Bone selective effect of an estradiol conjugate with a novel tetracycline-derived bone-targeting agent

Article abstract of DOI:10.1016/j.bmcl.2008.12.051

In this study a novel bone-targeting agent containing elements of the tricarbonylmethane system of ring A of tetracycline was developed and was shown to bind to the mineral constituent of bone, hydroxyapatite. Conjugation of this bone-targeting agent to estradiol resulted in a bone-targeted estrogen (BTE₂-A1) with an enhanced ability to bind to hydroxyapatite. In an ovariectomized rat model of osteoporosis a partial separation of the skeletal effects of estradiol from the uterine effects was observed following subcutaneous administration of BTE₂-A1. This novel bone-targeting estradiol delivery system has the potential to improve the safety profile of estradiol in the treatment of osteoporosis.

Full text of DOI:10.1016/j.bmcl.2008.12.051

Bioorganic & Medicinal Chemistry Letters 19 (2009) 680–683
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Bone selective effect of an estradiol conjugate with a novel
tetracycline-derived bone-targeting agent
Jason R. Neale a, Natali B. Richter b, Kevyn E. Merten c, K. Grant Taylor b, Sujan Singh b,
Leonard C. Waite a, Nicole K. Emery a, Ned B. Smith a, Jian Cai a, William M. Pierce Jr. a,b,c,
*
a Department of Pharmacology and Toxicology, University of Louisville, Louisville, KY 40292, USA
b Department of Chemistry, University of Louisville, Louisville, KY 40292, USA
c Pradama, Inc., Louisville, KY 40202, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
In this study a novel bone-targeting agent containing elements of the tricarbonylmethane system of ring
A of tetracycline was developed and was shown to bind to the mineral constituent of bone, hydroxyap-
atite. Conjugation of this bone-targeting agent to estradiol resulted in a bone-targeted estrogen (BTE2-A1)
with an enhanced ability to bind to hydroxyapatite. In an ovariectomized rat model of osteoporosis a par-
tial separation of the skeletal effects of estradiol from the uterine effects was observed following subcu-
taneous administration of BTE2-A1. This novel bone-targeting estradiol delivery system has the potential
to improve the safety profile of estradiol in the treatment of osteoporosis.
Received 23 October 2008
Revised 8 December 2008
Accepted 10 December 2008
Available online 24 December 2008
Keywords:
Bone-targeting
Bone-targeted estrogen
Tetracycline
Ó 2008 Elsevier Ltd. All rights reserved.
Hydroxyapatite
Osteoporosis
Osteotropic
Numerous agents including acidic oligopeptides and bisphos-
phonates display a high affinity for the mineral constituent of bone
(hydroxyapatite) and these agents have been used as bone-target-
ing moieties.1–5 Chemical conjugation of these agents to estradiol
has been investigated as a means to improve the therapeutic index
of this compound in the treatment of osteoporosis.6,7 These osteo-
tropic estradiol delivery systems exhibit preferential distribution
of estradiol to bone and show pharmacological activity in bone
comparable to estradiol without the systemic adverse effects of
estradiol.6–8 Clinically, however, these targeting agents are less
than ideal due to the inherent biological activities, potential side
effects, and routes of administration of the bone-targeting moieties
themselves.9,10 The polycyclic naphthacene carboxamide, tetracy-
cline, is known to avidly bind to hydroxyapatite and has previously
been evaluated as a bone-targeting moiety following conjugation
with therapeutic agents, including estradiol.11–13 As a bone-target-
ing agent, tetracycline offers the advantages of oral bioavailability
and of being relatively non-toxic, however, the clinical utility of
tetracycline conjugates is questionable due to the intrinsic antibi-
otic activity of tetracycline.11 Furthermore, the complex chemical
structure and poor stability of tetracycline during chemical conju-
gation limit the feasibility of using tetracycline as an osteotropic
agent.14
Structure–activity relationship (SAR) studies indicate that it
may be possible to separate the hydroxyapatite binding domain
from the minimum structural requirements for the biological
activity of tetracycline. These studies have revealed that many
changes can be made at positions 5, 6, and 7 of tetracycline with-
out affecting the antibacterial properties.15 However, the remain-
ing structural and stereochemical features of tetracycline,
including the dimethylamino group, the phenolic b-diketone sys-
tem, and the tricarbonylmethane system, are essential for in vivo
antibacterial activity.15,16 On the other hand, the structural
requirements for hydroxyapatite binding appear much less exten-
sive as the tricarbonylmethane grouping of ring A of tetracycline
(Fig. 1) has been suggested to be sufficient to allow binding to
hydroxyapatite.17 In this study we synthesized and characterized
a bone-targeting agent modeled after the tricarbonylmethane
grouping of ring A of tetracycline. This novel bone-targeting agent
was shown to possess significant hydroxyapatite binding affinity. A
bone-targeted estrogen was subsequently synthesized via chemi-
cal conjugation of estradiol and the pharmacological effects of this
conjugate were evaluated in vitro and in vivo in an ovariectomized
rat model of osteoporosis.
