Study of pharmacokinetics of prenylflavonoids and dynamics of estrogen action in sera following ingestion of epimedium using validated, ultra sensitive cell based bioassays 5

214
CHAPTER 5
CONCLUSION

5.1
Significance of findings
215
5.2
Limitations
218
5.3
Will Epimedium exert any beneficial effects in-vivo?
223
5.4
Future work
226


215
5.1 Significance of findings
This work and the bioassay tools developed in this thesis can serve as a model
framework for future projects which focus on development of estrogenic botanicals as
drugs for menopause, maintenance of bone health and other conditions requiring
estrogenic action in humans, from the lab bench to bedside. This thesis reported the
importance and value of the development and validation of appropriate tools with high
throughput, in this case, were cell-based assays that were needed to rapidly track and
measure global estrogenic activity in complex mixtures, from herbal preparations,
formulations to ex-vivo serum samples, which contain a myriad of bioactive compounds,
known and unknown. These bioassays can also be used to rapidly screen and select
compounds with optimal ERα- or ERβ-selective properties for estrogen replacement

therapy that has lower adverse effects such as breast cancer cell growth, thereby reducing
the number of expensive, larger-scale studies with negative clinical outcomes. Although a
compound or preparation is shown to be estrogenic when tested directly in the bioassays,
it cannot be assumed that it would exert similar activity when consumed.
Rigorous pharmacokinetics and pharmacodynamics studies on complex botanical
extracts are have been seriously constrained by the lack of suitable tools that can
simultaneously track serum levels and bioactivities of the many compounds, known and
unknown, that exert pharmacological effects in both animal models and humans. The
development of these cell-based assays enabled new insights into the pharmacokinetic
and pharmacodynamic interactions underlying the biological effects of a botanical
preparation like Epimedium. Two notable experiments were reported in this thesis include
a clinical trial where human volunteers ingested a traditional aqueous preparation of
Epimedium pubescens and another female ovariectomized rats being fed with an enriched
extract made from Epimedium brevicornu. Although they are not directly comparable due

216
to differences in their experimental design, it was found that only trace amounts of
unconjugated prenylflavonoids were absorbed into the bloodstream. From the rat study, it
was found that the vast majority of prenylflavonoids absorbed were conjugated which
were rendered relatively non-estrogenic during first pass metabolism. The appearance of
total amount of bioactive prenylflavonoids in sera corresponded broadly with estrogenic
activity that was measured after treatment of serum samples with β-glucuronidase and
sulfatase. Non-linear pharmacokinetics and prolonged effects for up to 72 h following
administration of a single dose of Epimedium extract were also observed. Coupled
processes of enteric and enterohepatic recycling may allow different prenylflavonoids to
be reabsorbed which resulted in longer than expected apparent plasma half-lives for some
Epimedium compounds and their conjugates. β-glucuronidases present in many tissues
such as bone, brain and mammary glands may convert conjugated prenylflavonoids in
these tissues into their bioactive form if they become deconjugated.
Gene expression profiling revealed the ability of estrogenic Epimedium

prenylflavonoids to induce the transcription of CYP1A1, which is a gene mediated by
AhR. AhR has been regarded as an inducible transcription factor that controls the
expression of enzymes that metabolize potentially dangerous xenobiotic chemicals.
However, there are AhR ligands, such as icaritin found in Epimedium, which have been
reported to be able to initiate the degradation of the ER and suppress estrogen signaling
and proliferation of breast cancer cells. By mediating with these pathways, these
beneficial AhR ligands can be used to influence breast cancer initiation and progression.
The ability of Epimedium compounds to act as dual activators for ERα and AhR
mechanistically represent a new class of SERMs. These compounds should be further
examined for development of a new family of drugs that can be co-administered during
hormone replacement therapy to reduce associated breast cancer risk.

217
Lastly, it is important to take note of the so m e experimental design pit-falls
presented in this thesis which were discussed at length in Section 5.2 – Limitations. Due
to a range of differences in experimental design, the results from the clinical trial where
human volunteers ingested a traditional aqueous preparation of Epimedium pubescens and
another where rats were fed with an enriched extract made from Epimedium brevicornu
are best interpreted in isolation as they were not easily and cannot be directly compared.
Two major differences in experimental design include the use of different botanical
extracts and model organisms for the two studies. Notably, it is vital to ensure consistency
in the range of parameters that are to be measured and conditions that need to be
employed throughout the project. Lessons learnt from these two studies would greatly
enhance the design and comparison of future similar studies on Epimedium and similar
herbal drugs. A good starting point for a pharmacokinetic and pharmacodynamic study on
a similar estrogenic botanical preparation would be to refer to the rat experiment reported
in this thesis which used the appropriate animal model, that is, female ovariectomized rats,
which were fed separately with increasing doses of an enriched extract made from
Epimedium brevicornu and treated over a period of 72 hours, with an inclusion of a
positive control of rats fed with a standard estrogen prodrug, such as, estradiol benzoate.

