Study of the Biotransformation of Tongmai Formula by Human Intestinal Flora and Its Intestinal Permeability across the Caco-2 Cell Monolayer

Tongmai formula (TMF) is a well-known Chinese medicinal preparation that contains isoflavones as its major bioactive constituents. As traditional Chinese medicines (TCMs) are usually used by oral administration, their fate inside the intestinal lumen, including their biotransformation by human intestinal flora (HIF) and intestinal absorption deserves study. In this work TMF extract was incubated with human intestinal bacteria under anaerobic conditions and the changes in the twelve main constituents of TMF were then investigated. Their intestinal permeabilities, i.e., the transport capability across the intestinal brush border were investigated with a human colon carcinoma cell line (Caco-2) cell monolayer model to predict the absorption mechanism. Meanwhile, rapid HPLC-DAD methods were established for the assay. According to the biotransformation curves of the twelve constituents and the permeability coefficients, the intestinal absorption capacity of the typical compounds was elevated from the levels of 10−7 cm/s to 10−5 cm/s from those of the original compounds in TMF. Among them the main isoflavone glycosides puerarin (4), mirificin (6) and daidzin (7) were transformed into the same aglycone, daidzein (10). Therefore it was predicted that the aglycone compounds might be the real active ingredients in TMF. The models used can represent a novel path for the TCM studies.


Introduction
Tongmai formula (TMF) is a popular compound prescription including three well-known traditional Chinese medicines (TCMs), namely Puerariae Lobatae Radix [roots of Pueraria lobata (Willd.) Ohwi, Family Leguminosae], Salviae Miltiorrhizae Radix (roots of Salvia miltiorrhiza Bge., Family Labiatae) and Chuanxiong Rhizoma (rhizomes of Ligusticum chuanxiong Hort., Family Umbelliferae) in a weight ratio of 1:1:1. TMF is widely used for the treatment of ischemic cerebrovascular and cardiovascular diseases, such as myocardial infarction and atherosclerosis, for lowering blood lipids, or improving blood viscosity [1][2][3]. Previous studies on TMF were mainly focused on its chemical constituents and 34 compounds were isolated from the water decoction of TMF by our group [4]. Reports also established the chemical constituent fingerprint of TMF and quantitative analysis of the major chemical constituents for quality control purposes [5][6][7][8]. These researches demonstrated that isoflavones (especially puerarin and daidzein) and their derivatives were the major bioactive constituents of TMF.
Conventionally TCMs are administered orally. In the intestine the constituents of TCM first meet the intestinal bacteria, a vast gut microecological community. Studies on the bacterial flora in the alimentary canals of mammalia have proved that 99% of the bacteria are anaerobic. The classification and distribution of anaerobes in the digestive tract have also been brought to light, and the metabolizing abilities of intestinal bacteria and biochemical interactions between a host and its intestinal flora have become the object of extensive research. As we all know, the intestinal bacteria play a significant role in the biotransformation of endogenous and xenobiotic substances, including diverse drug molecules [9,10]. However, in vivo/in vitro research on the biotransformation of TMF is still a neglected topic, due to its complicated multicomponent nature and the difficulty of preparing biotransformation products, so a systematic study on the biotransformation of TMF is necessary. Biotransformation studies can also assist in understanding the mechanism of the therapeutic benefits or adverse effects of drugs [11,12]. Therefore, the human intestinal flora (HIF) model in vitro was used to investigate the biotransformation of TMF.

Methodological Validation of the Quantitative Analysis for the Twelve Constituents of TMF
Chromatographic peaks of the twelve constituents and naringin used as an internal standard (I.S.) showed very good resolutions. The average retention times for the twelve constituents and the I.S. were 14.2, 15.2, 19.8, 25.2, 29.3, 32.5, 38.7, 77.8, 81.0, 85.1, 95.4, 99.5, and 71.1 min, respectively. The chromatographic profiles of the blank and heat-inactivated HIF were investigated, and no significant interferences were observed at the retention times of the twelve constituents and I.S. (Figure 2).
The tests of linearity, lower limits of detection (LLODs) and lower limits of quantitation (LLOQs), precision, repeatability, stability, and recovery were determined by using the optimized method of HPLC, which are shown in Supplementary Tables S1 and S2 (see Supplementary Materials). All of the correlation coefficient (r) values were above 0.99, which indicated a good linear correlation. The LLODs and LLOQs were 0.02-0.65 μg/mL and 0.07-1.96 μg/mL, respectively. The intra-and inter-day accuracies ranged from 91.29% to 109.54%, and the relative standard deviations (RSDs) of intra-day and inter-day precisions were below 12.67% and 11.88%, respectively. The extraction recovery rates ranged from 62.80% to 86.95% with RSD values less than 13.24%. The analytes were found to be stable at room temperature for 24 h as the accuracies of stability were in range of 97.15% to 106.26%. All these values were found to be in an acceptable range, indicating that the method was accurate, reproducible and reliable for assessing the quality of TMF according to the guidelines of the United State Food and Drug Administration (FDA) for bioanalytical method validation [15].

