Synthesis and biological evaluation of prostaglandin-F alkylphosphinic acid derivatives as bone anabolic agents for the treatment of osteoporosis

Article abstract of DOI:10.1021/jm010264b

A series of novel C₁ alkylphosphinic acid analogues of the prostaglandin-F family have been evaluated at the eight human prostaglandin receptors for potential use in the treatment of osteoporosis. Using molecular modeling as a tool for structur

Full text of DOI:10.1021/jm010264b

J . Med. Chem. 2001, 44, 4157-4169
4157
Syn th esis a n d Biologica l Eva lu a tion of P r osta gla n d in -F Alk ylp h osp h in ic Acid
Der iva tives a s Bon e An a bolic Agen ts for th e Tr ea tm en t of Osteop or osis
David L. Soper,† J ared B. J . Milbank,‡ Glen E. Mieling,*,† Michelle J . Dirr,† Andrew S. Kende,‡ Robin Cooper,†
Webster S. S. J ee,§ Wei Yao,§ J ian Liang Chen,§ Mark Bodman,† Mark W. Lundy,*,† Biswanath De,†
Mark E. Stella,† Frank H. Ebetino,† Yili Wang,† Mitchell A. deLong,† and J ohn A. Wos*,†
Procter & Gamble Pharmaceuticals, 8700 Mason-Montgomery Road, Mason, Ohio 45040, Department of Chemistry,
University of Rochester, Rochester, New York 14627, and Department of Radiobiology, University of Utah,
Salt Lake City, Utah 84112
Received J une 13, 2001
A series of novel C1 alkylphosphinic acid analogues of the prostaglandin-F family have been
evaluated at the eight human prostaglandin receptors for potential use in the treatment of
osteoporosis. Using molecular modeling as a tool for structure-based drug design, we have
discovered that the phosphinic acid moiety (P(O)(OH)R) behaves as an isostere for the C1
carboxylic acid in the human prostaglandin FP binding assay in vitro and possesses enhanced
hFP receptor selectivity when compared to the parent carboxylic acid. When evaluated in vivo,
the methyl phosphinic acid analogue (4b) produced a bone anabolic response in rats, returning
bone mineral density (BMD) to intact levels in the distal femur in the ovariectomized rat (OVX)
model. These results suggest that prostaglandins of this class may be useful agents in the
treatment of diseases associated with bone loss.
In tr od u ction
agent, it also possesses a considerable resorptive com-
ponent and complex pharmacology, in part due to an
inherent lack of receptor specificity of this ligand for
the corresponding receptor family, EP1-EP4.11 Within
a program of identifying other receptor selective pros-
taglandins which produce a bone anabolic effect, we
have evaluated a variety of prostaglandin-F-type com-
pounds at the eight human prostaglandin receptors and
have observed that potency and selectivity for the hFP
receptor was greatly influenced by substitution at the
Osteoporosis and related bone metabolic diseases
have emerged as leading health care issues worldwide.
In the United States alone, it is estimated that ap-
proximately 30% of postmenopausal women suffer from
osteoporosis, as based on the World Health Organization
definition using bone mineral density (BMD) T-score <
-2.5,1 with an estimated annual patient base growth
at 2%. It is anticipated that the estimated cost of hip
fracture alone in the United States will approach $250
billion dollars by year 2050.2 Currently, major emphasis
has been placed on the development of agents which
slow the progression of bone loss with antiresorptive
therapies which target inhibition of osteoclasts.3 These
therapies include bisphosphonates,4 calcitonins,5 estro-
gens,6and SERMs.7 More recently, interest has turned
to treatments which produce new bone growth via
targeting the osteoblast as a strategy toward treat-
ment of metabolic bone disease. These therapies include
PTH (parathyroid hormone)8 and fluoride.9 Due to the
inherent difficulties associated with these current ana-
bolic treatments (dose route, narrow therapeutic index,
bone quality), there remains considerable interest in
small molecules which produce new bone growth as
potential treatments for bone diseases including os-
teoporosis.
C
16-C20 region of the prostaglandin skeleton.12 To de-
velop a more thorough understanding of the other key
elements within the FP pharmacophore, we wished to
explore substitution at the C1 position in an attempt to
modify both the intrinsic binding affinity of ligands for
the FP receptor and the selectivity for this receptor rela-
tive to the other known prostaglandin receptors. We
have recently evaluated sulfonamide substitution at C1
and have found a significant effect of sulfonamide
structure on the binding affinity of ligands for the FP
receptor.13 On the basis of the results of this work, we
wished to evaluate replacement of the C1 carboxylic acid
with a functional group which would more precisely
mimic the parent COOH moiety and would allow for the
flexibility of substitution in order to probe the steric and
electron requirements necessary for receptor binding
affinity and selectivity for the prostaglandin FP receptor
at the 1-position. Using molecular modeling as guide,
we chose phosphorus as a suitable replacement at C1
and subsequently designed a series of phosphinate
analogues as potential surrogates for the carboxylic acid
at the 1-position. We now wish to report our findings
in this series; the result of which is the discovery of a
novel carboxylic acid mimic, the methyl phosphinic acid,
which possesses enhanced selectivity for the FP receptor
in vitro and a bone anabolic effect in vivo in the
ovariectomized rat model.
Prostaglandin E2, PGF2R, and analogues within the
prostaglandin-F (PGF) family have been reported to
produce a bone anabolic response in animal models,
including the ovariectomized rat model, and thus may
be useful as potential treatments for bone diseases such
as osteoporosis.10 Although PGE2 is a strong anabolic
* To whom correspondence should be addressed. For J .A.W.:
Phone: (513) 622-3638. Fax: (513) 622-2675. E-mail: wos.ja@pg.com.
† Procter & Gamble Pharmaceuticals.
‡ University of Rochester.
§ University of Utah.
10.1021/jm010264b CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/02/2001

4158 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
Soper et al.
Ta ble 1. Sequence Alignments of Receptor Site Helices for the
Prostaglandin Receptor Family
F igu r e 1. (a) Compound 4b docked in the surface model of
the hFP receptor. (b) Three-dimensional structural model of
compound 4b/hFP receptor binding site.
Molecu la r Mod elin g
We focused on the previously constructed homology
model of the hFP receptor that maximized the hypoth-
esized ionic, hydrogen bonding, and hydrophobic inter-
actions with the PGF2R prostaglandin.12 Although X-ray
coordinate information is currently unavailable for
prostaglandin receptors, they are members of the G-
protein-coupled receptor (GPCR) family type and there-
fore possess structural similarity to the Swiss Insti-
tute rhodopsin template, upon which this model is
based.14 The putative prostaglandin receptor site was
composed of residues from the third, fifth, sixth, and
seventh transmembrane helices. Sequence alignments
of these four transmembrane helices within the pros-
taglandin receptor family are shown in Table 1. In
previous work, we had focused on the C16-C20 region of
the ligands in order to identify the necessary contribu-
tions of the ligand/receptor interactions at the “tail”
region that supported the high degree of specificity
observed with the analogues described. In this work, we
focused on the C1 carboxylate region of the ligand, which
forms an ionic interaction with a conserved arginine
residue found across all prostaglandin receptors and
which has been shown previously via site-directed
mutagenesis studies to be a key component of the
prostaglandin ligand/receptor binding.15 The critical
receptor residues hypothesized to be involved in pros-
taglandin binding and function are highlighted in
Table 1.
Various hypotheses for achieving receptor selectivity
were considered, based on a model of the hFP receptor/
PGF2R complex and on sequence alignments of the
prostaglandin family of receptors. It was observed that,
adjacent to the common ionic interaction between the
C1 carboxylic acid and the arginine (amino acid 291 in
hFP) residue on the seventh transmembrane helix,
there appeared to be a unique serine residue in the third
transmembrane helix of the hFP receptor which dif-
ferentiated it structurally from the other prostaglandin
receptors. The sequence difference at this site suggested
that a larger C1 carboxylic acid isostere may be better
tolerated in the hFP receptor and provide improved
selectivity for the FP receptor relative to the other
receptors. Docking studies suggested a favorable dipole
interaction between the unique serine 118 and a phos-
phinic acid replacement at the 1-position, thus offering
differentiation between the FP receptor and the other
prostaglandin receptors, which all possessed a phenyl-
alanine at this location, and thus a potential for a less
favorable interaction. As an example, the key interac-
tions between analogue 4b and the hFP receptor site
are illustrated in the surface model Figure 1a and the
three-dimensional structure model Figure 1b. In these
models, the alkyl substitution on the phosphinic acid
moiety points directly toward helix 3 (not shown except
for serine 118 and methionine 115 for the hFP receptor)
and in close proximity to the guanidinium group of the

PGF Derivatives in Osteoporosis Treatment
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4159
Sch em e 1a
a
Key: (a) 30 equiv of Ac2O, Et3N, DMAP, CH2Cl2, room
temperature; then saturated aqueous Na2CO3, 99%. (b) (COCl)2,
DMF, CH2Cl2; then CF3CH2I, DMAP, sodium salt of N-hydroxy-
pyridine-2-thione, CH2Cl2, hν, reflux, 65%. (c) P(OEt)2R′, reflux,
50-85%. (d) NaOH, EtOH-H2O, 25-70%.
from helix 3 with both the R and ω chain of the
prostaglandin. Finally, the hypothesized sequence align-
ments of helix 5 residues suggest the importance of the
phenylalanine/tyrosine hydrophobic interaction in both
receptors with the “tail” of the prostaglandin, thus
stabilizing the active conformation of the receptor for
signal transduction (Figures 1a, 2a).
We subsequently designed a series of phosphinic acid
isosteres at the 1-position to test the validity of the
modeling hypothesis. We chose PGF2R and the 16-
phenoxy PGF2R series, based on the commercial avail-
ability of the starting materials, as well as previous
work supporting the selectivity and potency of the 16-
16
F igu r e 2. (a) Compound 4b docked in the surface model of
the EP1 receptor. (b) Three-dimensional structural model of
compound 4b/hEP1 receptor binding site.
phenoxy tail for the hFP receptor.
Ch em istr y
Compounds 10-15 and 4b were assembled in a
straightforward fashion using methodology previously
developed in conjunction with the synthesis of the
conserved arginine 291, thus implicating a potential
restriction of chain length to smaller substituents on
phosphorus. In the case of the other receptors modeled,
the positioning of the bulky phenylalanine (amino acid
120 in EP1) may result in an unfavorable interaction
between the FP-selective compound (4b) and the recep-
tor site (represented by the hEP1 in Figure 2a,b).
The other key interactions are also illustrated in
Figures 1 and 2. On helix 7, one turn away from the
arginine 291 in the hFP receptor, is a hydrogen bond
specificity site in most receptors for the C-9 hydroxyl
or carbonyl group of the prostaglandin ligand (Figure
1b). The structural difference between the serine (resi-
due 341 on EP1, Figure 2b) and threonine (residue 294
on hFP, Figure 1b) at this site in the receptor is
hypothesized to influence the ligand conformation of the
R chain and the resulting orientation of the headgroup
containing the C-9 hydrogen bonding site. Helix 6
contains two representative hydrogen bonding sites in
the FP (serine 258, 263 in Figure 1b) and EP1 (serine
306, 311 in Figure 2b) receptors for the C-11 and C-15
hydrogen bonding sites of the prostaglandin. In addition,
the other hypothesized hydrogen bonding locations/
interactions on helix 6 are also included (Table 1). The
important hydrophobic residues involved between the
respective receptors and ligand 4b are illustrated in
Figures 1a and 2a, including the methionine residue
phosphonic acid derivatives of prostaglandins F1R and
17
F2R
.
Accordingly, the commercially available prosta-
glandins (1a , 1b) were protected as the tris-acetates (2a ,
2b), and then subjected to a decarboxylation/iodination
protocol using a modified Barton-Hundsdieker reac-
tion18 to provide the primary iodides (3a , 3b). In a
subsequent step, the iodide was treated with a variety
of alkyl phosphinites via Arbuzov reaction19 to provide
the protected prostaglandin derivative, which was sa-
ponified to provide the appropriate C1 phosphinic acids
of 4a or 4b (Scheme 1).
For larger quantities of compound 4b needed for in
vivo studies, a more traditional approach involving
Wittig chemistry for installation of the C1 phosphinic
acid side chain proved to be the most efficient (Scheme
20
2). Consequently, commercially available 5 was pro-
tected with excess tert-butyldimethylsilyl trifluoro-
methanesulfonate21 and reduced with Dibal-H22 pro-
vided lactol (6). The lactol 6 subsequently underwent
Wittig condensation with (3-carboxypropyl)triphenylphos-
phonium bromide23 to provide, after esterification with
trimethylsilyl diazomethane, the protected prostaglan-
din nucleus as an undetermined mixture of C9-C15 and
C
11-C15 bis-silyl ethers due to transfer of the tert-

