Synthesis and preliminary pharmacological characterisation of a new class of nitrogen-containing bisphosphonates (N-BPs)

Article abstract of DOI:10.1016/j.bmc.2010.02.058

A new series of bisphosphonates bearing either the nitrogen-containing NO-donor furoxan (1,2,5-oxadiazole 2-oxide) system or the related furazan (1,2,5-oxadiazole) in lateral chain has been developed. pK_a values and affinity for hydroxyapatite were determined for all the compounds. The products were able to inhibit osteoclastogenesis on RAW 246.7 cells at 10 μM concentration. The most active compounds were further assayed on human PBMC cells and on rat microsomes. Unlike most nitrogen-containing bisphosphonates which target farnesyl pyrophosphate synthase, experimental and theoretical investigations suggest that the activity of our derivatives may be related to different mechanisms. The furoxan derivatives were also tested for their ability to relax rat aorta strips in view of their potential NO-dependent vasodilator properties.

Full text of DOI:10.1016/j.bmc.2010.02.058

Bioorganic & Medicinal Chemistry 18 (2010) 2428–2438
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and preliminary pharmacological characterisation of a new class
of nitrogen-containing bisphosphonates (N-BPs)
Marco L. Lolli a, Barbara Rolando a, Paolo Tosco a, Shilpi Chaurasia a, Antonella Di Stilo a, Loretta Lazzarato a,
,
*
Eva Gorassini b, Riccardo Ferracini b, Simonetta Oliaro-Bosso a, Roberta Fruttero a, Alberto Gasco a
a Dipartimento di Scienza e Tecnologia del Farmaco, Università degli Studi di Torino, Via Pietro Giuria 9, 10125 Torino, Italy
b CeRMS, Università degli Studi di Torino e ASO San Giovanni Battista, Corso Dogliotti 14, 10126 Torino, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A new series of bisphosphonates bearing either the nitrogen-containing NO-donor furoxan (1,2,5-oxadi-
azole 2-oxide) system or the related furazan (1,2,5-oxadiazole) in lateral chain has been developed. pKa
values and affinity for hydroxyapatite were determined for all the compounds. The products were able to
Received 29 September 2009
Revised 5 February 2010
Accepted 25 February 2010
Available online 1 March 2010
inhibit osteoclastogenesis on RAW 246.7 cells at 10 lM concentration. The most active compounds were
further assayed on human PBMC cells and on rat microsomes. Unlike most nitrogen-containing bisphos-
phonates which target farnesyl pyrophosphate synthase, experimental and theoretical investigations
suggest that the activity of our derivatives may be related to different mechanisms. The furoxan deriva-
tives were also tested for their ability to relax rat aorta strips in view of their potential NO-dependent
vasodilator properties.
Keywords:
Bisphosphonates
Nitric oxide
Osteoclastogenesis
Cardiovascular diseases
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
molecular mechanisms.3,6 The former, which is typical of products
devoid of nitrogen and containing short R and R0 side chains (first
Bisphosphonates (BPs) (1, Fig. 1) are a class of compounds
which contain the P–C–P substructure that, unlike the P–O–P moi-
ety present in pyrophosphate (2, Fig. 1), is resistant to chemical
and enzymatic hydrolysis. They are able to accumulate on the bone
mineral surface, as a consequence of their capacity of chelating cal-
cium ions, and, once there, to inhibit bone resorption following
internalisation by bone-resorbing osteoclasts (OCs). The affinity
to the mineral calcium is increased by the presence of a hydroxy
group on the carbon bearing the two phosphonic functions (1,
R0 = OH).1
The OCs derive from cells of monocyte–macrophage lineage,
which fuse to form bone-resorbing cells in the presence of macro-
phage colony stimulating factor (M-CSF) and receptor activator of
nuclear factor kB ligand (RANKL).2 A number of BPs are in clinical
practice or used in clinical trials for the treatment and prevention
of excessive osteoclast-mediated bone resorption typical of certain
pathologic states, including Paget’s disease, and post-menopausal
or glucocorticoid-induced osteoporosis.3,4 There is also extensive
in vitro and in vivo preclinical evidence that some BPs, particularly
the most potent compounds containing a nitrogen atom in the lat-
eral chain (N-BPs), possess antitumor activity and can reduce skel-
etal tumour burden and inhibit formation of bone metastases in
animal models.5 BPs can affect osteoclasts through two different
generation BPs, Fig. 1),6 is their incorporation into non-hydrolysa-
ble ATP analogues. These metabolites accumulate in the cell cyto-
plasm with consequent inhibition of numerous intracellular
metabolic enzymes, resulting in detrimental effects on cell func-
tion and survival. The latter, which is typical of second and third
generation N-BPs (Fig. 1, N-aliphatic and N-heteroaromatic, respec-
tively)7 is the blockade of protein prenylation due to their ability to
inhibit the mevalonate pathway. The main biochemical target of
bisphosphonates is farnesyl pyrophosphate synthase (FPPS), an en-
zyme of the presqualenic part of the sterol biosynthetic pathway.
