1521-009X/43/10/1590–1600$25.00 http://dx.doi.org/10.1124/dmd.115.064386DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 43:1590–1600, October 2015Copyright ª 2015 by The American Society for Pharmacology and Experimental Therapeutics

Special Section on Drug Metabolism and the Microbiome

Intestinal Absorption and Metabolism of Epimedium Flavonoids inOsteoporosis Rats s

Jing Zhou, Yi Hua Ma, Zhong Zhou, Yan Chen, Ying Wang, and Xia Gao

Multicomponent of Traditional Chinese Medicine and Microecology Research Center (J.Z., Y.H.M., Y.C., Y.W., X.G.) andDepartment of Orthopaedics (Z.Z.), Jiangsu Provincial Academy of Chinese Medicine, Nanjing, Jiangsu, People’s Republic of China

Received March 19, 2015; accepted July 1, 2015

ABSTRACT

Herba Epimdii is a traditional Chinese medicine used to treatosteoporosis. Its main pharmacological ingredients are flavonoids.In previous studies conducted in healthy animals, we showed thatepimedium flavonoids could be hydrolyzed into secondary glyco-sides or aglycon by intestinal flora or enzymes, thereby enhancingtheir absorption and antiosteoporosis activity. To study themedicine in the pathologic state, epimedium flavonoids wereincubated with intestinal mucosa and feces in vitro and intestinalperfusion in situ to explore the differences in absorption andmetabolism between sham and osteoporosis rats. For osteoporosisrats, the hydrolysis rates of icariin, epimedin A, epimedin B, andepimedin C incubated with intestinal flora for 1 hour were reducedby 0.19, 0.26, 0.19, and 0.14, respectively, compared with that in shamrats. Hydrolysis rates were reduced by 0.21, 0.24, 0.08, and 0.31

for icariin, epimedin A, epimedin B, and epimedin C incubatedwith duodenal enzymes for 1 hour and by 0.13, 0.09, 0.07, and 0.47for icariin, epimedin A, epimedin B, and epimedin C incubated withjejunum enzymes, respectively, compared with the sham group. Inaddition, the apparent permeability coefficient and eliminationpercentage of the four epimedium flavonoids in the duodenum,jejunum, ileum, and colon decreased by 29%–44%, 32%–50%,40%–56%, and 27%–53% compared with that in sham rats,respectively. The main metabolites of the four epimedium flavo-noids were the same for the two groups after intestinal perfusion,or flora and enzyme incubation. In conclusion, the amount andactivity of intestinal flora and enzymes changed in ovariectomizedrats, which affected the intestinal absorption and hydrolysis ofepimedium flavonoids whose structures contain 7-glucose.

Introduction

Osteoporosis is characterized by the reduction and deterioration of thebone microarchitecture, leading to increased bone frailty and suscep-tibility to fracture. The worldwide incidence of osteoporosis isincreasing, and osteoporosis can be an economic burden on bothfamilies and societies (Kanis et al., 1994; Johnell and Kanis, 2006).Currently, 200 million people worldwide have osteoporosis. A largenumber of people also have low bone mass, placing them at anincreased risk for developing osteoporosis. As the population ages, thesenumbers will increase. A majority of those with osteoporosis arewomen. Of people aged older than 50 years, one in two women and onein eight men are predicted to have an osteoporosis-related fracture intheir lifetime. Drugs being developed or used for treating osteoporosisinclude estrogen replacement therapy, calcitonin, selective estrogenreceptor modulators, and diphosphate. Although these drugs preventbone resorption, their effects on bone formation are extremely small

(Abrahamsen et al., 2014; Bandeira et al., 2014; Chen and Kubo, 2014;Kulak et al., 2014).Herba Epimdii, a popular traditional Chinese medicine, has been

used to treat osteoporosis in East Asian countries for over 2000 years(Nelson et al., 2002; Wu et al., 2003). In the perspective of modernmedicine, the main pharmacological ingredients of H. Epimdii werefound to be various prenylated epimedium flavonoids (including icariin,epimedin A, epimedin B, epimedin C, and baohuoside I) (Fig. 1).Modern research has also confirmed that prenylated epimediumflavonoids can regulate the balance between osteogenic and adipogenicdifferentiation of bone marrow stromal cells in ovariectomized rats bydownregulating the expression of DKK1 protein, thereby enhancingbone formation, and they can also prevent ovariectomy-induced boneloss (Zhang et al., 2007; Xu et al., 2011). However, the oralbioavailability of epimedium flavonoids is very low (Wei et al., 2012;Zhou et al., 2013; Chen et al., 2014).Our previous study found that there are several reasons for this low

bioavailability. First, epimedium flavonoids have low absorptivepermeabilities due to their physicochemical characteristics (Chenet al., 2008). Second, they are subject to efflux by the drug transportersin intestinal mucosa, such as P-glycoprotein, breast cancer resistanceprotein, and multidrug resistance–associated protein (Löscher andPotschka, 2005; Chen et al., 2014). Third, these flavonoids can be

This research was supported by the National Natural Science Foundation ofChina, funded by the Chinese government [Grant 81173557].

J.Z. and Y.H.M contributed equally to this work.dx.doi.org/10.1124/dmd.115.064386.s This article has supplemental material available at dmd.aspetjournals.org.

ABBREVIATIONS: CBG, cytosolic b-glucosidase; EP, eliminative percentage; HPLC, high-performance liquid chromatography; LPH, lactasephlorizin hydrolase; m/z, mass-to-charge ratio; P*app, apparent permeability coefficient; Q-TOF-MS, quadrupole time-of-flight mass spectrometry;RPI, rearrangement product of icariin; UPLC, ultra performance liquid chromatography.

1590

http://dmd.aspetjournals.org/content/suppl/2015/07/01/dmd.115.064386.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

hydrolyzed by hydrolase in the intestinal mucosa and flora. Lactasephlorizin hydrolase (LPH) and b-glucosidase were confirmed as themain hydrolases in intestinal mucosa and flora, which could hydrolyzethe epimedium flavonoids to secondary glycosides or aglycon (Zhaoet al., 2010; Qian et al., 2012; Cui et al., 2013). The above results wereobtained with healthy animals; however, the absorption and hydrolysisbehavior in osteoporotic animals remains unknown.Therefore, we studied the absorption and hydrolysis of epimedium

flavonoids under osteoporotic conditions. In this study, we exploredthe absorption and hydrolysis characteristics of icariin, epimedin A,epimedin B, epimedin C, and baohuoside I in osteoporosis rats. Toanalyze the hydrolysis rates of epimedium flavonoids and to detecttheir metabolites, they were incubated in vitro with rat intestinalmucosa and feces at 37�C. The absorption of epimedium flavonoidswas investigated using an in situ intestinal perfusion model. Withthese methods, the absorption and hydrolysis characteristics ofepimedium flavonoids in osteoporosis rats are revealed, and thedifferences in absorption and hydrolysis characteristics of epime-dium flavonoids between sham and osteoporosis rats are discussed.