4160 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
Soper et al.
Sch em e 2a
the evaluation of C1 sulfonamides of the PGE family.25
Based on the data obtained from these calculations,26
there appeared to be no clear correlation between in
vitro binding affinity and acidity at the P1 position of
the analogues tested (Table 2). This lack of correlation
suggests that other factors, such as steric and dipole
interactions originally proposed by modeling, may play
a more significant role in the binding and selectivity of
ligands of this type.
We chose to evaluate the methyl phosphinic acid
analogue 4b in vivo using the ovariectomized rat model
(OVX).27 Female Sprague-Dawley rats (Simonsen Labs,
Gilroy, CA) were received at 3 months of age and
allowed to age to 6 months. Rats were ovariectomized
at 6 months and allowed to lose bone for 60 days before
treatment with compound. Animals then treated for 60
days using doses of 10, 30, 100, 300, 1000, and 3000
µg/kg of compound 4b in a once daily subcutaneous
injection. The positive controls included two agents
known to increase bone formation, volume, and density;
the nonselective prostaglandin agonist PGE2, and the
FP-selective agonist Cloprostenol. In addition, the bis-
phosphonate Alendronate28 was used as a representa-
tive of an antiresorptive agent which maintains or
increases bone density through decreasing the rates of
both bone formation and resorption.
a
Key: (a) 2.2 equiv of TBDMSOTf, 2,6-lutidine, CH2Cl2, -78
°C-rt; then 1 N HCl, 91%; (b) 2.2 equiv of Dibal-H, THF, -78 °C,
80%; (c) 4.4 equiv of NaHMDS, 2.2 equiv of BrP(Ph)3+(CH2)3COOH,
THF, 88%; (d) 5 equiv of TBDMSOTf, 10 equiv of 2,6-lutidine,
CH2Cl2, 0 °C-rt; then 1 N HCl, 92%; (e) TMSCH2N2, Et2O, quant.;
(f) LiAlH4, Et2O, 95%; (g) Ph3P‚Br2, CH3CN, 95%; (h) CH3P(OEt)2,
toluene, 98%.
butyldimethylsilyl protecting group under the reaction
conditions. Only the C5-C6 (Z)-olefin geometry was
observed of each silyl regioisomer. After exhaustive
silylation of the product mixture to provide the ester 7,
reduction of the ester to the primary alcohol, bromina-
tion with triphenylphosphine-dibromide,24 and treat-
ment with diethylmethylphosphinite provided 8. Final
deprotection in sequential fashion using HF/pyridine
and then NaOH/EtOH/H2O provided the 16-phenoxy
analogue 4b in gram scale quantities.
The effect of 4b was measured at several sites in an
attempt to fully understand the anabolic response of
cancelous and cortical sites, and the tissue level mech-
anisms of the response.29 Initially, bone mineral density
(BMD) was measured in the distal femur using dual-
energy X-ray absorptiometry (Figure 3). The methyl
phosphinic acid (4b) produced a significant, dose-de-
pendent increase in bone mineral density at the four
highest doses (100-3000 µg/kg) as compared to final
OVX. At the two highest doses, the increase was
statistically significant when compared to the pretreat-
ment group as well (p < 0.005). The ED50 for compound
4b at this site was calculated to be 146 µg/kg (95% CI
) 58-364 µg/kg). Although the effect was not as strong
as with PGE2, compound 4b did return bone mineral
density to levels greater than the bisphosphonate and
similar to the intact vehicle treated group. In addition,
µCT 30 (microscopic computerized tomography, Table 3)
was used in the vertebrae to assess cancelous bone
volume and architecture. The methylphosphinic acid
induced a dose-dependent increase in bone volume (ED50
) 30 µg/kg, 95% CI ) 17-53 µg/kg), with increases in
trabecular thickness and number, but not connectivity
density (a more accurate index of the number of trabe-
culae) compared to the final OVX group (Tables 4 and
5). At 100 µg/kg the methyl phosphinic acid (4b)
returned bone volume to levels similar to the final intact
group, and 300 and 3000 µg/kg doses increased bone
volume above the starting values of the pretreatment
OVX group (p < 0.05). The methyl phosphinic acid also
increased cancelous bone volume in tibial metaphysis
with increases in trabecular thickness and number.
These increases were brought about by decreases in the
rate of bone turnover (BFR/BV) compared to the high
levels induced by ovariectomy.
Resu lts a n d Discu ssion
Compounds were tested in vitro in the human FP
receptor binding assay as described previously12 and
progressed for screening at the other prostaglandin
receptors when binding affinity < 250 nM at the FP
receptor was achieved. The results are shown in Table
1. As a direct comparison, PGF2R 9 and 16-phenoxy
tetranor PGF2R 16 were tested as benchmarks for these
studies. As seen from the compounds tested, the modi-
fication of substituents on phosphorus has a dramatic
effect on the binding affinity of individual compounds
for the FP receptor. For example, a comparison of 13
with either 12, 14, or 15 suggests that binding affinity
decreases with increasing steric bulk on phosphorus,
with the H-P substitution pattern being most potent
(H-P > CH3-P > Et-P > nBu-P). In addition, the
phosphinic acid series proved more potent than the
phosphonic acid, as seen by comparison of 12 or 13
versus 10. Most importantly, incorporation of the alkyl-
phosphinic acid resulted in a dramatic increase in
selectivity for the FP receptor when compared to binding
affinity at the other human prostaglandin receptors
tested. Thus, when comparing PGF2R 9 and the phos-
phinic acid 13, a marked increase in selectivity for the
FP receptor was seen for the later compound, while
PGF2R also possessed binding affinity for the EP1, EP3,
EP4, and DP receptors. This trend also was observed
when comparing the 16-phenoxy tetranor analogues 16
vs 4b, again with the latter compound being consider-
ably more selective for the FP receptor.
These results indicate that compound 4b produced an
anabolic response in the aged, ovariectomized rat model
and suggests that compounds of this class may be useful
We attempted to correlate FP receptor binding affinity
with ligand pKa, as has been implicated previously in

PGF Derivatives in Osteoporosis Treatment
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4161
Ta ble 2. In Vitro Binding and pKa Data Comparison of Prostaglandin F Derivatives and Phosphinic Acid Analoguesa
a
NT ) not tested. An asterisk marks values evaluated for functional activity in RAT-1 cells, transiently transfected with human
receptor having a constitutively expressed â-galactosidase reporter gene stably transfected within. Result is reported as EC50 shown
after timed hydrolysis of galactosidase pseudosubstrate. For details see: Messier, T.; Dorman, C. M.; Brauner, H.; Eubanks, D.; Brann,
M.R. Pharmacol. Toxicol. 1995, 76, 308.

4162 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
Soper et al.
formed on glass mounted silica gel plates (200-300 mesh;
Baker) and visualized using UV, 5% phosphomolybdic acid in
EtOH, or ammonium molybdate/cerric sulfate in 10% aqueous
H2SO4.
Ra d ioliga n d Bin d in g Assa y. COS-7 cells were transiently
transfected with a hFP recombinant plasmid using Lipo-
fectAMINE Reagent. The transfected cells were washed 48 h
later with Hank’s Balanced Salt Solution (HBSS, without
CaCl2, MgCl2, MgSO4, or phenol red). The cells were detached
with versene, and HBSS was added. The mixture was centri-
fuged at 200g for 10 min, at 4 °C to pellet the cells. The pellet
was resuspended in phosphate-buffered saline-EDTA buffer
(PBS; 1 mM EDTA; pH 7.4; 4 °C). The cells were disrupted by
nitrogen cavitation (Parr Model 4639), at 800 psi, for 15 min
at 4 °C. The mixture was centrifuged at 1000g for 10 min at
4 °C. The supernatant was centrifuged at 100000g for 60 min
at 4 °C. The pellet was resuspended to 1 mg of protein/mL of
TME buffer (50 mM Tris; 10 mM MgCl2; 1 mM EDTA; pH 6.0;
4 °C) based on protein levels measured using the Pierce BCA
Protein Assay kit. The homogenate was mixed using a Kine-
matica Polytron for 10 s. The membrane preparations were
then stored at -80 °C, until thawed for assay use.
The receptor competition binding assays were developed in
a 96 well format. Each well (n ) 3) contained 100 µg of hFP
membrane, 5 nM [3 H] PGF2R, and the various competing
compounds in a total volume of 200 µL. The plates were
incubated at 23 °C for 1 h. The incubation was terminated by
rapid filtration using the Packard Filtermate 196 harvester
through Packard UniFilter GF/B filters that were prewetted
with TME buffer. The filter was washed four times with TME
buffer. Packard Microscint 20, a high-efficiency liquid scintil-
lation cocktail, was added to the filter plate wells and the
plates remained at room temperature for 3 h prior to counting.
The plates were read on the Packard TopCount Microplate
Scintillation counter.
The hEP1 receptor was transiently expressed in COS-7 cells
and followed the above-mentioned transient transfection pro-
tocol. Five of the other receptors, hEP2, hEP3, hEP4, hTP, and
hIP, were stably expressed in the CHO cell line. The hDP
receptor was stably expressed in the HEK-293 cell line and
required a TME buffer pH of 7.4. The CHO cells and HEK-
293 cells expressing the receptors were collected 4-5 days after
seeding, were washed with PBS and detached with versene.
Cells were centrifuged at 100g for 10 min at 4 °C. Cells were
resuspended in lysis buffer (10 mM Tris-HCl; 5 mM EDTA;
10 µg/mL leupeptin and benzamidine; 2µg/mL aprotinin) and
the homogenate was centrifuged at 200g for 10 min at 4 °C.
The supernatant was centrifuged at 100000g for 60 min at 4
°C. The pellet was resuspended to 1 mg of protein/mL of TME
buffer based on measured protein levels. The homogenate was
mixed using a Kinematica Polytron for 10 s. The membrane
preparations were then stored at -80 °C, until thawed for
assay use. The radiolabeled ligands used for the competition
binding assays of the other seven receptors were as follows:
the EP1, EP2, EP3, and EP4 assays used [3H]PGE2, the DP
assay used [3H]PGD2, the IP assay used [3H]Iloprost, and the
TP assay used [3H]SQ-29548. All of the radiolabeled ligands
were run at 5 nM in the experiment and were purchased from
DuPont NEN and Amersham LIFE SCIENCE.
F igu r e 3. Vertebral cancelous bone volume for 4b.
in the treatment of bone degenerative disease, including
osteoporosis.
In conclusion, we have designed and prepared a series
of alkyl phosphinic acid analogues of PGF2R and have
evaluated them in vitro in an hFP receptor binding
assay and in vivo in the ovariectomized rat model. We
have found that, in general, binding affinity for the hFP
receptor decreases with increasing size of the alkyl
substituent on phosphorus. We have also found that
replacement at C1 with a phosphinic acid moiety results
in greater receptor selectivity for the FP receptor when
compared to the C1 carboxylic acid.31 Finally, we have
found that the methyl phosphinic acid (4b) produced a
dose-dependent increase in bone mineral density when
compared to both an antiresorptive agent and to the
final intact rat group. We are continuing to explore the
implications of these results within the realm of alter-
native treatments for bone degenerative diseases in-
cluding osteoporosis, as well as other therapeutic areas
where selective prostaglandin analogues may be desir-
able.
Exp er im en ta l Section
Gen er a l Meth od s. 1H NMR spectra were recorded on a
Varian Unity Plus 300 MHz spectrometer and are referenced
to either the deuteriochlororform singlet at 7.27 ppm or
deuteriomethanol singlet at 4.87 ppm. 13C spectra were
obtained on a Varian Unity Plus 300 MHz spectrometer and
are referenced at either the center line of the deuteriochloro-
form triplet at 77.0 ppm or the deuteriomethanol heptet at
49.15 ppm. Infrared absorption spectra were obtained on a
Perkin-Elmer Model 197 spectrophotometer and are referenced
to polystyrene (1601 cm-1). Mass spectra were obtained on
either a Fison Platform-II Quadrupole Mass Spectrometer or
2-Decar boxy-2-(P -m eth ylph osph in ico)pr ostaglan din F2r
(12). A mixture of 2-decarboxy-2-iodoprostaglandin F2R triac-
etate (3a , 46.1 mg, 82 µmol), diethyl methylphosphonite (0.2
mL, 1.32 mmol), and toluene (0.41 mL) was stirred at 100 °C
for 6 h. The solvents were evaporated, and the residue was
purified by dry-flash column chromatography (SiO2, 5-25%
2-propanol in 20% dichlormethane-hexane). Appropriate frac-
tions were concentrated, diluted with water (15 mL), and
extracted with ether (3 × 5 mL). The combined extracts were
washed with water (5 mL), dried (brine, Na2SO4), and evapo-
rated to give 2-decarboxy-2-(O-ethyl-P-methylphosphinico)-
prostaglandin F2R triacetate (38.1 mg, 86%). 1H NMR (300
MHz, CDCl3) δ 5.46-5.55 (m, 2 H), 5.27-5.38 (m, 2 H), 5.17-
5.25 (m, 1 H), 5.07 (t, J ) 4.6 Hz, 1 H), 4.87 (ddd, J ) 8.8, 7.6,
4.4 Hz, 1H), 4.03 (dq, J ) 7.0, 7.0 Hz, 2 H), 2.46-2.60 (m, 2
a
Fison Trio2000 Quadrupole Mass Spectrometer. High-
Resolution mass spectra were obtained from the Procter &
Gamble CRD Mass Spectrometry Lab (M. Lacey). Elemental
analyses were obtained from the Procter & Gamble EA Lab
in Norwich, NY. Melting points were determined in open Pyrex
capillary tubes on a Thomas-Hover Unimelt apparatus. Melt-
ing points and boiling points are uncorrected. All solvents were
purchased anhydrous (Aldrich Chemical) and used without
further purification. All air-sensitive reactions were performed
under an anhydrous nitrogen atmosphere. Compounds 9 and
16 were purchased from Cayman Chemical and used without
further purification. Compounds 10 and 11 were synthesized
previously.17 Flash chromatography was performed on silica
gel (70-230 mesh; Aldrich) or (230-400 mesh; Merck) as
appropriate. Thin-layer chromatography analysis was per-

