Drug-Herb Interaction Between Selective Estrogen Receptor Modulators and Estrogenic Herbal Medicine Er-Xian Decoction in Bone in Vivo and in Vitro

Background: Er-Xian decoction (EXD), a traditional Chinese Medicine for managing menopausal syndrome and osteoporosis in China, could exert osteoprotective action via activation of estrogen receptor (ERs) and regulation of serum estradiol without causing severe side effects. However, no fundamental studies have explored its potential interaction in the combined use of prescription drugs, Selective Estrogen Receptor Modulators (SERMs), regarding the osteogenic and uterotrophic effects. The present study evaluated the estrogenic effects of EXD and its potential interactions with tamoxifen and raloxifene in bone and uterus using a mature ovariectomized (OVX) Sprague-Dawley (SD) rat model and human osteoblastic MG-63 cells. Methods: Six-month-old female SD rats were randomly assigned to Sham-operated group or seven OVX groups: vehicle, 17ß-estradiol (E2, 1.0 mg/kg.day), Tamoxifen (Tamo, 1.0 mg/kg.day), Raloxifene (Ralo, 3.0 mg/kg.day), EXD (EXD, 1.6 g/kg.day), EXD+Tamoxifen (EXD+Tamo) and EXD+Raloxifene (EXD+Ralo). The effect of EXD on bodyweight, bone mineral density (BMD), bone microarchitecture, biochemical analysis of serum and urine, and uterus were evaluated. In addition, Alkaline phosphatase assay and activation of estrogen-responsive element mediated by EXD and in its combination with SERMs were investigated in MG-63 cells. Results: In vivo, EXD could interact with SERMs to modulate the serum estradiol, follicle-stimulating hormone, osteocalcin level as well as BMD and bone properties in OVX rats. Moreover, EXD could relieve the uterotrophic effect of SERMs. In vitro, EXD crude extract and EXD-treated serum could promote ALP activity. In particular, EXD-treated serum could interact with SERMs on regulating ALP activity in MG-63 cells. Conclusion: Our study demonstrated that EXD in vivo and EXD-treated serum in vitro did not weaken the osteogenic effect of SERMs. Interestingly, EXD seems to ameliorate the uterotrophic effects of SERMs. Therefore, the combined use of EXD and SERMs may be considered safe and effective in managing postmenopausal osteoporosis. analysis metabolites in EXD will not be high enough to compete with SERMs for binding towards ERs. Future studies will be needed to characterize major phytoestrogens' tissue distribution in OVX rats upon long-term treatment with EXD.


