A Comparative Study on the Metabolism of Epimedium koreanum Nakai-Prenylated Flavonoids in Rats by an Intestinal Enzyme (Lactase Phlorizin Hydrolase) and Intestinal Flora

The aim of this study was to compare the significance of the intestinal hydrolysis of prenylated flavonoids in Herba Epimedii by an intestinal enzyme and flora. Flavonoids were incubated at 37 °C with rat intestinal enzyme and intestinal flora. HPLC-UV was used to calculate the metabolic rates of the parent drug in the incubation and LC/MS/MS was used to determine the chemical structures of metabolites generated by different flavonoid glycosides. Rates of flavonoid metabolism by rat intestinal enzyme were quicker than those of intestinal flora. The sequence of intestinal flora metabolic rates was icariin > epimedin B > epimedin A > epimedin C > baohuoside I, whereas the order of intestinal enzyme metabolic rates was icariin > epimedin A > epimedin C > epimedin B > baohuoside I. Meanwhile, the LC/MS/MS graphs showed that icariin produced three products, epimedin A/B/C had four and baohuoside I yielded one product in incubations of both intestinal enzyme and flora, which were more than the results of HPLC-UV due to the fact LC/MS/MS has lower detectability and higher sensitivity. Moreover, the outcomes indicated that the rate of metabolization of flavonoids by intestinal enzyme were faster than those of intestinal flora, which was consistent with the HPLC-UV results. In conclusion, the metabolic pathways of the same components by intestinal flora and enzyme were the same. What’s more, an intestinal enzyme such as lactase phlorizin hydrolase exhibited a more significant metabolic role in prenylated flavonoids of Herba Epimedi compared with intestinal flora.


Introduction
Yinyanghuo (Herba Epimdii, YYH) is a popular Traditional Chinese Medicine tonic for kidney-reinforcing, used to invigorate the kidney-yang, strengthen the sinews and bones, dispel wind and eliminate dampness in clinical practice in East Asian countries for thousands of years [1][2][3]. Flavonoids as the phytoestrogens in medicinal plants have long been considered to exert beneficial effects on estrogen-related diseases by acting as selective estrogen receptor modulators. Modern pharmacological research has confirmed that flavonoids in YYH display beneficial anti-cancer, anti-depression, vasodilatation and immunoregulation properties. More than 60 flavonoids have been isolated from the herbal extracts, among which the prenylated flavonoids include icariin, epimedin A, epimedin B, epimedin C, baohuoside I ( Figure 1) and so on, which are considered the main pharmacological ingredients [4][5][6]. The symbol "glc" refers to glucose, "rha" to rhamnose, and "xyl" to xylose.
Recently, some investigations on the metabolism of flavonoids present in YYH by the intestinal flora of rats [7][8][9], rabbits [10] and humans [11] have been carried out. It has been reported that more than 50 metabolites are discovered in plasma, bile, urine, feces after oral administration of icariin, baohuoside I, epimedin C and the herbal extracts [12][13][14][15]. However, in plasma, bile, urine, and feces, it is generally believed that intestinal flora absolutely plays a main role in the metabolism of YYH flavonoids because the intestinal microflora comprises a complex ecosystem of a large variety of bacteria which can produce negative and positive effects on metabolism, but preliminary research in our laboratory has indicated that prenylated flavonoids might be hydrolyzed both by intestinal flora and intestinal enzymes, especially lactase phlorizin hydrolase (LPH). Thus, in this experiment, we sought to explore whether the enzyme can affect the hydrolysis of prenylated flavonoids in YYH or not and investigate the effects of intestinal enzyme hydrolysis on the prenylated flavonoids in YYH.
Since intestinal enzymes and intestinal flora can hydrolyze prenylated flavonoids, and the metabolism of orally administrated prenylated flavonoids by intestinal enzymes and intestinal flora may affect their bioavailability, it is necessary to study the metabolism of prenylated flavonoids by intestinal enzymes and intestinal flora. In this study, we investigated in detail the metabolism of icariin, epimedin A, epimedin B, epimedin C, baohuoside I by rat intestinal enzyme and intestinal flora in vitro. We scratched the intestinal mucosa and collected the fresh feces of rats to prepare intestinal enzyme and flora incubations, using HPLC-UV to determine the metabolic rates of parent drugs and applying LC/MS/MS to figure out the chemical structures of metabolites for the purpose of exploring the metabolic pathways.

