Potent Inhibition of Human Phosphodiesterase-5 by Icariin Derivatives

Abstract

Plant extracts traditionally used for male impotence (Tribulus terrestris, Ferula hermonis, Epimedium brevicornum, Cinnamomum cassia), and the individual compounds cinnamaldehyde, ferutinin, and icariin, were screened against phosphodiesterase-5A1 (PDE5A1) activity. Human recombinant PDE5A1 was used as the enzyme source. Only E. brevicornum extract (80% inhibition at 50 microg/mL) and its active principle icariin (1) (IC50 5.9 microM) were active. To improve its inhibitory activity, 1 was subjected to various structural modifications. Thus, 3,7-bis(2-hydroxyethyl)icaritin (5), where both sugars in 1 were replaced with hydroxyethyl residues, potently inhibited PDE5A1 with an IC50 very close to that of sildenafil (IC50 75 vs 74 nM). Thus, 5 was 80 times more potent than 1, and its selectivity versus phosphodiesterase-6 (PDE6) and cyclic adenosine monophosphate-phosphodiesterase (cAMP-PDE) was much higher in comparison with sildenafil. The improved pharmacodynamic profile and lack of cytotoxicity on human fibroblasts make compound 5 a promising candidate for further development.

Full text

Potent Inhibition of Human Phosphodiesterase-5 by Icariin Derivatives
Mario Dell’Agli,*
,†
Germana V. Galli,
†
Esther Dal Cero,
†
Federica Belluti,
‡
Riccardo Matera,
§
Elisa Zironi,
§
Giampiero Pagliuca,
§
and Enrica Bosisio
†
Department of Pharmacological Sciences, UniVersity of Milan, Via Balzaretti 9, 20133, Milan, Italy, Department of Pharmaceutical Sciences,
UniVersity of Bologna, Via Belmeloro 6, 40126, Bologna, Italy, and Department of Veterinary Public Health and Animal Pathology, UniVersity
of Bologna, Via Tolara di Sopra 30, 40064 Ozzano Emilia (BO), Italy
ReceiVed January 23, 2008
Plant extracts traditionally used for male impotence (Tribulus terrestris, Ferula hermonis, Epimedium breVicornum,
Cinnamomum cassia), and the individual compounds cinnamaldehyde, ferutinin, and icariin, were screened against
phosphodiesterase-5A1 (PDE5A1) activity. Human recombinant PDE5A1 was used as the enzyme source. Only E.
breVicornum extract (80% inhibition at 50 µg/mL) and its active principle icariin (1) (IC50 5.9 µM) were active. To
improve its inhibitory activity, 1was subjected to various structural modifications. Thus, 3,7-bis(2-hydroxyethyl)icaritin
(5), where both sugars in 1were replaced with hydroxyethyl residues, potently inhibited PDE5A1 with an IC50 very
close to that of sildenafil (IC50 75 vs 74 nM). Thus, 5was 80 times more potent than 1, and its selectivity versus
phosphodiesterase-6 (PDE6) and cyclic adenosine monophosphate-phosphodiesterase (cAMP-PDE) was much higher
in comparison with sildenafil. The improved pharmacodynamic profile and lack of cytotoxicity on human fibroblasts
make compound 5a promising candidate for further development.
The inability to achieve or maintain an erection sufficient for
satisfactory sexual function is an increasing problem with a
considerable impact on interpersonal relationships and quality of
life for men.
1
During erection, nitric oxide is released from the
axon terminals of the parasympathetic nerves and diffuses into
smooth muscle cells of the arterial walls of the corpus cavernosum.
The consequent activation of guanyl cyclase, converting guanosine
triphosphate (GTP) into cyclic guanosine monophosphate (cGMP),
causes smooth muscle relaxation, leading to dilation and increased
influx of blood into the penile tissue. The trapping of blood in the
penis results in an erection.
2
Selective inhibitors of cGMP-
phosphodiesterase-5 (PDE5) such as sildenafil (Viagra), tadalafil,
and vardenafil are currently used for erectile dysfunction (ED).
However, several adverse effects have been recorded in clinical
trials, including priapism and visual disturbances.
3
Furthermore,
therapy with PDE5 inhibitors is cost-effective. Thus, the search
for new compounds of this type for drug development could be
worthwhile. A variety of natural plant products, including berberine,
forskolin, papaverine, and yohimbine, are claimed to be useful for
improving sexual performance. Extracts from Lepidium meyenii
Walp. (maca), Panax ginseng C.A. Meyer, Ginkgo biloba L., Ferula
hermonis Boiss., and many other herbal remedies, alone or in
combination, have been promoted for the treatment of sexual
problems.