PGF Derivatives in Osteoporosis Treatment
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4163
Ta ble 3. Vertebral Bone Volume and Architecture Measured Using µCTa
bone vol/tissue
vol (%)
derived trabecular
thickness (um)
derived trabecular
no. (mm-1
directly measd
thickness (um)
connectivity
group
n
)
dens (mm-3
)
baseline
pretreatment intact
final intact
pretreatment OVX
final OVX
6
6
8
6
5
3
6
6
4
4
4
4
4
3
39.52 ( 0.96a
40.70 ( 2.60a
39.38 ( 2.17a
35.78 ( 3.12a
31.70 ( 1.78
43.79 ( 3.46a
36.67 ( 2.10a
42.04 ( 1.90a
31.04 ( 1.26
34.58 ( 2.49b
38.04 ( 2.72a
38.61 ( 2.33a
36.98 ( 0.94a
39.48 ( 0.52a
73.62 ( 4.22
78.29 ( 2.86b
78.09 ( 1.63b
73.85 ( 3.03
74.81 ( 1.81
95.14 ( 3.93a
76.62 ( 2.34
85.81 ( 1.77a
73.01 ( 2.09
77.77 ( 1.41
78.71 ( 1.94b
79.46 ( 3.07a
80.62 ( 1.45a
83.76 ( 1.44a
5.38 ( 0.24a
5.20 ( 0.22a
5.04 ( 0.21a
4.84 ( 0.30a
4.24 ( 0.26
4.60 ( 0.18
4.79 ( 0.31a
4.90 ( 0.16a
4.26 ( 0.17
4.44 ( 0.27
4.83 ( 0.26a
4.86 ( 0.25a
4.59 ( 0.19b
4.71 ( 0.12b
121.17 ( 5.46
127.50 ( 5.75b
127.63 ( 3.11c
121.33 ( 4.72
121.60 ( 3.91
149.00 ( 6.56a
123.83 ( 2.40
137.60 ( 3.91a
118.50 ( 2.65
125.50 ( 2.65
127.25 ( 3.40b
127.00 ( 4.08b
127.50 ( 1.73b
134.00 ( 2.00a
92.32 ( 19.33a
75.56 ( 3.19a
70.75 ( 7.53b
77.06 ( 8.24a
60.21 ( 8.08
47.81 ( 2.69b
68.78 ( 9.57
61.13 ( 7.00
60.28 ( 8.57
59.20 ( 5.21
68.10 ( 5.10
66.71 ( 7.59
59.66 ( 5.95
55.90 ( 5.49
PGE2, 6000 µg/kg
Alendronate, 3 µg/kg
Cloprostenol, 250 µg/kg
4b, 10 µg/kg
4b, 30 µg/kg
4b, 100 µg/kg
4b, 300 µg/kg
4b, 1000 µg/kg
4b, 3000 µg/kg
a
b
p e 0.005 vs final OVX. p e 0.05 vs final OVX. c p e 0.01 vs final OVX.
Ta ble 4. Bone Volume and Formation Rates in the Proximal Tibial Metaphysis Measured Histomorphometrically
mineralizing
surface/bone
surface (%)
mineral
apposition
rate (mm/day)
bone formation bone formation
bone vol/ tissue
vol (%)
trabecular
trabecular
rate/tissue
rate/bone
group
n
thickness (mm) no. (mm-1
)
vol (%/year)
vol (%/year)
baseline
pretreatment intact
final intact
pretreatment OVX
final OVX
6
6
8
6
6
6
6
5
4
4
4
4
4
3
12.35 ( 3.44a
11.71 ( 2.07a
11.15 ( 2.28a
5.74 ( 1.21c
2.44 ( 0.73
11.65 ( 2.30a
6.49 ( 3.16a
9.50 ( 1.63a
2.23 ( 1.48
3.63 ( 1.25
5.84 ( 1.05c
5.72 ( 2.28b
9.43 ( 3.33a
8.51 ( 1.69a
43.99 ( 4.06b
43.92 ( 2.75
45.10 ( 3.93b
44.21 ( 5.60b
36.73 ( 6.50
66.14 ( 3.17a
48.64 ( 7.60c
53.90 ( 5.95a
35.68 ( 8.12
41.23 ( 5.80
45.48 ( 5.70b
44.05 ( 3.77
56.68 ( 7.90a
55.84 ( 4.39a
2.77 ( 0.54a 27.19 ( 3.94b
2.66 ( 0.40a 29.52 ( 3.72
2.46 ( 0.40a 25.81 ( 4.02a
1.31 ( 0.29a 34.12 ( 3.51
0.66 ( 0.18 31.65 ( 3.89
1.76 ( 0.29a 36.23 ( 2.32b
1.29 ( 0.51a 17.72 ( 2.46a
1.77 ( 0.30a 22.99 ( 3.57a
0.58 ( 0.29 28.37 ( 2.45
0.87 ( 0.27 30.59 ( 5.44
1.28 ( 0.14a 27.11 ( 4.62b
1.27 ( 0.38a 23.57 ( 2.13a
1.64 ( 0.39a 22.04 ( 2.36a
1.51 ( 0.22a 23.66 ( 3.66a
0.93 ( 0.02b
0.91 ( 0.05c
0.81 ( 0.04a
1.09 ( 0.08
1.07 ( 0.11
1.19 ( 0.05
0.63 ( 0.09a
1.09 ( 0.10
0.99 ( 0.03
1.09 ( 0.22
0.87 ( 0.11a
0.98 ( 0.06
1.06 ( 0.22
0.94 ( 0.03
42.67 ( 10.80a 351.66 ( 55.24a
43.74 ( 10.43a 371.09 ( 41.06a
31.74 ( 9.55a
30.35 ( 5.99a
13.68 ( 3.91
283.23 ( 49.86a
538.26 ( 88.19
576.20 ( 126.60
PGE2, 6000 µg/kg
Alendronate, 3 µg/kg
Cloprostenol, 250 µg/kg
4b, 10 µg/kg
45.62 ( 12.58a 394.70 ( 13.53c
8.81 ( 3.62
26.44 ( 2.27a
9.81 ( 5.12
17.55 ( 7.45
18.23 ( 3.90
18.03 ( 6.25
23.44 ( 8.80b
20.13 ( 5.49
144.92 ( 47.06a
286.86 ( 63.28a
498.01 ( 133.83
495.65 ( 139.75
320.16 ( 84.37a
322.61 ( 52.05a
251.82 ( 56.88a
247.55 ( 28.77a
4b, 30 µg/kg
4b, 100 µg/kg
4b, 300 µg/kg
4b, 1000 µg/kg
4b, 3000 µg/kg
a
b
p e 0.005 vs final OVX. p e 0.05 vs final OVX. c p e 0.01 vs final OVX.
Ta ble 5. Bone Volume and Formation Rates in the Tibial Diaphysis Measured Histomorphometrically
periosteal
mineral
apposition
rate (mm/day)
endosteal
mineralizing
surface (%)
endosteal
eroded
surface (%) rate (mm/day)
mineral
apposition
periosteal
mineralizing
surface (%)
tissue
marrow
group
n
area (mm)
area (mm)
baseline
pretreatment intact
final intact
pretreatment OVX
final OVX
6
6
8
6
6
6
6
5
4
4
4
4
4
3
4.61 ( 0.37a 0.76 ( 0.09a 19.09 ( 4.59a
4.91 ( 0.21b 0.87 ( 0.04c 21.86 ( 3.73a
5.06 ( 0.22 0.92 ( 0.09b 23.83 ( 4.59c
2.62 ( 0.82a 0.20 ( 0.31a 50.82 ( 20.77a 0.99 ( 0.22a
2.84 ( 1.40a 0.43 ( 0.47a 31.12 ( 6.59
3.39 ( 1.70a 0.58 ( 0.37b 19.29 ( 15.74
0.36 ( 0.28
0.25 ( 0.36
4.96 ( 0.47b 0.95 ( 0.20 25.21 ( 8.74b 12.00 ( 7.47
0.81 ( 0.15
0.98 ( 0.08
42.49 ( 18.74a 0.89 ( 0.05a
5.30 ( 0.40 1.03 ( 0.14 36.87 ( 6.91
5.22 ( 0.30 0.76 ( 0.08a 61.22 ( 9.24a
5.01 ( 0.44 0.82 ( 0.10a 29.40 ( 4.74
4.78 ( 0.06c 0.70 ( 0.09a 55.18 ( 8.18a
9.52 ( 3.66
19.37 ( 4.77
0.28 ( 0.31
1.14 ( 0.23a
0.46 ( 0.36
0.00 ( 0.00
0.49 ( 0.35
0.31 ( 0.62
0.38 ( 0.44
0.19 ( 0.38
0.00 ( 0.00
0.00 ( 0.00
PGE2, 6000 µg/kg
Alendronate, 3 µg/kg
Cloprostenol, 250 µg/kg
4b, 10 µg/kg
0.79 ( 0.71a 1.33 ( 0.26
58.37 ( 7.13a
1.99 ( 1.36a 0.35 ( 0.39a 29.59 ( 12.38
2.22 ( 1.07a 1.60 ( 0.40a
0.71 ( 0.49
2.01 ( 0.98a 0.88 ( 0.07
9.91 ( 3.60
25.86 ( 10.11
26.23 ( 28.42
5.09 ( 0.27 0.93 ( 0.07 46.62 ( 15.10 11.83 ( 5.81
4b, 30 µg/kg
5.19 ( 0.11 0.87 ( 0.18b 28.72 ( 8.90
4b, 100 µg/kg
4.95 ( 0.17 0.82 ( 0.05a 22.28 ( 14.84c 2.30 ( 1.21a 0.42 ( 0.49a 15.21 ( 7.91
4b, 300 µg/kg
5.01 ( 0.29 0.89 ( 0.14b 29.02 ( 10.71
4.92 ( 0.17b 0.86 ( 0.11b 35.29 ( 9.19
4.88 ( 0.31b 0.74 ( 0.10a 29.40 ( 4.74
1.32 ( 0.49a 0.53 ( 0.63b
2.74 ( 1.22a 1.01 ( 0.13
2.20 ( 0.34a 1.03 ( 0.40
9.56 ( 4.93
10.35 ( 2.13
11.09 ( 6.03
4b, 1000 µg/kg
4b, 3000 µg/kg
a
b
p e 0.005 vs final OVX. p e 0.05 vs final OVX. c p e 0.01 vs final OVX.
H), 1.98-2.13 (m, 4 H), 2.04, 2.03, 2.00 (3 s, 3 H each), 1.50-
1.75 (m, 8 H), 1.42 (d, J ) 13.5 Hz, 3 H), 1.30 (t, J ) 7.0, 7.0
Hz, 3 H0, 1.20-1.30 (m, 6 H), 0.87 (t, J ) 6.6 Hz, 3 H); 13C
NMR (75 MHz, CDCl3) δ 170.5, 170.3, 170.2, 131.9, 131.5,
129.5, 128.2, 77.7, 74.2, 73.9, 59.9 (d, J ) 6 Hz), 51.8, 47.4,
38.8, 34.2, 31.4, 29.2 (d, J ) 94 Hz), 28.0 (d, J ) 16 Hz), 24.9,
24.7, 22.4, 22.0 (br), 21.3, 21.2, 21.0, 16.5 (d, J ) 6 Hz), 13.9,
13.5 (d, J ) 91 Hz); 31P NMR δ 55.3. m/z (FAB) 565 (M + Na,
100), 483 (95), 363 (55); Found (FAB) MNa+ 565.2905, C28H47O8-
PNa requires MNa+ 565.2905.
A mixture of 2-decarboxy-2-(O-ethyl-P-methylphosphinico)-
prostaglandin F2R triacetate (38.1 mg, 70 µmol), 95% ethanol
(2.8 mL), and 2.5 M aqueous sodium hydroxide (2.8 mL, 7.0
mmol) was stirred at reflux for 3 h. The ethanol was evapo-
rated and the residue was diluted with water (5 mL) and
washed with ethyl acetate (2 × 3 mL). The combined washes
were extracted with water (2 mL). The combined aqueous
phases were acidified with 2.5 M aqueous hydrochloric acid
and extracted with ethyl acetate (5 × 5 mL). The combined
extracts were dried (brine, Na2SO4) and evaporated to give
2-decarboxy-2-(P-methylphosphinico)prostaglandin F2R (12).
(24.2 mg, 89%). 1H NMR (300 MHz, CDCl3) δ 5.56 (dd, J )
15.2, 7.1 Hz, 1 H), 5.47 (dd, J ) 15.2, 8.5 Hz, 1 H), 5.32-5.44
(m, 2 H), 4.11 (br t, 1 H), 4.05 (q, J ) 6.7 Hz, 1 H), 3.94 (br s,
1 H), 3.73 (br s, 4 H), 2.30-2.42 (m, 3 H), 2.19-2.28 (m, 1 H),
1.99-2.06 (m, 2 H), 1.49-1.82 (m, 6 H), 1.45 (d, J ) 13.2 Hz,
3 H), 1.23-1.41 (m, 6H), 0.89 (t, J ) 6.7 Hz, 3 H); 13C NMR
(75 MHz, CDCl3) δ 135.5, 133.0, 129.8, 129.1, 77.5, 73.2, 72.1,
55.3, 49.9, 42.9, 36.9, 31.7, 27.9 (br d, J ) 12 Hz), 25.4, 25.3,
22.6, 22.1 (br), 14.0 (signals for carbons adjacent to P not