Materials and Methods

Ethics Statement. Animal welfare and experimental procedures werestrictly in accordance with the U.S. National Research Council 1996 Guide forthe Care and Use of Laboratory Animals and the related ethics regulations ofJiangsu Provincial Academy of Chinese Medicine.

Reagents and Chemicals. Icariin (purity .98%) was purchased from theNational Institute for the Control of Pharmaceutical and Biologic Products(Beijing, China). Epimedin A, epimedin B, epimedin C, and baohuoside I (allpurities.98%) were provided by the Laboratory of Pharmaceutical Preparation(Jiangsu Provincial Academy of Chinese Medicine, Nanjing, China). LPHantibody and b-actin were purchased from Acris Antibodies, Inc. (San Diego,CA). An enhanced chemiluminescence kit was obtained from KeyGen Biotech(Nanjing, China). Testosterone (purity .98%), K2HPO4, NaCl, (NH4)2SO4,CaCl2, MgSO4·H2O, Na2CO3, HCl, L-cysteine, L-ascorbic acid, erythritol,tryptone, and nutrient agar were purchased from Sigma-Aldrich (St. Louis,

MO). All other materials (typically analytical grade or better) were used asreceived.

Animals. Eight-week-old female Sprague-Dawley rats with a body weightof 170–250 g were obtained from the Shanghai Laboratory Animal Center ofShanghai (Shanghai, China) and were housed under standard conditions oftemperature, humidity, and light. Food and water were provided ad libitum. Therats were fasted overnight before the day of the experiment. Rats wererandomly assigned to two groups: the sham group (sham) or the ovariectomizedgroup (osteoporosis). After 1 week of acclimatization, rats were anesthetizedwith an intraperitoneal injection of 300 mg chloral hydrate per kg body weightand then both ovaries were removed. Rats were left untreated for 3 months toallow them to recover and develop osteopenia, according to a previouslypublished method (Wang et al., 2013). The sham rats underwent bilaterallaparotomy but the ovaries were left in place. Bone mineral densitymeasurements of the femur in rats confirmed that the osteoporosis modelwas available for subsequent experiments (Supplemental Table 1).

Intestinal Flora Hydrolysis of Epimedium Flavonoids. Anaerobic culturemedium was prepared as follows: K2HPO4 [37.5 ml, 0.78%], solution A [37.5 ml,0.47% KH2PO4, 1.18% NaCl, 1.2% (NH4)2SO4, 0.12% CaCl2, and 0.25%MgSO4·H2O], Na2CO3 [50 ml, 8%], L-cysteine [0.5 g], L-ascorbic acid [2 ml,25%], eurythrol [1 g], tryptone [1 g], and nutrient agar [1 g] were mixed togetherand diluted with distilled water to 1 liter. The solution was then adjusted to pH7.5–8.0 with 2 M HCl (Akao et al., 1996).

Fresh feces collected from Sprague-Dawley rats was immediately homog-enized in normal saline solution at a ratio of 1 g to 4 ml. The homogenate wasfiltered, and the filtrate of fresh feces (10 ml) was added to anaerobic culturemedium (90 ml) to obtain an intestinal flora cultural solution. Icariin, epimedinA, epimedin B, epimedin C, or baohuoside I (1 ml; 2 mM) was then added tothe intestinal flora cultural solution (9 ml), respectively.

After incubation for 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 4, 6, 8, 12, and 24 hours at37�C, the incubations were deproteinized by adding a 3-fold volume ofacetonitrile. The samples were vortexed for 1 minute and centrifuged for15 minutes at 15,000 rpm. For high-performance liquid chromatography(HPLC)-UV analysis, 400-ml aliquots of the supernatants were mixed with100 mM internal standard testosterone (100 ml), which was dissolved inacetonitrile. For ultra performance liquid chromatography (UPLC)/quadrupole time-of-flight mass spectrometry (Q-TOF-MS) analysis, thesupernatants (150 ml) were purified using a C18 solid phase extractioncartridge (Agilent, Santa Clara, CA). The cartridge was activated by methanol (3ml) and was then balanced using water (2 ml). Intestinal perfusion samples (0.6ml) were mixed with methanol (150 ml). The mixtures were centrifuged for 15minutes at 13,000 rpm. Supernatants (0.6 ml) were added into the cartridge, andthen the cartridge was eluted with water (5 ml) and methanol (5 ml). Theelutropic methanol was collected and evaporated to dryness under a gentle streamof nitrogen at 30�C before being reconstituted in acetonitrile (450 ml).

The experiments were divided into the sham group and the osteoporosis group,and fresh feces was collected from both groups at 3 months postoperation.

Intestinal Enzyme Hydrolysis of Epimedium Flavonoids. After overnightfood deprivation, rats were anesthetized by an intramuscular injection ofurethane (0.5 g/ml). An incision was made into the abdominal cavity to take outthe small intestine, and the intestine was immediately preserved in cold saline.After dividing the duodenum, jejunum, ileum, and colon into segments, thecontents of the four segments were removed by gently flushing them with saline(0�C). Intestinal mucosa was blunt scratched. Intestinal mucosa in different

Fig. 1. Chemical structures of five prenylated flavonoids in H. Epimdii and parts oftheir metabolites. glc, glucose; rha, rhamnose; xyl, xylose.

TABLE 1

Hydrolysis rates of epimedium flavonoids by intestinal flora

Data are presented as means 6 S.D.

CompoundSham Group Osteoporosis Group

LogCt R2 LogCt R2

Icariin 20.297 6 1.913 0.917 20.237 6 1.835 0.952Epimedin A 20.264 6 1.796 0.974 20.212 6 1.832 0.980Epimedin B 20.270 6 1.756 0.893 20.217 6 1.816 0.943Epimedin C 20.164 6 1.678 0.968 20.090 6 1.663 0.900

Ct, threshold cycle.

Absorption and Metabolism of Epimedium Flavonoids 1591

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

segments was immediately homogenized in normal saline solution at a ratio of 1g to 4 ml. Homogenates of intestinal mucosa (10 ml) were mixed with cold saline(90 ml) to prepare the intestinal enzyme cultural solution (Sesink et al., 2003).

The subsequent procedures and animal group assignment were the same as theexperiments of intestinal flora–hydrolyzing epimedium flavonoids. At the sametime, protein was extracted from frozen intestines by radioimmunoprecipitationassay buffer containing 1 mM phenylmethanesulfonylfluoride for the purpose ofexamining the LPH expression.