Some BPs were found to inhibit the conversion of farnesyl pyro-
phosphate to geranylgeranyl pyrophosphate through the blockade
of geranylgeranyl pyrophosphate synthase (GGPPS); the conse-
quent inhibition of protein geranylgeranylation would then lead
to apoptotic programmed cell death.8 Ibandronate and other deriv-
atives also behave as inhibitors of squalene synthase, although
their anti-resorptive potency does not parallel their degree of inhi-
bition of this enzyme.9 Interestingly, BPs have been shown to inhi-
bit atherogenesis.10 They accumulate extensively in arterial walls
where they suppress macrophages in atheromatous lesions, inhibit
arterial calcification and lipid accumulation. Recently, great inter-
est is being taken in BPs, also in view of the emerging relationship
between osteoporosis and cardiovascular disease (CD).11,12
In this paper we report synthesis, physico-chemical character-
isation and pharmacological profile of a new series of N-BPs bear-
ing the nitrogen-containing furoxan (1,2,5-oxadiazole 2-oxide)
* Corresponding author. Tel.: +39 0116707669; fax: +39 0116707687.
E-mail address: loretta.lazzarato@unito.it (L. Lazzarato).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.02.058

M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
2429
Figure 1. Structures of bisphosphonic acids (1), pyrophosphoric acid (2) and examples of first (etidronate, 3), second (alendronate, 4; ibandronate, 5) and third generation
(zoledronate, 6) bisphosphonates.
system or the related furazan (1,2,5-oxadiazole) in the lateral
chain. Since it is known that furoxans can produce NO under the
action of thiol cofactors,13 furoxan bisphosphonates herein de-
scribed represent a new class of NO-donor bisphosphonates (NO-
BPs). The corresponding furazan derivatives were taken as refer-
ences since they are unable to release NO. NO is involved in the
regulation of many physiological processes. In the cardiovascular
system it plays important roles in maintaining micro and macro-
vascular homeostasis through a number of mechanisms including
vascular relaxation, inhibition of platelet aggregation and smooth
muscle cell proliferation, modulation of monocyte and platelet
adhesion.14,15 Its impaired production is associated to many forms
of CD; among them, atrial fibrillation and atherosclerosis. It is also
an essential modulator of gastric mucosal defence.16 Nitric oxide
plays a role in bone remodelling, affecting osteoclast and osteo-
blast proliferation in a biphasic manner: high concentrations inhi-
bit formation of these bone cells, while low concentrations support
their development.17
furoxan 8 and 8a, respectively, through intermediates 11, 11a
(Scheme 1).
BPs 19, 19a, 20, 20a, bearing a hydroxyl group on the carbon
adjacent to phosphonic functions, were prepared starting from
the phenylsulfonyl substituted furazans 8, 14 and furoxans 8a,
14a by nucleophilic displacement of one phenylsulfonyl group un-
der the action of 1,5-pentanediol. This reaction afforded the inter-
mediate alcohols 15, 15a, 16, 16a which were subsequently
oxidised to the related acids 17, 17a, 18, 18a using Jones reagent.
