The individual plants required to produce Yajieshaba were identified by Professor Chao Zhang, Yunnan University of Traditional Chinese Medicine (Kunming, China). All the plants were purchased from Xishuangbanna Dai Autonomous Prefecture People's Hospital (Xishuangbanna, China). Specimens were deposited in the Herbarium of Pharmacognosy, Yunnan University of Traditional Chinese Medicine (Kunming, China). Yajieshaba powder was prepared using the following components: 30 g Pueraria lobata Ohwi, 15 g root of Dregea sinensis Hemsl, 15 g whole plant of Arundina graminifolia (D. Don) Hochr, 15 g stem of Fibraurea recisa Pierre, 30 g stem of Mappianthus iodioides Hand. Mazz, 10 g vine of Bousigonia mekongensis, 15 g root of Clerodendrum chinese (Osbeck) Mabb and 10 g root of Glycyrrhiza uralensis Fisch. The aforementioned herbal components were then ground and mixed. The powder of Yajieshaba was extracted three times at 100°C with water (12 volumes, 30 min; 8 volumes, 20 min; and 8 volumes, 20 min, respectively). Extracts were combined, condensed, lyophilized and stored at 4°C until use with a yield of 5.2%.

Palmatine and berberine (purity ≥98%) were purchased from Nanjing Jingzhu Bio-technology Co., Ltd. (Nanjing, China), lyophilized ovalbumin (OVA; purity ≥98%) was purchased from Sigma-Aldrich (Merck Millipore, Darmstadt, Germany). Loratadine was purchased from Xi'an Janssen Pharmaceutical Ltd. (Shaanxi, China). Alexa Fluor 647-anti-Forkhead box P3 (Foxp3) (catalog no., 560401), fluorescein isothiocyanate (FITC) anti-CD4 (catalog no., 553047), phycoerythrin (PE)-anti-CD25 (catalog no., 558642) and the mouse Foxp3 fixation and permeabilization buffers (catalog no., 560409) were purchased from BD Biosciences (Franklin Lakes, NJ, USA).

All experiments were performed on an Agilent 1200 series high performance liquid chromatography (HPLC) system (Agilent Technologies Inc., Santa Clara, CA, USA), which included a binary pump, an autosampler, a column oven and a diode array detector with an on-line degasser. The separations were achieved using a TSK gel octadecyl silane-100 V (internal diameter, 4.6×250 mm; particle diameter, 5 µm; Tosoh Corp., Tokyo, Japan).

The samples were ultrasonically extracted (53 KHz, room temperature) in 100 ml acetonitrile-0.4% phosphate buffer (32:68) for 30 min. The extracts were further eluted with acetonitrile-0.4% phosphate buffer (32:68) for 15 min during the mobile phase. The detection wavelength was set at 354 nm, the oven temperature was set at 40°C and the flow rate was set to 1.0 ml/min. A 10-µl injection volume was used in all analyses. Palmatine and berberine were dissolved in the mobile phase. Palmatine and berberine peaks were identified based on their retention time, as well as the ultraviolet spectral properties of the standards.

A total of 48 specific pathogen-free female BALB/c mice (age, 7–8 weeks; mean weight, 20±2 g), were purchased from Beijing HFK Bioscience Co., Ltd. (Beijing, China). The animals were housed in a stainless steel cage containing sterile paddy husk as bedding and were kept in a ventilated animal room. The mice had ad libitum access to water and commercial laboratory mice food and were kept at 22±1°C and 60±10% humidity under a 12-h light/dark cycle. The animals were given 10 days to acclimatize to the environment. The present study was approved by the Institutional Ethical Committee on Animal Care and Experimentations of Yunnan University of Traditional Chinese Medicine (Kunming, China; no. R-062014015). Reasonable efforts were made to minimize animal suffering.