In Situ Intestinal Perfusion Assay of Epimedium Flavonoids. The ratswere fasted overnight but were provided with deionized water. After overnightfasting, rats were anesthetized. The small intestine was exposed by midlineincision; the intestinal lumen was then gently flushed to remove intestinalcontents, and each of the four segments (duodenum, upper jejunum, terminalileum, and colon) of the intestine was cannulated with two cannulas. The outletof each segment was secured by ligation with a silk suture. The intestine wascarefully arranged and continuously monitored to avoid kinks, and a consistentflow was ensured after cannulation. Saline-soaked cotton was used to coveropened body cavities to prevent loss of fluids (Andlauer et al., 2000a,b,c; Liuand Hu, 2002; Chen et al., 2003; Higaki et al., 2004).

After 3-month ovariectomized postoperation, both sham rats and osteopo-rosis rats were operated as above steps. To keep the temperature of theperfusate constant, the inlet cannula was insulated and kept warm by a 37�Ccirculating water bath. A flow rate of 0.2 ml/min was used, and the perfusatesamples were collected every 30 minutes. The outlet concentration of drug inthe perfusate was determined by UPLC-UV, and the hydrolyzed products weredetected by UPLC/Q-TOF-MS. The preparation method of perfusate sampleswas the same as that of the intestinal flora and enzyme samples. The initialconcentration of the perfusate was 20 mM.

Analytical Methods. UPLC-UV is a rapid and efficient method fordetermining the contents of different samples. However, impurities in theintestinal flora and enzyme samples would block a UPLC-UV system. On theother hand, the perfusate samples were cleaner than the intestinal flora andenzyme samples. Therefore, HPLC-UV was used for determining the contents ofepimedium flavonoids in intestinal flora and enzyme samples, whereas UPLC-UVwas used for determining the contents of epimedium flavonoids in perfusatesamples. All of the metabolites were identified by the UPLC/Q-TOF-MS method.

The UPLC-UV method was performed with an Acquity UPLC system witha photodiode array detector and Empower software (Waters, Milford, MA) and anAcquity UPLC BEH C18 column (1.7 mm, 2.1 � 50 mm). The mobile phaseconsisted of acetonitrile (A) and water (B), with the following gradient program:0 to 0.5 minutes, 25% A; 0.5 to 1.5 minutes, 25% to 28% A; 1.5 to 2.5 minutes,28% A to 32%; 2.5 to 3.5 minutes, 32% A to 35% A; 3.5 to 4.5 minutes, 35% A to55%; 4.5 to 5.5 minutes, 55% A to 95%; and 5.5 to 6.0 minutes, 95% A. The flowrate was 0.4 ml/min, the column temperature was 30�C, and the injection volumewas 5 ml. The wavelength was 270 nm for epimedium flavonoids and 254 nm forinternal standard testosterone. In general, these methods were selective andreproducible with day-to-day variability of less than 2%. The accuracy and precisionwere greater than 92%. The tested linear response ranges for samples were2.5–40 mM. Epimedium flavonoids exhibited good linearity within the selectedconcentration ranges, with correlation coefficients (R2) between 0.9994 and 0.9999.

The following HPLC-UV method was performed with an Agilent 1260HPLC-UV system with a photodiode array detector and a ZORBAX SB-C18column (5 mm, 4.6 � 250 mm; Agilent). The mobile phase consisted of water(A) and acetonitrile (B), with the following gradient program: 0 to 6.5 minutes,70% A; 6.5 to 10 minutes, 45% A; 10 to 14 minutes, 45% A; and 14 to15 minutes, 70% A. The flow rate was 1 ml/min and the injection volume was 10 ml.The wavelength was 270 nm for epimedium flavonoids and 254 nm forinternal standard testosterone. In general, these methods were selective andreproducible, with day-to-day variability of less than 2%. The accuracy andprecision were greater than 97%. The tested linear response ranges for allflavonoids were 6.25–100 mM, respectively. Epimedium flavonoidsexhibited good linearity within the selected concentration ranges, with R2

values between 0.9992 and 0.9997.The UPLC/Q-TOF-MS method was performed with a Waters Synapt G2-S

Q-TOP mass spectrometer equipped with an electrospray ionization source.The samples were separated on an Acquity UPLC BEH C18 column (1.7 mm,2.1 � 50 mm; Waters). The mobile phase consisted of acetonitrile (A) andwater (B). The gradient program of epimedium flavonoids was as follows: 0 to1 minute, 5% A; 1 to 1.5 minutes, 5% to 25% A; 1.5 to 3 minutes, 25% A; 3 to 4minutes, 25% to 28% A; 4 to 5 minutes, 28% to 32% A; 5 to 6 minutes,32% to 35% A; 6 to 7 minutes, 35% to 55% A; 7 to 8 minutes, 55% to 95%A; 8 to 9.5 minutes, 95% A; and 9.5 to 10 minutes, 95% to 5% A. The flowrate was 0.4 ml/min, the column temperature was maintained at 30�C, andthe injection volume was 2 ml. The tested linear response ranges for sampleswere 1.25–40 mM. Epimedium flavonoids exhibited good linearity withinthe selected concentration ranges, with R2 values between 0.9992 and0.9996.

For mass spectrometry analysis, the electrospray ionization source wasoperated in the positive ion mode. The capillary and cone voltages were 3000and 20 V, respectively. The tapered bore gas was 50 L/h, the desolvation gas(nitrogen) was set to 600 L/h at 400�C, and the source temperature was 100�C.Mass range was scanned from 100 to 1200 Da. Multireaction monitoring modewas as follows: low energy, 6 V; and high energy, 20–30 V.

The monitored ion pairs of metabolites of the four flavonoids are shown inSupplemental Table 2. The metabolites of icariin were M1 and M2. M1 showedthe presence of [M + H]+ at a mass-to-charge ratio (m/z) of 515.18, and its mainfragment ion peaks contained m/z 313.07 and 369.13. M2 showed the presenceof [M + H]+ at m/z 677.24, and its fragment ion peaks were predominantlym/z 313.07 and 369.13. The metabolite of epimedin A showed the presence of[M + H]+ at m/z 677.24, and its main fragment ion peaks contained m/z 369.13and 313.07. The metabolite of epimedin B contained [M + H]+ at m/z 647.23,and its fragment ion peaks mainly included m/z 369.13 and 313.07. Themetabolite of epimedin C showed the presence of [M + H]+ at m/z 661.25, andits fragment ion peaks mainly included m/z 515.19, 369.13, and 313.07.

Data Analysis. To analyze the hydrolysis behavior of epimedium flavonoidsin intestinal flora and enzyme incubation, the hydrolysis rate of epimediumflavonoids was calculated using the following equation:

Hydrolysis rate ¼ At 2Ar

Atð1Þ

Fig. 2. Hydrolysis of epimedium flavonoids after coincubation with intestinal flora. The x-axis represents incubation time, and the y-axis represents the content of epimediumflavonoids at different time points. Data are means 6 S.D. (n = 3).