The bone-targeting agent was synthesized in four steps, as de-
picted in Scheme 1, starting with 2,6-dihydroxybenzoic acid (1).
* Corresponding author. Tel.: +1 502 852 7424; fax: +1 502 852 7868.
E-mail address: pierce@louisville.edu (W.M. Pierce).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.12.051

J. R. Neale et al. / Bioorg. Med. Chem. Lett. 19 (2009) 680–683
681
Alkylation of compound 1 with iodomethane via an intermediate
silver salt yielded the corresponding methyl ester 2. Treatment of
the ester with ammonium hydroxide led to the amide product 3.
Nitration at the 3-position was accomplished using the standard
combination of glacial acetic and concentrated nitric acids to yield
4. Subsequently, compound 5 was attained by reduction of the
3-nitro group.
The ability of compounds 1, 3, 4, and 5 to bind to the mineral
constituent of bone (hydroxyapatite, Ca10(PO4)6ÁOH2) was evalu-
ated using a hydroxyapatite (HA) binding assay.18 For comparative
purposes tetracycline was included as a control and was defined as
having a HA binding index of 100. Compound 1 did not bind to
hydroxyapatite (data not shown), while compounds 3 and 4
showed little to no appreciable binding (Fig. 2). Compound 5,
3-amino-2,6-dihydroxybenzamide, on the other hand, bound to
hydroxyapatite approximately 50% as well as tetracycline. Interest-
ingly, neither 1-carboxamidocyclohexane-2,6-dione (data not
shown) nor 2,6-dihydroxybenzamide (3), both similar in structure
to the tricarbonylmethane system of ring A of tetracycline, bound
to hydroxyapatite in vitro indicating that this system may not be
sufficient in and of itself to allow binding to bone mineral. Our data
suggest that in addition to the tricarbonylmethane system of tetra-
cycline, an amino group at position 3 of compound 5 (position 4 of
the A ring of tetracycline) is necessary to allow binding to hydroxy-
apatite. This conclusion is further supported by the finding that the
corresponding 3-nitro analog (4) does not appreciably bind to
hydroxyapatite.
As indicated above, 3-amino-2,6-dihydroxybenzamide (5) binds
to hydroxyapatite approximately 50% as well as tetracycline. Com-
pound 5 was therefore chosen as the prototype bone-targeting
agent and was subsequently conjugated with estradiol to produce
a bone-targeted estrogen, BTE2-A1. The synthesis of BTE2-A1 was
carried out as shown in Scheme 2. Prior to nitration, compound 3
was protected asymmetrically using one molar equivalent of ben-
zyl bromide under basic conditions to give product 6. The com-
pound was nitrated at the 3-position, yielding 7, which was then
selectively reduced with hydrazine, ferric chloride hexahydrate,
and activated carbon to give compound 8. Carbodiimide-mediated
coupling of the bone-targeting moiety to estradiol-17-O-hemisuc-
cinate (Steraloids, Inc., Newport, R.I.) yielded product 9. Finally,
catalytic hydrogenation of 9 gave the desired conjugate 10,
(13S,17S)-3-hydroxy-13-methyl-7,8,9,11,12,13,14,15,16,17-deca-
hydro-6H-cyclopenta[a]phenanthren-17-yl 4-(3-carbamoyl-2,4-
dihydroxyphenylamino)-4-oxobutanoate. Detailed synthetic pro-
tocol, as well as NMR and mass spectral data, for compounds 5
(Scheme 1) and 10 (Scheme 2) can be found in the Supplementary
material.
Figure 1. Structure of tetracycline. The region outlined is the tricarbonylmethane
region of tetracycline and is the template for synthesis of the novel bone-targeting
agent.
The ability of BTE2-A1 (10) to bind to crystalline hydroxyapatite
in vitro was used as a predictive measure of this compounds ability
to bind to bone in vivo. Tetracycline, estradiol, and BTE2-A1 (10)
were evaluated using the hydroxyapatite binding assay.18 Surpris-
ingly, BTE2-A1 (10) had a HA binding index of 105 (tetracy-
cline = 100), indicating that the addition of the estrogenic group
to the bone-targeting agent (5) enhanced the ability of this com-
pound to bind to the mineral constituent of bone. This data lends
support to our bone-targeting strategy, as non-targeted estradiol
did not bind to hydroxyapatite (HA binding index = 0).
