kit (Mouse adiponectin/Acrp30 DuoSet^® ELISA, cat. No. DY1119-05, R&D Systems, Inc.) according to the manufacturer's instructions. Cells were fixed with 10% formaldehyde for 1 h at room temperature. Following incubation in 2-propanol for 1 min at room temperature, the lipid droplets in the cells were stained with 3 mg/ml ORO in 60% 2-propanol solution for 20 min at room temperature. After washing with 60% 2-propanol, 4% triton-X-100/2-propanol solution was added to the well for 5 min at room temperature to extract ORO in lipid droplets of adipocytes. The absorbance of the extracted solution was measured at 492 nm with a spectrophotometer and the lipid accumulation value was quantified using an ORO standard curve. Pioglitazone (FUJIFILM Wako Pure Chemical Corporation) was used as a positive control.

A total of 12 6-week-old male BALB/c mice were divided into three groups (n=4/group) and allowed free access to 0.1% water solution of methanol extract of LH or LT for 14 days. Four days later, mice were subcutaneously injected with 50 µl 1 mg/ml concanavalin A suspended in complete Freund's adjuvant (Merck KGaA) through the tail to stimulate T cells. After 7 days, splenocytes were prepared as aforementioned. A total of 5x10⁶ splenocytes and 4 µg concanavalin A were added to the individual wells of a 24-well plate. The plate was incubated at 37˚C in a humidified atmosphere containing 5% CO₂. The medium in the individual wells was collected following 25 and 50 h of incubation and stored at -80˚C until measurement of the IL-10, IL-17 and TNF-α concentrations in medium using an ELISA kit (Mouse IL-10, cat. no. DY417-05, IL-17, cat. No. DY421-05, and TNF-α, cat. No. DY410-05, Quantikine ELISA kits, respectively, R&D Systems, Inc.) according to the manufacturer's instructions.

A total of 15 6-week-old male BALB/c mice were divided into three groups (n=5/group) and allowed free access to 0.1% or 0.2% water solution of methanol extract of LH to 0.1% LH-loaded group or 0.2% LH-loaded group for 14 days, respectively, to determine the dose-dependent effect of LH. To determine low-(0.1%) and high-dose (0.2%) concentrations of LH, voluntary water intake/day of mice was observed so that the mice ingested 100 or 200 mg/day sample extract. After fasting for 15 h, the blood glucose levels of mice in the tail vein were measured using a blood glucose meter and test paper (ACCU-CHECK^® Aviva Strip, Roche Diagnostics Co., Ltd.). After mouse euthanasia, the blood was collected from the heart by heparin-treated syringe and centrifuged at 15,000 x g for 5 min at room temperature to obtain plasma and adiponectin concentration was measured using an ELISA kit (Mouse adiponectin/Acrp30 DuoSet^® ELISA, cat. No. DY1119-05, R&D Systems, Inc.) according to the manufacturer's instructions.

A total of 15 6-week-old male BALB/c mice were divided into three groups (n=5/group) and allowed free access to 0.1% or 0.2% water solution of LH for 14 days. Furthermore, obese 13-week-old male BALB/c mice established by the administration of high-fat diet (D12492, RESEARCH DIETS Inc.) were divided into two groups (n=6/group) and allowed free access to 0.1% water solution of LH for 30 days. Four days later, the mice were subcutaneously injected with 50 µl 1 mg/ml concanavalin A in complete Freund's adjuvant (Merck KGaA) through the tail to stimulate T cells. After 7 days, spleen and subcutaneous fat were extirpated under sterile conditions following mouse euthanasia via inhalation of isoflurane. The splenocytes were prepared as aforementioned. The subcutaneous fat was minced in RPMI-1640 medium containing 0.5% collagenase type 4 (Worthington Biochemical Corporation) using dissecting scissors. After incubation at 37˚C for 30 min under shaking, DMEM containing 2% FBS was added to further mince the tissue by pipetting and filtered through a 40-µM mesh. The tissue suspension was centrifuged at 800 x g for 5 min at room temperature to separate the adipocyte and the stromal vascular cell fraction. Lymphocytes were isolated from stromal vascular cells using Lympholyte^®-M Cell Separation Media (Cedarlane).