The reaction of these products with thionyl chloride generated
the corresponding acylchlorides A, which, without further purifica-
tion, were treated with trimethyl phosphite in dry THF to yield the
related acylphosphonates B (Michaelis–Arbuzov reaction). The
products were immediately transformed into the expected tetra-
methyl 1-hydroxybisphosphonates C by action of dimethylphos-
phite in dry THF, in the presence of diethylamine. These esters
were purified by flash chromatography and immediately trans-
formed into the final BPs under the action of TMSBr and methanol
(Scheme 2).
Therefore, NO-BPs could represent interesting tools in therapy
to limit bone loss and atherogenesis. In addition, they could display
reduced gastric irritation1 and atrial fibrillation risk,18 two well
known drawbacks of BPs.
The analogues of alendronate 24, 24a, 25, 25a bearing furazan
and furoxan substructures were prepared from bromomethylsub-
stituted furazans and furoxans 21, 21a, 22, 22a by action of N-
methylalendronate 23 (Scheme 3).
2. Results and discussion
2.1. Chemistry
2.2. Dissociation constants
In Table 1 the pKa values of the final bisphosphonic acids are re-
ported. They were measured by pH-metric method using a Sirius
GLpKa instrument. In the case of BPs not bearing an amino nitro-
gen in the lateral chain, four dissociation constants were detected.
This is in keeping with the very low basicity of the furazan and
furoxan rings that, in the pH range studied (1.2–12), do not under-
go protonation.19 The pKa1 value, corresponding to the first ionisa-
tion of the bisphosphonic group, is not detectable by this method
because it is too low. All the pKa values are in keeping with those
found for similar compounds.20,21 In the case of compounds 24,
24a, 25 and 25a the picture is complicated by the presence of a
fifth pKa associated with the protonation of the basic nitrogen.
The synthetic route to the target compounds is outlined in
Schemes 1–3. Phenylfurazan and phenylfuroxan BPs 12 and 12a
were easily obtained starting from nitrofurazan and nitrofuroxan
derivatives 7 and 7a, respectively. These products were reacted
with tetraethyl 5-hydroxypentylidene bisphosphonate 9 in the
presence of t-BuOꢀK+ to afford intermediate phosphonates 10,
10a, which were transformed into the final acids by action of
bromotrimethylsilane (TMSBr) and methanol. Similarly, phen-
ylsulfonylfurazan and phenylsulfonylfuroxan BPs 13, 13a were ob-
tained starting from bis(phenylsulfonyl) substituted furazan and
Scheme 1. Reagents and conditions: (a) t-BuOꢀK+, dry THF, ꢀ15 °C for compounds 10, 10a, ꢀ45 °C for compounds 11, 11a; (b) TMSBr, CH2Cl2; (c) MeOH, 0 °C to rt.

2430
M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
Scheme 2. Reagents and conditions: (a) 1,5-pentanediol, NaOH 50% w/w, THF; (b) Jones reagent, acetone, 0 °C to rt; (c) SOCl2; (d) P(OMe)3, dry THF, 0 °C to rt; (e) HPO(OMe)2,
Et2NH, dry THF, 0 °C to rt; (f) TMSBr, CH2Cl2; (g) MeOH, 0 °C to rt.
2.3. Hydroxyapatite affinity
BPs are trapped by hydroxyapatite (HAP) calcium present at the
sites of new bone formation, giving rise to the formation of stable
complexes. After incorporation in the bone they interact with oste-
oblasts and osteoclasts through a variety of biochemical pathways
inducing inhibition of bone resorption. In order to evaluate their
affinity for HAP, BPs described in the present work were incubated
with HAP, according to a procedure described elsewhere.23 After
0-h, 6-h and 24-h incubation, the residual concentration of BPs in
solution was determined by reverse phase HPLC. The results
obtained are collected in Table 1. As expected, ibandronate (5),
1-hydroxy-substituted BPs 19, 19a, 20, 20a and the analogues of
alendronate 24, 24a, 25 and 25a, due to their ability to form triden-
tate complexes with Ca2+, display higher affinity for HAP than their
analogues 12, 12a, 13, 13a deprived of the hydroxy function. These
data are in keeping with previous results reported in literature.6
2.4. Osteoclastogenesis inhibition
All the compounds were tested for their ability to inhibit osteo-
Scheme 3. Reagents and conditions: (a) 2 M NaOH, pH 12, H2O, CH3CN.