4164 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
Soper et al.
observed); 31P NMR δ 56.6; MS m/z 411 (M + Na, 100), 393
(5), 371 (10), 335 (10); FAB 433 (M - H + 2Na); Found (FAB)
MNa+ 411.2264, C20H37O5P + Na+ requires 411.2276.
H), 4.88 (ddd, J ) 8.8, 7.5, 4.4 Hz, 1 H), 4.05 (dq, J ) 7.1, 7.1
Hz, 2H), 2.47-2.60 (m, 2 H), 1.99-2.16 (m, 4 H), 2.06, 2.05,
2.02 (3 s, 3 H each), 1.49-1.75 (m, 12 H), 1.36-1.47 (m, 2 H),
1.19-1.35 (m, 6 H), 1.30 (t, J ) 7.1 Hz, 3 H), 0.93 (t, J ) 7.3
Hz, 3 H), 0.88 (t, J ) 6.7 Hz, 3 H); 13C NMR (75 MHz, CDCl3)
δ 170.6, 170.3, 170.2, 132.0, 131.5, 129.6, 128.22, 77.7, 74.3,
73.9, 59.9 (d, J ) 6 Hz), 51.8, 47.4, 38.9, 34.3, 31.5, 28.1 (d, J
) 16 Hz), 27.7 (d, J ) 92 Hz), 27.6 (d, J ) 91 Hz), 25.0, 24.8,
24.0 (d, J ) 8 Hz), 23.9 (d, J ) 4 Hz), 22.5, 21.3, 21.2, 21.0,
16.7 (d, J ) 5 Hz), 14.0, 13.6; 31P NMR δ 58.8; m/z (FAB) 607
(M + Na, 100%), 525 (80), 405 (35); Found (FAB) MNa+
607.3400, C31H53O8PNa requires MNa+ 607.3376.
Sod iu m Sa lt of 2-Deca r boxy-2-(P -eth ylp h osp h in ico)-
p r osta gla n d in F 2r (14). A mixture of 2-decarboxy-2-iodopros-
taglandin F2R triacetate (3a , 24.1 mg 0.43 mmol), diethyl eth-
ylphosphonite (72 µL, 0.43 mmol), and toluene (0.21 mL) was
stirred at 100 °C for 8 h. The solvents were evaporated, and
the residue was purified by dry-flash column chromatography
(SiO2), 5-25% 2-propanol in 20% dichloromethane-hexane).
Appropriate fractions were concentrated, diluted with water
(15 mL), and extracted with ether (3 × 5 mL). The combined
extracts were washed with water (5 mL), dried (brine,
Na2SO4), and evaporated to give 2-decarboxy-2-(O-P-dieth-
ylphosphinico)prostaglandin F2R triacetate (18.0 mg, 79%).
1H NMR (300 MHz, CDCl3) δ 5.47-5.57 (m, 2 H), 5.29-5.38-
(m, 2 H), 5.18-5.24 (m, 1 H), 5.07 (t, J ) 4.6 Hz, 1 H, 4.87
(ddd, J ) 8.9, 7.5, 4.4 Hz, 1 H), 4.04 (dq, J ) 7.1, 7.1 Hz, 2 H),
2.48-2.59 (m, 2 H), 2.05-2.16 (m, 4 H), 2.05, 2.04, 2.01 (2 s,
2 H each), 1.50-1.77 (m, 10 H), 1.22-1.34 (m, 6 H), 1.30
(t, J ) 7.0 Hz, 3 H), 1.13 (dt, J ) 17.8, 7.7 Hz, 3 H), 0.86 (t, J
) 6.7 Hz, 3 H). 13C NMR (75 MHz, CDCl3) δ 170.5, 170.3,
170.2, 132.0 131.5, 129.6, 128.2, 77.7, 64.2, 73,0, 59.5 (d, J )
6 Hz), 51.8, 47.4, 38.9, 34.3, 31.5, 28.1 (d, J ) 15 Hz), 26.9 (d,
J ) 90 Hz), 24.9, 24.7, 22.5, 21.7 (br), 21.3, 21.2, 21.0, 20.9 (d,
J ) 91 Hz), 16.7 (d, J ) 5 Hz), 13.9, 6.0 (d, J ) 3 Hz); 31P
NMR δ 60.0; m/z (FAB) 579 (M + Na, 100%), 497, (45), 377
(45); Found (FAB) MNa+ 579.3067, C29H49O8PNa requires
MNa+ 579.3063.
A mixture of 2-decarboxy-2-(P-butyl-O-ethylphosphinico)-
prostaglandin F2R triacetate (11.7 mg, 21 µmol), 95% ethanol
(0.9 mL), and 2.5 M aqueous sodium hydroxide (0.9 mL, 2.3
mmol) was stirred at reflux for 2 h. The ethanol was evapo-
rated, and the residue was diluted with water (7 mL) and
washed with ethyl acetate (2 × 3 mL). The combined washes
were extracted with water (1 mL). The combined aqueous
phases were acidified with 2.5 M aqueous hydrochloric acid
and extracted with ethyl acetate (4 × 5 mL). The combined
extracts were dried (brine, Na2SO4) and evaporated. The
residue was taken up in methanol (2 mL) and stirred with
Amberilite CG-50 (Na+ form, 0.45 g) for 30 min. The resin was
removed by filration, and the filtrate was concentrated to give
the sodium salt of 2-decarboxy-2-(P-butyphosphinico)prosta-
1
glandin F2R (15, 7.0 mg, 74%). H NMR (300 MHz, CD3OD) δ
5.41-5.55 (m, 3 H), 5.31-5.39 (m, 1 H), 4.08 (td, J ) 5.3, 2.2
Hz 1 H), 3.99 (q, J ) 6.5 Hz, 1 H), 3.82 (ddd, J ) 8.4, 7.4, 5.4
Hz, 1 H), 2.34 (ddd, J ) 14.4, 8.6, 5.8 Hz, 1 H), 2.10-2.29 (m,
3 H), 2.10 (sept., J ) 6.8 Hz, 2 H), 1.25-1.63 (m, 20 H), 0.92
(t, J ) 7.1 Hz, 3 H), 0.91 (t, J ) 7.2 Hz, 3 H); 13C NMR (75
MHz, CD3OD) δ 136.5, 134.3, 130.8, 130.0, 77.9, 74.0, 72.1,
56.2, 50.9, 44.3, 38.4, 33.0, 31.4 (d, J ) 91 Hz), 31.3, (d, J )
91 Hz), 29.9 (d, J ) 16 Hz), 26.4, 26.3 (d, J ) 3 Hz), 26.2, 25.6
(d, J ) 16 Hz), 24.2 (d, J ) 3 Hz), 23.8, 14.4, 14.2; 31P NMR δ
43.7; m/z (negative ion FAB) 429 (M - Na, 100), 329 (15);
Found (FAB) (M+) 429.2749, C23H42O5P requires (M+) 429.2770.
A mixture of 2-decarboxy-2-(O, P-diethylphosphinico)pros-
taglandin F2R triacetate (17.8 mg, 32 µmol), 95% ethanol (1.3
mL), and 2.5 M aqueous sodium hydroxide (1.3 mL, 3.2 mmol)
was stirred at reflux for 3 h. The ethanol was evaporated, and
the residue was diluted with water (5 mL) and washed with
ethyl acetate (2 × 3 mL). The combined washes were extracted
with water (1 mL). The combined aqueous phases were
acidified with 2.5 M aqueous hydrochloric acid and extracted
with ethyl acetate (4 × 5 mL). The combined extracts were
dried (brine, Na2SO4) and evaporated. The residue was taken
up in methanol (3 mL) and stirred with Amberlite CG-50 (Na+
form, 0.4 g) for 25 min. The resin was removed by filtration
and the filtrate was concentrated to give the sodium salt of
2-decarboxy-2-(P-ethylphosphinico)prostaglandin F2R (14, 13
mg, 100%). 1H NMR (300 MHz, CD3OD) δ 5.31-5.50 (m, 4 H),
4.09 (td, J ) 16.4, 7.7 Hz, 3 H), 3.99 (q, J ) 6.4 Hz, 1H), 3.82
(ddd, J ) 8.4, 7.3, 5.3 Hz, 1H), 2.32 (ddd, J ) 14.4, 8.5, 7.8
Hz, 1H), 2.30-2.03 (m, 3H), 1.62-1.26 (m, 16H), 1.06 (dt, J )
16.4, 7.7 Hz, 3H), 0.90 (t, J ) 6.9 Hz, 3 H); 13C NMR (75 MHz,
CD3OD) δ 136.5, 134.2, 130.8, 130.0, 77.9, 73.9, 72.1, 56.2, 50.0,
44.3, 38.4, 33.0, 30.6 (d, J ) 91 Hz), 29.9 (d, J ) 16 Hz), 26.4,
26.3, 24.2 (d, J ) 3 Hz), 24.0 (d, J ) 92 Hz), 23.8, 14.4, 7.4 (d,
J ) 5 Hz); 31P NMR δ 45.0; m/z (negative ion FAB) 423 (M,
10%), 401 (M - Na, 100), 361 (10), 329 (10); Found (FAB) (M
- Na)- 401.2457.
Sod iu m Sa lt of 2-Deca r boxy-2-(P -bu tylp h osp h in ico)-
p r osta gla n d in F 2r (15). A mixture of 2-decarboxy-2-iodopros-
taglandin F2R triacetate (3a , 30.1 mg, 55 µmol), diethyl butyl-
phosphonite (0.11 mL, 0.58 mmol), and toluene (0.30 mL) was
stirred at 100 °C for 7 h. The solvents were evaporated, and
the residue was purified by dry-flash column chromatography
(SiO2, 5-25% 2-propanol-hexane). Appropriate fractions were
concentrated, diluted with water (10 mL), and extracted with
ether (3 × 5 mL). The combined extracts were washed with
water (2 × 5 mL), dried (brine, Na2SO4), and evaporated. The
residue was purified further by dry flash column chromatog-
raphy (SiO2, 5-45% tetrahydrofuran-dichloromethane) and
preparative TLC (SiO2 2% methanol-dichloromethane). Ap-
propriate fractions were concentrated, taken up in 20% aque-
ous methanol (15 mL), and extracted with hexane (3 × 5
mL). The combined extracts were washed with water, dried
(brine, Na2SO4), and evaporated to give 2-decarboxy-2-(P-butyl-
O-ethylphosphinico)prostaglandin F2R triacetate (11.7 mg,
36%). 1H NMR (300 MHz, CDCl3) δ 5.47-5.57 (m, 2 H),
5.29-5.39 (m, 2 H), 5.18-5.25 (m, 1 H), 5.08 (t, J ) 4.7 Hz, 1
Sod iu m Sa lt of 2-Deca r b oxy-2-p h osp h in icop r ost a -
gla n d in F 2r (13). A solution of 2-decarboxy-2-iodoprostaglan-
din F2R triacetate (3a , 69.4 mg, 0.12 mmol) in dichloromethane
(2.1 mL) was degassed through five freeze-pump-thaw cycles
and was then placed under argon. Bis(trimethylsilyl) phos-
phonite (Ca u tion ! Pyrophoric) (2.1 mL, 2.3 mmol) was added,
and the flask was sealed tightly and allowed to stand at room
temperature for 25.5. h. The mixture was transferred, by
means of a syringe, into saturated aqueous sodium hydrogen
carbonate (15 mL) below the surface. The mixture was acidified
with 2.5 M aqueous hydrochloric acid and stirred for 10 min,
before being neutralyzed with saturated aqueous sodium
hydrogen carbonate. The mixture was washed with ethyl
acetate (4 × 7 mL). The combined washes were extracted with
water (10 mL). The combined washed were dried (brine,
Na2SO4) and evaporated to give unchanged starting material
(40.2 mg, 58%). The combined aqueous phases were acidified
with 2.5 M aqueous hydrochloric acid and extracted with ethyl
acetate (5 × 10 mL). The combined extracts were dried (brine,
Na2SO4) and evaporated to give 2-decarboxy-2-phosphinico-
prostaglandin F2R triacetate (14.5 mg, 23%). 1H NMR (300
MHz, CDCl3) δ 9.89 (br s, 1 H), 7.15 (br d, J ) 541.6 Hz, 1 H),
5.47-5.59 (m, 2 H), 5.29-5.41 (m, 2 H), 5.19-5.25 (m, 1 H),
5.09 (t, J ) 4.6 Hz, 1 H), 4,89 (ddd, J ) 8.9, 7.5, 4.4 Hz, 1 H),
2.49-2.61 (m, 2 H), 2.05-2.20 (m, 4 H), 2.06, 2.05, 2.02 (3 s 3
H each), 1.49-1.81 (m, 8 H), 1.23-1.34 (m, 6 H), 0.88 (t, J )
6.7 Hz, 3 H); 13C NMR (75 MHz, CDCl3) δ 170.6, 170.4, 170.3,
131.9, 131.6, 129.3, 128.5, 77.7, 74.3, 74.0, 51.8, 47.4, 38.9, 34.3,
31.5, 27.7 (br), 24.9, 24.8, 22.5 (br), 21.3, 21.2, 21.0, 20.4 (br),
14.0, signal for the carbon adjacent to P not observed; 31P NMR
δ 40.9; m/z (negative ion FAB) 499 (M - H, 100), 354 (5); Found
(FAB) (M - H) 499.2484, C25H41O8P requires (M - H)
499.2461.
A mixture of 2-decarboxy-2-phosphinicoprostaglandin F2R
triacetate (14.5 mg, 29 µmol), methanol (1.2 mL), water (1.2
mL, and 1.08 M aqueous sodium hydroxide (0.2 mL, 0.17

