The Chinese medicinal herbs in this study were A: Rhizoma Alismatis (Zexie), B: Fructus Crataegi (Shanzha), C: Semen Coicis (Yiyiren), D: Rhizoma Atractylodis Macrocephalae (Baizhu), E:Rhizoma Atractylodis (Cangzhu), F:Sclerotium Poriae Cocos (Fuling), G: Semen Cassiae (Juemingzi), H: Folium Sennae (Fanxieye), I: Radix Angelica Sinensis (Danggui), J: Rhizoma Curumae (Ezhu), K: Flos Chrysanthemi (Juhua), L: Radix Notoginseng (Sanqi), M: Folium Nelumbinis (Heye), N: Herba Taraxaci (Pugongying), O: Pericarpium Citri Reticulatae (Chenpi), P: Fructus Schisandrae Chinensis (Wuweizi) and Q: Fructus Mori (Sangshen). All the herbs were purchased from Eu Yan Sang International Ltd and Tung Fong Hong Medicine Co Ltd in Hong Kong and were authenticated by organoleptic characteristics according to the Pharmacopoeia of the People's Republic of China (2005 edition, volume I). Each experimental species (300 g) was deposited in the Herbarium of the Department of Biology, Hong Kong University of Science and Technology (Additional file 1).

These herbs were divided into two groups. The first group includes the herbs that treat obesity or obesity-related diseases according to the Pharmacopoeia of the People's Republic of China and other literature. The second group includes common dietary herbs not documented to have functional effects on obesity.

Two methods, namely water and ethanol extractions, were used to prepare herbal extracts in the present study. In water extraction, each herb was ground and boiled twice in eight units of water for one hour. In ethanol extraction, each grounded herb was immersed in eight units of 95% ethanol for two hours and reflux for further two hours. Both water and ethanol extracts were dried into powder and stored at -80°C.

The stable cell line of Caco-2/TC7 transfected with the human ApoA-IV promoter was provided by Prof M Lacasa (Université Pierre et Marie Curie, France). TC7 is the selected clone from Caco-2 cells 28. The cells were grown at 37°C in a water-saturated incubator containing 5% CO₂ in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 20% heat-inactivated fetal bovine serum (HI-FBS), 100 U/ml penicillin and 100 μg/ml streptomycin. In all experiments, cells were maintained at about 90% confluence. The high confluence condition allowed cell differentiation, which increases the expression of ApoA-IV. Prior to drug treatment, Caco-2/TC7 cells were seeded on 24-well microtiter plates (40000 cells/well) for 24 hours. Mouse 3T3-L1 fibroblast cells (ATCC no.CL-173) were obtained from American Type Culture Collection (USA) and maintained at 37°C in a water-saturated incubator containing 5% CO₂ and in DMEM supplemented with 4.5 g/L glucose, 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin. Induction of lipogenic differentiation was detailed in a previous study 29. Briefly, the confluent cultures were treated with a differentiation cocktail containing dexamethasone (1 μM, Sigma, USA), insulin (1.8 μM, Sigma, USA) and dibutryl-cAMP (300 μM, Sigma, USA) for 72 hours to induce lipogenesis. The cultures were set as day 0, and replaced with the culture medium containing insulin (1.8 μM) for every two days. At day 10, about 80% of cultures were induced to contain triglyceride (TG). Treatments including serum starvation (DMEM only), insulin (1.8 μM), Radix Angelica Sinensis and Rhizoma Alistmatis (10, 1 and 0.1 mg/ml water extracts) were given to differentiated cultures (on day 10) for 72 hours. Unless described otherwise, all the culture reagents were purchased from Invitrogen Technologies (USA).

The lipid micelles was used to mimic the duodenal micelles resulting from digestion of lipid 30, and prepared according to the method of Carrière et al. 27. The stock solution contained 0.6 mM oleic acid, 0.2 mM L-α-lysophosphatidylcholine, 0.05 mM cholesterol, 0.2 mM 2-monooleoylglycerol and 2 mM taurocholic acid, and used in the dilutions from 1:1000 to 1:3000.