1592 Zhou et al.

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

In eq. 1, At is the total amount of epimedium flavonoids before hydrolysis, andAr is the residual amount of epimedium flavonoids after hydrolysis.

To analyze the absorption behavior of epimedium flavonoids in intestinalperfusion, drug absorption was measured by the rate of its disappearance, andthe apparent permeability coefficient (P*app) was determined through the rate ofdrug disappearance. The P*app was calculated using the following equation:

P*app ¼ 2 v� 12prl

lnroutVout

rinVinð2Þ

In eq. 2, Vin and Vout (in milligrams) are volumes of poured and collected testingsolution, respectively; v is the perfusion velocity; rout and rin (in grams per liter)

are concentrations of inlet and outlet solutions, respectively; l (in centimeters) is thelength of the perfused bowel (typical length is 10 cm); and r (in centimeters) is thecross-sectional radius of the perfused bowel. Values are indicated as means6 S.D.

Statistical Analysis. The t test (Microsoft Excel; Microsoft Corporation,Redmond, WA) was used to analyze the data. The prior level of significancewas set at P , 0.05.

Results

Hydrolysis of Epimedium Flavonoids by Intestinal Flora. Whenincubated with intestinal flora of sham and osteoporosis rats, icariin,epimedin A, epimedin B, and epimedin C were rapidly hydrolyzed.

Fig. 3. Hydrolysis behaviors of epimedium flavonoids after coincubation with intestinal enzymes. The x-axis represents incubation time, and the y-axis represents the contentof epimedium flavonoids at different time points. (A–D) Content-time profiles of icariin, epimedin A, epimedin B, and epimedin C in intestinal enzyme incubations,respectively. Data are means 6 S.D. (n = 3).

Absorption and Metabolism of Epimedium Flavonoids 1593

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

HPLC-UV analysis demonstrated that each of the flavonoids had onlyone metabolite (Supplemental Figs. 1 and 2). UPLC/Q-TOF-MSanalysis revealed that the metabolites of the four flavonoids were the7-deglucosed products baohuoside I (icariin), sagittatoside A (epi-medin A), sagittatoside B (epimedin B), and 299-O-rahmonosylicarsideII (epimedin C). The metabolites of the sham and osteoporosis groupswere the same.Although the metabolites in the two groups were similar, the

hydrolysis rates were different. To measure the hydrolysis rate ofepimedium flavonoids, rates measured at logarithmically varyingconcentrations of flavonoids (Y) and time (X/h) were fit to regressionequations. Larger slopes indicate a more rapid hydrolysis. Thehydrolysis velocities of icariin, epimedin A, epimedin B, and epimedinC in sham rats (slope values = 0.297, 0.264, 0.279, and 0.164,respectively)were greater than in osteoporosis rats (slope values =0.237, 0.212, 0.217, and 0.090, respectively) (Table 1). The hydrolysisof icariin was the most rapid of the four epimedium flavonoids in bothgroups, indicating that the 7-glucose in icariin, which has twoglycosylations, was more easily hydrolyzed by intestinal flora thanthat in epimedin A, epimedin B, and epimedin C, all of which havethree glycosylations.The extent of hydrolysis of epimedium flavonoids incubated

with intestinal flora also varied between sham and osteoporosisrats. After incubation for 1 hour with intestinal flora, 0.65, 0.63,0.59, and 0.51 of icariin, epimedin A, epimedin B, and epimedin C,respectively, was metabolized in sham rats. In osteoporosis rats, theextent of hydrolysis of icariin, epimedin A, epimedin B, andepimedin C was 0.56, 0.37, 0.40, and 0.37, respectively. Icariin,epimedin A, and epimedin B could not be detected after 4 hours ofincubation in intestinal flora from sham rats, whereas they stillcould be detected in the samples of intestinal flora fromosteoporosis rats after 6 hours. The hydrolysis of epimedin Cwas slower than the other three compounds and it could still bedetected after 8 hours of incubation with intestinal flora from bothsham and osteoporosis rats. Nevertheless, the extent of hydrolysisof epimedin C incubated with intestinal flora of sham rats was22.03% higher than that of osteoporosis group. Baohuoside I wasnot hydrolyzed by the intestinal flora of either sham rats orosteoporosis rats (Fig. 2). Overall, the intestinal flora of theosteoporosis group exhibited slower hydrolysis than did the shamgroup.Hydrolysis of Epimedium Flavonoids by the Intestinal Enzymes

from Four Different Intestinal Segments. Epimedium flavonoidswere mainly hydrolyzed by duodenum and jejunum enzymes in thesham and osteoporosis groups. The metabolites were identified byUPLC/Q-TOF-MS. The metabolites in the two groups were thesame. Furthermore, after intestinal enzyme incubation, the metab-olites of epimedin A, epimedin B, and epimedin C were the same asthat after intestinal flora incubation. However, icariin had anadditional metabolite after intestinal enzyme incubation (Supple-mental Fig. 2). This metabolite has the same m/z of protonatedmolecular ions and daughter ions, suggesting that this metabolite oficariin might be a rearrangement product of icariin (RPI). Theseresults indicate that hydrolase might be different in intestinal fecesand mucosa.The hydrolysis rates of epimedium flavonoids incubated with

four different intestinal segments from sham and osteoporosis ratswere calculated. As shown in Fig. 3 and Table 2, the hydrolysisrates of epimedium flavonoids incubated with intestinal segmentsfrom osteoporosis rats were lower than those from sham rats. After1-hour incubation with duodenum, jejunum, ileum, or colon enzymes,the hydrolysis rates of icariin for the sham group were 0.89, 0.85, 0.040,

and 0.0050, respectively. The hydrolysis rates of icariin for theosteoporosis group were 0.68, 0.72, 0.0030, and 0, respectively. Withduodenum and jejunum enzyme incubation, hydrolysis of icariin by thesham group was completed at 1.5 hours, whereas hydrolysis in theosteoporosis group was not complete until 2 hours. The hydrolysis ratesof epimedin A after 1-hour incubation with duodenum, jejunum, ileum,and colon enzymes were 0.39, 0.33, 0.040, and 0.020, respectively, forthe sham group and 0.15, 0.24, 0, and 0, respectively, for theosteoporosis group. For epimedin B, the rates were 0.30, 0.16, 0.21, and0, respectively, for the sham group and 0.22, 0.090, 0, and 0,respectively for the osteoporosis group. For epimedin C, the rates were0.63, 0.70, 0.090, and 0, respectively, for the sham group and 0.32,0.23, 0.070, and 0, respectively, for the osteoporosis group.The metabolites generated were also measured to evaluate the

hydrolysis capability of intestinal enzymes. As shown in Fig. 4, theamounts of metabolites of epimedium flavonoids incubated with fourdifferent intestinal segments were higher in the sham rats than in theosteoporosis rats. Metabolites of icariin, baohuoside I, and RPI weredetected after incubation with enzymes from the duodenum, jejunum,and ileum for both groups. After incubation with colon enzymes, onlybaohuoside I could be found in the sham group, and RPI was notdetected in either group. Moreover, the amount of baohuoside Iincreased with time; however, the amount of RPI increased, thendeclined, and finally disappeared, indicating that RPI might betransferred to baohuoside I over time. Epimedin A, epimedin B, andepimedin C have similar structures, suggesting that their hydrolysismechanisms might be similar. Their metabolites, sagittatoside A,sagittatoside B, and 299-O-rahmonosylicarside II, respectively, wereobserved upon incubation with duodenum, jejunum, and ileum enzymesin the sham group, but they could only be detected upon incubation withduodenum and ileum enzymes in the osteoporosis group. No hydrolysisof baohuoside I was observed in any intestinal segment.Absorption and Metabolism of Epimedium Flavonoids in an In