Transport of the Eleven Compounds across the Caco-2 Cell Monolayer
The linearity, precision, accuracy, and stability recovery of the HPLC methods were all validated for the assay (see Supplementary Materials Tables S3 and S4). To validate the Caco-2 cell monolayer system, apparent permeability coefficient (Papp) values of propranolol, a well-transported marker by passive diffusion, and atenolol, a poor-transported marker by passive diffusion [16], from the apical (AP) to basolateral (BL) chamber across the Caco-2 monolayer were determined as 3.37 × 10 −5 cm/s and 6.17 × 10 −7 cm/s, respectively, which were comparable to the report [17].
The bidirectional Papp values of the eleven compounds have been summarized in Table 1. In general, well-absorbed drugs were found to have high Papp (>1.0 × 10 −5 cm/s), moderately absorbed drugs were found to have a Papp value of 1-10 × 10 −6 cm/s, whereas poor-absorbed drugs had low Papp (<1.0 × 10 −6 cm/s) in the Caco-2 cell monolayer model [18]. The Papp values of isoflavone aglycones such as 10, 11 and 12 were at a level of 10 −5 cm/s and similar to that of propranolol, so they were assigned to the well-absorbed compound group. The Papp values of the isoflavone glycosides such as puerarin (4) and daidzin (7) both belonged to a level of 10 −7 cm/s, two orders of magnitude lower than those of the isoflavone aglycones and therefore compounds 4 and 7 could be thought as poorly absorbed compounds. The Papp value of the aglycone daidzein (10) was much higher than that of its glycoside daidzin (7), suggesting that daidzein was taken up and transported more effectively than daidzin. It has been reported that the inhibitory effects of daidzein (10) on lipopolysaccharide-induced nitric oxide production in RAW 264.7 murine macrophages cells was stronger than that of daidzin [19]. The Papp value of puerol aglycone (9) was higher than that of puerol glycoside (8), though they were at a same level of 10 −6 cm/s, which may be attributed to the extra methoxylation of 8 making it more lipophilic and resulting in a larger Papp value despite the glycosylation.
There was no indication of efflux or active transport because the ratios of Papp BL→AP/Papp AP→BL for eleven compounds were between 0.55-1.32 according to the net efflux criterion proposed by the FDA Guidance (Table 1). This result suggests that passive diffusion is the main transport mechanism of all eleven compounds.

Mass Balance of the Eleven Compounds in the Caco-2 Cell Model
In order to check the mass balance, the recovery of each compound at the end of transport experiment was determined as the total amount in both AP and BL sides of the Caco-2 cells monolayer model. The recoveries of the eleven compounds were from 76% to 102% in both bidirectional transport studies as shown in Table 2, which suggests a high stability during the transport across the intestinal barrier. The incubation time was up to 90 min. Data are means (n = 3).

Structure-Intestinal Permeability Relationship
Physicochemical characters, such as log P, log D and polar surface area etc., have recently been used to study the relationship between chemical structure and intestinal permeability. Among them, we have found that log D (at pH 7.4) is a simple but potent indicator for the prediction of the biomembrane permeability [20,21]. Here the corrected Papp AP→BL by molecular weight (MW), log (Papp AP→BL*MW 0.5 ) vs. log D (at pH 7.4), was again used to investigate whether the potential rule is reproducible for the isoflavonoids and the analogues in TMF. A similar sigmoid dependency between the log D (at pH 7.4) and the corrected log Papp of the eleven compounds was found ( Figure 4). All isoflavone glycosides (1, 3-7) with low log D values (-1.61 to 0.23), indicating hydrophilic compounds, clustered within the region of low permeability. While the log D values of the isoflavone aglycones 10-12 are 1.65 to 2.92, the curve reached a plateau, implying good permeabilities. The results further confirmed a hypothesis that only medium lipophilicity with a log D range of about 1.5-4.0 can result in a good permeability, and both excessive hydrophilicity and lipophilicity will lead to a weak permeability. The other two puerol derivatives 8 and 9 were roughly in line with the trend of the curve. It seems that a single log D may be more precise and predictable for derivatives of the same skeleton.
A Thermo Scientific 1029 Forma Anaerobic System (Forma Scientific, Inc., Marietta, OH, USA) was applied to create anaerobic conditions and Oxoid BR0055 anaerobic indicator was used. A mode HZ-2111K-B shaking incubator (Hualida Laboratory Equipment Company, Jiangsu, China) was used for biological sample incubation.

Preparation of GAM and HIF
The general anaerobic medium (GAM) of HIF was prepared as follows: 10.0 g of tryptone, 10.0 g of proteose peptone, 13.5 g of digested serum powder, 5.0 g of yeast extract, 2.2 g of beef extract, 1.2 g of beef liver extract powder, 3.0 g of glucose, 2.5 g of KH2PO4, 3.0 g of NaCl, 5.0 g of soluble starch, 0.3 g of L-cysteine hydrochloride, and 0.3 g of sodium thioglycolate were dissolved in H2O. Adjust the pH of GAM solution to 7.1-7.2 with 1M NaOH aqueous solution and supply the total volume to 1 L. Ten grams of fresh feces obtained from a Chinese healthy male volunteer, who had not taken any medicine in three months, were immediately homogenized and suspended in GAM (50 mL) solution. The HIF was undertaken according to the previous paper of our laboratory [22]. Preparation of the biotransformation products of TMF by HIF was followed by the method of previous research [23].