PGF Derivatives in Osteoporosis Treatment
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4165
mmol) was stirred at room temperature for 22 h. Amberilite
CG-50 (H+ form, 0.17 g) was added, and the mixture was
stirred a further 1 h. The resin was removed by filtration, and
the filrate was concentrated and reevaporated from acetonitrile
(2 × 5 mL). The residue was taken up in methanol (3 mL)
and stirred with Amberilite CG-50 (Na+ form, 50 mg) for 30
min. The resin was removed by filtration, and the filtrate was
concentrated to give the sodium salt of 2-decarboxy-2-phos-
phinicoprostaglandin F2R (9, 12 mg, 100%). 1H NMR (300 MHz,
CD3OD) δ 6.96 (dt, J ) 487.1, 1.7 Hz), 5.42-5.55 (m, 3 H),
5.30-5.38 (m, 1 H), 4.08 (td, J ) 5.3, 2.2 Hz 1 H), 3.99 (q, J )
6.4 Hz, 1 H), 3.83 (ddd, J ) 8.5, 7.2, 5.3 Hz, 1 H), 2.03-2.37
(m, 6 H), 1.24-1.64 (m, 14 H), 0.90 (t, J ) 6.9 Hz, 3 H); 13C
NMR (75 MHz, CD3OD) δ 136.5, 134.2, 130.6, 130.2, 77.9, 73.9,
72.2, 56.2, 51.0, 44.3, 38.4, 33.4 (d, J ) 91 Hz), 33.0, 29.4 (d,
J ) 17 Hz), 26.4, 26.3, 23.8, 23.2, 14.4; 31P NMR δ 29.2; m/z
(negative ion FAB) 373 (M - Na, 100%), 329 (25); Found (FAB)
(M + Na)- 373.2142, C19H34NaO5P requires (M + Na)-
373.2144.
A mixture of 2-decarboxy-2-iodo-16-phenoxytetranorpros-
taglandin F2R triacetate (3b, 18.2 mg, 30 µmol), diethyl
methylphosphonite (0.1 mL, 0.66 mmol), and toluene (0.2 mL)
was stirred at 100° C for 7 h. The solvents were evaporated
and the residue was purified by dry-flash column chromatog-
raphy (SiO2, 5-25% 2-propanol in 20% dichloromethane-
hexane). Appropriate fractions were concentrated, diluted with
water (15 mL), and extracted with ethyl acetate (3 × 5 mL).
The combined extracts were washed with water (10 mL), dried
(brine, Na2SO4), and evaporated to give 2-decarboxy-2-(O-
ethyl-P-methylphosphinico)-16-phenoxytetranorprostaglan-
1
din F2R triacetate (10.8 mg, 61%). H NMR (300 MHz, CDCl3)
δ 7.28 (td, J ) 8.6, 7.3 Hz, 2 H), 6.97 (t, J ) 7.3 Hz, 1 H), 6.90
(d, J ) 8.6 Hz, 2 H), 5.65-5.73 (m, 2 H), 5.57-5.62 (m, 1 H),
5.29-5.49 (m, 2 H), 5.09 (t, J ) 4.6 Hz, 1 H), 4.92 (ddd, J )
8.7, 7.5, 4.3 Hz, 1 H), 4.08 (dd, J ) 10.2, 6.1 Hz, 1 H), 4.04
(dq, J ) 7.1, 7.1 Hz, 2 H), 4.02 (dd, J ) 10.2, 4.6 Hz, 1 H),
2.56-2.65 (m, 1 H), 2.53 (ddd, J ) 15.8, 9.02, 5.7 Hz, 1 H),
2.10, 2.06, 2.01 (3 s, 3 H each), 1.99-2.21 (m, 4 H), 1.56-1.74
(m, 6 H), 1.42 (d, J ) 13.4 Hz, 3 H), 1.11 (t, J ) 7.0 Hz, 3 H);
13C NMR (75 MHz, CDCl3) δ 170.6, 170.3, 170.0, 158.4, 134.2,
129.6, 129.5, 128.2, 127.7, 121.2, 114.7, 77.6, 74.3, 71.9, 69.0,
59.9 (d, J ) 6 Hz), 52.0, 47.4, 38.9, 29.3 (d, J ) 93 Hz), 28.1
(d, J ) 16 Hz), 25.0, 22.0 (br), 21.2, 21.1, 21.0, 16.6 (d, J ) 6
Hz), 13.7 (d, J ) 91 Hz); 31P NMR δ 55.3; mz (FAB) 579 (M +
H, 100), 519 (90), 399 (50); Found (FAB) MH+ 579.2746,
C30H43O9P requires MH+ 579.2723.
Sod iu m Sa lt of 2-Deca r boxy-2-(P -m eth ylp h osp h in ico)-
16-p h en oxytetr a n or p r osta gla n d in F 2r (4b). A mixture of
16-phenoxytetranorprostaglandin F2R (1b, 81 mg, 208 µmol),
DMAP (5 mg, 41 µmol), dichloromethane (1.9 mL), triethyl-
amine (1.2 mL, 8.5 mmol), and acetic anhydride (0.70 mL, 7.3
mmol) was stirred at room temperature for 9.5 h and was then
stored in a freezer overnight. The mixture was concentrated,
and the residue was diluted with saturated aqueous sodium
carbonate (12 mL) and stirred at room temperature for 3 h.
The mixture was acidified with 1 M aqueous hydrochloric acid
and was then extracted with ethyl acetate (5 × 10 mL). The
combined extracts were dried (brine, Na2SO4) and evaporated
to give 16-phenoxytetranorprostaglandin F2R triacetate (2b,
A mixture of 2-decarboxy-2-(O-ethyl-P-methylphosphinico)-
16-phenoxytetranorprostaglandin F2R triacetate (9.9 mg, 17
µmol), 95% ethanol (0.7 mL), and 2.5 M aqueous sodium
hydroxide (0.7 mL, 1.7 mmol) was stirred at reflex for 3 h.
The mixture was diluted with water (6 mL) and washed with
ethyl acetate (2 × 3 mL). The combined washes were extracted
with water (2 mL). The combined aqueous phases were
acidified with 1 M aqueous hydrochloric acid and extracted
with ethyl acetate (4 × 5 mL). The combined extracts were
dried (brine, Na2SO4) and evaporated. The residue was taken
up in methanol (2 mL) and stirred with Amberilite CG-50 (Na+
form, 0.3 g) for 15 min. The resin was removed by filtration,
and the filtrate was concentrated to give the sodium salt of
2-decarboxy-2-(P-methylphosphinico)-16-phenoxytetranorpros-
1
95.6 mg, 89%). H NMR (300 MHz, CDCl3) δ 9.0 (br s, 1 H),
7.28 (td, J ) 8.6, 7.3 Hz, 2 H), 6.96 (t, J ) 7.3 Hz, 1 H), 6.90
(d, J ) 8.6 Hz, 2 H), 5.67-5.74 (m, 2 H), 5.59-5.65 (m, 1 H),
5.30-5.49 (m, 2 H), 5.11 (t, J ) 4.5 Hz, 1 H), 4.92 (ddd, J )
8.8, 7.4, 4.3 Hz, 1 H), 4.08 (dd, J ) 10.2, 6.0 Hz 1 H), 4.03
(dd, J ) 10.2, 4.4 Hz, 1 H), 2.58-2.64 (m, 1 H), 2.54 (ddd,
J ) 15.9, 8.9, 5.7 Hz, 1 H), 2.31 (t, J ) 7.3 Hz, 2 H), 2.11,
2.07, 2.02 (3 s, 3 H each), 1.96-2.16 (m, 4 H), 1.59-1.74 (m,
4 H); 13C NMR (75 MHz, CDCl3) δ 178.5, 170.6, 170.4, 170.2,
158.3, 134.1, 129.8, 129.4, 127.8, 127.4, 121.1, 114.6, 77.6, 74.2,
71.9, 68.9, 51.9, 47.3, 38.8, 31.1, 26.3, 24.7, 24.3, 21.1, 21.0,
20.9.
1
taglandin F2R (4b, 4.4 mg, 58%). H NMR (300 MHz, CD3OD)
δ 7.22-7.28 (m, 2 H), 6.87-6.95 (m, 3 H), 5.63-5.72 (m, 2 H),
5.41-5.51 (m, 1 H), 5.29-5.37 (m, 1 H), 4.44 (ddd, J ) 7.3,
3.7, 1.2 Hz, 1 H), 4.09 (td, J ) 5.2, 2.0 Hz, 1 H), 3.95 (dd, J )
9.7, 4.7 Hz, 2 H), 3.92 (dd, J ) 9.7, 6.6 Hz, 1 H), 3.86 (td, J )
7.9, 5.5 Hz, 1 H), 2.02-2.39 (m, 6 H), 1.34-1.63 (m, 6 H), 1.14
(d, J ) 13.2 Hz, 3 H); 31P NMR δ -43.4; m/z (ESI) 425.3 (M +
H); Found (FAB) (M + Na)+ 447.1927, C22H32NaO6P requires
(M + Na)+ 447.1912. Anal. Calcd for C22H33O6P‚H2O: C, 59.72;
H, 7.97. Found: C, 59.83; H, 8.21.
Oxalyl choride (80 µL, 0.92 mmol) was added to a solution
of 16-phenoxytetranorprostaglandin F2R triacetate (12, 50.6
mg, 98 µmol) and dimethylformamide (0.5 µL, 6.3 µmol) in
dichloromethane (0.20 mL). The mixture was allowd to stand
at room temperature for 30 min and was then concentrated.
The residue was taken up in dichloromethane (0.4 mL).
Meanwhile, a mixture of the sodium salt of N-hydroxypyridine-
2-thione (16 mg, 0.11 mmol), DMAP (0.8 mg, 6.5 µmol), and
dichloromethane (0.5 mL) under argon was brought to reflux
by irradiation with a 250 W General Electric floodlamp. To
this mixture was added 1,1,1-trifluoro-2-iodoethane (0.10 mL,
1.0 mmol) followed by the solution of acid chloride which was
added over 15 min. Irradiation was continued for a further 45
min and the mixture was concentrated and purified by dry-
flash column chromatography (SiO2, 5-45% ethyl acetate-
hexane) to give 2-decarboxy-2-iodo-16-phenoxytetranorpros-
taglandin F2R triacetate (3b, 40.2 mg, 69%). 1H NMR (300
MHz, CDCl3) δ 7.26 (td, J ) 8.6, 7.2 Hz, 2 H), 6.97 (t, J ) 7.2
Hz, 1 H), 6.90 (d, J ) 8.6 Hz, 2 H), 5.67-5.74 (m, 2 H), 5.57-
5.64 (m, 1 H), 5.26-5.41 (m, 2 H), 5.11 (t, J ) 4.6 Hz, 1 H),
4.94 (ddd, J ) 8.6, 7.6, 4.3 Hz, 1 H), 4.10 (dd, J ) 10.2, 6.1
Hz, 1 H), 4.04 (dd, J ) 10.2, 4.7 Hz, 1 H), 3.15 (t, J ) 6.7 Hz,
2 H), 2.59-2.67 (m, 1 H), 2.54 (ddd, J ) 15.9, 8.9, 5.7 Hz, 1
H), 2.10, 2.09, 2.02 (3 s, 3 H each) 1.99-2.21 (m, 4 H), 1.64-
1.87 (m, 4 H); 13C NMR (75 MHz, CDCl3) δ 170.6, 170.4, 170.0,
158.4, 134.3, 129.5, 128.9, 128.5, 127.7, 121.2, 114.7, 77.6, 74.3,
72.0, 69.0, 52.0, 47.5, 38.9, 33.0, 27.8, 25.1, 21.3, 21.2, 21.0,
6.7; m/z (DEI) 598 (M, 5%), 581 (10), 538 (65), 522 (25), 505
(30), 478 (30); Found (DEI) M+. 598.1425, C27H35IO7 requires
M+. 598.1428.
Alter n a tive P r ep a r a tion of 2-Deca r boxy-2-(P -m eth -
ylp h osp h in ico)-16-p h en oxyt et r a n or p r ost a gla n d in F 2r
(4b). 5-(ter t-Bu tyld im eth ylsila n yloxy)-4-[3-(ter t-bu tyld i-
m et h ylsila n yloxy)-4-p h en oxyb u t -1-en yl]h exa h yd r ocy-
clop en ta [b]fu r a n -2-ol (6). In a round-bottom flask at 0 °C
was placed 3.5 g (11.5 mmol) of 16-phenoxy Corey lactone diol
(5) in 80 mL of dichloromethane. To this was added 10.0 mL
(86.4 mmol) of 2,6-lutidine and 10.0 mL (43.5 mmol) of TBDMS
triflate; the reaction was allowed to warm to room temperature
and then stirred overnight. The solution was added to 150 mL
of dichloromethane and washed with 1 N HCl (2×) and brine.
The organics were dried with MgSO4 and reduced. The residue
was chromatographed on SiO2 gel using 5% EtOAc-hexane
as the eluent. Appropriate fractions were combined and
evaporated giving 5.58 g of the bis-silyl ether (91%) as a white
1
solid. HNMR (300 MHz, CDCl3) δ 7.31 (m, 2 H), 6.96 (t, J )
7.3 Hz, 1 H), 6.89 (m, 2 H), 5.65 (m, 2 H), 4.96 (dt, J ) 2.5, 6.9
Hz, 1 H), 4.52 (m, 1H), 4.01 (q, J ) 5.5 Hz, 1 H), 3.85 (d, J )
6.8 Hz, 2 H), 2.61-2.82 (m, 3 H), 2.46-2.53 (m, 2 H), 2.26-
2.38 (m, 2 H), 2.00 (ddd, J ) 2.6, 5.5, 14.7 Hz, 1 H), 0.89 (s, 9
H), 0.85 (s, 9 H), 0.08 (s, 3 H), 0.07 (s, 3 H), 0.03 (s, 3 H), 0.00
(s, 3 H); 13C NMR (75 MHz, CDCl3) δ 181.8, 163.5, 136.8, 135.7,
134.3, 125.6, 119.2, 88.0, 81.5, 76.9, 76.1, 61.6, 47.1, 45.5, 39.7,