Figure 2. Hydroxyapatite binding capacity. The ability of tetracycline and com-
pounds 3, 4, and 5 to bind to hydroxyapatite was evaluated using a hydroxyapatite
binding assay. Compound 5, similar to the A ring of tetracycline, has approximately
50% of the binding ability of tetracycline in this surrogate bone-targeting system.
Scheme 1. Synthesis of bone-targeting agent. Reagents: (a) NH4OH (28% NH3), AgNO3, CH3I; (b) NH4OH (28% NH3); (c) glacial acetic acid, concentrated HNO3; (d) 10% Pd/C,
H2.

682
J. R. Neale et al. / Bioorg. Med. Chem. Lett. 19 (2009) 680–683
Scheme 2. Synthesis of bone-targeted estrogen. Reagents: (a) K2CO3, benzyl bromide; (b) glacial acetic acid, concentrated HNO3; (c) activated carbon, FeC13Á6H2O,
NH2NH2ÁH2O; (d) estradiol-17-O-hemisuccinate, HOBT, DCC; (e) 10% Pd/C, H2.
Aspartic acid conjugated estradiol has previously been shown to
have a significantly lower affinity for the estrogen receptor than
free estradiol.6 A competitive estrogen receptor binding assay
was performed using purified recombinant ERa and ERb in order
to determine the binding affinity of BTE2-A1.19 Data for BTE2-A1
and an estradiol (E2) standard curve were plotted as percent
[3H]-E2 bound versus molar concentration and the IC50 (50% inhibi-
tion of [3H]-E2 binding) for each was determined. The relative
binding affinity (RBA) of BTE2-A1 was calculated as IC50 E2/IC50
BTE2-A1 Â 100%. BTE2-A1 was found to have a RBA of 16% for
ERa and 9.4% for ERb indicating that BTE2-A1 has the ability to bind
to both estrogen receptors though with a significantly lower affin-
ity than free estradiol.
The ovariectomized rat is commonly used as a model of post-
menopausal bone loss as the two systems share many characteris-
tics, including similar skeletal and systemic responses to estrogen
therapy.20 This model was therefore used to determine whether or
not BTE2-A1 demonstrates selective effects on bone, in comparison
to 17b-estradiol. Three month old, bilaterally ovariectomized
(OVX) or sham-operated Sprague–Dawley female rats were admin-
istered vehicle, 17b-estradiol or BTE2-A1 three times per week via
subcutaneous injection. Drug administration was initiated 4 days
following surgery and lasted for 6 weeks. Uterine and femoral mass
were determined directly using excised tissue and normalized to
body mass. Ovariectomy significantly decreased the normalized
femoral mass (femoral mass/body weight) by 17.1% and the nor-
malized uterine mass (uterine mass/body weight) by 90.5% in com-
parison to sham-operated controls. Subcutaneous administration
of 17b-estradiol suppressed the OVX-induced decrease in normal-
ized femoral mass and normalized uterine mass in a dose-depen-
dent manner (Fig. 3A). 17b-Estradiol appeared to be equipotent
in both bone and uterus as demonstrated by the overlapping
dose–response curves in these tissues (Fig. 3A). In fact, the calcu-
lated uterine and femoral ED50 of 17b-estradiol were 9 and
10 nmol/kg, respectively. On the other hand, subcutaneous admin-
istration of BTE2-A1 resulted in a partial separation of the bone and
uterine effects of this compound (Fig. 3B) with calculated uterine
and femoral ED50 of 90 and 8 nmol/kg, respectively. The separation
of skeletal and systemic effects following subcutaneous adminis-
Figure 3. Effects of subcutaneous administration of 17b-estradiol (A) or BTE2-A1
(B) on normalized femoral mass and normalized uterine mass in OVX rats. Percent
effect is expressed as the ratio of the difference between treatment and OVX control
groups to the difference between sham-operated and OVX control groups (n = 6–
10). 0% effect represents control OVX rats and 100% effect represents sham-
operated control rats. The data are expressed as mean SEM. Significantly different
from OVX rats: *p < 0.05.
tration of BTE2-A1 is similar to the dose-dependent suppression
of bone loss and mitigation of uterine stimulation previously ob-
served with other osteotropic estradiol delivery systems.6,7 Collec-
tively, the preservation of femoral mass and the partial separation
of uterine and skeletal effects achieved by BTE2-A1 indicate that
this compound may produce its effects through the selective deliv-
ery and accumulation of estradiol in bone. Alternatively, the partial
separation of skeletal and uterotrophic effects following adminis-

J. R. Neale et al. / Bioorg. Med. Chem. Lett. 19 (2009) 680–683
683
tration of BTE2-A1 could be attributable to a differential selectivity
for ER subtypes as BTE2-A1 has an ER selectivity (ER RBA/ERb
References and notes
a
a
1. Fujisaki, J.; Tokunaga, Y.; Takahashi, T.; Hirose, T.; Shimojo, F.; Kagayama, A.;
Hata, T. J. Drug Target. 1995, 3, 273.