Introduction
Osteoporosis is a metabolic bone disease characterized by decreased bone mineral density (BMD) and compromised bone microarchitecture that result in an increased risk of fracture. Over 200 million postmenopausal women worldwide are currently estimated to suffer from this disease, causing more than 8.9 million fractures annually [1]. Dramatic decrease in circulating estrogen level due to ovarian dysfunction is believed to be the major cause of bone loss in postmenopausal women [2].
Selective estrogen receptor modulators (SERMs) are estrogen receptor (ERs) ligand and behave as agonists or antagonists depending on tissue type [3]. They are commonly prescribed to postmenopausal women to manage estrogen-related diseases, tamoxifen for ER-positive breast cancer treatment and raloxifene for osteoporosis treatment [3,4]. Indeed, both tamoxifen and raloxifene act as ER agonist and exert extensive bone protective effects [3]. However, administration of tamoxifen is associated with the occurrence of endometrial polyps (8-36%) and endometrial hyperplasia (1-20%) in women [5]. Therefore, the modi cation of SERMs usage should be considered.
Traditional Chinese medicine (TCM) have been clinically prescribed as an alternative approach for treating bone diseases with a long history of safe use. The development of postmenopausal osteoporosis is due to kidney de ciency according to the principles of TCM. Er Xian decoction (EXD), one of the most popular kidneytonifying Chinese formulas, has been clinically used to relieve postmenopausal osteoporosis for more than 60 years with high effectiveness, fewer side effects on reproductive organs, and at relatively low cost [6].
EXD contains six herbs, including Herba epimedii (HEP, ) and Curculigo orchioides Gaertn (XM, ) as the principal drugs, while the other four herbs as the adjuvant drugs, Morinda o cinalis (BJT, ), Rhizoma Anemarrhenae Bunge (ZM, ), Phellodendron amurense Rupr (HB, ), and Radix Angelicae Sinensis (DG, ) [7]. The main active components of EXD were avonol phytoestrogens, ER ligands exhibiting estrogenic activities which account for its e cacy in treating osteoporosis [8]. A systemic review and meta-analysis involving 677 patients in 5 clinical trials indicated that EXD was clinically effective in relieving menopausal syndrome via increasing circulating estradiol (E2) [9]. A study clearly revealed the modulatory impact of EXD on the hypothalamic-pituitary axis [10], which might partially account for its bone protective effects through hormonal regulation. Our in vivo study demonstrated the inhibitory effects of EXD on bone turnover in mature ovariectomized (OVX) rats [11] while our in vitro study found that EXD dramatically exerted ER-dependent cell proliferation and differentiation and activated estrogen response element-dependent transcriptional activity in rat osteoblastic UMR-106 cells [12], indicating the estrogen-like actions of EXD. These ndings suggest that phytoestrogen-containing EXD could safely exert bone protective effects via ERs.
With the increasing popularity of TCM in treating menopausal symptoms, it is anticipated that EXD might be a good candidate to use in combination with SERMs exacerbating the side effect of SERMs without weakening SERMs' osteoblastic effects and. However, there is no fundamental study on the interaction between SERMs and EXD in combination regarding the estrogenic effects of SERMs at bone as well as reproductive tissues, especially uterus. The present study aimed to systematically investigate the estrogenic activity of EXD in bone and uterus as well as the interactions between EXD and SERMs (tamoxifen and raloxifene) using a mature OVX rat model and human osteosarcoma MG-63 cells.

Methods
Authentication, extraction and quality control of EXD Six herbal medicines used in EXD, including HEP, XM, BJT, ZM, HB, and DG at a ratio of (9:9:9:6:6:9) were purchased from mainland China and authenticated by High-performance liquid chromatography (HPLC) assays. HPLC was conducted to ensure that these herbs' quality ful lls the requirement of the China Pharmacopoeia and/or the Hong Kong Chinese Materia Medica Standard. Upon authentication, herbs were delivered to Xi'an Pincredit Bio-tech Co., Ltd for water extract preparation as descripted previously [11]. The quantities of chemical markers in EXD extract was determined by Liquid chromatography-mass spectrometry (LC-MS).

Experimental design and animal treatment
The animal experiment protocol was approved by the Hong Kong Polytechnic University Animal Subjects Ethics Sub-committee (ASESC Case: 15-16/31-ABCT-HMRF). Eighty six-month-old female Sprague Dawley (SD) rats were given OVX or sham operation. Upon two-week recovery, the OVX rats were orally administrated with vehicle, 17ß-estradiol (E2, 1.0 mg/kg.day), tamoxifen (1.0 mg/kg.day), raloxifene (3.0 mg/kg.day), EXD (1.6 g/kg.day) or its combinations with tamoxifen (EXD+Tamo) and raloxifene (EXD+Ralo) for 12 weeks (n=10/group). Sham-operated rats treated with vehicle were used as control. Dosages of E2, tamoxifen, raloxifene, and EXD were chosen based on equivalent human dosages and/or previous preclinical studies [11,13]. During the whole treatment, the animals were pair-fed with phytoestrogen-free diet (AIN-93M, Research diet, New Jersey, USA) to remove the in uence of phytoestrogens in the diet. Bodyweight was measured every two weeks. Urine, serum and uterus were collected and stored at -80℃ for further detection. The left leg and lumbar spine were collected for micro-CT analysis. Sample size (n = 10/group) was determined at alpha of 5 % and power of 90 % based on our previous data of serum osteocalcin in rats (1 % suppression) [14].