4,5
With the aim of looking for new leads for selective
PDE5 inhibitors, plant extracts and their putative active principles
were selected for screening against human PDE5 activity in vitro.
Our attention focused on Tribulus terrestris L., Ferula hermonis,
Epimedium breVicornum Maxim., and Cinnamomum cassia L.,
since these extracts are claimed traditionally to improve sexual
performance. T. terrestris caused vasodilating and antihypertensive
effects in rats
6
and a pro-erectile effect on the rabbit corpus
cavernosum;
7
F. hermonis has been studied for its effects on sexual
behavior in male and female rats;
8,9
C. cassia and “Epimedii Herba”
are components of Chinese herbal products patented for the
treatment of sexual dysfunction.
10-14
“Epimedii herba” is the
common name for the dried aerial parts of E. breVicornum,E.
sagittatum Maxim., or E. koreanum Nakai, collected in the
summer.
15,16
Among the extracts, only “Epimedii Herba” was active
against PDE5A1, for which the presence of icariin (1), the major
pharmacologically active constituent,
17-19
was considered a lead
compound for chemical modifications in order to improve inhibitory
activity. Modifications applied at the hydroxyl groups at C-3, C-7,
and C-8 included partial or complete removal of the sugar moieties,
partial or complete sugar replacement with a hydroxyethyl residue,
and cyclization of the prenyl group (Scheme 1). All compounds
produced (1-6) were tested for PDE5A1 activity. Also, selectivity
versus human retina PDE6C and human platelet cAMP-PDE, and
cytotoxicity on human fibroblasts were investigated.
Results and Discussion
The activity of plant extracts and individual compounds against
human recombinant PDE5A1 is shown in Figure S1 (Supporting
Information). Cinnamaldehyde, icariin (1), and ferutinin were
considered as the putative active principles of C. cassia,E.
breVicornum, and F. hermonis, respectively, since the compounds
represent the most abundant secondary metabolites of those species.
Only E. breVicornum and icariin (1) strongly inhibited PDE5A1
(-80% and -72%, respectively), whereas the other test materials
were much less active (-15 to -23%). Inhibition by cinnamalde-
hyde (-16%) and ferutinin (-7%) was not significant. The
medicinal plants tested in the present study had a reputation for
aphrodisiac effects and therefore represented the start of a screening
program to search for compounds to be developed as a new natural
drug alternative to sildenafil. The observation that only E. breVi-
cornum and its active principle 1inhibited PDE5 in a significant
manner, in agreement with previous results,
17-20
suggests that the
other plant extracts may interfere with erectile function through
mechanisms other than PDE5 inhibition.
Compound 1was a good PDE5 inhibitor (IC50 of 5.9 µM), but
required improvement in order to have equivalent potency to
sildenafil, which gave an IC50 of 75 nM. The inhibitory effects of
icariin derivatives 2-6on PDE5A1 is shown in Table 1. Since
aglycons might be expected to possess higher activity than the
corresponding glycosides, the first general structural modification
to 1was the removal of one or both of the sugar moieties at the
hydroxyl groups at positions C-3 and C-7 of the flavone scaffold.
Enzymatic hydrolysis of 1with cellulase and naringinase allowed
the partial or total removal of the sugar moieties, respectively,
* To whom correspondence should be addressed. Tel: +39-02-50318345.
Fax: +39-02-50318391. E-mail: mario.dellagli@unimi.it.
†
University of Milano.
‡
Department of Pharmaceutical Sciences, University of Bologna.
§
Department of Veterinary Public Health and Animal Pathology,
University of Bologna.
J. Nat. Prod. 2008, 71, 1513–1517 1513
10.1021/np800049y CCC: $40.75 2008 American Chemical Society and American Society of Pharmacognosy
Published on Web 09/09/2008

affording the known compounds 2and 3. Indeed, the removal of
the glucose at the hydroxyl group in C-7, thus furnishing icariside
II (2), improved drastically the enzyme inhibition, attaining an IC50
value on the nanomolar order (IC50 156 nM). Conversely, icaritin
(3), where both sugars were removed, was only around 3-fold more
potent than 1(IC50 2.2 µM).
To investigate if the prenyl moiety is essential for enzyme
inhibition activity, β-anhydroicaritin (6) was tested. The cyclization
Scheme 1. Outline of the Synthetic Route Followed for the Synthesis of Icariin (1) Derivatives
a
a
Reagents and conditions: (a) cellulase, 37 °C, Na acetate pH )5 buffered hydroalcoholic solution, 6 days; (b) naringinase, 37 °C, Na acetate pH )5 buffered
hydroalcoholic solution, 11 days; (c) H2SO4, dioxane, reflux, 24 h; (d) 2-bromoethanol, K2CO3, acetone, reflux, 8 h.