Splenocytes and lymphocytes isolated from stromal vascular cell fraction were fixed and permeabilized using Fixation/Permeabilization Concentrate and Diluent kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions and which divided into 100 µl to four tubes (tubes A-D). To analyze the percentage of CD4⁺ cells in the lymphocytes, the splenocytes or the lymphocytes in Tube B were stained with 180-fold diluted Super Bright 600-conjugated CD4 antibody [monoclonal (RM4-5), Super Bright^™ 600, eBioscience^™, cat. no. 63-0042-82, Thermo Fisher Scientific Inc.]. To analyze the percentage of Tregs (CD4⁺ CD25⁺ Foxp3⁺ T cells), IL-10⁺ Tregs (CD4⁺ Foxp3⁺ IL-10⁺ T cells), cytotoxic T-lymphocyte antigen (CTLA)-4⁺ Tregs (CD4⁺ Foxp3⁺ CTLA-4⁺ T cells) and T helper (Th)17 (CD4⁺ IL-17A⁺ T cells), the splenocytes or the lymphocytes in Tube C were stained with 180-fold diluted Super Bright 600-conjugated CD4 (cat. no. 63-0042-82) and 160-fold diluted PE-Cyanine5-conjugated CD25 antibody [CD25 monoclonal antibody (PC61.5), PE-Cyanine5, cat. no. 15-0251-82)], 20-fold diluted Alexa Fluor 488-conjugated Foxp3 [(cat. no. 53-4774-42)], 80-fold diluted Alexa Fluor 700-conjugated IL-10 [IL-10 monoclonal antibody (JES5-16E3), Alexa Fluor 700, cat. no.56-7101-82)], 80-fold diluted phycoerythrin-conjugated CTLA-4 antibody [CD152 (CTLA-4 monoclonal antibody (UC10-4B9), PE, cat. no. 12-1522-82] and 80-fold diluted eFluor 450-conjugated IL-17A antibody [IL-17A monoclonal antibody (eBio17B7), eFluor 450, cat. no. 48-7177-82; all eBioscience^™, Thermo Fisher Scientific, Inc.]. To analyze the percentage of CD8⁺ T cells in the lymphocytes, the splenocytes or the lymphocytes in Tube D were stained with 80-fold diluted Super Bright 702-conjugated CD8 antibody (cat. no. 67-0081-82; Thermo Fisher Scientific, Inc.). The splenocytes or the lymphocytes in Tube A were not stained. These tubes were incubated on ice for 1 h. Flow cytometric analysis was conducted using Attune NxT Acoustic Focusing Cytometer and the data were analyzed using the Attune NxT Software version 2.6 (Thermo Fisher Scientific, Inc.).

Methanol extracts of LH and LT were sonicated with methanol at 40 kHz, 40˚C for 1 h and filtrated through a 0.45-µm filter membrane to obtain solutions with a concentration of 1 mg/ml. Formononetin (0.05 mg/ml) was purchased from Tokyo Chemical Industry Co., Ltd. Lespedezaflavanone H (0.025 mg/ml) was gifted by Professor Toshio Miyase, School of Pharmaceutical Sciences, University of Shizuoka (Shizuoka, Japan). HPLC analysis of methanol extracts of LH and LT was performed by LC-20AD Solvent delivery unit (Shimadzu Scientific Instruments) with COSMOSIL 5PE-MS packed column (4.6x250.0 mm, 5 µm particle size), column oven (cat. no. CO631C, GL Sciences) and UV detector (UV-VIS detector S-3170, Soma Optics, Ltd.). Analysis was performed using the following conditions: Mobile phase, acetonitrile:water=60:40; flow rate, 1 ml/min; detection wavelength, 214 and 280 nm; injection volume, 10 µl and column temperature, 40˚C. The peak areas were analyzed by Chromato-PRO version 5.0.0.199 (Runtime Instruments Co., Ltd.).