Hematoxylin-eosin staining
The morphological change of uterus was examined by Hematoxylin-eosin staining (H&E). Uterus samples were collected and xed in 4% paraformaldehyde for 6 hours. After dehydration (Leica TP1020, Leica biosystem, Buffalo Grove, USA), tissues were embedded in para n. Sections for histology were cut at 5 microns from the para n blocks using a standard microtome (Leica biosystem, Buffalo Grove, USA) before mounting and heatxed on glass slides. Standard H&E staining was performed. A minimum of 5 sections from each sample were observed using 400× magni cation and photographed using a photoscope (Olympus BX51, Olympus corporation, Tokyo, Japan)

Real-time PCR
The mRNA expression histone H3, a marker of proliferation for endometrium, was measured in rat uterus. Total RNA was isolated from the uterus of rats in Trizol reagent (Thermo sher, MA, USA) by using Precellys 24 homogenizer (Bertin Technologies SAS, France). RNA was reverse-transcribed into cDNA by using High-Capacity cDNA Reverse Transcription Kits (Applied Biosystems, MA, USA) in Veriti™ 96-Well Thermal Cycler (Applied Biosystems, MA, USA). The speci c primer for histone H3 and GAPDH (histone H3 forward 5' CTACCAGAAGTCGACCGAGC 3' , reverse 5' TCCTTGGGCATGATGGTGAC 3' ; GAPDH forward 5' CAAGTTCAACGGCACA GTCAAGG 3'; reverse 5' ACATACTCAGCACCAGCATCACC 3') were used to perform RT-PCR with TB Green detection (TaKaRa Bio, Kyoto, Japan) using QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems, MA, USA). For each gene, a standard curve was established to determine the relative quantity of mRNA, and the melting curve was used to assess the speci city of the ampli cation.

Preparation of EXD-treated serum
The metabolites of TCM usually considered as the functional component, which require biological activation.
To observe the direct estrogenic actions of EXD in vitro, EXD-treated serum (EXD-Ts), a biologically activated form of EXD, were prepared in OVX rats. OVX rats were given vehicle or EXD at 16.0 g/kg.day (n=10/group) for three consecutive days and pair-fed with phytoestrogen-free diet. Upon the last oral administration on day 3, the rats were fasted overnight and given drugs one more time in the following morning. Rats were then sacri ced after an hour, and serum was prepared and stored at -80 o C. LC-MS analysis of the serum was performed to con rm the presence of major chemical markers from each herb. Methanol extract of serum was prepared, and extract of 1ml serum was dissolved in 1ml of ethanol, and the concentration of this solution was de ned as "1".

Statistical analysis
Data were reported as mean ± SEM. Inter-group differences of in vivo experiment were analyzed by one-way ANOVA with Tukey's as post hoc test. Inter-group differences in in vitro experiment were analyzed by t-test. Interactions between drugs were analyzed by two-way ANOVA with Bonferroni as a post hoc test. A value of p 0.05 was considered statistically signi cant. The correlation between serum hormone level and BMD at tibia in rats were analyzed using Pearson's correlation.