Chart 1
1514 Journal of Natural Products,2008, Vol. 71, No. 9 Dell’Agli et al.

led to a dramatic drop in inhibitory activity. The IC50 value for 6
was 45.5 µM, indicating that a free prenyl group at position C-8 is
important for enzyme inhibition. To confirm that the prenyl group
is required for enzyme inhibition, the 8-prenyl derivatives of
naringenin (8-PN), quercetin (8-PQ), and apigenin (8-PA) were
tested and their activity compared to that of the corresponding free
flavonoid. As shown in Figure S2 (Supporting Information), all
prenylflavonoids inhibited PDE5A1 with the following order of
potency: 8-PQ (IC50 0.70 (0.10 µM) >8-PA (IC50 1.29 (0.11
µM) .8-PN (IC50 16.23 (1.16 µM). Quercetin, apigenin, and
naringenin (10 µM) showed 23%, 12%, and 6% inhibition,
respectively, much lower than the corresponding prenyl derivatives.
Data from the literature confirm the importance of the prenyl group:
sophoflavescenol, a prenylated flavonol from Sophora flaVescens
Ait. (Leguminosae), and osthole, a prenyl coumarin from Angelica
pubescens Maxim., are two additional examples of PDE-5 inhibitors
in the class of prenylated phenolic compounds.
21,22
The last modification to 1was the replacement of one or both
sugar moieties with the hydroxyethyl side chain, representing a
simplification of the sugar residue. The substitution for Glc by a
hydroxyethyl group at C-7 gave 7-(2-hydroxyethyl)-3-O-rhamno-
sylicariin (4), which was less potent than 2(IC50 363 vs 156 nM,
respectively). When both hydroxyls at C-3 and C-7 were derivatized
with hydroxyethyl moieties, as in 3,7-bis(2-hydroxyethyl)icaritin
(5), PDE5A1 was potently inhibited, with an IC50 of 74 nM, almost
identical to that of sildenafil (75 nM).
The selectivity against human PDE5A1 was investigated by
testing all compounds against human PDE6C, at concentrations 10-
fold higher than their PDE5A1 IC50 values. Compounds 1and 4
inhibited PDE6 activity (45% and 86%, respectively), while 2,3,
5, and 6were inactive. For 5, the best PDE5 inhibitor among the
icariin derivatives tested, concentration inhibition curves for PDE6
and cAMP-PDE were performed. The results were compared with
those obtained for sildenafil (Table 2). The PDE6C/PDE5 IC50 ratio
was 418 for 5and 2.2 for sildenafil, while the cAMP-PDE/PDE5
IC50 ratio was 1300 for 5and 367 for sildenafil. These results
indicate that the selectivity of 5for PDE5 was improved with
respect to that of sildenafil. Compound 5was not cytotoxic for
human fibroblasts even at the highest concentration tested (100 µM).
Thus, the inhibitory potency of 5was 80-fold higher than that of
the parent compound icariin (1). Its selectivity and lack of
cytotoxicity make 5a candidate worthy of further study.
Experimental Section
General Experimental Procedures. Melting points were determined
in open glass capillaries using a Bu¨chi apparatus and are uncorrected.
Nuclear magnetic resonance (1H NMR) spectra were recorded on a
Varian VXR 200 or Varian VXR 300 spectrometer equipped with
VNMR software. Chemical shifts (δ) are reported in ppm with
tetramethylsilane (TMS) as the internal standard, and spin multiplicities
are given as s (singlet), d (doublet), t (triplet), q (quartet), dd (double
doublet), dt (double triplet), m (multiplet), or br (broad). ESIMS were
obtained on a Finnigan MAT LCQ ion trap mass spectrometer or Waters
Micromass ZQ 4000 apparatus equipped with a Microsoft Windows
NT data system and an ESI interface. HPLC-MS analysis was carried
out with a Waters 600 MS liquid chromatograph equipped with an
Agilent Zorbax SB C18 column (4.6 mm ×2.5 cm) held at 35 °C and
a Waters 486 tunable detector set at 289 nm. Analytical conditions
were as follows: elution gradient 0.01% trifluoroacetic acid in CH3CN
(A) and 0.01% trifluoroacetic acid (v/v) in water (B) eluting in gradient
mode starting from 10% (A) up to 60% (A) in 40 min at a flow rate of
1.0 mL/min.