clastogenesis on RAW 246.7 cells, a murine monocyte/macrophage
cell line with pre-osteoclastic attributes (TIB-71) at 10 M concen-
l
tration, according to the protocol reported in Section 4; the results
are collected in Figure 2. All the compounds were able to signifi-
cantly decrease the number of developed osteoclasts compared
to the controls, in a manner not significantly different from ibandr-
onate. Furoxan and related furazan derivatives behave likewise,
but a trend of higher activity of the latter compared to the former
can be observed. Since the congener furazan derivatives, which are
devoid of the capacity to release NO, are endowed with anti-
resorptive activity near that of the related furoxans, we have to
conclude that in the latter NO does not play crucial roles in their
antiosteoclastogenic action, unlike what happens in a series of
nitrooxy-substituted BPs we recently developed.11 The most active
pairs of products, namely 20, 20a, 12, 12a, 25, 25a, were also tested
These pentaprotic BPs have three dissociation constants (pKa1
,
pKa2, pKa5) corresponding to the first, second, and fourth ionisation
of the bisphosphonic group. In the physiological pH range there are
two ionisation steps described by pKa3 and pKa4 macroconstants
that result from overlapping ionisations of the third bisphosphonic
group and the amino nitrogen in the lateral chain. To assign pKa
values to these centres we measured the ionisation constants in
the presence of methanol. The pKa3 and pKa4 were obtained by
the plots of the apparent psKa values assessed by potentiometric
titration in co-solvent mixtures (water/methanol) versus the %
methanol (% w/w) in the mixtures. Where straight lines with neg-
ative slopes were obtained the value was attributed to the preva-
lent ionisation of the amine substructure of each compound
(pKa4). This behaviour is typical of the basic centres.22
at 10
lM concentration on PBMC cells collected from healthy con-
trols (Fig. 3). The results are similar to those obtained working with

M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
2431
Table 1
pKa values and affinity for HAP of ibandronate and of final bisphosphonates
Compd
Dissociation constants
Affinity to HAPa
6 h
pKa1
pKa2 SD
pKa3 SD
pKa4 SD
pKa5 SD
0 h
24 h
5b
12
12a
13
13a
19
19a
20
20a
24
24a
25
25a
—
2.84
6.08
9.80
10.43
—
—
—
—
—
—
—
71
75
76
64
70
56
59
56
59
69
69
70
62
2
3
1
2
2
1
4
2
2
2
2
2
1
56
54
51
48
57
40
36
41
42
54
51
51
49
1
3
2
4
2
4
3
1
1
1
4
2
2
56
44
44
40
42
33
27
34
33
46
46
52
46
3
1
2
5
3
2
2
1
2
1
2
2
1
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
<1.5
2.81 0.02
2.67 0.02
2.69 0.04
2.64 0.03
2.69 0.04
2.90 0.02
2.80 0.05
2.52 0.03
2.05 0.06
2.26 0.07
2.04 0.09
2.16 0.06
6.88 0.01
6.86 0.01
7.00 0.03
6.87 0.02
6.66 0.02
6.86 0.01
6.64 0.03
6.65 0.01
6.05 0.01
5.89 0.01
6.02 0.02
6.03 0.01
11.4 0.1
11.2 0.1
10.7 0.1
11.6 0.1
11.0 0.2
11.2 0.1
10.6 0.1
11.3 0.1
7.45 0.02
7.13 0.01
7.65 0.03
7.36 0.01
—
11.1 0.1
11.2 0.1
11.0 0.1
10.8 0.1
a
Percentage of residual BP in solution; all values are expressed as mean SD, n = 4–6.