Statistical analysis was conducted using the SigmaStat statistical software ver. 2.03 (SPSS Inc.). Student's unpaired t test, and one-way ANOVA followed by Dunnett's multiple comparison or Bonferroni's test were performed for comparisons. P<0.05 was considered to indicate a statistically significant difference. Data are represented the mean ± SD of two independent experiments.

PPARγ binding activity of LB, LH, LM, and LT was determined by ELISA. Pioglitazone, a PPARγ agonist, had significant binding activity on PPARγ (Fig. 1A). LB had significantly reduced PPARγ binding activity at 0.038, 0.150, 0.300 mg/ml. LM exhibited no change in binding activity; LH exhibited a significant increase at 0.300 but decrease in 0.038 mg/ml compared with that of the control. LT had significantly increased binding activity at 0.075-0.300 mg/ml (Fig. 1B). Trolox exhibited a dose-dependent increase in DPPH radical scavenging activity (Fig. 1C). LH and LT exhibited a dose-dependent increase in these DPPH scavenging, with LH having a stronger radical scavenging activity than LT at 6.25-500.00 µg/ml (Fig. 1D).

PPARγ expression in T cells decreases the production of inflammatory cytokines (19). To investigate whether PPARγ-agonistic LH suppressed the production of inflammatory cytokine IL-2, murine splenocytes were incubated with LH, and the splenocyte toxicity and IL-2 secretion from splenocytes were assessed. Splenocyte toxicity was not observed in the presence of LH whereas LH showed significant proliferation of splenocytes at 0.0625 and 0.1250 compared with that at 0 mg/ml (Fig. 1E). IL-2 secretion from splenocytes was evaluated following the addition of LH. LH significantly suppressed IL-2 secretion at 0.15625 mg/ml compared with the control (Fig. 1F).

Activating of PPARγ expressed in adipocytes increases lipid accumulation (20) and adiponectin secretion (21). Therefore, LH and LT, which showed PPARγ binding activity, were evaluated for lipid accumulation and adiponectin secretion of 3T3-L1-derived adipocytes. Pioglitazone, significantly decreased lipid accumulation at 0.28 µM (Fig. 2A). LH exhibited significantly increased lipid accumulation at 0.00001 and 0.00100 mg/ml, whereas LT showed significantly increased accumulation at 0.00100 mg/ml (Fig. 2B). Pioglitazone significantly increased adiponectin concentration at 28-280 µM (Fig. 2C). Further-more, LH showed increased adiponectin concentration at 0.00001-0.00100 mg/ml (Fig. 2D).

To determine whether LH and LT influence the levels of anti-inflammatory cytokine IL-10 as well as proinflammatory cytokines IL-17 and TNF-α secreted from the spleen, splenocytes of 0.1% LH- or LT-exposed mice were cultured. At 25 h, the splenocytes of 0.1% LH-exposed mice exhibited significantly increased IL-10 levels, whereas those of 0.1% LT-exposed mice exhibited significantly decreased IL-10 levels (Fig. 3A). Furthermore, splenocytes of LH- and LT-exposed mice exhibited decreased IL-17 levels, but the decrease was had not significant (Fig. 3B). Similarly, splenocytes of 0.1% LH-exposed mice decreased exhibited TNF-α levels, but the decrease was not significant (Fig. 3C).

PPARγ agonists increase levels of glucose transporter (GLUT)-4 in adipocytes of peripheral tissues such as muscle and liver to facilitate glucose uptake (22) and promote translocating fatty acids in the peripheral tissues to adipose tissues, leading to suppression of blood glucose levels (23). In white adipose tissue, PPARγ agonists promote adiponectin transcription, synthesis and secretion (21). Therefore, it was determined whether LH, which exerted PPARγ agonistic activity, affects blood glucose and adiponectin levels. Mice exposed to 0.1 and 0.2% LH exhibited significantly decreased fasting blood glucose levels compared with the control mice (Fig. 3D). 0.1% LH-exposed mice showed lower fasting adiponectin levels whereas 0.2% LH-exposed mice showed higher fasting blood adiponectin levels than the control mice, but these levels were not significantly different (Fig. 3E).