4166 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
Soper et al.
30.7, 30.6, 23.2, 22.8, 1.9, 0.2, 0.1, 0.0; M/S m/z (ESI) 550.3
(M + NH4); Found (FAB) MH+ 533.3138, C29H49O5Si2 requires
MH+ 533.3119; Anal. Calcd for C29H48O5Si2: C, 65.37; H, 9.08.
Found: C, 65.25; H, 9.09.
5.33 (m, 1 H), 4.58 (m, 1 H), 4.14 (m, 1 H), 3.88 (m, 2 H), 3.68
(s, 3 H), 3.46 (m, 1 H), 2.35 (m, 3 H) 2.24 (m, 1 H), 2.18 (m, 2
H), 1.62 (ddd, J ) 2.9, 5.2, 13.9 Hz, 1 H) 1.44 (m, 1 H), 0.95 (s,
9 H), 0.94 (s, 9 H), 0.90 (s, 9 H), 0.14 (s, 3 H), 0.14 (s, 3 H),
0.09 (s, 3 H), 0.09 (s, 3 H), 0.06 (s, 3 H), 0.04 (s, 6 H); 13C NMR
(75 MHz, CDCl3) δ 173.8, 159.1, 133.5, 131.6, 130.6, 129.6,
127.9, 120.8, 114.6, 77.54, 72.8, 72.0, 71.7, 55.2, 51.7, 49.9, 45.1,
34.3, 26.1, 25.2, 23.2, 18.6, 18.3, -3.9, -4.2, -4.2, -4.4, -4.7;
m/z (ESI) 750.4 (M + NH4); Found (FAB) (M + Na)+ 755.4522,
In a flame dried round-bottom flask at -78 °C under argon
was placed 5.40 g of the bis-silyl ether (10.2 mmol) in 50 mL
of anhydrous dichloromethane. To this was added 13.2 mL
(13.2 mmol) of 1 M DiBAL in dichloromethane. The solution
was stirred for 2 h and then quenched with NaHCO3(aq). The
organics were washed quickly with 1 N HCl (2 × 50 mL) and
brine. By drying the organics over Na2SO4 and evaporating
the solvent, the crude lactol (6) was used immediately. 1H
NMR (300 MHz, CDCl3) δ 7.31 (m, 2 H), 6.89-6.99 (m, 3 H),
5.64 (m, 2 H), 4.52 (m, 1 H), 4.02-4.28 (m, 1 H), 3.69-3.92
(m, 4 H), 1.62-2.37 (m, 7 H), 0.93-0.97 (m, 18 H), 0.05-0.18
(m, 12 H); m/z (ESI) 517.2 (M + H - H2O).
6-{3,5-Bis(ter t-bu tyld im eth ylsila n yloxy)-2-[3-(ter t-bu -
tyldim eth ylsilan yloxy)-4-ph en oxybu t-1-en yl]cyclopen tyl}-
h ex-4-en oic Ester (7). In a flame dried round-bottom flask,
9.57 g of (3-carboxypropyl)triphenylphosponium bromide (22.3
mmol) was dissolved in 80 mL of anhydrous THF under argon
at -78 °C. A 44.7 mL aliquot of 1 M sodium bis(trimethylsilyl)-
amide in THF (44.7 mmol) was added in one portion, and the
solution was allowed to warm to 0 °C over 2 h. The solution is
recooled to -78 °C, and lactol 6 in 10 mL of anhydrous THF
is added over 10 min. The reaction is allowed to warm to room
temperature and is stirred overnight. The reaction is quenched
with water and stirred for 15 min. A 150 mL aliquot of EtOAc
is added, and the solution is reduced to one-third the volume.
The solution is taken up in 1:1 EtOAc-hexane (150 mL) and
washed 3× with 1 N HCl (100 mL) and brine and then dried
over MgSO4. The solvent was removed, and the residue was
taken up in methanol. The solution was treated with TMS
diazamethane until complete esterification by TLC (10%
EtOAc/hexane). The methanol was evaporated, and the residue
was chromatographed on SiO2 gel (10% EtOAc/hexane) A small
portion of the C9-C15 and C11-C15 bis-silyl ether mixture was
separated. The total combined yield was 5.58 g (88%) of
colorless oil.
Less polar fraction: Rf 0.40 in 10% EtOAc/hexane. 1H NMR
(300 MHz, CDCl3) δ 7.30 (t, J ) 7.4 Hz, 2 H), 6.96 (m, 1 H),
6.91 (m, 2 H), 5.68 (m, 2 H), 5.33-5.49 (m, 2 H), 4.52
(overlapping dt, J ) 4.9, 5.5 Hz, 1 H), 4.28 (m, 1 H), 3.93 (m,
1 H), 3.89 (d, J ) 5.9 Hz, 2 H), 3.69 (s, 3 H), 2.57 (d, J ) 9.1
Hz, 1 H), 2.42 (m, 2 H), 2.36 (m, 3 H), 2.25 (t, J ) 6.9 Hz, 2
H), 2.04 (m, 1 H), 1.82 (m, 1 H), 1.55 (m, 1 H), 0.95 (s, 9 H),
0.94 (s, 9 H), 0.14 (s, 3 H), 0.13 (s, 6 H), 0.11 (s, 3 H); 13C NMR
(75 MHz, CDCl3) δ 173.8, 159.1, 133.8, 130.8, 130.2, 129.7,
128.3, 120.9, 114.7, 79.0, 74.9, 72.5, 72.0, 57.0, 51.8, 51.7, 43.6,
34.2, 26.4, 26.1, 23.1, 18.6, 18.2, -4.1, -4.3, -4.4, -4.8; M/S
m/z 618 (M + H) 619.3.
More polar fraction: Rf 0.30 in 10% EtOAc-hexane. 1H
NMR (300 MHz, CDCl3) δ 7.29 (m, 2 H), 6.95 (m, 1 H), 6.91
(m, 2 H), 5.65 (m, 2 H), 5.34-5.53 (m, 2 H), 4.55 (m, 1 H),
4.14 (m, 1 H), 4.04 (m, 1 H), 3.87 (d, J ) 5.9 Hz, 2 H), 3.67 (s,
3 H), 2.79 (d, J ) 8.1 Hz, 1 H), 2.43 (m, 4 H) 2.39 (t, J ) 5.9
Hz, 2 H), 2.18 (m, 1 H), 2.04 (m, 2 H), 1.82 (d, J ) 4.3 Hz, 1
H), 1.53 (sept, J ) 4.8 Hz, 1 H), 0.95 (s, 9 H), 0.91 (s, 9 H),
0.13 (s, 6 H), 0.08 (s, 6 H); 13C NMR (75 MHz, CDCl3) δ 173.9,
159.1, 133.4, 130.9, 130.2, 129.6, 128.5, 120.9, 114.7, 79.6, 73.9,
72.6, 71.8, 56.3, 51.7, 51.3, 43.5, 34.1, 26.5, 26.1, 22.9, 18.6,
18.2, -4.3, -4.4, -4.5; M/S m/z 618 (M + H) 619.2.
C
40H72NaO6Si3 requires (M + Na)+ 755.4534.
(6-{3,5-Bis(ter t-bu tyld im eth ylsila n yloxy)-2-[3-(ter t-bu -
tyldim eth ylsilan yloxy)-4-ph en oxybu t-1-en yl]cyclopen tyl}-
h ex-4-en yl)m eth ylp h osp h in ic Acid Eth yl Ester (8). In a
round-bottom flask under argon at -78 °C was placed 5.80
(7.92 mmol) of 7 in 80 mL of THF. A 10.3 mL aliquot of 1 M
Lithium aluminum hydride (10.3 mmol) was added in one
portion. The solution was stirred for 1 h at -78 °C and then
quenched with water. Dichloromethane was added (100 mL),
and the solution was acidified to pH ) 1. The aqueous layer
is extracted 3× with dichloromethane, and the combined
organics were dried over MgSO4. The solvent was removed,
and the residue was chromatographed on SiO2 gel using 5%
EtOAc-hexane. Appropriate fractions were combined giving
5.29 g of the alcohol (95%) as colorless oil. 1H NMR (300 MHz,
CDCl3) δ 7.31 (m, 2 H), 6.99 (m, 1 H), 6.90 (m, 2 H), 5.65 (m,
2 H), 5.49 (m, 1 H), 5.40 (m, 1 H), 4.57 (m, 1 H), 4.14 (m, 1 H),
3.88 (m, 3 H), 3.64 (overlapping dt, J ) 5.5, 6.2 Hz, 2 H), 2.45
(m, 1 H), 2.24 (m, 1 H), 2.15 (m, 3 H), 1.64 (m, 1 H), 1.64 (t, J
) 6.9 Hz, 2H), 1.46 (m, 1H), 1.40 (t, J ) 5.5 Hz, 1H), 0.95 (s,
9H), 0.94 (s, 9H), 0.90 (s, 9H), 0.15 (s, 6H), 0.09 (s, 3 H), 0.08
(s, 3 H), 0.05 (s, 6 H); 13C NMR (75 MHz, CDCl3) δ 133.6, 131.5,
129.8, 129.7, 129.5, 77.6, 72.8, 72.0, 71.7, 62.8, 55.1, 49.9, 45.0,
32.8, 26.1, 25.3, 24.0, 18.3, -3.9, -4.2, -4.3; M/S m/z 704 (M
+ NH4) 722.4; Found (FAB) (M + Na)+ 727.4594, C39H72NaO5-
Si3 requires (M + Na)+ 727.4585.
In a flame dried round-bottom flask under argon is placed
5.2 g (7.38 mmol) of the alcohol, 3.11 g (10.3 mmol) of
dibromotriphenylphosphorane, and 0.60 mL (11.8 mmol) of
pyridine in 40 mL of anhydrous acetonitrile. The solution was
monitored by TLC until complete. The mixture was then
poured with stirring into 400 mL of 5% EtOAc-hexane. A
precipitate formed and was filtered through a 6 in. plug of SiO2
gel with additional 5% EtOAc-hexane to wash. The solvent
is removed, giving 5.39 g (95%) of the bromide as a colorless
oil.
1H NMR (300 MHz, CDCl3) δ 7.32 (m, 2 H), 6.97 (t, J ) 7.4
Hz, 1 H), 6.90 (m, 2 H), 5.67 (m, 2 H), 5.51 (m, 1 H), 5.31 (m,
1 H), 4.58 (m, 1 H), 4.16 (overlapping dt, J ) 2.2, 5.1 Hz, 1 H),
3.89 (m, 3 H), 3.40 (t, J ) 6.6 Hz, 2 H), 2.47 (m, 1 H), 2.25 (m,
1 H), 2.19 (m, 2 H), 1.92 (qnt, 6.9 Hz, 2 H), 1.63 (ddd, J ) 2.6,
5.1 14.3 Hz, 1 H), 1.45 (s, 1 H), 0.96 (s, 9 H), 0.95 (s, 9 H), 0.91
(s, 9 H), 0.15 (s, 6 H), 0.10 (s, 3 H), 0.08 (s, 3 H), 0.05 (s, 6 H).
13C NMR (75 MHz, CDCl3) δ 159.1, 133.5, 131.6, 130.8, 129.7,
127.9, 120.8, 114.6, 77.6, 72.8, 72.0, 71.7, 55.2, 49.9, 45.1, 335,
32.9, 26.2, 25.4, 18.6, 18.3, -3.9, -4.2, -4.3, -4.7; m/z (ESI)
767.3 (M + H+); Found (FAB) (M + H)+ 767.3955, C39H71O4-
BrSi3 requires (M + H)+ 767.3922. Anal. Calcd for C39H71O4-
BrSi3: C, 60.98; H, 9.32. Found: C, 58.88; H, 8.37.
In a flame dried 200 mL high-pressure flask was added 30
mL of toluene and 20 mL (excess) of diethoxy methylphos-
phinite. The solution was purged with argon for 5 min, and
4.90 g (6.40 mmol) of the bromide was added. The flask was
sealed and heated to 145 °C for 22 h. The mixture was
evaporated to a thick oil, which was placed on a SiO2 column
and eluted with 1:20:0.1 MeOH/dichloromethane/acetic acid.
The appropriate fractions were combined and reduced to give
4.99 g (98%) of compound 8 as a glassy oil.
In a round-bottom flask at 0 °C was placed 5.5 g (8.88 mmol)
of the mixture of silyl ethers in 60 mL of dichloromethane. To
this was added 10.0 mL (86.4 mmol) of 2,6-lutidine and 10.0
mL (43.5 mmol) of TBDMS triflate; the reaction was allowed
to warm to room temperature and then stirred overnight. The
solution was added to 100 mL of dichloromethane and washed
with 1 N HCl (2×) and brine. The organics were dried with
MgSO4 and reduced. The residue was chromatographed on
SiO2 gel using 5% EtOAc-hexane as the eluent. Appropriate
fractions were combined and evaporated giving 5.97 g of 7
1H NMR (300 MHz, CDCl3) δ 7.28 (t, J ) 7.6 Hz, 2 H), 6.94
(t, J ) 7.3 Hz, 1 H), 6.87 (d, J ) 8.1 Hz, 1 H), 5.64 (m, 2 H),
5.50 (m, 1 H), 5.29 (m, 1 H), 4.56 (bs, 1 H), 4.13 (m, 1 H), 3.86
(m, 3 H), 2.44 (m, 1 H), 2.20 (m, 2 H), 2.11 (m, 2 H), 1.73 (m,
2 H), 1.63 (m, 2 H), 1.46 (d, J ) 13.5 Hz, 3 H), 1.32 (t, J ) 7.0
Hz, 3 H), 1.27 (m, 1 H), 0.94 (s, 9 H), 0.93 (s, 9 H), 0.89 (s, 9
H), 0.13 (s, 6 H), 0.08 (s, 3 H), 0.05 (s, 3 H), 0.03 (s, 6 H); 13C
1
(92%) as a white solid. H NMR (300 MHz, CDCl3) δ 7.30 (m,
2 H), 6.96 (m, 1 H), 6.89 (m, 2 H), 5.66 (m, 2 H), 5.49 (m, 1 H),