b
Dissociation constants reported in literature.19
compounds of the series (12, 12a, 20, 20a) were tested for their abil-
ity to inhibit FPPS and other enzymes of the pre-squalene section of
sterol biosynthesis, for example, squalene synthase (SQS). A preli-
minary test carried out by incubating a rat liver S10 supernatant
with labelled (R,S)-[2-14C]-mevalonic acid showed that in the pres-
ence of the oxidosqualene cyclase inhibitor U14266A and in the ab-
senceof bisphosphonate mostradioactivity accumulatedinto aband
co-eluting with squalene. In the presence of bisphosphonates 12, 20,
20a (10 lM) or ibandronate 5 (1 lM) the peak of radioactivity co-
eluting with squalene completely disappeared, indicating a possible
inhibition of one or more enzymes of the pre-squalenic sterol bio-
synthetic pathway. When tests were repeated with the S100 super-
natant, a subcellular fraction deprived of most squalene synthase
activity, radioactivity mainly accumulated into a chromatographic
fraction corresponding to unresolved metabolites derived from the
hydrolysis of prenyl pyrophosphates. This radioactive fraction
Figure 2. TRAP+ RAW-OCs number. The number of TRAP+ RAW-OCs containing
three or more nuclei/cell was quantified from three independent experiments. Data
are presented as the mean SD; statistical differences from controls not treated
with N-BPs (10
is 0.0001.
diminished by 50% in assays with 1 lM ibandronate, while a similar
l
M): *p <0.001; **p <0.00001. The result of an ANOVA statistical test
effect was not observed with bisphosphonates 12, 12a, 20, 20a. The
decrease of radioactivity in the prenyl intermediates fraction in the
presence of ibandronate was in agreement with inhibition of FPPS,
a known target of this molecule. The bisphosphonates 12, 12a, 20,
20a, causingan inhibitionof thepre-squalenic biosyntheticpathway
following the incubation of (R,S)-[2-14C]-mevalonic acid in the S10,
but not in the S100 cellular fraction, are more likely inhibitors of
squalene synthase. The inhibition of squalene synthase by com-
pounds 12, 12a, 20, 20a and ibandronate was tested by incubating
microsomes with the labelled substrate farnesyl pyrophosphate.
The values of IC50 are listed in Table 2. Compounds 20 and 20a are
the most active compounds against squalene synthase, showing
IC50s of 3.71 0.42
pounds 12 and 12a have a lower activity, with IC50 ramping around
20 M. All bisphosphonates tested are less active than ibandronate
(IC50 = 0.64 0.16 M), but, from the results obtained with the
lM and 7.95 2.23 lM, respectively. Com-
l
l
S100 supernatant, appear to be more specific inhibitors of squalene
synthase. On the basis of these assays, it is not possible to rule out a
Figure 3. TRAP+ MNC number. The number of TRAP+ OCs containing three or more
nuclei/cell was quantified from nine independent wells per condition. Data are
presented as the mean SD; statistical differences from controls not treated with
N-BPs (10 l
M): *p <0.05; **p <0.002. The result of an ANOVA statistical test is 0.0001.
Table 2
Inhibitory effect of compounds 5, 12, 12a, 20, 20a on squalene synthase enzyme in rat
liver microsomes
RAW 246.7 cells. The most relevant differences are a slightly higher
activity of derivatives 12 and 20a.
a
Compd
IC50
(
lM)
5
0.64 0.16
17.68 1.72
22.95 3.57
3.71 0.42
7.95 2.23
12
12a
20
20a
2.5. Enzymatic assays and molecular modelling
OneofthemainpharmacologicaltargetsofBPsisknowntobefar-
nesyl pyrophosphate synthase (FPPS);4 therefore, the most active
a
Data expressed as mean SE (three different experiments in duplicate).

2432
M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
contribution by GGPPS blockade to the antiosteoclastogenic action
exerted by compounds 12, 12a, 20, 20a, since an eventual inhibitory
effect on this enzyme would not result in a significant reduction of
squalene or isoprenoid metabolites.
In the attempt to rationalise the activity profile of the most ac-
tive compounds 20, 20a compared to ibandronate, a molecular
modelling investigation was carried out. Taking advantage of the
existence of crystal structures of human FPPS,24 SQS25 and
GGPPS,26 flexible docking of the aforementioned bisphosphonates
was accomplished in the active site of these three enzymes. In
the case of FPPS, since a co-crystal with ibandronate has been ob-
tained,24 we had the opportunity to validate our approach, verify-
ing that AUTODOCK 4.0 is able to reproduce the crystal pose of the
reference inhibitor within the limits of experimental error (RMS
deviation on heavy atoms 0.52 Å). Additionally, the pose that
reproduces the experimental binding mode also belongs to the
most populated cluster, characterised by the best docking score
(average
D
Gbinding = ꢀ11.34 kcal molꢀ1). Analysing the results for
the newly synthesised bisphosphonates 20 and 20a the situation
is radically different (Fig. 4a). Respectively, only 7 and 3 poses over
200 are located in the enzyme cavity, with average
DGbind-
ing = ꢀ4.48 and ꢀ3.22 kcal molꢀ1, while in the remaining runs the
inhibitor was placed in the bulk water outside the enzyme. This
outcome is amenable to the steric bulk of the oxadiazole ring,
which hinders these compounds from entering the active site.