HPLC-UV analysis was carried out with a Kontron 325 pump/system
controller equipped with a Merck-Hitachi UV-vis detector set to 278
nm. The analyses were performed on Phenomenex Luna RP C18 (3
µm, 4.6 mm ×1.5 cm) columns. Analytical conditions were as follows:
elution gradient CH3CN (A) and 0.01% trifluoroacetic acid (v/v) in
water (B) according to the following profile: 0-60 min, 15-100% A,
85-0% B; flow rate 1.0 mL/min.
All solvents and reagents were obtained from commercial sources
and used without further purification unless otherwise noted. Reaction
courses and product mixtures were routinely monitored by TLC and
HPLC. TLC was carried out on precoated silica gel F254 (Merck) plates
or on silica gel 60 (Merck) plates (visualizing developed chromatograms
by spraying plates with 20% CH2O/H2SO4followed by heating at 100
°C for 3 min). Column chromatography was carried out with silica gel
(Kieselgel 40, 0.040-0.063 mm; Merck) using the flash technique. For
the semisynthetic derivatives, yields are reported after chromatographic
purification and crystallization.
Dulbecco’s modified Eagle’s medium, trypsin, protease inhibitors,
naringinase (from Penicillium decumbens, 596 units/g solid; β-glu-
cosidase activity: 69 units/g solid), and all chemical reagents for cell
culture were purchased from Sigma Aldrich (Milan, Italy). Cellulase
(from Aspergillus niger) was from Fluka (Milan, Italy). Penicillin,
streptomycin, and L-glutamine were from GIBCO (Grand Island, NY);
fetal calf serum was provided by Mascia Brunelli SpA (Milan, Italy).
The COS-7 cell line was purchased from ATCC (Manassas, VA).
Superfect reagent for transient transfections was obtained from Qiagen
GmbH (Hilden, Germany). The expression plasmid pcDNA3 containing
the full-length cDNA of PDE5A1 was a kind gift of Prof. C. S. Lin
(Department of Urology, University of California, San Francisco, CA).
Human recombinant PDE6C, cloned from the human retina and
expressed in S. frugiperda insect cells using a baculovirus expression
system, was purchased from Scottish Biomedical (Glasgow, UK). [3H]-
cGMP and [3H]-cAMP were from Amersham Pharmacia Biotech
(Amersham Place, Little Chalfont, Buckinghamshire, UK). DEAE-
Sephadex A25 was from Pharmacia (Uppsala, Sweden). cGMP, cAMP,
AMP, and Crotalus adamanteus snake venom were purchased from
Sigma Aldrich. Sildenafil was provided by Sequoia Research Products
(Oxford, UK). Cinnamaldehyde and ferutinin were supplied by Indena
Spa (Milan, Italy). 8-Prenylnaringenin, 8-prenylquercetin, and 8-pre-
nylapigenin (purity >98%) were donated by Prof. Giovanni Appendino
(Universita` del Piemonte Orientale, Italy).
Plant Material. T. terrestris L. dried extract (44% furostanolic
saponins) was from Farmbio Ltd. (Sofia, Bulgaria); the ethanolic extract
of the aerial parts of E. breVicornum Maxim. (20.9% icariin) was from
Chengdu Wagott Natural Products Co. Ltd., Xian City, People’s
Republic of China. The root methanolic extract from F. hermonis Boiss.
(26.3% ferutinin) and C. cassia L. oil extract (73.4% cinnamaldehyde)
were supplied by Indena Spa (Milan, Italy). Plant material was identified
against a crude drug standard and/or authoritative literature source by
a suitable qualified person. A voucher of each plant is kept at the
botanical laboratory of the company. Extracts were quantified by HPLC,
and the chromatographic profiles are shown in Figures S3-S6
(Supporting Information).
Extraction and Isolation of Icariin (1). A dried extract of E.
sagittatum as a greenish-brown residue (4 g) was dissolved in a mixture
of CH3OH/H2O (1:1) (200 mL). The solution was stirred for 20 min
and then washed with CH2Cl2(3 ×80 mL). Methanol was evaporated
under vacuum, and the remaining aqueous solution was diluted with
water to 400 mL. The solution was extracted with EtOAc (5 ×400
mL), and the organic phase was taken to dryness. The extract was
resuspended with CH2Cl2(200 mL) and filtered under vacuum to yield
1.08 g of extract (A), from which icariin (1) was purified by
precipitation with methanol (50 mL) as a yellow powder (purity 95.3%)
(530 mg; 13% yield on the dry extract); mp 224-226 °C;
22
1H NMR
(DMSO-d6, 300 MHz, 30 °C) δ0.80 (3H, d, J)5.4 Hz, rha CH3),
1.60 (3H, s, CH3-14), 1.70 (3H, s, CH3-15), 3.05-3.20 (4H, m, H-11
and sugar protons), 3.40-3.80 (7H, m, sugar protons), 3.87 (3H, s,
OCH3), 4.00 (1H, m, sugar proton), 4.55-4.78 (3H, m, OH), 4.85-5.22
(6H, m, sugar protons and OH), 5.30 (1H, t, J)6.9 Hz, H-12), 6.60
(s, 1H, H-6), 7.15 (2H, d, J)8.4 Hz, H-3′, H-5′), 7.85 (2H, d, J)8.4
Table 1. IC50 Values of Icariin Derivatives and Sildenafil on
Human PDE5A1
compound PDE5A1 (IC50 µM(SD)
15.9 (1.1
20.16 (0.02
32.2 (0.09
40.36 (0.06
50.074 (0.007
645.5 (4.6
sildenafil 0.075 (0.004
Inhibition of Phosphodiesterase-5 by Icariin DeriVatiVes Journal of Natural Products,2008, Vol. 71, No. 9 1515

Hz, H-2′, H-6′), 12.60 (s, 1H, OH-5); ESIMS (positive-ion mode) m/z
677 [M +H]+, 699 [M +Na]+.