2. Sekido, T.; Sakura, N.; Higashi, Y.; Miya, K.; Nitta, Y.; Nomura, M.; Sawanishi, H.;
Morito, K.; Masamune, Y.; Kasugai, S.; Yokogawa, K.; Miyamoto, K. J. Drug
Target. 2001, 9, 111.
RBA) of 1.74. However, these data appear to support this bone-tar-
geting strategy, as the bone protective effects of BTE2-A1 and estra-
diol were observed at similar ED50 despite the fact that BTE2-A1
binds to the estrogen receptor with a much lower affinity than
estradiol.
3. Gil, L.; Han, Y.; Opas, E. E.; Rodan, G. A.; Ruel, R.; Seedor, J. G.; Tyler, P. C.; Young,
R. N. Bioorg. Med. Chem. 1999, 7, 901.
Additional studies will be required to determine the mechanism
of action of BTE2-A1. The ester bond linking the bone-targeting agent
to estradiol may be hydrolyzed to generate free estradiol following
the selective delivery of the compound to bone. Alternatively, intact
BTE2-A1 may mediate the selective effects on bone. Future SAR stud-
ies will be performed in which the bond linking the bone-targeting
agent to estradiol is modified from an ester to a less labile linkage
such as an amide or ether. Additionally, drug activation mechanisms
which will allow for the generation of estradiol specifically at bone
will be explored. Modifications will also be made to the bone-target-
ing agent in order to determine the effects that steric hindrance may
have on the ability of the moiety to bind to hydroxyapatite. Addi-
tionalstudieswill also be requiredtodemonstrate thatour bone-tar-
geting agent is biologically inactive and devoid of the inherent
antibacterial properties of tetracycline. However, from this study it
is apparent that our single ring bone-targeting agent maintains an
affinity for hydroxyapatite similar to that of the complete four ring
system of tetracycline and that conjugation of the bone-targeting
agent to estradiol results in a compound with selectively efficacious
effects on bone.
In summary, we have successfully synthesized and characterized
a novel bone-targeting agent modeled after the tricarbonylmethane
grouping of ring A of tetracycline. This bone-targeting agent, 3-ami-
no-2,6-dihydroxybenzamide (5), was shown to possess significant
hydroxyapatite binding affinity. A bone-targeted estrogen, BTE2-
A1 (10), was subsequently synthesized via chemical conjugation of
estradiol and the pharmacological effects of this conjugate were
evaluated in an ovariectomized rat model of osteoporosis. Using this
model, a partial separation of the skeletal effects of estradiol from
the uterine effects was observed following subcutaneous adminis-
tration of BTE2-A1, indicating that our novel osteotropic estradiol
delivery system has the potential to improve the safety profile of
estradiol in the treatment of osteoporosis.
4. Uludag, H. Curr. Pharm. Des. 2002, 8, 1929.
5. Erez, R.; Ebner, S.; Attali, B.; Shabat, D. Bioorg. Med. Chem. Lett. 2008, 18,
816.
6. Yokogawa, K.; Miya, K.; Sekido, T.; Higashi, Y.; Nomura, M.; Fujisawa, R.;
Morito, K.; Masamune, Y.; Waki, Y.; Kasugai, S.; Miyamoto, K. Endocrinology
2001, 142, 1228.
7. Fujisaki, J.; Tokunaga, Y.; Takahashi, T.; Shimojo, F.; Kimura, S.; Hata, T. J. Drug
Target. 1997, 5, 129.
8. Fujisaki, J.; Tokunaga, Y.; Takahashi, T.; Kimura, S.; Shimojo, F.; Hata, T. Biol.
Pharm. Bull. 1997, 20, 1183.
9. Stepensky, D.; Kleinberg, L.; Hoffman, A. Clin. Pharmacokinet. 2003, 42, 863.
10. Hirabayashi, H.; Fujisaki, J. Clin. Pharmacokinet. 2003, 42, 1319.
11. Skinner, H. C.; Nalbandian, J. Yale J. Biol. Med. 1975, 48, 377.
12. Pierce, W. M.; Waite, L. C. Proc. Soc. Exp. Biol. Med. 1987, 186, 96.
13. Orme, M. W.; Labroo, V. M. Bioorg. Med. Chem. Lett. 1994, 4, 1375.
14. Wang, D.; Miller, S. C.; Kopecková, P.; Kopecek, J. Adv. Drug Delivery Rev. 2005,
57, 1049.