The Metabolic Rates of Flavonoids by Intestinal Enzyme and Intestinal Flora of Rats
Five prenylated flavonoids were incubated with rat intestinal enzyme and intestinal flora solution and the degradation products were analyzed with time by HPLC-UV and LC/LC/MS. The HPLC-UV chromatograms of five prenylated flavonoids are shown in Figure 2. We can see that there were two metabolites and one metabolite created from icariin after hydrolysis by intestinal enzyme and intestinal flora, respectively, whereas epimedin A, epimedin B and epimedin C gave only one metabolite with both intestinal enzyme and intestinal flora. As to baohuoside I, no hydrolysis products were found with either intestinal enzyme or intestinal flora.
To intuitively calculate the drug metabolic rates, the logarithmic concentrations of flavonoids (Y) and time (X/h) were used to get the corresponding regression equations. Figure 3 shows the metabolic results for intestinal enzyme and Figure 4 illustrates the metabolic results for intestinal flora. A lower slope value indicates a higher metabolic rate. Epimedin A, epimedin B, epimedin C and icariin all had higher metabolic rates with intestinal enzyme (the slope values were −0.0706 ± 0.00010, −0.0248 ± 0.00021, −0.0438 ± 0.00015, −0.2551 ± 0.00025, respectively) than with intestinal flora (the slope values were −0.0098 ± 0.00025, −0.0158 ± 0.00011, −0.0085 ± 0.00050, −0.0176 ± 0.00015, respectively). Baohuoside I had a similar metabolic rate with intestinal enzyme and flora, and the slope values were −0.0019 ± 0.00015 and −0.0018 ± 0.00011, respectively. Except for baohuoside I, the slope values for epimedin A, epimedin B, epimedin C and icariin showed significant differences between intestinal enzyme and intestinal flora (p < 0.05). The sequence of metabolic rates for intestinal enzyme was icariin > epimedin A > epimedin C > epimedin B > baohuoside I, and the order of metabolic rates for intestinal flora was icariin > epimedin B > epimedin A > epimedin C > baohuoside I.     Our data suggest that intestinal hydrolysis of glycosides by intestinal enzymes is rapid. Even icariin was completely metabolized in 6 h and the epimedin A was totally metabolized in 12 h in incubations with intestinal enzyme. These results inform that intestinal glycosidases are brush-border enzymes. LPH is the only mammalian brush border-glucosidase, thus, we hypothesize that LPH is responsible for the observed hydrolysis of prenylated flavonoids in YYH. LPH has two distinct catalytic active sites, one for the hydrolysis of lactose and flavonoid glucosides and another, phlorizin hydrolase, for the hydrolysis of phlorizin and -glucosylceramides [16]. As reported by Wilkinson et al. [17], LPH plays a major role in the deglycosylation of daidzin. In our previous research, we used gluconolactone, a LPH enzyme inhibitor, to see if LPH was involved in the hydrolysis of the flavonoids. If LPH plays an important role in absorption of icariin, epimedin A, epimdin B, epimedin C, and baohuoside I, then low activity of LPH would result in a reduced rate of metabolism.
The data of HPLC-UV show that flavonoid metabolic rates with rat intestinal enzyme were higher than those with intestinal flora. Moreover, LPH is located in the mammalian small intestine, and the intestinal flora usually exists in the large intestine. After oral administration, the drug lagged in the small intestine for a long time before it reached the large intestine, so if this situation occurs as fast in humans as we observed in the rats, the role played by intestinal flora in hydrolyzing glycosides would be significantly diminished. The product ion at m/z 369 was produced by the loss of rahmnose and glucose (308 Da). The product ion at m/z 515 was generated by the loss of glucose. The MS 2 spectra of epimedin A shows three identical fragment ions at m/z 369, 531, 677. The fragment ion at m/z 369 was attributed to the loss of rahmnose and bimolecular glucose. The product ion at m/z 531 was generated through the reduction of bimolecular glucose and the ion at m/z 677 was produced via the loss of glucose. The daughter ions appearing at m/z 369, 531, 677 in the MS 2 spectra of epimedin B were produced by the neutral loss of 440, 278, 132 Da, corresponding to the losses of glucose, rhamnose and xylose, glucose and xylose, xylose, respectively. Epimedin C yielded five main daughter ions at m/z 313, 369, 515, 531, 677. The fragment ions at m/z 369, 515, 531 and 677 were generated by the reduction of glucose and bimolecular rhamnose, glucose and rhamnose, bimolecular rhamnose, rhamnose respectively. Baohuoside I gave two product ions at m/z 313 and 369, in which the latter was attributed to the loss of rhamnose. All the product ions at m/z 313 were 56 Da less than m/z 369 by the loss of C 4 H 7 (56 Da), which are produced by the rearrangement of the isopentene group at the position 8 of the A-ring. Above all, these characterisitic product ions and neutral losses were a sound basis to identify the metabolites of icariin, epimedin A, epimedin B, epimedin C and baohuoside I (Figures 6-10).            (Figure 14). Comparatively speaking, the 7-O-glucose could be removed more easily and 3-O-rhamnose was difficult to hydrolyze. M 1 and M 2 replacing icariin could be detected in the incubated sample at 24 h demonstrating that icariin was quite easily hydrolyzed further. This phenomenon also ocurred in epimedin A and B. Simultaneously, the content of baohuoside I in samples after incubation of epimedin C was significantly higher than that of epimedin A and B. However, the contents of M 1 and M 2 were so low that they could be barely be measured. Presumably these were two rhamnosyls in the molecular structure of epimedin C which made epimedin C hydrolyze more difficultly in the intestinal enzyme and intestinal flora incubations. The metabolic pathway of baohuoside I by intestinal flora was consistent with that by intestinal enzyme, which was only 3-O-rhamnose hydrolysis ( Figure 15). After incubation for 24 h with both intestinal flora and intestinal enzyme, a high content of baohuoside I could be detected. It indicated that 3-O-rhamnose was hard to hydrolyze by either intestinal flora or intestinal enzyme.     Icariin is a dual glucoside, it was thought to be metabolized by hydrolysis to first produce baohuoside I after removal of 7-O-glucose and then baohuoside I was metabolized to produce M 2 after further removal of 3-O-rhamnose. Meanwhile, icariin could be metabolized to generate M 1 after only removal of 3-O-rahamnose. Epimedin A, B, C were triple glucosides with similar structures, and besides the M 3 for epimedin A, M 4 for epimedin B, M 5 for epimedin C, they had other four common metabolites: icariin, M 1 , M 2 and baohuoside I. Epimedin A was considered to be metabolized by hydrolysis to first produce M 3 after removal of the 2"-O-glucose and then M 3 was metabolized to produce baohuoside I after further removal of 3-O-rhamnose. Epimedin A also can be metabolized by removal of 7-O-glucose to then produce icariin. The generated icariin had the same metabolic pathway as above. Epimedin B was thought to be metabolized by hydrolysis to first produce icariin which had the same metabolic pathway as above after removal of 7-O-glucose. Epimedin B also can be metabolized by removal of 2"-O-oxylose to next produce M 4 and then M 4 was metabolized to produce baohuoside I after further removal of 3-O-rhamnose. Epimedin C was thought to be metabolized by hydrolysis to first produce icarrin which had the same metabolic pathway as above after removal of 7-O-glucose. Epimedin C also can be metabolized by removal of 3-O-rahmnose to secondly produce M 5 and then M 5 was metabolized to produce baohuoside I after further removal of 3-O-rhamnose. Presumably it was the presence of two rhamnosyls in the molecular structure of epimedin C which made epimedin C difficultly hydrolyzed by intestinal flora and intestinal enzyme. It can be deduced from the above data that icariin, epimedin A, B, C were generally absorbed as metabolites. On the other hand, as a single glucoside, it was difficult for baohuoside I to hydrolyze in the intestine and usually it was absorbed as the precursor.