The presence of the heterocycle at one end of the lipophilic tail also
appears to restrict the conformational freedom of the bisphospho-
nate head, which is not able to establish interactions as effectively
as ibandronate does. In particular, each of the phosphonate moie-
ties of ibandronate gives rise to electrostatic interactions with the
metal ions present in the active site, in addition to charge-en-
hanced hydrogen bonds with Arg112, Lys200 and Lys257.24 In-
stead, in the case of compounds 20 and 20a the bisphosphonate
head is rotated, which results in only one of the phosphonate moi-
eties coming into close contact with metal ions and Lys200, while
Arg112 and Lys257 are not within hydrogen-bonding distance. The
loss of these interactions sums up to the lack of a protonated nitro-
gen atom capable of hydrogen bonding with the side chain of
Thr201.24 Finally, the phenylsulfonyl substituent on the oxadiazole
ring protrudes out of the cavity, as evidenced by the representation
of the solvent-excluded surface of the cavity in Figure 4a; the expo-
sure to water of such a hydrophobic moiety further penalizes the
binding of these inhibitors. Overall, this picture yields a reasonable
interpretation of the lack of affinity for FPPS displayed by com-
pounds 20, 20a. The docking results on SQS and GGPPS appear
markedly different from the ones on FPPS. While these enzymes
share some degree of structural homology, the solvent accessible
volume of the FPPS cavity is only about 260 Å3, compared to
608 Å3 for GGPPS and 1240 Å3 for SQS. These differences are due
to the fact that FPPS is designed to catalyse the condensation be-
tween dimethyl allyl pyrophosphate (DMAPP) and isopentenyl
pyrophosphate (IPP), which are both small substrates. On the con-
trary, GGPPS needs to accommodate FPP and IPP, while the SQS
cavity is even broader to allow the condensation of two FPP mole-
cules. Consequently, both ibandronate and the oxadiazole bisphos-
phonates are well accommodated in the cavities of SQS and GGPPS.
On SQS both ibandronate and newly synthesised BPs display fair
Figure 4. Ibandronate (thicker lines) and compound 20 (thinner lines) docked in
FPPS (a), SQS (b) and GGPPS (c), respectively. Polar hydrogens have been omitted for
clarity. The solvent-excluded surface of the cavity has been included to evidence
how bisphosphonates 20 and 20a partially protrude from the FPPS cavity, while
they are well accommodated in the active sites of SQS and GGPPS.
docking scores (ibandronate: average
D
Gbinding = ꢀ7.07 kcal molꢀ1
;
ibandronate establishes a charge-enhanced hydrogen bond be-
tween its protonated nitrogen and the carboxylate group of
Asp188, compounds 20 and 20a stack their phenylsulfonyl substi-
tuent between the aromatic side chains of His57, Phe156 and
Phe184. All ligands display similar, good in silico affinity (ibandro-
20 and 20a: ꢀ8.26 and ꢀ8.21 kcal molꢀ1, respectively; Fig. 4b). The
phosphonate heads bind similarly, establishing electrostatic inter-
actions with three arginine residues (Arg52, Arg77 and Arg218),
while the lipophilic tails lie in a pocket lined by mostly hydropho-
bic residues (Phe72, Tyr73, Leu76, Val179, Leu183, Leu211 and
Phe288; Fig. 4b). Coming to GGPPS, the ligands’ phosphate groups
interact with both metal ions present in the active site, as well as
with positively charged Arg73, Lys151, Lys202, Lys212. While
nate: average
D
Gbinding = ꢀ11.15 kcal molꢀ1; 20 and 20a: ꢀ10.84
and ꢀ11.55 kcal molꢀ1, respectively; Fig. 4c). These figures indicate
that there is a possibility that binding at GGPPS may contribute to
the antiosteoclastogenic effect displayed by 20 and 20a.