15. Albert, A. In Selective Toxicity; John Wiley and Sons: New York, 1979, pp 431–
433.
16. Korst, J. J.; Johnston, J. D.; Butler, K.; Bianco, E. J.; Conover, L. H.; Woodward, R.
B. J. Am. Chem. Soc. 1968, 90, 439.
17. Myers, H. M.; Tochon-Danguy, H. J.; Baud, C. A. Calcif. Tissue Int. 1983, 35, 745.
18. Hydroxyapatite binding assay. A 10À3 M solution for each analyte was made in
100% dimethylsulfoxide (DMSO). A 100-fold dilution was then made to form a
10À5 M solution in 50 mM Tris–HCl buffer, pH 7.4, 1% DMSO. Tetracycline was
used as a reference analyte and approximately 50% was adsorbed to HA at the
concentration of 10À5 M. The HA slurry was 0.5 g/100 mL 50 mM Tris–HCl
buffer, 1% DMSO. For each analyte, two samples were prepared. For one
sample, 1 mL of 10À5 M analyte and 100
was pipetted into a microcentrifuge tube. For the second sample, 1 mL of
10À5 M analyte and 100
of the HA slurry was pipetted into
lL 50 mM Tris–HCl buffer, 1% DMSO
lL
a
microcentrifuge tube. The samples were mixed gently by inversion for 4 min
and then centrifuged at 12,000g for 3 min to sediment the HA contained in
those samples. The supernatant was transferred to another microcentrifuge
tube. An electronic spectral scan (ultraviolet–visible) from 220 to 520 nm was
obtained for each analyte using a Varian Cary 300 Bio Scan. The blank was
50 mM Tris–HCl buffer, 1% DMSO. The wavelength of maximum absorbance
(kmax) was determined, and the extinction coefficient (
the Beer–Lambert Law. The absorbance of the samples incubated with HA was
measured at kmax and the molar concentration of the analyte was then
e) was calculated using
,
determined using the Beer–Lambert Law and the previously calculated
extinction coefficient. The fraction adsorbed to HA for each sample was
calculated.
19. Estrogen receptor competition binding assays. Purified ERa and ERb were
obtained following expression in Sf21 insect cells and were generously
provided by Dr. Carolyn Klinge of the University of Louisville. Reaction
mixtures containing 5 nM ER
concentrations of competitor were added to 40 mM Tris buffer, pH 7.4,
containing 1 mM ethylenediamine tetraacetic acid (EDTA), 111 mM KCl,
0.5 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mM dithiothreitol
Acknowledgments
a
or ERb, 5 nM [3H]-estradiol, and increasing
We thank Carolyn Klinge, Ph.D. of the University of Louisville
for generously providing purified ERa and ERb for the competitive
estrogen receptor binding assay. This work was supported in part
by grants from the Kowa Co., Ltd, the Kentucky Science and Tech-
nology Corporation and the University of Louisville. J.R. Neale, K.G.
Taylor, L.C. Waite, and W.M. Pierce have a financial interest in
Pradama, Inc. as does author K.E. Merten as an employee of the
company. Pradama, Inc. has licensed the technology included here-
in from the University of Louisville Research Foundation, Inc.
(DTT). Total reaction volume was 60 lL. A 200-fold excess of unlabeled
estradiol was used to determine non-specific binding (0% [3H]-estradiol bound)
and 0 nM competitor represented total binding (100% [3H]-estradiol bound).
Each sample was performed in triplicate and incubations were conducted for
16 h at 4 °C. Bound and free radioligand were separated by a hydroxyapatite
(HA) slurry consisting of 10% (w/v) HA, 40 mM Tris buffer, pH 7.4, 111 mM KCl,
0.5 mM PMSF, and 1 mM DTT. A volume of 50 lL of the HA slurry was added to
the reaction mixture and vortexed briefly to precipitate the ER and stop the
binding reaction. The reaction mixture containing the HA solution and the
precipitated ER with bound ligand was centrifuged, the supernatant was
discarded, and the HA pellet was resuspended in buffer to rinse the HA pellet of
any unbound [3H]-estradiol. Scintillation cocktail was added to the HA pellet. A
scintillation counter was used to determine the amount of [3H]-estradiol
present.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2008.12.051.
20. Kalu, D. N. Bone Miner. 1991, 15, 175.