Preparation of Icariside II (2). A solution of 1(500 mg) in DMSO
(1 mL) was added dropwise for 48 h to a Na acetate-buffered
hydroalcoholic solution at 37 °C (0.25 M, pH 5.0, in EtOH/H2O, 30:
70) (50 mL) containing cellulase (210 mg). The suspension obtained
was stirred at 37 °C for 4 days. Then, a further amount of cellulase
(100 mg) was added, and the mixture was stirred under the same
conditions for a further 2 days. EtOH was then removed under vacuum
and the residue was diluted to 200 mL with H2O and extracted with
EtOAc (3 ×200 mL). The organic layer was dried over anhydrous
Na2SO4and evaporated under reduced pressure to afford 2(290 mg);
yield 76%; mp 208-210 °C;
23
1H NMR (DMSO-d6, 300 MHz, 30 °C)
δ0.90 (3H, s, rha CH3), 1.82 (3H, s, CH3-14), 1.87 (s, 3H, CH3-15),
3.05-3.60 (4H, m, rha protons, H-11), 3.85 (3H, s, OCH3), 4.22-4.24
(1H, m, rha proton), 4.55-4.80 (3H, m, sugar OH), 4.90 (1H, m, rha
proton), 5.20 (1H, t, J)6.8 Hz, H-12), 5.52 (1H, d, J)1.5 Hz, rha
proton), 6.37 (1H, s, H-6), 7.15 (2H, d, J)8.4 Hz, H-3′, H-5′), 7.83
(2H, d, J)8.4 Hz, H-2′, H-6′), 10.60 (1H, s, OH-7), 12.80 (1H, s,
OH-5); ESIMS (positive-ion mode) m/z537 [M +Na]+.
Preparation of Icaritin (3). A solution of icariin (1) (526 mg) in
DMSO (1 mL) was added dropwise for 72 h to a Na acetate-buffered
hydroalcoholic solution at 37 °C (0.25 M, pH 5.0, in EtOH/H2O, 30:
70) (50 mL) containing naringinase (207 mg). The obtained suspension
was allowed to stir at 37 °C for 7 days. Then, a further amount of
naringinase (97 mg) was added and the mixture was stirred under the
same conditions for a further day. EtOH was removed by evaporation
and the aqueous suspension was filtered under vacuum and dried. The
residue obtained was washed with H2O and dried to give icaritin (3,
290 mg; purity 95%) as a yellow powder. The mother liquors were
diluted with H2O and extracted with EtOAc (2 ×200 mL). The organic
phase was dried over anhydrous Na2SO4and evaporated under reduced
pressure to afford an additional amount of 3(20 mg); quantitative yield,
mp 232-233 °C;
24
1H NMR (CDCl3, 300 MHz, 30 °C) δ1.78 (3H, s,
CH3-14), 1.87 (3H, s, CH3-15), 2.70 (2H, s, OH), 3.61 (2H, d, J)6.8
Hz, H-11), 3.89 (3H, s, OCH3), 5.36 (1H, t, J)6.8 Hz, H-12), 6.32
(1H, s, H-6), 7.04 (2H, d, J)8.4 Hz, H-3′, H-5′), 8.16 (2H, d, J)8.4
Hz, H-2′, H-6′); ESIMS (positive-ion mode) m/z369 [M +H]+.