To determine whether Tregs essential for suppressing adipose tissue inflammation were increased in subcutaneous fat of mice, the percentage of CD25⁺ Foxp3⁺ Tregs in the CD4⁺ T cell population was analyzed via flow cytometry (Fig. 4A). Th17 cells mutually maintain a balance with Tregs to regulate the immune system (24); therefore, the percentage of IL-17A⁺ CD4⁺ Th17 cells in CD4⁺ T cells was analyzed via flow cytometry (Fig. 4B). Mice exposed to 0.1% LH had significantly increased Tregs but there was no effect in mice exposed to 0.2% LH compared with the control (Fig. 4C). Although mice exposed to 0.2% LH exhibited significantly decreased Th17 cells and Th17/Treg ratio, 0.1% LH exhibited no significant decrease in Th17 cells and Th17/Treg ratio compared with control mice (Fig. 4D).

The proportion of CD25⁺ Foxp3⁺ and IL-17A⁺ CD4⁺ T cells was analyzed (Fig. 5A and B). Mice exposed to 0.1 and 0.2% LH had significantly decreased Tregs compared with the control (Fig. 5C). Mice exposed to 0.2% LH exhibited significantly increased Th17 cells (Fig. 5D), whereas mice exposed to 0.1 and 0.2% LH had significantly increased Th17/Treg ratio (Fig. 5E).

To determine whether LH creates an immunosuppressive milieu in subcutaneous adipose tissue by increasing suppressive cytokine IL-10 secretion from Tregs of the adipose tissue, IL-10⁺ Tregs in subcutaneous adipose tissue and spleen were analyzed via flow cytometry (Fig. 6A and B). No significant difference was observed in IL-10⁺ Tregs in subcutaneous adipose tissue (Fig. 6C); however, in the spleen, mice exposed to 0.1 and 0.2% LH exhibited significantly decreased IL-10⁺ Treg levels (Fig. 6D).

To determine whether LH creates an immunosuppressive milieu in the subcutaneous adipose tissue by increasing CTLA-4 expression on Tregs to induce antigen-presenting cell (APC) dysfunction, CTLA-4⁺ Tregs in the subcutaneous adipose tissue and spleen were analyzed via flow cytometry (Fig. 7A and B). Mice exposed to 0.1% LH exhibited significantly increased CTLA-4⁺ Treg levels but there was no effect on mice exposed to 0.2% LH in the subcutaneous adipose tissue (Fig. 7C); no significant difference was observed in CTLA-4⁺ Tregs in the spleen for either group (Fig. 7D).

To determine whether the significant increase in CTLA-4⁺ Tregs induced by LH in subcutaneous adipose tissue affected expression of CD4⁺ and CD8⁺ T cells, the proportion of CD4⁺ and CD8⁺ T cells in was analyzed via flow cytometry. Mice exposed to 0.1% LH showed decreased CD4⁺ T cells in the subcutaneous adipose tissue (Fig. 8A); however, in the spleen, CD4⁺ T cells were increased in LH-exposed mice, significantly so in 0.2% LH-exposed mice (Fig. 8B). 0.1% LH-exposed mice had significantly decreased CD8⁺ T cells in subcutaneous adipose tissue (Fig. 8C); however, in the spleen, no significant difference was observed in CD8⁺ T cells (Fig. 8D).

(Fig. S1), peak areas of No. 1 and 2, 4, 8, 10, and 11 of LH were significantly higher than those of LT (Table SI). Formononetin, which is hydroxylated at the 7-position and methylated at 4'-position of isoflavone, showed approximately 4 min of the retention time. Lespedezaflavanone H, which is prenylated at the 6-, 8-, and 2'-positions of flavanone, showed approximately 12 min of the retention time. Based on these results, LH and LT might contain low-polarity flavonoids, such as isoflavone or prenylated flavonoids. Although LH and LT contain almost similar components in the low-polarity flavonoids targeted in this study, the properties are thought to differ because of the difference in the content of these components.