Chemical analytical methods coupled with cell-based bioassays that can measure global
estrogenicity of sera of animals were employed to understand the levels and bioactivity of
conjugated and unconjugated active compounds respectively.



218
5.2 Limitations
The first study, which involved a human clinical trial, was a pilot project to
validate the cell-based bioassays in a clinical setting. These bioassays were developed for
use to m e asure global activity in serum and to evaluate the effects of ligands in complex
mixtures on ERα, ERβ bioactivity and MCF-7 breast cancer cell growth. In this human
study, healthy male subjects, instead of females, were enrolled to reduce interference
from endogenously estrogens encountered in the latter group, which may mask the effects
of administered estrogenic drugs. The subjects were administered separately with
estradiol valerate or Epimedium pubescens decoction. Serum samples were obtained and
assayed ex-vivo for levels of estrone and estradiol by tandem mass spectrometry, for ERα
and ERβ bioactivity and MCF-7 breast cancer cell proliferative effects. The
concentrations of icaritin and desmethylicaritin and bioactivity in sera after ingestion of
Epimedium pubescens decoction by human subjects were also measured as they are
prenylflavonoids unique to Epimedium which exerted stronger estrogenic activities than
their glycosides and flavonoids such as apigenin, kaempferol, luteolin and quercetin.
In the view of the low bioactivity measured in a traditionally prepared, aqueous
extract, a brand new study was done in collaboration with Dr. Willmar Schwabe
Pharmaceuticals (Karlsruhe, Germany). A new standardized, prenylflavonoid-enriched
extract based on Epimedium brevicornu was formulated and fed in three increasing doses
to female ovariectomized Sprague–Dawley rats, with an additional group of rats which
are fed with estradiol benzoate included that served as the positive control. The use of
female ovariectomized Sprague–Dawley rats is a more appropriate model for testing
drugs for use to enhance post-menopausal bone health.

219
In contrast to the earlier study, the serum bioactivity and concentrations of both
unconjugated and conjugated levels of ingested prenylflavonids were monitored. The
levels of conjugated prenylflavonoids were included to see whether lack of estrogenic
activity of Epimedium was due to conjugation of active compounds during first pass
metabolism. The quantification of prenylflavonoid glycosides, such as, icariin, icariside I
and icariside II was also undertaken to understand the metabolism of Epimedium
glycosides.
Although one may want to compare whether the prenylflavonoid-enriched extract
had greater bioavailability compared to the aqueous decoction due to higher amounts of
compounds present, meaningful comparison is hampered by the lack of consistency in the
design of the two experiments. Table 22 lists these differences.

Table 22: Methodological differences present in the human and rat studies.
Reference
Li et al., 2009
Wong et al., 2009
Model organism
Human males
Ovariectomized female rats
Sample size
8
4 rats per time-point
Herb species
Epimedium pubescens
Epimedium brevicornu
Extraction method
Water
70% EtOH

Dose(s)
mg/kg body weight
90.1
100, 300 & 600
Study Duration (h)
48
72
Compounds
analyzed
Unconjugated

icaritin & desmethylicaritin
Unconjugated

Icariin, icariside I, icariside II,
icaritin, desmethylicaritin

Total

Icariside II, icaritin,
desmethylicaritin


220
In the human clinical trial, male human subjects were recruited. Male subjects,
instead of females, were used so as to reduce interference from endogenously produced
estrogens, which may mask the effects of administered estrogenic drugs. With the
intention to develop Epimedium as a botanical drug for use to maintain bone health after
menopause, the more ideal candidates to use for such a study would be post-menopausal
women as they will be a more relevant physiological model. Differences in metabolism