EXD extract preparation and authentication
Raw herbs were authenticated according to the presence and quantities of speci c chemical markers ( see Additional le 1 and 2). They are icariin for HEP, curculigoside for XM, ferulic acid for DG, berberine hydrochloride for HB, timosaponin BII for ZM, and nystose for BJT. HPLC pro le con rmed that the quality of all raw herbs had ful lled their respective requirements. After water extraction, the amount of each chemical marker in EXD extract was quanti ed by LC-MS (Table 1).
EXD could interact with SERMs to increase BMD and improve bone properties in OVX rats Figure 1A clearly indicated that ovariectomy signi cantly reduced BMD at distal femur (111.88 ± 11.89 mgHA/ccm 2 ), proximal tibia (114.49 ± 12.40 mgHA/ccm 2 ), and lumbar vertebra (199.78± 21.05 mgHA/ccm 2 ) in rats when compared to Sham group. Treatment with E2 , EXD, SERMs alone and in combination with EXD signi cantly increased BMD in OVX rats at all three sites (p<0.01 vs. OVX). EXD, SERMs, and their combinations dramatically improved bone microstructure at the distal femur ( Figure 1B) and trabecular bone properties at the distal femur (Table 3), proximal tibia, and lumbar spine (see Additional le 3) in OVX rats. As expected, OVX in rats signi cantly decreased BV/TV, Conn.D, Tb.N, and Tb.Th while signi cantly increased SMI and Tb.Sp at all the three bone sites (p<0.05 vs. Sham). Treatment of OVX rats with E2, tamoxifen, raloxifene, and EXD alone signi cantly increased BV/TV, Conn.D, Tb.N, and Tb.Th as well as decreased SMI and Tb.Sp of bone at all three sites (p<0.05 vs. OVX). Treatment of OVX rats with EXD combined with tamoxifen or raloxifene also signi cantly improved bone properties at three sites (p<0.05 vs. OVX). BMD and bone properties at all three sites in OVX rats treated with SERMs alone were not statistically different from those in OVX rats treated with respective SERMs combined with EXD. Two-way ANOVA analysis indicated that EXD interacted with tamoxifen and with raloxifene to alter BMD at all three sites in OVX rats ( Figure 1A, Tamoxifen Î EXD, distal femur: p=0.0003, proximal tibia: p=0.0022, lumbar vertebra: p=0.0005; Raloxifene Î EXD, distal femur: p=0.0199, proximal tibia: p=0.0208, lumbar vertebra: p=0.0014). And also, EXD interacted with both SERMs on improving trabecular bone properties at all three sites in OVX rats (Table 3, p<0.05).
EXD could interact with SERMs to serum OCN, but not urinary DPD As shown in gure 3, serum OCN (28.93± 0.77 ng/ml) and urinary DPD (142.12 ± 5.92 nmol/mmol) were signi cantly increased in OVX rats while the treatment of OVX rats with E2 signi cantly restored changes in OCN (14.31 ± 1.09 ng/ml) and DPD (84.92 ± 4.78 nmol/mmol). Treatment with both tamoxifen and raloxifene also signi cantly suppressed the increase in OCN and DPD levels in OVX rats (p<0.001 vs. OVX). However, the treatment with EXD alone signi cantly restored E2 de ciency-induced change in the OCN level (21.13 ± 2.45 ng/ml), but not urinary DPD (109.12 ± 5.08 nmol/mmol) in OVX rats. Co-treatment of EXD and SERMs in OVX rats also markedly reduced both serum OCN and urinary DPD (p<0.01 vs. OVX). Two-way ANOVA analysis suggested no interaction between EXD and tamoxifen while EXD interacted with raloxifene on suppressing serum OCN (raloxifene Î EXD, p=0.0334).

EXD relieved the uterotrophic effect of SERMs
In gure 4A, the mRNA expression of histone H3 was signi cantly increased by E2 (1.36 ± 0.16) and raloxifene (1.39 ± 0.18). The combination of EXD and raloxifene (0.8 ± 0.19) could further suppress raloxifene-induced mRNA expression of H3 in rat uterus. Two-way ANOVA analysis suggested the interaction between EXD and SERMs on suppressing H3 mRNA expression (tamoxifen Î EXD: p=0.0005, raloxifene Î EXD: p=0.0493). In gure 4B, OVX in rats obviously reduced the uterus epithelial cell layer's thickness (red line) compared to the sham group. Treatment of E2 and tamoxifen alone could apparently reverse OVX-induced change of it. Raloxifene alone, EXD alone, and EXD combined with SERMs did not to increase the thickness of the epithelial cell layer in the uterus when compared to the OVX group. Indeed, there was an observable difference in layer's thickness between the group treated with tamoxifen alone and EXD combined with tamoxifen.

EXD interacted with SERMs on ALP activity in human osteosarcoma MG-63 cells
In gure 5A, both crude EXD extract and EXD-Ts signi cantly increased the ALP activity in MG-63 cells in 48 hours (p<0.05 vs. control). Crude EXD extract at the concentrations of 10 μg/ml or above exerted impressive e cacy in increasing ALP activity in MG-63 cells (1.37 ± 0.04), which was more potent than E2 (1.26 ± 0.03).