Luciferase assay was performed with a commercial kit (Tropix, USA). Briefly, the treated Caco-2/TC7 cells were collected and re-suspended by 0.2% Triton X-100, 1 mM dithiothreitol and 100 mM potassium phosphate buffer (pH7.8). The lysate was subjected to luciferase assay and protein assay. The luminescent reaction was measured by Tropix TR717 microplate luminometer, while the protein concentrations were measured according to the Bradford method 31 with a protein assay kit (Bio-Rad Laboratories, USA). The luciferase activity reading was normalized by protein amount in the sample.

Total RNAs, isolated by TRIzol reagent (Invitrogen, USA) from treated Caco-2/TC7 cultures, were reverse-transcribed to cDNAs by Moloney murine leukemia virus reverse transcriptase (Invitrogen, USA) according to the manufacturer's instructions. Quantitative PCR was performed with SYBR Green Master mix and Rox reference dye according to the manufacturer's instructions (Applied Biosystems, USA). The primers used for human ApoA-IV (NM_000482) were 5'-ATG TTC CTG AAG GCC GTG G-3' and 5'-TGC AGG TCA CCT GCG TAA G-3' (-105 to -334), and human18S rRNA (NR_003286) 5'-TGT GAT GCC CTT AGA TGT CC-3' and 5'-GAT AGT CAA GTT CGA CCG TC-3'(-1494 to -1813). The SYBR green signal was detected by a quantitative PCR (Mx3000p multiplex, Stratagene, USA). The relative transcript expression levels were quantified according to the ΔΔCt (cycle threshold) method 32. The calculation was done with the Ct value of 18S rRNA to normalize the Ct value of target gene in each sample to obtain the ΔCt value, which was then used to compare different samples. The PCR products were analyzed by gel electrophoresis, while the specificity of amplification was confirmed by melting curve.

Oil Red O at 0.2% in isopropanol was mixed with water (3:2, v/v) and filtered. Experimental cultured cells were washed with PBS, fixed by paraformaldehyde (4% in PBS, Sigma, USA) for 5 minutes, incubated with filtered Oil Red O for 30 minutes, and washed twice with PBS. The stained TG was extracted by isopropanol and its quantity was measured at 490 nm absorbance 28.

One-way analysis of variance (ANOVA) was carried out with SPSS software (version 13.0, SPSS, USA). The levels of statistical significance were P < 0.05 (*), P < 0.01 (**) and P < 0.001 (***).

ApoA-IV was first chosen for the investigation of anti-obesity effect due to its potential role in modulating food intake 171819. Accordingly, a promoter-reporter system containing a human ApoA-IV promoter (about 230 bp) tagged with a luciferase reporter gene was employed 27. This reporter construct was stably transfected into cultured Caco-2/TC7 cells for the screening of potential drugs that regulate the transcriptional activity of ApoA-IV promoter in gut cells. The functionality of this reporter construct was validated by its responsiveness to high concentration of lipid. Cultured Caco-2/TC7 cells were treated with lipid micelle (an artificial mixture of lipids mimicking the duodenal micelles after ingestion) at concentrations of 1:1000 and 1:2500. After a 24-hour treatment, total RNAs were extracted to quantify the amount of ApoA-IV mRNA by a quantitative PCR. Results showed that the expression of ApoA-IV mRNA was not changed by the lipid micelle at the concentration of 1:2500, possibly due to the insufficient amount of lipid micelle to stimulate gene transcription (Figure 1A). However, the induction effect was observed at a higher concentration of 1:1000; the ApoA-IV mRNA was up-regulated to nearly 6 folds compared with the buffer-treated control (Figure 1A). These results confirmed the previous findings that high-fat diet increases the ApoA-IV expression in gut cells 30.