M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
2433
2.6. Vasodilator activity
experimental work is necessary to draw definitive conclusions on
the mechanism of action. NO-BPs are able to relax contracted vas-
cular tissue in a concentration-dependent manner. A decrease in
potency was observed in the presence of ODQ, proving the involve-
ment of NO in the vasodilator mechanism. This new series of NO-
BPs is worthy of additional studies as potential drugs in view of the
emerging relationship between osteoporosis and CD.
The NO-BPs 12a, 13a, 19a, 20a, 24a, 25a and ibandronate 5 were
tested for their ability to relax rat aorta strips precontracted with
phenylephrine, in view of their potential NO-dependent vasodilator
properties. All furoxan derivatives were capable of dilating the con-
tracted tissue in a concentration-dependent manner, while ibandro-
nate resulted completely inactive. An example of this behaviour is
reported in Figure 5. The vasodilator potencies (EC50) of the products
are collected in Table 3. When the experiments were repeated in the
4. Experimental section
4.1. Chemistry
presence of 1 lM ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-
one), a strong decrease in potency was observed, in keeping with a
NO-mediated, sGC-dependent vasodilator mechanism. The poten-
ciesofthesehybridsparalleltheonesofthesimplereferencefuroxan
models,27 while being consistently lower. This is most likely due to
the highly hydrophilic character of NO-BPs, which reduces their
capacity to enter into vascular smooth muscle cells where they are
metabolised to NO.
1H and 13C NMR spectra were recordedon a BrukerAC-200 and on
a Bruker Avance 300; abbreviations: s, singlet; d, doublet; t, triplet;
tt, triplet of triplets; m, multiplet; br, broad; Fx, furoxan; Fz, furazan.
Low resolution mass spectra were recorded with a Finnigan-Mat
TSQ-700. Meltingpointsweredeterminedwitha capillaryapparatus
(Büchi 540). Flash column chromatography was performed on Silica
Gel (Merck Kieselgel 60, 230–400 mesh ASTM); PE stands for 40–60
petroleum ether. The progress of the reactions was followed by thin
layer chromatography (TLC) on 5 ꢁ 20 cm plates with a layer thick-
nessof 0.25 mm. Anhydrousmagnesiumsulfate was usedas the dry-
ing agent for the organic phases. Organic solvents were removed
undervacuumat30 °C. PreparativeHPLCwasperformedonaLichro-
3. Conclusion
A new series of bisphosphonates bearing either the nitrogen-
containing NO-donor furoxan (1,2,5-oxadiazole 2-oxide) or the re-
lated furazan (1,2,5-oxadiazole) systems in the lateral chain has
been developed. Their pKa profile is in keeping with that of similar
compounds. As expected, 1-hydroxy substituted products show a
higher affinity for HAP than their methylene analogues. All the
new BPs are able to inhibit osteoclastogenesis on RAW 246.7 and
spherÒ C18 column (250 ꢁ 25 mm, 10
lm) (Merck Darmstadt, Ger-
many) with a Varian ProStar mod-210 with Varian UV detector
mod-325. When necessary they were developed with iodine, KMnO4
and ammonium thiocyanate–iron(III) chloride reagent.28 Elemental
analyses (C, H, N) were performed by REDOX(Monza) and the results
are within 0.4% of the theoretical values. Compounds 7,29 7a,29 8,30
8a,31 9,21 14,29 14a,29 21,32 21a,33 22,32 22a,34 2333 were synthesised
according to literature. Tetrahydrofuran (THF)wasdistilledimmedi-
ately before use from Na and benzophenone under a positive atmo-
sphere of N2.
PBMC cells at 10 lm concentration. The most active compounds
were tested for their ability to inhibit the mevalonate pathway.