Discussion
Er Xian decoction (EXD) has been prescribed for the treatment of osteoporosis for decades, possibly via modulating the hypothalamus-pituitary-gonadal (HPG) axis and ER-mediated estrogen signaling. The present study con rmed the anti-osteoporotic activity in OVX rats and human osteoblastic MG-63 cells of EXD. Moreover, we are the rst to address EXD at its clinical dose could interact with two clinically prescribed SERMs (tamoxifen and raloxifene) to attenuate estrogen de ciency-induced bone loss in mature ovariectomized rats and relieving the uterotrophic effect of SERMs. Our study also suggested that the inhibition on FSH and LH were correlated with the osteoprotection of EXD and in vitro study con rmed the direct estrogenic activity of both crude EXD extract and biologically activated EXD and further reported the existence of interaction between biologically activated EXD and SERMs at certain concentrations in human osteosarcoma MG-63 cells.
The dose of EXD, tamoxifen and raloxifene used in the present was selected based on our previous study and converted from its clinical dosages [11], while dosages of tamoxifen and raloxifene used in the present study were chosen based on our previous study in OVX rats [13], which are equivalent to 11 mg/kg.day and 33.3 mg/kg.day in human, respectively. These dosages are close to the clinical dosages of them (10-40 mg/kg.day of tamoxifen for breast cancer treatment, 60mg/kg.day of raloxifene for treatment and prevention of osteoporosis as well as prevention for breast cancer). As expected, EXD and SERMs was effective in attenuating bone loss at three bone sites regarding BMD and bone microarchitecture in mature ovariectomized rats. EXD and SERMs could restore the serum OCN level, a speci c osteoblast product, and is considered a biomarker for bone formation [15] while only SERMs could restore urinary DPD level, a breakdown product of collagen during bone resorption and is a biomarker for bone resorption in OVX rats [13]. Afterward, we demonstrated that EXD could interact with SERMs to exert bone protection. This interaction could be explained by the similar underlying mechanisms of EXD and SERMs mediated by ERs. Estrogen receptors α (ER-α) have been detected in osteoblasts, osteocytes, and osteoclasts that mediate the direct effects of estrogen, resulting in decreased bone resorption and formation activity [16]. Previously, our group [17] and others [18] address that EXD, tamoxifen, and raloxifene could activate estrogenic signaling, exciting osteoblastic activity of osteoblasts.
Therefore, it is possible that they could interact with each other in bone by sharing the similar mechanisms. Further study is needed to investigate the estrogenic pathway which EXD interact with SERMs in bone remodeling.
Our study was the rst to compare the bone protective activities between EXD and SERMs and found that EXD at its clinical dose exerted a similar but weaker bone protective effect than SERMs in OVX rats. Although EXD and SERMs could activate ER-α, their binding a nity to ER-α might illustrate their bone protection effectiveness.
EXD is a mixture of phytoestrogen or phytoestrogen-like natural compounds. For example, HEP, as one of the principal herbs in EXD, has been reported to possess phytoestrogens (e.g., icariin, icaritin, and baohuoside) and act via ERs [19]. Phytoestrogens, as alternatives to estrogen, also exert estrogen-like activities, possibly via their direct but weak a nity for ERs [20] when compared to SERMs. The weaker estrogenic bone protection of EXD might be a possible reason that EXD did not alter SERMs' action in bone.
During menopause, the decrease in E2 level and increase in FSH level in serum accounts for bone loss [21,22].
Thus, we studied the hormonal regulation by EXD and SERMs in OVX rats, mimicking the endocrinological condition in menopause. We demonstrated EXD, SERMs, and their combination could dramatically reverse OVXinduced suppression of E2 level and upregulation of the FSH level in rats. These results are consistent with previous studies EXD was found to promote the E2 synthesis through stimulating aromatase activity in the ovary, liver, and fats as well as decrease FHS production by hypothalamic-pituitary axis. It is anticipated that the increase of E2 level by EXD is resulted from the promoted synthesis in liver and fats [7]. Also, raloxifene and E2 was proved to decrease the FSH level in postmenopausal women, while tamoxifen [23] was reported to increase the E2 level in women. These might explain the interactive regulation of hormone level in OVX rats after treatment of EXD with SERMs in the present study. Most importantly, these treatments' hormonal changes likely contributed to the bone protection of EXD and SERMs. A recent study reports that FSH could directly stimulate osteoclastogenesis and bone resorption, and its circulating level is associated with changes in bone turnover biomarkers in postmenopausal women [24]. Besides, E2 could govern female bone remodeling through ERs [21,22]. In agreement with the report, a dramatic elevation was also observed in the present study in OVX rats in which the changes in BMD and bone turnover markers appeared to be associated with the changes in circulating FSH and E2 in OVX rats. We also showed that the BMD in rats was positively correlated with serum E2 and negatively correlated with serum FSH and LH level. In sum, our results might suggest that the modulation of HPG axis, including stimulating E2 and suppressing FSH, might be involved in mediating the in vivo bone protective actions of EXD and SERMs. Further study is needed to systemically review the regulation of brain-bone axis by EXD and SERMs.