Mice were divided into 8 groups, each consisting of 6 animals (Table I) and treated as follows: A, mice were not sensitized (control); B, mice were sensitized but not treated; C, mice were treated with palmatine (0.1048 g/kg); D, mice were treated with berberine (0.0352 g/kg); E, mice were treated with loratadine (1.37 g/kg); F, mice were treated with 2.35 g/kg body weight Yajieshaba; G, mice were treated with 3.50 g/kg body weight Yajieshaba; H, mice were treated with 4.70 g/kg body weight Yajieshaba. Food allergy was induced by subcutaneous injection of 3 mg OVA in 150 µl aluminum hydroxide (Xilong Chemical Co., Ltd., Shantou, China) gel and 150 µl 0.9% sodium chloride solution in the neck region. This was repeated at 48 and 96 h (14–16). Treatments were administered by oral gavage at 30 min following the second and third OVA injections. Half of the animals were sacrificed under diethyl ether (ShanDian Pharmaceutical Co., Ltd., Yunnan, China) inhalation anesthesia by cervical dislocation 2 h following the second OVA injection and the remaining ones were sacrificed 2 h following the third OVA injection to collect samples. A schematic diagram of the treatment schedule is presented in Fig. 1.

Scratching behavior was observed in a quiet experimental room. Each animal was placed in a plastic observation cage (29×18×16 cm) for ~10 min prior to scoring of scratching behavior. One observer, who was blinded to the experimental design, counted the total number of scratches on the injection site in each treatment group. A scratch was defined as lifting of the forelimb towards the injected area followed by placing of the limb back to the floor, regardless of how many scratching strokes took place between those two movements. However, if the scratching bout extended beyond 1 min, it was counted as two events (17).

Blood samples (1–1.5 ml) were collected from each mouse via the retro-orbital vein under diethyl ether inhalation anesthesia and centrifuged at 393 × g for 10 min at room temperature. The serum was stored at −20°C until further analysis. The amounts of IgE (ml037602), IgG (ml037601), interleukin (IL)-4 (ml002149), IL-10 (ml037873), IL-17 (ml037866), IL-21 (ml037878), interferon (IFN)-γ (ml002277), and tumor necrosis factor (TNF)-α (ml002095) were measured using respective ELISA kits (Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China).

At the end of the treatment, all mice were sacrificed under diethyl ether inhalation anesthesia by cervical dislocation. Spleens were collected in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA). The spleens were minced and dissociated for flow cytometry by pressing them through a 70-µm filter membrane to produce a single-cell suspension. The red blood cells were lysed (BD Pharm Lyse™ lysing solution; BD Biosciences) and the remaining white blood cells were diluted with staining buffer (FBS; BD Biosciences) to a concentration of 1×10⁶ cells/ml. The cells were incubated with FITC-conjugated anti-mouse CD4 (1:20) and PE-conjugated anti-mouse CD25 (1:40) at room temperature in the dark for 20 min. The cells were fixed with Foxp3 fixing solution (1:20) at 4°C for 30 min and permeabilized with Foxp3 permeabilization solution (1:20) at 37°C for another 30 min according to manufacturer's protocol, and were stained with Alexa Fluor 647 anti-mouse Foxp3 (1:20) prior to analysis using a flow cytometer (BD FACSCalibur, BD Biosciences) to quantify Treg frequncies.

Values are expressed as the mean ± standard deviation. One-way analysis of variance was performed to compare the treatment effects on dependent variables. P<0.05 was considered to indicate a statistically significant difference.

The linear association between injection quantity (mg/10 µl, X-axis) and peak area (Y-axis) was Y=34.31X+21.82 for palmatine (R²=0.999) and Y=33.51X+0.2250 for berberine (R²=0.999). The limits of detection for palmatine and berberine were 0.1496 and 0.06432 µg/ml, respectively. The limits of quantification of palmatine and berberine were 0.3990 and 0.3216 µg/ml, respectively. The average recovery rates of palmatine and berberine were 99.37 and 97.15%, respectively.

The HPLC elution times for palmatine and berberine were 10.2 and 11.0 min, respectively (Fig. 2). The palmatine and berberine content in Yajieshaba formulation were 1.49 and 0.50 mg/g, respectively.