Identification of Flavonoid Metabolites with Intestinal Flora and Intestinal Enzyme of Rats
Most of the previous studies consider that the deglycosylation of flavonoids may be caused by intestinal flora and hepatic biotransformation enzymes [18,19]. However, our study indicated that intestinal enzymes play an important role in the metabolism of the prenylated flavonoids present in YYH. Meanwhile, based on our analysis and literature data, the metabolic pathways of icariin, epimedin A, epimedin B, epimedin C and baohuoside I in intestinal flora and enzyme solution of rats were basically the same, and the absorption pathways of Epimedium flavonoids in rat intestine initially involved deglycosylation of flavonoid glycosides, yielding the major hydrolytic metabolites. It demonstrated that the enzyme existing in flora was a β-enzyme which had the same function as LPH in hydrolyzing Epimedium flavonoids. The metabolic effects in intestinal enzyme were far higher than those in intestinal flora. It might because of the expression and distribution of LPH in the upper gastrointestinal tract were comparatively large and there were no bacterial growth in the upper gastrointestinal tract.

Drugs and Chemicals
Icariin (purity > 98%) was purchased from National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Epimedin A, epimedin B, epimedin C, and baohuoside I (all purity > 98%) were provided by the Laboratory of Pharmaceutical Preparation (Jiangsu Provincial Academy of Chinese Medicine, Nanjing, China). Testosterone (purity > 98%) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Epimedium koreanum Nakai was purchased from a drug store in Nanjing (China) and was identified to be the correct species by Professor Dekang Wu, a pharmacognosy researcher at Nanjing University of Chinese Medicine (Nanjing, China). All other materials (typically analytical grade or better) were used as received.

Animals
Eight weeks old Male Sprague-Dawley rats with body weight of 170-250 g were obtained from the SLEK Lab Animal Center of Shanghai (Shanghai, China), housed under standard conditions of temperature, humidity, and light. Food and water were provided ad libitum. The rats were fasted overnight before the day of the experiment. The procedures were approved by the Animal Ethics Committee of Jiangsu Provincial Academy of Chinese Medicine.

Preparation of Intestinal Enzyme and Intestinal Flora Cultural Solution
After overnight food deprivation, rats were anesthetized by intramuscular injection of urethane (0.5 g/mL). An incision was made into the abdominal cavity to take out the small intestine and the intestine immediately was preserved in cold saline. The contents of the small intestine were removed by flushing gently with saline (0 °C) after being opened. Intestinal mucosa was blunt scratched. At the same time, fresh feces were obtained from SD rats. Intestinal mucosa and fresh feces of rats were homogenized in normal saline solution at the ratio of 1 g to 4 mL immediately, respectively. Ten mL of the filtrate of intestinal mucosa was mixed with 90 mL of cold saline to prepare 100 mL of intestinal enzyme cultural solution. Ten mL of the filtrate of fresh feces was added to 90 mL anaerobic culture medium to obtain 100 mL intestinal flora cultural solution.