As EXD could increase the circulating E2 in OVX rats, we studied the interaction between EXD and SERMs in the uterus, one of the sensitive target tissues to estrogen and the tissue that the side effects of HRT and SERMs are frequently reported [3,12]. Indeed, our study con rmed the stimulatory effects of E2 and tamoxifen on the uterus indicated by the signi cantly increased uterus weight and uterine epithelial cell growth in OVX rats while E2 and raloxifene increase a uterine proliferation marker histone H3 mRNA expression in rat uterus. In contrast, EXD did not stimulate the growth of the uterus and histone H3 expression suggesting a tissue-selective action of EXD. Interestingly, EXD was shown to relieve the side effects of tamoxifen in the uterus which might be resorted to its direct interaction with ER or indirect modulation of the E2 level. Our results agree with the previous studies that tamoxifen causes endometrium cancer via modulating endogenous E2 level and being an ER-α agonist promoting the uterus's growth [25]. EXD and its abundant iso avone, icariin [26] were found to maintain a healthy epithelial layer of uterus in OVX rats, probably via suppressing the protein expression of cancer-promoting ER-α and increasing the protein expression of cancer-inhibiting ER-beta (ER-β) in uterus. Also, a phenolic glycoside (curculigoside) [27], and an alkaloids (berberine) [28] in EXD, were found to improve histopathological lesion of uterus in mice and suppress the growth of endometrium cancer, respectively.
Despite tamoxifen could stimulate the growth of the uterus via ER-α, EXD and its phytochemical seems to counteract it through targeting the same responder and regulating uterine growth. Even though EXD increased the E2 level, phytochemicals in EXD (icariin, ferulic acid and berberine) might exert estrogen-like activities as phytoestrogens, which have been reported to facilitate the clearance of estrogens from local tissues, such as the uterus and breast tissue, and catabolize the estrogens to more benign 2-hydroxylated metabolites, regulating the E2 level in local tissue [29]. Therefore, it is possible that EXD might enhance estrogens' clearance from reproductive tissues and (or) the catalysis of E2, thereby reducing the potential side effect of tamoxifen in the uterus. Taken together, EXD probably reduce the uterotrophic effect of tamoxifen via modulation of ERs and local E2 level.
In vitro study con rmed the estrogen-like activities of crude EXD extract in promoting ALP activity and EREdependent transcriptional activities in human osteosarcoma MG-63 cells. As mentioned above, EXD is a complex mixture of chemicals that might not be completely bioavailable, and its estrogenic actions might require metabolic activation in vivo. To observe the direct estrogenic activities of EXD and possible interactions between SERMs, biologically activated EXD in the form of serum derived from EXD-treated OVX rats were also applied to MG-63 cells. Additional le 4 shows the presence of major polyphenol and alkaloid groups in EXDtreated serum extract and the EXD crude extract, suggesting that TCMs extract and their metabolites were being absorbed and transported in rat circulation. Our results showed that EXD-Ts dose-dependently induced ALP activity while the stimulatory effect seemed less potent than that of crude EXD extract. Moreover, unlike crude EXD extract, EXD-Ts at 10 -3 dilution did not stimulate ERE -dependent transcriptional activities. These results suggest that the estrogenic activity of biologically activated EXD was different from that of crude EXD extract.
Two-way ANOVA analysis indicated that EXD-Ts interacted with SERMs (tamoxifen at certain concentrations and raloxifene at all concentrations applied) and altered the effects of SERMs at lower concentrations. Our previous study showed that serums obtained from these TCM-treated animals were too dilute, and the major markers and metabolites of TCM were undetectable. Therefore, EXD at high dosage, which was ten times to its clinical dose, was used for preparing biologically activated EXD for use in the in vitro studies. Our observation that biologically activated EXD could alter the effects of SERMs in MG-63 cell, but not in OVX rats, are likely due to the high dosages of EXD being used for harvesting EXD-treated serum. Thus, it is possible that upon metabolic activation and clearance, the level of phytoestrogens and their metabolites in EXD will not be high enough to compete with SERMs for binding towards ERs. Future studies will be needed to characterize major phytoestrogens' tissue distribution in OVX rats upon long-term treatment with EXD.
As summarized in gure 6, bone protective TCMs formula EXD at its clinical equivalent dose selectively exert estrogenic actions in bone without causing undesirable effects in the uterus. EXD interacts with SERMs while co-treatment with EXD does not signi cantly suppress the responses of bone tissue to SERMs in OVX rats and tends to relieve the side effect of tamoxifen in the uterus. Crude EXD extract, but not EXD-treated serum, directly increases ERE activities in human osteoblastic MG-63 cells, indicating actions of crude EXD extract might be different from metabolically activated EXD. The present study provided evident that the combined use of EXD and SERMs could be a modi ed therapeutic approach to treat postmenopausal osteoporosis while relieving the side effect of SERMs.