Situ Intestinal Perfusion Model. The above studies focused on the invitro hydrolysis of epimedium flavonoids. In this section, we used anin situ intestinal perfusion model to study whether osteoporosisaffected the in situ intestinal absorption and hydrolysis of epimediumflavonoids. The epimedium flavonoids could be absorbed andmetabolized in two ways: either directly, as the unchanged molecule,or after hydrolysis by intestinal enzymes, such as LPH, and thenabsorbed as metabolites. The apparent permeability coefficient (P*app)and eliminative percentage (10 cm% EP; 10-cm intestinal segments)were used to characterize the absorption and hydrolysis of epimediumflavonoids for sham and osteoporosis rats.The metabolites in perfusates for the two groups were the same

and were consistent with the metabolites observed after incubationswith intestinal enzymes (Supplemental Fig. 2), further confirming

TABLE 2

Hydrolysis of epimedium flavonoids by four different intestinal segments for 1 hour

Compound GroupHydrolysis Rate

Duodenum Jejunum Ileum Colon

Icariin Sham 0.89 0.85 0.040 0.0050Osteoporosis 0.68 0.72 0.0030 0

Epimedin A Sham 0.39 0.33 0.040 0.020Osteoporosis 0.15 0.24 0 0

Epimedin B Sham 0.30 0.16 0.21 0Osteoporosis 0.22 0.090 0 0

Epimedin C Sham 0.63 0.70 0.090 0Osteoporosis 0.32 0.23 0.070 0

1594 Zhou et al.

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

Fig. 4. Metabolites of epimedium flavonoids after coincubation with intestinal enzymes. The x-axis represents incubation time, and the y-axis represents relativeconcentrations of metabolites (the ratio of peak areas of metabolites and testosterone). (A–E) Peak area-time profiles of baohuoside I, RPI, sagittatoside A, sagittatoside B,and 299-O-rahmonosylicarside II in intestinal enzyme incubations, respectively. Data are means 6 S.D. (n = 3).

Absorption and Metabolism of Epimedium Flavonoids 1595

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

that the hydrolases might be different in intestinal feces andmucosa.The results of absorption and hydrolysis of epimedium flavonoids

in the in situ intestinal perfusion model are shown in Figs. 5–8. Asshown in Fig. 5, in both groups, the P*app and 10 cm% EP values oficariin became progressively smaller from the duodenum to the colon.In the sham group, the P*app values of icariin in the duodenum,jejunum, ileum, and colon were 5.59 6 0.57, 5.26 6 0.74, 1.30 60.18, and 0.50 6 0.09, respectively, and the 10 cm% EP values oficariin were 72.26% 6 7.89%, 69.09% 6 8.29%, 19.66% 6 2.93%,and 9.48% 6 1.50%, respectively. For the osteoporosis group, theP*app values of icariin in the duodenum and jejunum as well as the10 cm% EP values of icariin in the four intestinal segments weresignificantly lower than those of sham group (versus the correspond-ing bowel, P , 0.05). The P*app values of icariin in four intestinalsegments were 3.876 0.46 (duodenum), 3.576 0.55 (jejunum), 0.7360.60 (ileum), and 0.36 6 0.29 (colon), and the 10 cm% EP values were50.20% 6 5.34%, 44.17% 6 6.18%, 9.01% 6 1.35%, and 4.72% 60.76%, respectively. At the same time, the amounts of metabolites weresignificantly different. The metabolites baohuoside I and RPI were foundin all four intestinal segments of the sham group. However, althoughbaohuoside I could be detected in all four intestinal segments in theosteoporosis group, and only RPI could be found in the first threeintestinal segments. Furthermore, the amounts of baohuoside I and RPIin the osteoporosis group were less than that in the sham group.The P*app values of epimedin A in the duodenum, jejunum, ileum,

and colon (3.79 6 0.18, 2.72 6 0.08, 0.88 6 0.19, and 0.36 6 0.06,respectively) of the sham group exceeded the values for theosteoporosis group (2.66 6 0.36, 1.55 6 0.16, 0.41 6 0.04, and0.17 6 0.04, respectively) (versus the corresponding bowel, P ,0.05). The 10 cm% EP values of epimedin A in the four segments

(69.01%6 2.65%, 50.30%6 2.63%, 15.70%6 2.45%, and 6.36%61.12%, respectively) of sham rats were also higher than that ofovariectomized rats (48.61% 6 6.76%, 29.16% 6 3.05%, 7.18% 61.70%, and 3.34% 6 0.68%, respectively) (versus the correspondingbowel, P , 0.05). Under the two conditions, from the duodenum tothe colon, the P*app and 10 cm% EP decreased sequentially. In turn,the amounts of epimedin A in the four intestinal segments weredecreased from the duodenum to the colon (Fig. 6). The amounts ofsagittatoside A in the sham group were much greater than that of theosteoporosis group at the relevant intestinal segments.In sham rats’ duodenum, jejunum, ileum, and colon, the P*app

values of epimedin B were 3.52 6 0.17, 2.44 6 0.19, 0.60 6 0.11,and 0.42 6 0.04, respectively, and the 10 cm% EP values were66.35% 6 3.15%, 45.18% 6 2.43%, 11.17% 6 2.43%, and 7.83% 60.66%, respectively. In osteoporosis rats’ duodenum, jejunum, ileum,and colon, the P*app values of epimedin B were 2.11 6 0.13, 1.20 60.13, 0.36 6 0.04, and, 0.23 6 0.04, respectively, and the 10 cm% EPvalues were 39.75% 6 3.07%, 23.09% 6 2.47%, 6.91% 6 1.63%,and 4.33% 6 0.73%, respectively. These findings show that for bothsham and osteoporosis rats, the P*app of epimedin B in the duodenum,jejunum, ileum, and colon fell and, in turn, so did the 10cm% EP ofepimedin B. Compared with the sham group, the P*app and 10 cm%EP values of epimedin B in the osteoporosis group decreasedremarkably (versus the corresponding bowel, P , 0.05). The amountof sagittatoside B was highest in the duodenum and lowest in thecolon. The sagittatoside B content was reduced in the osteoporosisgroup (Fig. 7).In osteoporosis rats’ duodenum, jejunum, ileum, and colon, the P*app

values of epimedin C were 1.726 0.11, 1.556 0.12, 0.446 0.06, and0.28 6 0.05, respectively, and the 10 cm% EP values were 32.48% 62.17%, 26.22% 6 2.36%, 7.90% 6 0.96%, and 4.50% 6 0.90%,