and plasma transport of steroids between men and women would likely to also influence
the results of the study. Results obtained from this human study may not be appropriate
for comparison with those obtained from ovariectomized Sprague–Dawley rat study due
to species differences in the model organism used. Both studies also suffered from sm all
sample sizes as shown in Table 22. Blood at the various time-points was taken from the
s am e human subjects over time whereas four rats were sacrificed at each time interval.
Inter-individual differences may affect the results of both studies more significantly due
to small sample sizes used.
Further difficulties arise when different species of Epimedium and extraction
methods employed in the two studies, which result in two vastly different herbal extracts.
From chemical profiles obtained from quantitative analysis via chromatographic tools
used in this study, the dried residue from the aqueous decoction made from Epimedium
pubescens was found to contain 2 mg/g icariin, 0.119 mg/g icaritin and 0.031 mg/g
desmethylicaritin. The amounts of icariside I and icariside II were not determined then.
The enriched Epimedium brevicornu extract generally comprised more prenylflavonoids,
especially in terms of icariin but it had a lesser amount of icaritin - 143.3 mg/g icariin,
0.009 mg/g icaritin and 0.122 mg/g desmethylicaritin. The amounts of icariside I and
icarside II were determined to be 0.157 mg/g and 18.68 mg/g respectively. The increased

221
amount of prenylflavonoids in the latter extract was due to the increased solubility of the
compounds in 70% ethanol that was used for extraction.

Table 23: Comparison of extract dosages and amounts of prenylflavonoids ingested
from traditionally prepared aqueous Epimedium pubescens decoction and
prenylflavonoid-enriched extract from Epimedium brevicornu.
Preparation
Aqueous
Epimedium pubescens
decoction

1
Prenylflavonoid-enriched
Epimedium brevicornu extract
(70% ethanol extracted)
Extract dosage
2
90.1
100
300
600
Icariin
2

0.180
14.3
42.9
85.8
Icariside I
2

Not determined
0.0157
0.0471
0.0942
Icariside II
2

Not determined
1.9
5.6

11.2
Icaritin
2

0.0107
0.0009
0.0027
0.0054
Desmethylicaritin
2

0.00279
0.0122
0.0366
0.0732

1
Decoction was prepared from 50 g of dried leaves from Epimedium pubescens. Yield after extraction was
12.6% and mean weight of human subjects was 70 kg.

2
Amount of extract and compounds ingested are expressed in mg / kg body weight.

On a dry weight basis, the lowest dose of Epimedium brevicornu employed in the
rat study (100 mg/kg) closely approximates the amount of Epimedium pubescens extract
(90.1 mg/kg) that was ingested by human subjects (Table 23). As a result of the use of
different extraction methods, the profiles of prenylflavonoids ingested were markedly
different, especially for icariin, which was nearly 80 times more abundant in the
prenylflavonoid-enriched Epimedium brevicornu extract (see Table 23). With reference
to the pharmacokinetic data from both studies, unconjugated desmethylicaritin was not

detected in human sera in all time-points whereas it could be detected in low levels in sera
from rats that were fed with Epimedium extract at a dose of 100 mg/kg. The in-vivo

222
hydrolysis of glycosides like icariin can lead to the production and absorption of
aglycones like icaritin and desmethylicaritin into the blood stream and this could have
occurred more significantly rats which were fed with the Epimedium brevicornu extract
due to a greater amount of icariin and similar precursor glycosides that were present. As a
result of such differences in starting composition of constituents, the bioavailability of
Epimedium compounds cannot be accurately ascertained from the results from both
studies.
The duration of two studies and the number of metabolites that were examined
also differed. The rat study reported the levels of total amounts of prenylflavonoids
absorbed (sum of conjugated and unconjugated metabolites) and a depot effect was
observed as the duration of study was prolonged to 72 h. On the other hand, the human
study ended at 48 h and the total amounts of prenylflavonoids absorbed were not
determined. It is also not certain whether a similar depot effect would occur in humans.
The results from both studies are hence best interpreted in isolation due to the
differences present in their design as discussed above. Lessons learnt from these two
studies would greatly enhance the design and comparison of future similar studies on
Epimedium and similar herbal drugs.