Conclusions
To our knowledge, this is the rst study to examine the drug-herb interaction between EXD and SERMs using an

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.     followed by Tukey's test for post hoc comparison. Interactions between EXD and SERMs were determined by Two-way ANOVA followed by Bonferroni test as a post-test. ***p<0.001 vs sham; ^^p<0.01 vs OVX, ^^^p<0.001 vs OVX. Correlation analyses were performed to study the correlation between E2 (D), FSH (E) or LH (F) with BMD at tibia of rats. Pearson's correlation coe cients (r) and p values were shown.

Figure 3
Estrogenic effects of EXD, SERMs, and their combinations on bone turnover biomarkers in ovariectomized rats Upon treatment, serum and urine samples were collected. Serum level of osteocalcin (OCN) and urinary deoxypyridinoline (DPD) were measured by ELISA kits by following manufacturers' instruction. A. Serum level of OCN; B. Urinary level of DPD. Data was expressed as mean ± SEM. n=5 to 12. Differences between groups were determined by one-way ANOVA followed by Tukey's test for post hoc comparison. Interactions between EXD and SERMs were determined by Two-way ANOVA followed by Bonferroni test as a post-test. ***p<0.001 vs sham; ^p<0.05, ^^p<0.01, ^^^p<0.001 vs OVX.

Figure 4
Uterotrophic effect of SERMs and EXD in ovariectomized rats. Uterus of OVX rats were isolated from OVX rats.  Estrogenic effects of EXD, SERMs and their combinations on ALP and ERE-luciferase activity in vitro Human osteosarcoma MG-63 cells were routinely cultured and treated with EXD (crude EXD extract and EXD-treated serum), tamoxifen, raloxifene and their combinations for 48 hr. Upon treatment, ALP activity and estrogen response element [12] luciferase activity were measured by ALP assay and Dual Luciferase® Reporter Assay System, respectively. Results were expressed as a ratio to control. n=3 or more. Differences between groups were determined by independent t-test. Interactions between EXD and SERMs were determined by Two-way ANOVA followed by Bonferroni test as a post-test. *p<0.05, **p<0.01, ***p<0.001 vs control; ^p<0.05 vs SERM alone; #p<0.05 vs EXD-Ts alone.

Figure 6
Bone protective activity of EXD alone and in combination with SERMs In vivo study suggests that EXD promotes bone formation by regulating the release of E2 and FSH and suppressing bone resorption. In vitro study indicates that crude EXD extract or SERMs could increase ALP activity in osteoblasts via modulating ERE while biologically activated EXD promotes ALP activity without ERE activation. Both in vivo and in vitro studies demonstrate that EXD might interact with SERMs but do not alter the bone protective activity of SERMs.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Supplementary20201028.docx