Preparation of 7-(2-Hydroxyethyl)-3-O-rhamnosylicariin (4). A
stirred suspension of 2(200 mg, 0.39 mmol), 2-bromoethanol (50 mg,
0.43 mmol), and anhydrous K2CO3(60 mg, 0.43 mmol) in dry acetone
(15 mL) was refluxed for 8 h. The hot reaction mixture was filtered,
and the solvent was evaporated under reduced pressure. The residue
was purified by flash chromatography on silica gel (EtOAc/CH3OH,
9.5:0.5) to give a yellow crystalline compound (4, 112 mg; purity
93.0%); 55% yield; mp 194-196 °C (EtOH); 1H NMR (acetone-d6+
D2O, 300 MHz, 30 °C) δ0.88 (3H, d, J)5.4 Hz, rha CH3), 1.64 (3H,
s, CH3-14), 1.75 (3H, s, CH3-15), 3.2-3.8 (3H, m, rha protons), 3.7
(2H, d, J)8.7, H-11), 3.91 (3H, s, OCH3), 4.21-4.24 (2H, m,
OCH2O), 4.40 (2H, t, J)6.0 Hz, CH2OH), 4.22-4.24 (1H, m, rha
proton), 5.25 (1H t, J)6.9 Hz, H-12), 5.52 (1H, d, J)1.5 Hz, rha
proton), 6.50 (1H, s, H-6), 7.14 (2H, d, J)6.9 Hz, H-3′, H-5′), 7.96
(2H, d, J)6.9 Hz, H-2′, H-6′); ESIMS (positive-ion mode) m/z581
[M +Na]+;anal. C 62.31%, H 6.18%, calcd for C29H34O11, C 62.36%,
H 6.14%.
Preparation of 3,7-Bis(2-hydroxyethyl)icaritin (5). A stirred
suspension of 3(250 mg, 0.7 mmol), 2-bromoethanol (210 mg, 1.7
mmol), and anhydrous K2CO3(240 mg, 1.7 mmol) in dry acetone (75
mL) was refluxed for 8 h. The hot reaction mixture was filtered, and
the solvent was evaporated under reduced pressure. The residue was
purified by flash column chromatography on silica gel (CH2Cl2/acetone,
9:1) and crystallized from EtOH to give the desired compound as a
yellow crystalline powder (70 mg, purity 96.0%); 20.2% yield; mp
152-153 °C; 1H NMR (CDCl3, 300 MHz, 30 °C) δ1.77 (3H, s, CH3-
14), 1.87 (3H, s, CH3-15), 3.61 (2H, d, J)6.8 Hz, H-11), 3.78-3.83
(2H, m, 7-OCH2O), 3.90 (3H, s, OCH3), 3.95-4.05 (4H, m, CH2OH),
4.15-4.22 (2H, m, OCH2O-3), 5.19 (1H, t, J)6.8 Hz, H-12), 6.32
(1H, s, H-6), 7.04 (2H, d, J)8.4 Hz, H-3′, H-5′), 8.16 (2H, d, J)8.4
Hz, H-2′, H-6′); ESIMS (positive-ion mode) m/z479 [M +Na]+;anal.
C 65.82%, H 6.22%, calcd for C25H28O8, C 65.78%, H 6.18%.
Preparation of β-Anhydroicaritin (6). A solution of extract A, used
in the isolation of 1(200 mg) in dioxane (25 mL), was added to 1 M
H2SO4(12.5 mL) and refluxed for 24 h. After cooling, the reaction
mixture was adjusted to pH 7-8 with NaHCO3and extracted with
EtOAc (3 ×50 mL). The organic phase was dried over anhydrous
Na2SO4and evaporated under vacuum to afford β-anhydroicaritin as a
yellow powder (purity 96.5%; 105 mg, 23% yield of the dry extract);
25
1H NMR, (DMSO-d6,30°C) δ1.40 (6H, s, 14, CH3-15), 1.90 (2H, t,
H-12), 2.85 (2H t, H-11), 3.90 (3H, s, OCH3), 6.20 (1H, s, H-6), 7.15
(2H, d, J)8.9 Hz, H-3′, H-5′), 8.20 (2H, d, J)8.9 Hz, H-2′, H-6′),
9.57 (1H, s, OH-3), 12.20 (1H, s, OH-5); ESIMS (positive-ion mode)
m/z369 [M +H]+.
Human Recombinant PDE5A1 Expression. Human recombinant
PDE5A1 was prepared by expression of the full-length cDNA of
PDE5A1 into COS-7 cells, as previously described.
26
PDE5A1 and PDE6C Enzyme Assays. PDE5A1 activity was
determined according to the method of Kincaid and Manganiello
27
with
minor modifications.
28
Screening of plant extracts was performed at
50 µg/mL, whereas the individual compounds were tested at 10 µM.
PDE6C activity was evaluated under the same conditions used for
PDE5A1 activity, with 0.5 U enzyme/sample being used. Screening of
the individual compounds against PDE6C activity was performed at
concentrations 10-fold higher than each IC50 obtained against PDE5A1.