To assess the transcriptional activity of the ApoA-IV promoter, we treated Caco-2/TC7 cells with various concentrations of lipid micelles (1:1000 to 1:3000) for 48 hours and then collected them for luciferase activity. The addition of lipid micelles increased the promoter's activity in a dose-dependent manner. Induction of over six folds was observed in the Caco-2/TC7 cells treated with 1:1000 lipid micelles (Figure 1B). The concentrations of lipid micelles from 1:2000 to 1:1000 were effective in activating the promoter, which was consistent with the findings that ApoA-IV mRNA expression with a concentration at 1:2500 did not produce any response (Figure 1A and Figure 1B). Finally, the optimal treatment time was determined for inducing ApoA-IV promoter activity by lipid micelles. Cultures were treated with two concentrations of lipid micelles (1:2500 and 1:2000) at various time points (i.e. 24, 48 and 72 hours). The promoter activity induced by lipid micelles at 48 hours was similar to that at 72 hours, suggesting that treatment time of 48 hours was sufficient for activation (Figure 1C). Activation was not observed at 1:2500. Therefore, the ApoA-IV promoter is a suitable screening tool in Caco-2/TC7 enterocytes.

Water and ethanol extracts of 17 Chinese medicinal herbs in water and ethanol extractions were screened for their effect in regulating ApoA-IV promoter activity in cultured Caco-2/TC7. The herbs were divided into two groups, namely those that are used to treat obesity (A-H) and those that are not used to treat obesity (I-Q). The powders of water and ethanol extracts of these herbs were dissolved in water and DMSO respectively. The pH value of each solution in the medium was measured to ensure that the cell culture condition was not affected by the addition of the herb itself. The results showed that more than half of the herbs in both groups induced the ApoA-IV promoter activity (Figure 2). In the anti-obesity herb group, Rhizoma Alismatis (A), Fructus Crataegi (B), Semen Coicis (C), Rhizoma Atractylodis Macrocephalae (D) and Rhizoma Atractylodis (E) increased the ApoA-IV promoter activity by more than two folds (Figure 2). Moreover, both water and ethanol extracts of the herbs demonstrated similar effects, suggesting high availability of active ingredients in the herbs. In the group of herbs not documented for anti-obesity treatment, Radix Angelica Sinensis (I), Rhizoma Curcumae (J), Flos Chrysanthemi (K), Radix Notoginseng (L), Folium Nelumbinis (M) and Herba Taraxaci (N) significantly up-regulated the transcriptional activity of ApoA-IV promoter after 48 hours of treatment (Figure 2).

Water extracts of Rhizoma Alismatis (A) and Radix Angelica Sinensis (I) were further examined for their dose-dependent effect in Caco-2/TC7. After 48 hours of treatment at various concentrations (0 to 10 mg/ml), the luciferase activity was stimulated gradually in response to the increase concentrations of extracts (Figure 3A). Lastly, the same activation effect of Rhizoma Alismatis and Radix Angelica Sinensis in up-regulating the ApoA-IV mRNA expression was revealed in treated Caco-2/TC7 cells (Figure 3B). Lipid micelles served as the positive control for mRNA analysis. These results showed that Rhizoma Alismatis and Radix Angelica Sinensis stimulated the ApoA-IV promoter activity.

The 3T3-L1 pre-adipocyte model for adipogenesis studies 283334 was employed to further determine the anti-obesity activity of Rhizoma Alismatis and Radix Angelica Sinensis. Induced by a chemical cocktail, the pre-adipocytes were differentiated, indicated by morphological changes and accumulation of triglyceride (TG). The TG vesicles inside the cells were stained by Oil Red O dye for visualization and quantification. The differentiated 3T3-L1 cells were serum-starved or treated with insulin for three days to confirm that TG formation did respond to changes. The TG content decreased about 50% in the serum starvation group, and increased to 160% in the insulin group (Figure 4A). The differentiated 3T3-L1 cells were treated with various concentrations of Rhizoma Alismatis and Radix Angelica Sinensis for three days. With the addition of 10 mg/ml and 1 mg/ml water-extracts, both Rhizoma Alismatis and Radix Angelica Sinensis reduced the TG levels to varying extents (Figure 4B); the most significant effect (over 30% TG reduction) was observed in Radix Angelica Sinensis treatment at 10 mg/ml. These results showed that both Rhizoma Alismatis and Radix Angelica Sinensis inhibited the formation of TG.