223
5.3 Will Epimedium exert any beneficial effects in-vivo?
Sera samples from human subjects who ingested the aqueous decoction contained
low amounts of unconjugated icaritin and no desmethylicaritin. The dose was likely to be
too low and any absorbed aglycones would have been conjugated during extensive first
pass metabolism. Any unconjugated compounds would be present in trace levels and
become bound tightly to albumin in sera during circulation. This could be the reason why

estrogenicity of sera was not detected in human serum samples as levels of unconjugated
prenylflavonoids that w ere present in nanomolar quantities that are well below the
detection limits of the panel of cell-based bioassays (see dose-response curves in Fig. 15).
Although the amount of prenylflavonoids was enriched in the alcoholic
Epimedium brevicornu extract that was fed in three increasing doses in the rat study, due
to extensive first pass metabolism, the peak serum levels of both unconjugated aglycones,
namely, icaritin and desmethylicaritin, as well as, two estrogenic monoglucosides,
icariside I and icariside II, were in the low nanomolar levels, even at the highest dose of
Epimedium pubescens extract in rats. The estrogenicity of rat serum samples for 100 and
200 mg/kg dosages were also not significant when measured the ERα cell-based bioassay.
Interestingly, estrogenicity equivalent to 10 pM of estradiol was detected in sera at the 8 h
time-point from rats fed with the highest dose of Epimedium brevicornu extract (600
mg/kg dose) (Fig. 39) which was likely to be conferred by estrogenic metabolites not
studied in this work. This is indicative that Epimedium can exert estrogenicity in-vivo and
future work entails optimization of the human equivalent dosage of extract that need to be
taken to ensure efficacy.
Although the results for a single dose of Epimedium extract in this study may
imply that the preparations would not exert significant estrogenicity in-vivo due to the low

224
amounts of unconjugated prenylflavonoids in circulation, it is important to note that in
Traditional Chinese Medicine, medicinal herbs are rarely taken via a single dose. In the
Bencao Gangmu (本草纲目), a Chinese materia medica work written by Li Shizhen (李
时珍) in the Ming dynasty, Epimedium formulations are largely noted to require repeated
dosings (Li, circa 1500). As a result of repeated dosings, Epimedium prenylflavonoids
and their conjugates may be accumulated in the body over time and these may potentially
be bioactive if they reach sufficiently high levels. Conjugates can get hydrolyzed into
bioactive aglycones in target organs and tissues (Penza et al., 2007).
Epimedium is used as part of complicated formulations with other herbs that
include rehmania root, curculigo rhizome and dogwood and wolfberry fruits. In this study,

Epimedium is administered as a single herb preparation and the bioavailability of
prenylflavonoids in it may be enhanced when taken in combination with phytochemicals
from other plants. Lambert et al. (2004) reported the co-treatment with piperine from
black pepper, enhanced the bioavailability of (-)-epigallocatechin-3-gallate found in green
tea in mice. Moon & Morris (2007) reported that the co-administration of quercetin and
(-)-epigallocatechin-3-gallate significantly increased the biochanin A area under the
plasma concentration versus time curve in rats in both intravenous and oral administration
of biochanin A.
According to the ancient text, Shen Nong Ben Cao Jing (神农本草经), Epimedium
is documented to strengthen the bones and tendons, rheumatic conditions, impotence,
seminal emission, weakness of the limbs, rheumatoid arthralgia with numbness and
muscle contracture and climacteric hypertension. In particular, much work has been done
on the effects of Epimedium on bone health and these have been reviewed in the
Introduction of this thesis. Like many flavonoids, prenylflavonoids in Epimedium that

225
have been absorbed after ingestion are largely present in the conjugated form in
circulation as observed in the rat study. The conjugated forms that are present in large
amounts in circulation, although rendered non-estrogenic, may still be able to exert
beneficial effects on the bone. This is supported by the observation that hesperetin-7-O-
glucuronide that was reported to enhance primary rat osteoblast proliferation and
differentiation, and down-regulated the expression of receptor activator of nuclear factor-
kappa B with no change in osteoprotegerin expression (Trzeciakiewicz et al., 2010).
Hesperetin-7-O-glucuronide may be able to limit the activation of osteoclasts, similar to
what have been reported for isoflavones such as daidzein and genistein (Chen et al., 2002).
Osteoblasts express receptor activator of nuclear factor-kappa B ligand and
osteoprotegerin. In this case, osteoprotegerin can act as a decoy receptor for receptor
activator of nuclear factor-kappa B ligand which then neutralizes its function in
osteoclastogenesis. Hesperetin, like prenylflavonoids in Epimedium, is conjugated mainly
into glucuronides when ingested.