Fig. 5. Absorption and metabolism of icariin in situ intestine model. (A) P*app of icariin in different intestines. (B) 10 cm% EP of icariin in different intestines. (C) Peakareas of baohuoside I in different intestines. (D) Peak areas of RPI in different intestines. Significant differences of P*app and 10 cm% EP of icariin and amounts ofmetabolites in different intestines existed between the sham and osteoporosis groups (versus the corresponding bowel, *P , 0.05). Data are means 6 S.D. (n = 4).

1596 Zhou et al.

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

respectively. At the corresponding bowel level, the P*app and 10 cm%EP values in the sham group were significantly increased comparedwith the values of the osteoporosis group (versus the correspondingbowel, P , 0.05). The P*app values of epimedin C were 3.10 6 0.16,2.67 6 0.11, 1.02 6 0.17, and 0.48 6 0.14, respectively, and the10 cm% EP values were 58.47% 6 3.09%, 48.65% 6 2.06%,17.55% 6 4.82%, and 8.38% 6 1.36%. Sequential reduction was alsoobserved for epimedin C in both the sham and osteoporosis groups.The amounts of 299-O-rahmonosylicarside II in the sham group werehigher than that in the osteoporosis group. In the two groups, theamount of 299-O-rahmonosylicarside II diminished sequentially fromthe duodenum to the colon (Fig. 8).

Discussion

This study focused on the influence of osteoporosis on absorptionand metabolism of epimedium flavonoids. We chose ovariectomizedrats as an osteoporotic model. In one series of experiments, weinvestigated the effect of osteoporosis on hydrolysis of epimediumflavonoids in vitro by incubating epimedium flavonoids with intestinalmucosa and feces from osteoporosis and sham rats. In another series,we explored the effect of osteoporosis on the absorption andmetabolism of epimedium flavonoids in vivo by conducting intestinalperfusion experiments in osteoporosis and sham rats.With the exception of baohuoside I, whose absorption was

unchanged, we found that after oral administration of epimediumflavonoids, most of them were hydrolyzed to secondary glycoside oraglycon by enzymes in the intestine, which led to increasedabsorption. The enzymes in the intestine have two main sources:intestinal mucosa and intestinal bacteria. When epimedium flavonoidsare orally administered, they reach the stomach first. Our earlierstudies showed that the main flavonoids in epimedium could not be

hydrolyzed by gastric juice (Gao et al., 2013). Hence, the mainabsorption site of epimedium flavonoids is in the small intestine. Inhuman and rat small intestines (from the duodenum to the ileum),there are two kinds of b-glucosidase that could hydrolyzeepimedium flavonoids: LPH and cytosolic b-glucosidase (CBG)(Németh et al., 2003). Both could hydrolyze the b-glycosidic bondsof glycosides, albeit by substantially different hydrolysis mecha-nisms (Németh et al., 2003). For the CBG hydrolysis pathway, theglycosides must be transported into cells and then hydrolyzed,because CBG is an intracellular enzyme. In our preliminaryresearch, we found that prototypes of epimedium flavonoids hadpoor membrane permeability and had a slow uptake by intestinalcells. In addition, it was reported that the transporter of glycosidesubstances such as SLTGT1 could not transport flavonoids intocells (Dongmei et al., 2012). Therefore, it could be inferred thatepimedium flavonoid hydrolysis was not the role of CBG. LPH,a type of extracellular enzyme that is located in the brush border ofthe small intestinal epithelium, is the only b-glucosidase in themammalian intestinal brush border. There are two distinct activesites of LPH for catalytic hydrolysis: one for hydrolyzing lactoseand flavonoids, and the other for hydrolyzing phlorizin andb-glucosylceramidase (Tseung et al., 2004). Because LPH islocated in the brush border of the small intestinal epithelium(Németh et al., 2003), the flavonoids could be hydrolyzed by LPHonce they enter the small intestine. Our previous experimentsalso confirmed that the hydrolysis of epimedium flavonoidsdecreased after adding LPH inhibitors (Chen et al., 2014), indicatingthat LPH in intestinal mucosa is the major hydrolase thathydrolyzes epimedium flavonoids. However, the above experimentswere carried out in healthy rats, and it was not possible to determinewhether the hydrolysis capability of LPH would be different in apathologic state.

Fig. 6. Absorption and metabolism of epimedin A in situ intestine model. (A) P*app of epimedin A in different intestines. (B) 10 cm% EP of epimedin A in differentintestines. (C) Peak areas of sagittatoside A in different intestines. Significant differences of P*app and 10 cm% EP of epimedin A and amounts of sagittatoside A in differentintestines existed between the sham and osteoporosis groups (versus the corresponding bowel, *P , 0.05). Data are means 6 S.D. (n = 4).

Absorption and Metabolism of Epimedium Flavonoids 1597

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

Because epimedium flavonoids have a strong effect againstosteoporosis, we studied the hydrolysis capability of LPH in theosteoporotic animals. Our experimental results showed that thehydrolysis rates of epimedium flavonoids in osteoporosis rats wasslower than in sham rats, whether the flavonoids were incubated withintestinal enzymes in vitro or were conducted by in situ intestinalperfusion. These results indicated that the amount and activity ofLPH might be lower in osteoporosis rats. Later, the experiment ofLPH expression affirmed our hypothesis that LPH expression wasindeed downregulated in the ovariectomized rats. In addition, ithas been reported that the high incidence of fractures and low bonemineral density in menopausal women are correlated withdecreased activity and expression of LPH (Bacsi et al., 2007),and the activity of LPH in senile osteoporosis patients is greatlyreduced, or even completely inactivated (Bacsi et al., 2007). Bothour results and the literature confirm that low LPH expression isthe major factor that decreases epimedium flavonoid hydrolysisand affects the absorption of flavonoids and their utilization by thebody.In addition to enzymes from intestinal mucosa, enzymes from

microorganisms are involved in the hydrolysis of epimediumflavonoids in the intestinal tract. There are about 100 trillion bacteriain the bowel, of which more than 99% are anaerobic bacteria,including species from the Bacteriodacae, Erysipelotrichaceae (Cat-enibacterium), Clostridiaceae (Peptostreptococcus), and Spirillaceaefamilies. Different species of bacteria secrete different metabolicenzymes and are thereby involved in different types of drugmetabolism. For example, there are many bacteria (e.g., Escherichiacoli, Enterococcus, etc.) that secrete b-glucosidase, which canhydrolyze b-glucoside bonds for the purpose of detoxification oreasier absorption. Epimedium flavonoids that are not fully hydrolyzedin the small intestine travel into the large intestine, where they are