Subcutaneous injection of 3 mg OVA into the neck region elicited a clear scratching response in mice. The scratching frequencies of mice in different treatment groups at 50 (Fig. 3A) and 98 h (Fig. 3B) were evaluated. Scratching frequencies increased from 0.325 and 1.43 scratches/minute in the control group to 2.58 and 2.23 scratches/minute in the untreated OVA-sensitized mice at 50 and 98 h, respectively. Mice treated with loratadine, palmatine, berberine or Yajieshaba at 3.50 or 4.70 g/kg body weight exhibited a markedly lower scratching frequency than the OVA-sensitized mice at both time points. Mice receiving a low dose of Yajieshaba (2.35 g/kg body weight) did not have a markedly different scratching frequency compared with the mice in OVA-sensitized group at 50 and 98 h. Mice in the palmatine (0.1048 g/kg) and Yajieshaba (3.50 and 4.70 g/kg) treatment groups had a markedly lower scratching frequency compared with mice in the loratadine group at 50 h. Palmatine treatment reduced the scratching frequency by ~40%, indicating that it may be the major active anti-allergic component of Yajieshaba.

The serum levels of IgE and IgG (Fig. 4) as well as IL4, IL-10, IL-21, IFN-γ and TNF-α (Table II) were measured at 50 and 98 h to identify the mechanism through which Yajieshaba suppressed the scratching frequency in OVA-sensitized mice.

IgE levels were similar in the normal group and the model group at 50 h (Fig. 4A); however, IgE levels in Groups D, F and G were higher than the normal group at 50 h (Fig. 4A). Notably, the IgE level in the model group was significantly lower than that in the normal group at 98 h (P<0.05; Fig. 4B). Palmatine, loratadine, and low- and middle-dose Yajieshaba were able to reverse this OVA-induced reduction at 98 h (Fig. 4B). IgE was increased in group D after 50 h; however, it reduced after 98 h. Similarly, at 50 h, IgE levels were higher in group D than group C, whereas the opposite was observed after 98 h.

OVA sensitization decreased serum IgG levels at 50 h (Fig. 4A), which was partially reversed in mice treated with Yajieshaba; however, no significant difference was detected. At 98 h, berberine significantly (P<0.01) increased serum IgG levels. The amount of serum cytokines, including IL4, IL-10, IL-17, IL-21, IFN-γ and TNF-α, did not significantly change in OVA-sensitized mice compared with those in the control group at 50 or 98 h (Table II). Yajieshaba treatment in the high dosage group significantly increased (P<0.05 vs. OVA-sensitized mice) serum IL-4, IL-10, IL-21, IFN-γ and TNF-α at 50 h. However, at 98 h, Yajieshaba treatment in high dosage group only significantly increased serum IFN-γ (P<0.05 vs. OVA-sensitized mice).

As indicated in Fig. 5 and Table III, OVA sensitization markedly decreased the frequency of CD4⁺CD25⁺ Tregs in the spleen (6.80%) compared with that in the control group (12.10%; Table III). This decreased CD4⁺CD25⁺ Treg frequency in the OVA-sensitized mice may impair Treg-mediated immune inhibition, which is involved in suppressing the development of allergic-effector T cells from naive T cells and may therefore lead to an increase in the severity of allergic symptoms (18).

The high dose of the Yajieshaba prescription markedly increased the frequency of CD4⁺CD25⁺ Tregs in the spleen from 6.80%, in the OVA-sensitized group to 12.50% (Table III). Mice in the high-dose Yajieshaba group had an increased frequency of CD4⁺CD25⁺ Tregs in the spleen compared with mice in the loratadine group. However, mice in the palmatine and berberine group exhibited only a marginal increase in the frequency of CD4⁺CD25⁺ Tregs. This suggested that Yajieshaba contains a number of active components other than palmatine and berberine, which may contribute to its anti-allergic effects.

Mice in the OVA-sensitized group had a 44.6% decrease in Foxp3 expression in splenic CD4⁺CD25⁺ cells compared with that in the control group (Table III). Yajieshaba treatment reversed the OVA-induced Foxp3 suppression in a dose-dependent manner. Mice in the high-dose Yajieshaba group had a CD4⁺CD25⁺Foxp3⁺/CD4⁺ ratio of 5.50%, which was markedly higher than that in the OVA-sensitized group (1.03%; Table III).