IC50 values were calculated using Graph Pad Prism 4 for sigmoidal
curves. Sildenafil was used as reference compound. Each result is the
mean (SD of at least two experiments in triplicate.
Platelet Homogenate Preparation and Assay for cAMP-PDE
Activity. The blood fraction enriched in platelets, obtained from healthy
volunteers, was submitted to two centrifugations at 160gfor 10 min at
room temperature. The pellet was removed, and platelet-rich plasma
(PRP) was centrifuged at 1000gfor 15 min. The resulting pellet was
suspended in 10 mM Tris/HCl, pH 7.4 (2/5 of the initial volume). The
suspension was centrifuged at 1000gfor 15 min and the pellet
suspended in the Tris/HCl buffer, pH 7.4 (1/12 of the initial volume).
All these steps were performed at 4 °C. Cells were disrupted by freezing
and thawing three times, obtaining the homogenate,
29
and cell lysate
was stored at -80 °C. Total protein concentration was measured
according to Bradford.
30
cAMP-PDE activity was determined according to the method of
Kincaid and Manganiello
27
with minor modifications. Briefly, platelet
lysate (64 µg of protein/mL) was incubated with 0.5 µM cAMP and
63 nCi [3H]-cAMP suspended in 30 mM Tris-HCl, pH 7.4, 4 mM
MgCl2; final reaction volume was 250 µL. After 5 min of incubation
at 30 °C, the reaction was stopped with 0.1 N HCl. Samples were then
incubated for a further 4 min at 70 °C with AMP (5 mM) and cAMP
(5 mM), and the pH was adjusted to 7 on ice with 0.1 N NaOH. Samples
were then added with 50 µL of nucleotidase from Crotalus adamanteus
snake venom (1 mg/mL in Tris-HCl 0.1 M, pH 8.0) and incubated for
20 min at 37 °C. The reaction was stopped with 50 µL of 200 mM
NaEDTA containing 5 mM adenosine. The nucleoside formed during
the incubation was separated from the unreacted substrate by DEAE-
Sephadex A25 column chromatography. The eluted [3H]-adenosine was
counted in a β-scintillation counter. Compound 5and sildenafil were
tested in a range of 1-250 µM, and IC50 values calculated using Graph
Pad Prism 4 for sigmoidal curves. Inhibition (%) by aminophylline
(100 µM) used as reference compound was 74.5 (1.3 (mean (SD,
n)11). Each result is the mean (SD of three experiments in triplicate.
Cytotoxicity Assay. Cellular toxicity was assessed using a 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) (MTT) colo-
rimetric assay.
31
Human skin fibroblasts were treated with increasing
concentrations (0.25-100 µM) of 5for 24 h in DMEM-F12 supple-
mented with 10% heat-inactivated FBS, 1% penicillin, and 1%
L-glutamine. The medium was removed, and cells were incubated with
a solution containing MTT 0.5 mg/mL in PBS at 37 °Cfor3h.The
MTT solution was removed, the formazan was extracted with 2-pro-
Table 2. IC50 Values of 5and Sildenafil on Human PDE6 and cAMP-PDE
compound
PDE6C
(IC50 µM(SD)
PDE6C/PDE5A1
(IC50 ratio)
cAMP-PDE
(IC50 µM(SD)
cAMP-PDE/PDE5A1
(IC50 ratio)
530.9 (2.6 418 96.3 (12.9 1301
sildenafil 0.16 (0.007 2.2 27.5 (5.3 367
1516 Journal of Natural Products,2008, Vol. 71, No. 9 Dell’Agli et al.

panol/DMSO (9:1; 500 µL/well) for 15 min at 37 °C, and aliquots of
100 µL were read on a plate reader (Bio-Rad Laboratories) at 560 nm.
Acknowledgment. The authors gratefully acknowledge Prof. C. S.
Lin for the supply of PDE5A1 cDNA, and Prof. G. Appendino for the
8-prenyl flavonoids.
Supporting Information Available: Figures showing the HPLC
traces of the extracts under study and the effects of plant extracts and
pure compounds on the inhibition of PDE5A1. This information is
available free of charge via the Internet at http://pubs.acs.org.
References and Notes
(1) Anonymous. J. Am. Med. Assoc. 1993,123,23-27.
(2) Drewes, S. E.; George, J.; Khan, F. Phytochemistry 2003,62, 1019–
25.
(3) Boyce, E. G.; Umland, E. M. Clin. Ther. 2001,23, 2–23.
(4) McKay, D. Altern. Med. ReV.2004,9, 4–16.