reabsorbed after hydrolysis by bacteria located there. However, owingto the complexity of the intestinal flora, the relationship of intestinalflora and osteoporosis needs to be further studied (Claesson et al.,2012; Koren et al., 2012).As the body’s largest and most sophisticated microecosystem,

maladjusted intestinal bacteria might cause disease, and the diseasecould in turn then change the function of bacteria (Sekirov et al.,2010; Murphy et al., 2013; Shen et al., 2013; Zhao et al., 2013).This could explain the reduced hydrolysis rates of icariin, epimedinA, epimedin B, and epimedin C in the intestinal flora of osteoporosisrats compared with that observed for sham rats. The above resultsindicated that intestinal bacteria activity decreased because of thepathologic conditions of osteoporosis. Some previous reports inthe literature are consistent with our results. It has been reportedthat the number of anaerobic bacteria, especially Lactobacillus,was significantly lower than normal in osteoporosis rats, and thecolonization ability of anaerobic bacteria in the intestine wasdecreased. With respect to metabolism, the ability of bacterialhydrolysis of the glycosidic bonds in general is reduced dueto changes in flora and preference for carbon source (Ishiharaet al., 2002).Diseases can alter the hydrolytic capability of intestinal flora and

enzymes, which further affects drug bioavailability. Hence, it isnecessary to understand the absorption and metabolism of drugs inpathologic states. This experiment explored the influence of osteopo-rosis on the absorption and metabolism of epimedium flavonoids. Theresults showed that the absorption of epimedium flavonoids waslower in osteoporosis rats than in sham rats. This difference should notbe ignored when new therapeutic strategies are developed for thetreatment of osteoporosis or similar diseases. The deficiency ofabsorption and hydrolysis of epimedium flavonoids in osteoporoticstates should be offset, in order for epimedium flavonoids to have

Fig. 7. Absorption and metabolism of epimedin B in situ intestine model. (A) P*app of epimedin B in different intestines. (B) 10 cm% EP of epimedin B in differentintestines. (C) Peak areas of sagittatoside B in different intestines. Significant differences of P*app and 10 cm% EP of epimedin B and amounts of sagittatoside B in differentintestines existed between the sham and osteoporosis groups (versus the corresponding bowel, *P , 0.05). Data are means 6 S.D. (n = 4).

1598 Zhou et al.

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

stable antiosteoporosis efficacy. These results might give rise to newresearch ideas for similar diseases, providing foundations for deeplyexploring the mechanism of absorption and hydrolysis of epimediumflavonoids in osteoporosis rats and for developing new therapeuticstrategies in the future.

Conclusions

In this study, we found that the intestinal absorption and hydrolysisof epimedium flavonoids was slower in osteoporosis rats than in shamrats. Ovariectomization of rats resulted in osteoporosis, which mightaffect the activities of intestinal enzymes and the number of intestinalflora. These changes can then affect the intestinal absorption andhydrolysis of epimedium flavonoids whose structures containglycosylations.

Authorship ContributionsParticipated in research design: Chen, Z. Zhou.Conducted experiments: J. Zhou, Ma, Chen, Wang.Contributed new reagents or analytic tools: Wang, Gao.Performed data analysis: J. Zhou, Ma, Chen.Wrote or contributed to the writing of the manuscript: J. Zhou, Ma, Z. Zhou,

Chen.

References

Abrahamsen B, Brask-Lindemann D, Rubin KH, and Schwarz P (2014) A review of lifestyle,smoking and other modifiable risk factors for osteoporotic fractures. Bonekey Rep 3:574.

Akao T, Che QM, Kobashi K, Hattori M, and Namba T (1996) A purgative action of barbaloin isinduced by Eubacterium sp. strain BAR, a human intestinal anaerobe, capable of transformingbarbaloin to aloe-emodin anthrone. Biol Pharm Bull 19:136–138.

Andlauer W, Kolb J, and Fürst P (2000a) Absorption and metabolism of genistin in the isolatedrat small intestine. FEBS Lett 475:127–130.

Andlauer W, Kolb J, and Fürst P (2000b) Isoflavones from tofu are absorbed and metabolized inthe isolated rat small intestine. J Nutr 130:3021–3027.

Andlauer W, Kolb J, Stehle P, and Fürst P (2000c) Absorption and metabolism of genistein inisolated rat small intestine. J Nutr 130:843–846.

Bacsi K, Kosa JP, and Balla B (2007) The decreased activity of lactase phlorizin hydrolase andbone mineral density in postmenopausal women. J Bone Miner Res 22:186–198.

Bandeira F, Costa AG, Soares Filho MA, Pimentel L, Lima L, and Bilezikian JP (2014) Bonemarkers and osteoporosis therapy. Arq Bras Endocrinol Metabol 58:504–513.

Chen H and Kubo KY (2014) Bone three-dimensional microstructural features of the commonosteoporotic fracture sites. World J Orthop 5:486–495.

Chen J, Lin H, and Hu M (2003) Metabolism of flavonoids via enteric recycling: role of intestinaldisposition. J Pharmacol Exp Ther 304:1228–1235.

Chen Y, Wang Y, Zhou J, Gao X, Qu D, and Liu C (2014) Study on the mechanism of intestinalabsorption of epimedins a, B and C in the Caco-2 cell model. Molecules 19:686–698.

Chen Y, Zhao YH, Jia XB, and Hu M (2008) Intestinal absorption mechanisms of prenylatedflavonoids present in the heat-processed Epimedium koreanum Nakai (Yin Yanghuo). PharmRes 25:2190–2199.

Claesson MJ, Jeffery IB, Conde S, Power SE, O’Connor EM, Cusack S, Harris HMB, CoakleyM, Lakshminarayanan B, and O’Sullivan O, et al. (2012) Gut microbiota composition corre-lates with diet and health in the elderly. Nature 488:178–184.

Cui L, Sun E, Zhang Z, Qian Q, Tan X, Xu F, and Jia X (2013) Metabolite profiles of epimedin Bin rats by ultraperformance liquid chromatography/quadrupole-time-of-flight mass spectrom-etry. J Agric Food Chem 61:3589–3599.

Dongmei C, Guofeng L, and Weilin P (2012) Serum cytosolic b-glucosidase levels in neonatalnecrotizing enterocolitis. Iran J Pediatr 22:452–456.

Gao X, Chen Y, Wang Y, Sun WJ, and Jia XB (2013) [Study on different factors affecting thebionic enzymatic hydrolysis of icariin]. Yao Xue Xue Bao 48:1716–1721.