PGF Derivatives in Osteoporosis Treatment
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4167
NMR (75 MHz, CDCl3) δ 175.0, 159.1, 133.5, 131.6, 130.6,
129.4, 128.4, 120.8, 114.6, 77.6, 72.7, 71.9, 71.6, 60.7, 60.6, 55.2,
48.9, 45.0, 29.8, 28.5, 28.3, 26.1, 25.3, 22.1, 21.2, 18.6, 18.3,
16.8, 16.8, 16.7, 14.2, 13.0, -3.9, -4.2, -4.2, -4.3, -4.7; MS
m/z 795 (M + Na) 818.4.
mg/kg) and Xylazine (10 mg/kg) and sacrificed by cardiac
puncture. Changes of bone mass were measured in the tibia
by bone histomorphometry; in the femur by dual-energy X-ray
absorbtiometer (DEXA, Hologic QDR-2000 plus bone densito-
meter. Hologic, Inc., Waltham, MA) and in the fifth lumbar
vertebra by microcomputed tomography system (µCT 20, serial
no. 96-2004, Scanco Medical, AG).
(a ) Bon e Histom or p h om etr y. The proximal tibiae and the
middle third of the right tibiae were stained with Villanueva
bone stain, dehydrated in graded concentrations of ethanol,
defatted in acetone, and embedded in methyl methacrylate
(Fisher Scientific, Fairlawn, NJ ). Longitudinal sections of
proximal tibiae (PT) and cross-sections at the tibiofibular
junction of the tibial shafts (TX) were cut to 230 µm thickness
using a low-speed metallurgical saw and then ground to 20
µm (PT) and 30 µm (TX) for histomorphometric measurements.
Histomorphometry was done using a semiautomatic image
analysis system (OsteoMeasure, OsteoMetrics Inc., Decatur,
GA) linked to a microscope equipped with transmission and
fluorescent light.
The region of the proximal tibial metaphysis that was
studied was from 1 to 4 mm distal to the growth plate-
metaphyseal junction. Static measurements included total
tissue area, bone area, and bone perimeter. Dynamic measure-
ments included single and double-labeled perimeter, eroded
perimeter, and interlabel width. These indices were used to
calculate percent bone volume per tissue volume, trabecular
number, trabecular thickness, percent eroded surface, mineral
apposition rate, percent mineralizing surface, and bone forma-
tion rate per unit of either bone volume or total tissue
volume.32 Cortical bone measurements included total cross-
sectional area, marrow area, eroded perimeter, single- and
double-labeled perimeter, and interlabeled width. These pa-
rameters were used to calculate the percent mineralizing
surface, eroded surface, and mineral apposition rate of the
periosteal and endocortical bone surfaces.33
2-Deca r boxy-2-(P -m eth ylp h osp h in ico)-16-p h en oxytet-
r a n or p r osta gla n d in F 2r (4b). In a round-bottom flask under
argon at 0 °C was added 4.82 g (6.06 mmol) of 8 and 18 mL of
HF/pyridine in anhydrous acetonitrile. The solution was
allowed to warm to room temperature and stirred for 5 h. The
solvent was removed, and the residue was taken up in
methanol and dichloromethane. The solution was placed on a
SiO2 column and eluted with 1:20:0.1 MeOH:dichloromethane:
acetic acid. Appropriate fractions were collected and evapo-
1
rated to give 2.31 g (84%) of the triol as an oil. H NMR (300
MHz, CDCl3) δ 7.31 (m, 2 H), 6.98 (m, 1 H), 6.94 (m, 2 H),
5.73 (m, 2 H), 5.32-5.52 (m, 2 H), 4.54 (m, 1 H), 4.16 (m, 1
H), 4.04 (m, 2 H), 4.98 (m, 3 H), 3.14 (bs, 3 H), 2.35 (m, 3 H),
2.15 (m, 3 H), 1.82 (m, 1 H), 1.74 (m, 3 H), 1.53 (m, 1 H), 1.46
(d, J ) 13.6 Hz, 3 H), 1.34 (t, J ) 6.9 Hz, 3 H); 13C NMR (75
MHz, CDCl3) δ 158.9, 135.1, 130.2, 129.8, 129.3, 121.3, 114.9,
78.2, 72.9, 72.1, 71.0, 60.5, 56.2, 50.8, 43.4, 29.6, 28.2, 25.8,
22.3, 16.8, 14.7, 13.5; MS m/z 452 (M + H) 453.3.
In a round-bottom flask was placed 1.7 g (3.76 mmol) of the
triol and 20 mL of 2.5 N sodium hydroxide in 40 mL of 1:1
ethanol:water. The solution was refluxed for 3 h, and then
solvent was removed. The residue was diluted with 20 mL of
water. The aqueous layer was extracted 2× with 20 mL of
EtOAc, and the organic layer was discarded. The aqueous layer
was then acidified to pH ) 1 and extracted 4× with EtOAc
(30 mL). Organics were combined and quickly washed with
brine (20 mL), followed by drying over Na2SO4. The solvent
was evaporated and placed on high vacuum to dry. A total of
1.43 g (90%) of 2-decarboxy-2-(P-methylphosphinico)-16-phe-
noxytetranorprostaglandin F2R (4b) was isolated as a colorless
1
oil. H NMR (300 MHz, CD3OD) δ 7.22-7.28 (m, 2 H), 6.87-
(b) Micr ocom p u ted Tom ogr a p h y. The fifth lumbar ver-
tebral bodies were removed from all animals and were cleaned
of soft tissue. The processes were removed and the vertebral
bodies placed in 70% ethanol. Each lumbar vertebral body was
imaged using a microcomputed tomography system (µCT 20,
serial no. 96-2004, Scanco Medical AG) by stacking at least
five vertebrae in a single holder. The caudal end of the vertebra
was placed on the left side of the holder alignment line to aid
in consistent positioning of the bone. A sponge material
moistened with 70% ethanol, which acts to secure vertebra in
position and keep the sample moist, separated the samples.
Image acquisition parameters for the vertebra included stan-
dard resolution (300 projections), 26 µm slice increment, and
150 ms integration time. Approximately 186 slices were
scanned per vertebra. Once acquisition was complete, the
images were sent to a SGI Octane Workstation for all sub-
sequent analyses. The image analysis involved the following:
(a) setting threshold of the images to bone and background;
(b) determining of the volume of interest (VOI); (c) separating
of the cortical from the trabecular bone; and (d) measuring of
structural parameters (5). Measurements made on the 3D
datasets included trabecular bone volume, surface area, the
derived trabecular thickness and trabecular number according
to Parfitt using the plate model,32 and the directly measured
connectivity density and trabecular thickness.29
6.95 (m, 3 H), 5.63-5.72 (m 2 H), 5.41-5.51 (m, 1 H), 5.29-
5.37 (m, 1 H), 4.46 (ddd, J ) 7.3, 3.7, 1.2 Hz, 1 H), 4.09 (td, J
) 5.2, 2.0 Hz, 1 H), 3.95 (dd, J ) 9.7, 4.7 Hz, 1 H), 3.92 (dd, J
) 9.7, 6.6 Hz, 1 H), 3.86 (td, J ) 7.9, 5.5 Hz, 1 H), 2.02-2.39
(m, 6 H), 1.34-1.63 (m, 6 H), 1.14 (d, J ) 13.2 Hz, 3 H); 13C
NMR (75 MHz, CD3OD) δ 174.2, 159.2, 134.7, 131.3, 129.8,
129.5, 128.8, 120.9, 114.6, 76.6, 71.9, 71.2, 70.9, 55.0, 49.5, 43.2,
30.2, 28.9, 28.1, 27.9, 25.2, 22.1, 22.0, 19.9, 14.3, 13.0; 31P NMR
δ -43.4; m/z (ESI) 425.3 (M + H); Found (FAB) (M + Na)+
447.1927, C22H32NaO6P requires (M + Na)+ 447.1912. Anal.
Calcd for C22H33O6P‚H2O: C, 59.72; H, 7.97. Found: C, 59.83;
H, 8.21.
In Vivo Assa ys. Seventy-two female 3-month-old Sprague-
Dawley rats were acclimated to local vivarium conditions
(Simonsen Laboratories, Gilroy, GA). They were pair-fed in
cages with the room temperature maintained at 72 °F and 12:
12 light/dark cycles. The rats were allowed free access to water
and pelleted commercial natural diet (Teklad Rodent Labora-
tory Chow no. 8604, Harlan Teklad, Madison, WI) that
contains 1.46% calcium, 0.99% phosphorus, and 4.96 IU/g of
vitamin D3. At 6 months of age, the rats were divided into 10
body-weight-matched groups with 6-8 rats per group. One
group was killed as baseline control (Basal); the others were
sham or bilateral ovariectomized. After 60 days of operation,
pretreatment sham (60-d sham) or ovariectomized (60-d OVX)
animals (6 per group) were euthanized as additional pretreat-
ment controls. The remaining rats were treated daily with
0.01, 0.03, 0.1, 0.3, 1.0, and 3.0 mg/kg methyl phosphinic acid
analogue for 60 days or with vehicle of acetate buffers
physiologic saline, methylcellulose, and polyoxyethylenesor-
bitan monooleate. A group was injected with 6 mg/kg/day of
PGE2 (Cayman Chemicals, Ann Arbor, MI) to serve as positive
control. All the rats received 90 mg/kg of Xylenol orange before
treatments and 10 mg/kg of Calcein (Sigma Chemical Co., St.
Louis, MO) 14 and 4 days before sacrifice. At necropsy the final
sham (120-d sham) and ovariectomized (120-d OVX) vehicle-
treated and methyl phosphinic acid analogue-treated rats were
anesthetized by an intraperitoneal injection of Ketamine (50
Ack n ow led gm en t. We would like to thank Arthur
Grieb, Greg Bosch, and Kenetha Stanton of the Custom
Chemistry Scale-up Unit for synthetic support.
Refer en ces
(1) (a) Gowen, M.; Emery, J . G.; Kumar, S. Emerging Therapies for
Osteoporosis. Emerging Drugs 2000, 5 (1), 1-43. (b) Melton, L.
J ., 3rd. How Many Women have Osteoporosis Now? J . Bone
Miner. Res. 1995, 10, 175-177.
(2) Cummings, S. R.; Rubin, S. M.; Black, D. The Future of Hip
Fractures in the United States. Numbers, Costs, and Potential
Effects in Postmenopausal Estrogen. Clin. Orthop. Relat. Res.
1990, 252, 163-166.