(5) Colman-Saizarbitoria, T.; Boutros, P.; Amesty, A.; Bahsas, A.;
Mathison, Y.; Garrido, M. D.; Israel, A. J. Ethnopharmacol. 2006,
106, 327–332.
(6) Phillips, O. A.; Mathew, K. T.; Oriowo, M. A. J. Ethnopharmacol.
2006,104, 351–5.
(7) Adaikan, P. G.; Gauthaman, K.; Prasad, R. N.; Ng, S. C. Ann. Acad.
Med. Singapore 2000,29, 22–26.
(8) Zanoli, P.; Zavatti, M.; Rivasi, M.; Baraldi, M. Physiol. BehaV.2005,
86, 69–74.
(9) Zanoli, P.; Rivasi, M.; Zavatti, M.; Brusiani, F.; Vezzalini, F.; Baraldi,
M. Int. J. Impot. Res 2005,17, 513–518.
(10) Zhang, J.; Li, H.; Fang, L.; Zhou, Z.; Ai, J. Peop. Rep. China Patent
1814119, 2006.
(11) Chen, F.; Wang, D. Peop. Rep. China Patent 1927326, 2007.
(12) Zhang, W. Peop. Rep. China Patent 1903244, 2007.
(13) Zhao, Z. A. Peop. Rep. China Patent 1903310, 2007.
(14) Tu, C.; Liao, N. Peop. Rep. China Patent 1872244, 2006.
(15) Kim, J.-H.; Mun, Y.-J.; Im, S.-J.; Han, J.-H.; Lee, H.-S.; Woo, W.-H.
Int. Immunopharmacol. 2001,1, 935–944.
(16) He, W.; Sun, H.; Yang, B.; Zhang, D.; Kabelitz, D. Arzneim.-Forsch.
1995,45, 910–913.
(17) Xin, Z. C.; Kim, E. K.; Lin, C. S.; Liu, W. J.; Tian, L.; Yuan, Y. M.;
Fu, J. Asian J. Androl. 2003,5, 15–18.
(18) Liu, W. J.; Xin, Z. C.; Xin, H.; Yuan, Y. M.; Tian, L.; Guo, Y. L.
Asian J. Androl. 2005,7, 381–388.
(19) Ning, H.; Xin, Z. C.; Lin, G.; Banie, L.; Lue, T. F.; Lin, C. S. Urology
2006,68, 1350–1354.
(20) Chiu, J. H.; Chen, K. K.; Chien, T. M.; Chen, C. C.; Wang, J. Y.;
Lui, W. Y.; Wu, C. W. Int. J. Impot. Res. 2006,18, 335–342.
(21) Shin, H. J.; Kim, H. J.; Kwak, J. H.; Chun, H. O.; Kim, J. H.; Park,
H.; Kim, D. H.; Lee, Y. S. Bioorg. Med. Chem. Lett. 2002,12, 2313–
2316.
(22) Teng, C. M.; Lin, C. H.; Ko, F. N.; Wu, T. S.; Huang, T. F. Naunyn-
Schmiedeberg’s Arch. Pharmacol. 1994,349, 202–208.
(23) Mizuno, M.; Hanioka, S.; Suzuki, N.; Iinuma, M.; Tanaka, T.; Liu,
X.; Min, Z. Phytochemistry 1987,26, 861–863.
(24) Mizuno, M.; Iinuma, M.; Tanaka, T.; Sakakibara, N.; Fujikawa, T.;
Hanioka, S.; Ishida, Y.; Liu, X.; Murata, H. Phytochemistry 1988,
27, 3645–3647.
(25) Kang, S. S.; Kang, Y. J.; Lee, M. W. J. Nat. Prod. 1991,54, 542–
546.
(26) Lin, C. S.; Lau, A.; Tu, R.; Lue, T. F. Biochem. Biophys. Res. Commun.
2000,268, 628–635.
(27) Kincaid, R. L.; Manganiello, V. C. Methods Enzymol. 1988,159, 457–
470.
(28) Dell’Agli, M.; Galli, G. V.; Vrhovsek, U.; Mattivi, F.; Bosisio, E. J.
Agric. Food Chem. 2005,53, 1960–1965.
(29) Giovanazzi, S.; Accomazzo, M. R.; Letari, O.; Oliva, D.; Nicosia, S.
Biochem. J. 1997,325 (Part 1), 71–77.
(30) Bradford, M. M. Anal. Biochem. 1976,72, 248–254.
(31) Denizot, F.; Lang, R. J. Immunol. Methods 1986,89, 271–277.
NP800049Y
Inhibition of Phosphodiesterase-5 by Icariin DeriVatiVes Journal of Natural Products,2008, Vol. 71, No. 9 1517