Higaki K, Sone M, Ogawara K, and Kimura T (2004) Regulation of drug absorption from smallintestine by enteric nervous system I: a poorly absorbable drug via passive diffusion. DrugMetab Pharmacokinet 19:198–205.

Ishihara M, Homma M, Kuno E, Watanabe M, and Kohda Y (2002) [Combination use of kampo-medicines and drugs affecting intestinal bacterial flora]. Yakugaku Zasshi 122:695–701.

Johnell O and Kanis JA (2006) An estimate of the worldwide prevalence and disability associatedwith osteoporotic fractures. Osteoporos Int 17:1726–1733.

Kanis JA, Melton LJ, 3rd, Christiansen C, Johnston CC, and Khaltaev N (1994) The diagnosis ofosteoporosis. J Bone Miner Res 9:1137–1141.

Koren O, Goodrich JK, Cullender TC, Spor A, Laitinen K, Bäckhed HK, Gonzalez A, Werner JJ,Angenent LT, and Knight R, et al. (2012) Host remodeling of the gut microbiome and met-abolic changes during pregnancy. Cell 150:470–480.

Kulak CA, Borba VZ, Kulak Júnior J, and Custódio MR (2014) Bone disease after trans-plantation: osteoporosis and fractures risk. Arq Bras Endocrinol Metabol 58:484–492.

Liu Y and Hu M (2002) Absorption and metabolism of flavonoids in the caco-2 cell culture modeland a perused rat intestinal model. Drug Metab Dispos 30:370–377.

Löscher W and Potschka H (2005) Blood-brain barrier active efflux transporters: ATP-bindingcassette gene family. NeuroRx 2:86–98.

Murphy EF, Cotter PD, Hogan A, O’Sullivan O, Joyce A, Fouhy F, Clarke SF, Marques TM,O’Toole PW, and Stanton C, et al. (2013) Divergent metabolic outcomes arising from targetedmanipulation of the gut microbiota in diet-induced obesity. Gut 62:220–226.

Fig. 8. Absorption and hydrolysis of epimedin C in situ intestine model. (A) P*app of epimedin C in different intestines. (B) 10 cm% EP of epimedin C in different intestines.(C) Peak areas of 299-O-rahmonosylicarside II in different intestines. Significant differences of P*app and 10 cm% EP of epimedin C and amounts of 299-O-rahmonosylicarside II in different intestines existed between the sham and osteoporosis groups (versus the corresponding bowel, *P , 0.05). Data are means 6 S.D. (n = 4).

Absorption and Metabolism of Epimedium Flavonoids 1599

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from

Nelson HD, Humphrey LL, Nygren P, Teutsch SM, and Allan JD (2002) Postmenopausal hor-mone replacement therapy: scientific review. JAMA 288:872–881.

Németh K, Plumb GW, Berrin JG, Juge N, Jacob R, Naim HY, Williamson G, Swallow DM,and Kroon PA (2003) Deglycosylation by small intestinal epithelial cell beta-glucosidases isa critical step in the absorption and metabolism of dietary flavonoid glycosides in humans. EurJ Nutr 42:29–42.

Qian Q, Li SL, Sun E, Zhang KR, Tan XB, Wei YJ, Fan HW, Cui L, and Jia XB (2012)Metabolite profiles of icariin in rat plasma by ultra-fast liquid chromatography coupled totriple-quadrupole/time-of-flight mass spectrometry. J Pharm Biomed Anal 66:392–398.

Sekirov I, Russell SL, Antunes LCM, and Finlay BB (2010) Gut microbiota in health and disease.Physiol Rev 90:859–904.

Sesink AL, Arts IC, Faassen-Peters M, and Hollman PC (2003) Intestinal uptake of quercetin-3-glucoside in rats involves hydrolysis by lactase phlorizin hydrolase. J Nutr 133:773–776.

Shen J, Obin MS, and Zhao L (2013) The gut microbiota, obesity and insulin resistance. MolAspects Med 34:39–58.

Tseung CW, McMahon LG, Vázquez J, Pohl J, and Gregory JF, 3rd (2004) Partial amino acid sequenceand mRNA analysis of cytosolic pyridoxine-beta-D-glucoside hydrolase from porcine intestinalmucosa: proposed derivation from the lactase-phlorizin hydrolase gene. Biochem J 380:211–218.

Wang X, Guo B, Li Q, Peng J, Yang Z, Wang A, Li D, Hou Z, Lv K, and Kan G, et al. (2013)miR-214 targets ATF4 to inhibit bone formation. Nat Med 19:93–100.

Wei Y, Li P, Fan H, Sun E, Wang C, Shu L, Liu W, Xue X, Qian Q, and Jia X (2012) Metaboliteprofiling of four major flavonoids of Herba Epimedii in zebrafish. Molecules 17:420–432.

Wu H, Lien EJ, and Lien LL (2003) Chemical and pharmacological investigations of Epimediumspecies: a survey. Prog Drug Res 60:1–57.

Xu YX, Xu B, Wu CL, Wu Y, Tong PJ, and Xiao LW (2011) Dynamic expression of DKK1protein in the process whereby Epimedium-derived flavonoids up-regulate osteogenic anddown-regulate adipogenic differentiation of bone marrow stromal cells in ovariectomized rats.Orthop Surg 3:119–126.

Zhang G, Qin L, and Shi Y (2007) Epimedium-derived phytoestrogen flavonoids exert beneficialeffect on preventing bone loss in late postmenopausal women: a 24-month randomized, double-blind and placebo-controlled trial. J Bone Miner Res 22:1072–1079.

Zhao H, Fan M, Fan L, Sun J, and Guo D (2010) Liquid chromatography-tandem mass spec-trometry analysis of metabolites in rats after administration of prenylflavonoids from Epime-diums. J Chromatogr B Analyt Technol Biomed Life Sci 878:1113–1124.

Zhao XB, Wu WJ, Li DD, Li XY, Zhang HL, Tan ZJ, and Cai GX (2013) The effect of modelingspleen-deficiency constipation on the intestinal microbiota and enzyme activities in mice.Chinese J Microeco 25:993–996.

Zhou J, Chen Y, Wang Y, Gao X, Qu D, and Liu C (2013) A comparative study on themetabolism of Epimedium koreanum Nakai-prenylated flavonoids in rats by an intestinal en-zyme (lactase phlorizin hydrolase) and intestinal flora. Molecules 19:177–203.

Address correspondence to: Yan Chen, Multicomponent of Traditional ChineseMedicine and Microecology Research Center, Jiangsu Provincial Academy ofChinese Medicine, 100 Shizi Road, Nanjing 210028, Jiangsu, China. E-mail:ychen202@hotmail.com

1600 Zhou et al.

at ASPE

T Journals on M

ay 26, 2021dm

d.aspetjournals.orgD

ownloaded from