4168 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
Soper et al.
(3) Lopez, F. J . New Approaches to the Treatment of Osteoporosis.
Curr. Opin. Chem. Biol. 2000, 4 (4), 383-393.
(4) Francis, R. M. Oral Bisphosphonates in the Treatment of
Osteoporosis: A Review. Curr. Ther. Res. Clin. Exp. 1995, 56,
831-851.
(5) Civitelli, R.; Gonnelli, S.; Zacchei, F.; Bigazzi, S.; Valtimo, A.;
Avioli, L. V.; Gennari, C. Bone Turnover in Postmenopausal
Osteoporosis. Effect of Calcitonin Treatment. J . Clin. Invest.
1988, 82, 1268-1274.
(6) Turner, R. T.; Riggs, B. L.; Spelsberg, T. C. Skeletal Effects of
Estrogen. Endocr. Rev. 1994, 15, 275-300.
24, 4979-4982. (b) Barton, D. H. R.; Crich, D.; Motherwell, W.
B. The Invention Of New Radical Chain Reactions. Part VIII.
Radical Chemistry of Thiohydroxamic Esters: A New Method
for the Generation of Carbon Radicals from Carboxylic Acids.
Tetrahedron 1985, 41, 3901-3924.
(19) Engel, R. Synthesis of Carbon-Phosphorus Bonds; CRC: Boca
Raton, FL, 1988; pp 21-76.
(20) Commercially available from Cayman Chemical Co., Ann Arbor,
MI.
(21) Corey, E. J .; Cho, H.; Rucker, C.; Hua, D. Studies with Trialkyl-
silyltriflates: New Syntheses and Applications. Tetrahedron
Lett. 1981, 22, 3455-3548.
(22) (a) Stjernschantz, J .; Resul, B. Phenyl substituted prostaglandin
analogues for glaucoma treatment. Drugs Future 1992, 17, 691-
704. (b) Resul, B.; Stjernschantz, J .; No, K.; Liljebris, C.; Selen,
G.; Astin, M.; Karlsson, M.; Bito, L. Z. Phenyl-Substituted
Prostaglandins: Potent and Selective Antiglaucoma Agents. J .
Med. Chem. 1993, 36, 243-248. (c) Liljebris, C.; Selen, G.; Resul,
B.; Stjernschantz, J .; Hacksell, U. Derivatives of 17-Phenyl-18,-
19,20-trinorprostaglandin F2R Isopropyl Ester: Potential Anti-
glaucoma Agents. J . Med. Chem. 1995, 38, 289-304. (d) Liljebris,
C.; Nilsson, B.; Resul, B.; Hacksell, U. J . Org. Chem. 1996, 61,
4028-4034.
(7) Mitlak, B. H.; Cohen, F. J . In Search of Optimal Long-Term
Female Hormone Replacement: The Potential of Selective
Estrogen Receptor Modulators. Horm. Res. 1997, 48, 155-163.
(8) (a) Lindsay, R.; Nieves, J .; Henneman, E.; Shen, V.; Cosman, F.
Subcutaneous Administration of the Amino-Terminal Fragment
of Human Parathyroid Hormone (1-34): Kinetics and Biochemi-
cal Response in Estrogenized Osteoporotic Patients. J . Clin.
Endocrinol. Metab. 1993, 77, 1535-1539. (b) Mohan, S.; Kutilek,
S.; Zhang, C.; Shen, H. G.; Kodama, Y.; Srivastava, A. K.;
Wergedal, J . E.; Beamer, W. G.; Bayling, D. J . Comparison of
Bone Formation Responses to Parathyroid Hormone (1-34), (1-
31), and (2-34) in Mice. Bone 2000, 27 (4), 471-478.
(9) Ericksen, E. F.; Mosekilde, L.; Melsen, F. Effects of Sodium
Fluoride, Calcium Phosphonate, and Vitamin D2 on Bone Bal-
ance and Remodelling in Osteoporosis. Bone 1985, 6, 381-390.
(10) (a) Ueda, K.; Saito, S.; Najano, H.; Aoshima, M.; Yokata, M.;
Muraoka, R.; Iwaya, T. Cortical hyperostosis following long-term
administration of prostaglandin E-1 in infants with cyanotic
congential heart disease. J . Pediatr. 1980, 97, 834-836. (b)
Ueno, K.; Haba, T.; Woodbury, D.; Price, P.; Anderson, R.; J ee,
W. S. S. The effects of prostaglandin E2 in rapidly growing rats:
Depressed longitudinal and radial growth and increased meta-
physeal hard tissue mass. Bone 1985, 6, 79-86. (c) Ma. Y. F.;
Li, X. J .; J ee, W. S. S.; Mcosker J .; Liang, X. G.; Setterberg, R.;
Chow, S. Y. Effects of Prostaglandin E2 and F2R on the Skeleton
of Osteopenic Ovariectomized Rats. Bone 1995, 17, 549-554. (d)
J ee, W. S. S.; Ma, Y. F. The In Vivo Anabolic Actions of
Prostaglandins in Bone. Bone 1997, 21, 297-304. (e) Ke, H. Z.;
Shen, V. W.; Qi, H.; Crawford, D. T.; Wu, D. D.; Liang, X. G.;
Chidsey-Frink, K. L.; Pirie, C. M.; Simmons, H. A.; Thompson,
D. D. Prostaglandin E2 Increases Bone Strength in Intact Rats
and Ovariectomized Rats with Established Osteoporosis. Bone
1998, 23, 249-255.
(23) Commercially available from Aldrich Chemical Co. The phos-
phonium salt was dried under vacuum at 145 °C for 24 h before
use.
(24) (a) Sandri, J .; Viala, J . Convenient Conversion of cis-Homoallylic
Alcohols into Corresponding Bromides with Ph3PBr2. Synth.
Commun. 1992, 22 (20), 2945-2948. (b) Bricklebank, N.; God-
frey, S. M.; Mackie, A. G.; MacAuliffe, C. A.; Pritchard, R. B.
The Structure of Triphenylphosphorus-Dibromine, the First
Crystallographically Characterised Bromophosphorane, a Com-
pound which has the Novel Four-Coordinate Molecular Ph3P-
Br-Br Geometry. J . Chem. Soc., Chem. Commun. 1992, 355-
356.
(25) (a) Schaaf, T. K.; Hess, H. J . Synthesis and Biological Activity
of Carboxyl-Terminus Modified Prostaglandin Analogues. J .
Med. Chem. 1979, 22, 1340-1346. (b) Schaaf, T. K.; Bindra, J .
S.; Eggler, J . F.; Plattner, J . J .; Nelson, A. J .; J ohnson, M. R.;
Constantine, J . W.; Hess, H. J . N-(Methanesulfonyl)-16-phenoxy-
prostaglandincarboxamides: Tissue-Selective Uterine Stimu-
lants. J . Med. Chem. 1981, 24, 1353-1359.
(11) Machwate, M.; Harada, S.; Leu, C. T.; Seedor, G.; Labelle, M.;
Gallant, M.; Hutchins, S.; Lachance, N.; Sawyer, N.; Slipertz,
D.; Metters, K. M.; Rodan, S. B.; Young, R.; Rodan, G. A.
Prostaglandin Receptor EP4 Mediates the Bone Anabolic Effects
of PGE2. Mol. Pharmacol. 2001 60 (1), 36-41, and references
therein.
(12) Wang, Y.; Wos, J . A.; Dirr, M. A.; Soper, D. L.; deLong, M. A.;
Mieling, G. E.; De, B.; Amburgey, J . S.; Suchanek, E. G.; Taylor,
C. J . Design and Synthesis of 13,14-Dihydro Prostaglandin F1R
Analogues as Potent and Selective Ligands for the Human FP
Receptor. J . Med. Chem. 2000, 43, 945-953.
(13) Wang, Y.; Soper, D. L.; Dirr, M. J .; deLong, M. A.; De, B.; Wos,
J . A. The Synthesis and Human FP Receptor Binding Affinity
of 13,14-Dihydro Prostaglandin F1R Sulfonamides: Potential
Treatments for Osteoporosis. Chem. Pharm. Bull. 2000, 48 (9),
1332-1337.
(26) Using ACD/pKa DB Software (Version 4.0; Advanced Chemistry
Development Inc., W. Toronto, Canada), the pKa at the P1
position was calculated for each of the ligands tested.
(27) Taken in part from internal study PG-14; Dose Response of
Anabolic Effects of Prostaglandins in the Ovariectomized Rat;
Project No. CR-0132, Study No. B99-5022, May 8, 2000; Procter
& Gamble Pharmaceuticals: Mason, OH, 2000.
(28) The bisphosphonate Alendronate is the common name for
Fosamax, currently marketed by Merck & Co. for treatment of
osteoporosis.
(29) Statistical analysis was performed using analysis of variance
followed by least significant differences test comparing each
treatment to vehicle treated animals (SAS statistical software;
SAS Institute Inc.: Cary, NC, 1996). The ED50 and 95% confi-
dence intervals based on the six dose levels was calculated using
GraphPad Prism (GraphPad Software, Inc., San Diego, CA,
1999), with dose-response curves analyzed using a sigmoidal
nonlinear fit algorithm.
(30) Borah, B.; Gross, G. J .; Dufresne, T. E.; Smith, T. S.; Cockman,
M. D.; Chmielewski, P. A.; Lundy, M. W.; Hartke, J . R.; Sod, E.
W. Three-dimensional microimaging (MRµI and µCT), finite
element modeling, and rapid prototyping provide unique insights
into bone architecture in osteoporosis. Anat. Rec. (New Anat.)
2001, 265, 101-110.
(31) For a leading reference on the use of phosphinate isosteres for
Grb2-SH2 domain inhibitors, see: (a) Furet, P.; Caravatti, G.;
Denholm, A. A.; Faessler, A.; Fretz, H.; Garcia-Echevrria, C.;
Gay, B.; Irving, E.; Press: N. J .; Rahuel, J .; Schoepfer, J .;
Walker, C. V. Structure-Based Design and Synthesis of Phos-
phinate Isosteres of Phosphotyrosine for Incorporation in Grb2-
SH2 Domain Inhibitors. Part 1. Bioorgan. Med. Chem. Lett.
2000, 10, 2337-2341. (b) Walker, C. V.; Caravatti, G.; Denholm,
A. A.; Egerton, J .; Faessler, A.; Furet, P.; Garcia-Echeverria, C.;
Gay, B.; Irving, E.; J ones, K.; Lambert, A.; Press: N. J .; Woods,
J . Structure-Based Design and Synthesis of Phosphinate Iso-
steres of Phosphotyrosine for Incorporation in Grb2-SH2 Domain
Inhibitors. Part 2. Bioorgan. Med. Chem. Lett. 2000, 10, 2343-
2346.
(14) (a) Peitsch, M. C.; Herzyk, P.; Wells, T. N. C.; Hubbard, R. E.
Automated Modeling of the Transmembrane Region of G-
Coupled Protein Receptor by Swiss-Model. Recept. Channels
1996, 4, 161-164. (b) Using TRIPOS computational tools (Sybyl
6.6), several variants of the hFP receptor model were created
with the mapping procedure described in the text. Several
orientations of various receptor helices were investigated, and
different conformations of known FP agonists were docked into
the putative models and evaluated. The homology model pre-
sented in this paper best explains the SAR data discussed.
(15) (a) Funk, C. D.; Furci, L. Moran, N.; Fitzgerald, G. A. Point
Mutation in the Seventh Hydrophobic Domain of the Human
Thromboxane A2 Receptor Allows Discrimination between Ago-
nist and Antagonist Binding Sites Mol. Pharmacol. 1993, 44,
934-939. (b) Cascierci, M. A.; Fong, T. M.; Strader, C. D.
Identification of critical functional groups within binding pockets
of G-protein-coupled receptors. Drugs Future 1996, 21, 521-527.
(16) deLong, M. A.; Amburgey, J .; Taylor, C.; Wos, J . A.; Soper, D.
L.; Wang, Y.; Mieling, G.; Suchanek, E.; Hicks, R. Synthesis and
in Vitro Evaluation of Human FP-Receptor Selective Prostag-
landin Analogues Bioorg. Med. Chem. Lett. 2000, 10, 1519-1522.
(17) Kende, A. S.; Milbank, J . B. J .; Ebetino, F. H.; deLong, M. A.
Prostaglandin phosphonic acid acids through homolytic halode-
carboxylation of prostaglandins F1R and F2R Tetrahedron Lett.
1999, 40, 8189-8192.
(32) (a) Parfitt, A. M.; Drezner, M. K.; Glorieux, F. H.; J anis, J . A.;
Malluche, H.; Meunier, P. J .; Ott, S. M.; Recker, R. R. Bone
histomorphometry: Standardization of nomenclature, symbols
(18) (a) Barton, D. H. R.; Crich, D.; Motherwell, W. B. A Practical
Alternative to the Hundsdieker Reaction Tetrahedron Lett. 1983,

PGF Derivatives in Osteoporosis Treatment
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4169
and units. Report of the ASBMR Histomorphometry Committee.
J . Bone Miner. Res. 1987, 2, 595-610. (b) Parfitt, A. M.;
Matthews, C. H. E.; Villanueva, A. R.; Kleerekoper, M.; Frame,
B.; Rao, D. S. Relationships between surface, area, and thickness
of iliac trabecular bone in aging and in osteoporosis. J . Clin.
Invest. 1983, 72, 1396-1409.
(33) J ee, W. S. S.; Inoue, J .; J ee, K.; Haba, T. Histomorphometric
assay of the growing long bone. In Handbook of Bone Morphol-
ogy; Takahashi, H., Ed.; Niigata City: Nishimura, J apan, 1983;
pp 101-103.
J M010264B