Time Course the Twelve Constituents of TMF by HIF
A biotransformation system consisting of HIF (10 mL) and TMF (10 mg) was anaerobically incubated at 37 °C. All experiments were carried out in sextuplicate at each time point. The biotransformation reaction was stopped at 0. 5,1,2,4,8,12,24,36,48, 60 h. The biotransformation solution was extracted with n-butanol (each 10 mL) three times. The combined n-butanol layer was evaporated to dryness under a gentle stream of nitrogen. MeOH (400 μL) was used to dissolve the residue which acted as a deproteinization agent. The resultant solution was vortex-mixed for 1 min and centrifuged at 16,000 g for 10 min at 4 °C. The supernatant was collected and filtered through a 0.22 μm filter as the test solution. The blank sample was prepared using TMF-free medium solution with the same processing steps. The test and blank solutions were analyzed by HPLC using a gradient mobile phase consisting of solvent A (MeCN) and B (0.4% acetic acid aqueous solution) using a gradient elution of 11%-12% A at 0-28 min, 12%-15% A at 28-30 min, 15%-17% A at 30-45 min, 17%-22% A at 45-65 min, 22%-28% A at 65-66 min, 28%-35% A at 66-85 min, 35%-50% A at 85-86 min and 50% A at 86-98 min. The column temperature was set at 30 °C and detection wavelength was set at 254 nm.

Caco-2 Permeation Experiment of the Eleven Compounds
The Caco-2 cells were seeded at a density of 1.0 × 10 5 cells/mL on a 12-wells Transwell™ insert filter and left to grow for 21 d to reach confluence and differentiation. Differentiation of Caco-2 cells was assayed by determining the activity of alkaline phosphatase with an assay equipment on the 3rd and 14th days [24] and detecting the cellular morphology by transmission electron microscopy on the 21st day [17]. The integrity and transportation ability of the Caco-2 cell monolayer were examined by measuring the transepithelial electrical resistance (TEER) with an epithelial voltohmmeter (EVOM, World Precision Instrument, Sarasota, FL, USA) before and after the transport. Only cell monolayers with TEER values above 500 Ω·cm 2 were qualified for the transport assays. Standard compounds, propranolol and atenolol were run as the active and passive transport marker, respectively [25,26].
Ten mM stock solutions of test compounds were prepared in DMSO and diluted with HBSS to 50 μM before transport experiments. After washing the Caco-2 cell monolayer twice with prewarmed HBSS medium (pH 7.4), the transport experiments were carried out by adding the test samples to either the apical (AP, 0.5 mL) or basolateral side (BL, 1.5 mL) while the receiving chamber contained the corresponding volume of HBSS medium. After shaking at 55 rpm for 1 h at 37 °C in a water bath, samples were collected from both sides of Caco-2 cell monolayer and immediately frozen, lyophilized and preserved below −20 °C for subsequent HPLC analysis.
To determine the eleven compounds, the lyophilized samples from both sides of the Transwell were dissolved in 200 μL of MeOH and centrifuged at 15,000 g for 10 min. Twenty μL aliquot of the supernatant solution was used for HPLC analysis. The cells were extracted after transport assays with 300 μL of MeOH, which was used to measure the amount of the intracellular accumulation for the test compounds. The HPLC mobile phase consisted of solvent A (MeCN) and B (0.4% acetic acid aqueous solution) using isocratic elutions (Supplementary Materials Table S3).
In the meantime, the apparent permeability coefficients, Papp (cm/s) were calculated on the basis of the following equation: where ΔQ/Δt is the appearance rate of the test compound on the receiver side (μmol/s); A is the membrane surface area (cm 2 ); C0 is the initial concentration of the test compound at the donor side (μM). All the experiments were conducted in triplicate and the data were expressed as mean ± S.D.

Conclusions
The TMF biotransformation results showed that its glycoside compounds were converted into the corresponding aglycones (isoflavone and puerol) and as time went by, the relative content of aglycones obviously increased. Daidzein (10), the principal isoflavone aglycone in TMF, has been marketed as a prescription drug in the treatment of hypertension, coronary heart disease, and cerebral thrombosis in China. Accordingly we predict that the aglycone compounds might be the real active ingredients in TMF, and a permeability test was needed to determine the intestinal absorption capacity of the compounds.
From the permeability result, isoflavone aglycones were well-absorbed compounds, which was consistent with the previous result [27]. In addition, the intestinal absorption capacity of puerol aglycones was better than puerol glycosides, and the methylation reaction could promote the intestinal absorption capacity of isoflavone glycosides. In conclusion, the constituents of TMF can be transformed by HIF, and the intestinal absorption capacities of biotransformation products were better than the original compounds. The methodology in this paper will be useful for studying the TCMs and clarifying the molecular basis of therapeutics.