226
5.4 Future work

In this study, ultra-sensitive, high throughput, cell-based bioassays have been
developed that can be used to distinguish the ERα - or ERβ-selective properties, as well as,
cell proliferative effects on MCF-7 breast cancer cells of pure compounds and for
measurement of global estrogenicity of complex mixtures. To enable the study of the
whole spectrum of SERM activities of Epimedium compounds on critical estrogen-
responsive tissues such as bone, endometrium and breast, the next step is to develop and
validate a panel of such cell-based bioassays. Even within the same cell line, differences
exist in the ability of different compounds to induce estrogen-regulated gene expression
as seen in Ishikawa cells where different levels of cofactors were induced when exposed
to tamoxifen, ICI 182,780 and LY 117,018 (Bramlett & Burris, 2003). This can be done
by introducing natural ERα or ERβ-driven promoters in bone, endometrium, and breast
cell lines to obtain promoter specific markers of SERM activity. Examinations of various
end-points in these cells are necessary to fully characterize the activity of compounds that
may be SERMs.
With the intention to develop Epimedium as a botanical drug for use to maintain
bone health after menopause, future work can be built on the relatively well-designed
study which feature female ovariectomized Sprague–Dawley rats that were fed with
increasingly doses of a standardized Epimedium brevicornu extract. This model organism
is a cheaper, more appropriate and relevant physiological model for evaluating
Epimedium as a drug that can be ingested to enhance post-menopausal bone health than to
enroll post-menopausal females at this initial stage of research.
Future investigation would involve the use of analytical methods to quantify
Epimedium prenylflavonoids and bioassays can also be used to determine if these

227

compounds accumulate in target tissues such as bone of female ovariectomized Sprague–
Dawley rats that were fed with a single dose and repeated doses standardized Epimedium
brevicornu extract and whether these conjugated compounds can be deconjugated and
become bioactive. A similar study has been published on hesperitin, the aglycone of
hesperidin, found in citrus, was reported to be unequally distributed and accumulated
mostly in the aorta of rats that were orally fed with a 0.2% of the aglycone for 4 weeks
(Takumi et al., 2011). This observation is in good agreement with published work that
reported hesperitin had the ability to strengthen blood vessels and prevent capillary
permeability which also supported the idea that hesperitin effectively accumulates in the
aorta to enhance vascular function.
At the same time, to investigate the preventive effect of Epimedium on
osteoporosis induced by ovariectomy in rats, female ovariectomized Sprague–Dawley rats
after the induction of osteoporosis are divided into three groups, each being fed with the
standardized Epimedium extract, estradiol or the negative control. To analyze the
osteoprotective effects of the various treatments, bone mineral density, bone turnover
biochemical markers, serum estrogen and prenylflavonoid levels and bioactivity, and
uterine wet weight will be monitored to assess whether the administration of standardized
Epimedium extract could prevent bone loss induced by estrogen deficiency.

The last
parameter will provide insights on whether Epimedium can exert any beneficial effect in
preventing bone loss without resulting in hyperplasia of the uterus.
Another worthwhile work to embark on is to ascertain whether the ingestion of
estrogenic prenylflavonoids from Epimedium for the purpose of maintaining bone health
can increase risk of ERα-positive breast cancer. Interestingly, icaritin, a prenylflavonoid
found in Epimedium, stimulated the growth of MCF-7 breast cancer cells at low
concentrations like an estrogenic ligand on its own and exerted a growth suppressive

228
effect on estrogen-stimulated breast cancer cell proliferation at higher concentrations

(Tiong, 2010). Similar to estradiol, icaritin was also found to dose-dependently destabilize
ERα protein (Tiong, 2010). Results for microarray gene expression analyses reported in
this thesis implicated the aryl hydrocarbon receptor (AhR) signaling for this suppressive
effect of icaritin. To find out whether the ingestion of estrogenic prenylflavonoids from
Epimedium can increase risk of ERα-positive breast cancer, the athymic nude mouse
model can be used where the mouse can be implanted with MCF-7 breast cancer cells and
estradiol given to stimulate their growth. Icaritin will be administered to see if the
compound can restrict estradiol-stimulated breast cancer xenograft growth and whether
there is any reduced ERα protein levels and related signaling. This work will facilitate the
development of dual agonists like icaritin which are estrogenic but yet, through activating
AhR-signaling, can destabilize ERα protein to restrict ERα-positive breast cancer cell
growth.
Data obtained from proposed future work above will provide further insights that
would be useful for the development of Epimedium extract and prenylflavonoids as drugs
for menopause, maintenance of bone health and other conditions requiring estrogenic
action in humans.