The assay indicates that they are more potent inhibitors of SQS
compared to FPPS, while an involvement of other targets (e.g.,
GGPPS) cannot be ruled out. This finding is in keeping with the
results of a molecular modelling investigation, but additional
4.1.1. General procedure for the preparation of esters 10, 10a,
11, 11a
A
solution of potassium tert-butylate (0.416 M; 11 mL;
4.58 mmol) in dry THF was added over 90 min to a solution of 9
(1.50 g; 4.16 mmol) and the appropriate furazan (7, 8) or furoxan
(7a, 8a) (4.58 mmol) in dry THF (15 mL), stirred under inert atmo-
sphere between ꢀ45 and ꢀ15 °C. The mixture was allowed to
reach room temperature, then diluted with saturated NH4Cl
(20 mL) and extracted with Et2O (25 mL). The organic phases were
collected, washed with saturated NH4Cl (10 mL), dried and concen-
trated under reduced pressure. The crude product was purified by
flash chromatography using the appropriate eluent obtaining the
desired compound as pale yellow oil.
Figure 5. Concentration–response curves of 20a in the absence and in the presence
of 1 lM ODQ.
4.1.1.1. Tetraethyl
(5-(4-phenylfurazan-3-yloxy)pentylidene)
bisphosphonate (10). Eluent from CH2Cl2 to CH2Cl2/MeOH 98:2
v/v; pale yellow oil; 86% yield. 1H NMR (200 MHz, CDCl3, TMS):
d = 1.27–1.37 (m, 12H, –PO(OCH2CH3)2), 1.72–2.17 (m, 6H, –
3
2
Table 3
CH2CH2CH2CH–), 2.29 (tt, 1H, JHH = 5.8 Hz, JPH = 24.0 Hz,
–
Vasodilating activity of compounds 5, 12a, 13a, 19a, 20a, 24a, 25a
CH2CH[PO(OCH2CH3)2]2), 4.07–4.23 (m, 8H, –PO(OCH2CH3)2), 4.46
a
b
3
(t, 2H, JHH = 6.6 Hz, –OCH2–), 7.44–8.00 ppm (m, 5H, C6H5); 13C
Compd
EC50
(l
M)
EC50ODQ (lM)
NMR (50 MHz, CDCl3, TMS): d = 16.2 (d), 25.2 (t), 28.5, 36.6 (t,
5
Inactive
Not tested
1JPC = 133 Hz), 62.4 (m), 72.3, 125.1, 127.2, 128.8, 130.4, 145.0,
c
12a
13a
19a
20a
24a
25a
73 18
0.20 0.03
108 20
(M+H)+.
Anal.
Calcd
for
193 36
163.4 ppm.
MS
m/z
505
c
C21H34N2O8P2ꢂ0.5H2O: C, 49.12; H, 6.87; N, 5.46. Found: C, 48.92;
0.23 0.03
192 27
c
H, 6.99; N, 5.34.
45
35
3
2
c
4.1.1.2. Tetraethyl
(5-(3-phenylfuroxan-4-yloxy)pentylidene)
a
b
c
Data expressed as mean SE, n = 3–15.
bisphosphonate (10a). Eluent: from CH2Cl2 to CH2Cl2/MeOH
98:2 v/v; pale yellow oil; 60% yield. 1H NMR (200 MHz, CDCl3,
TMS): d = 1.27–1.36 (m, 12H, –PO(OCH2CH3)2), 1.81–2.02 (m, 6H,
Experiments performed in the presence of 1 lM ODQ.
At the maximal concentration tested (1 ꢁ 10ꢀ4 M) vasodilation does not reach
50%.

2434
M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
3
2
–CH2CH2CH2CH–), 2.31 (tt, 1H, JHH = 5.8 Hz, JPH = 23.8 Hz,
–
Calcd for C13H18N2O9P2: C, 38.25; H, 4.44; N, 6.86. Found: C, 38.28;
H, 4.48; N, 6.85.
3
CH2CH[PO(OCH2CH3)2]2), 4.52 (t, 2H, JHH = 6.1 Hz, –OCH2–),
4.09–4.25 (m, 8H, –PO(OCH2CH3)2), 7.46–8.16 ppm (m, 5H, C6H5);
13C NMR (50 MHz, CDCl3, TMS): d = 16.2 (d), 25.2 (t), 28.3, 36.6
4.1.2.3. 5-(4-Phenylsulfonylfurazan-3-yloxy)pentylidenebis(phos-
phonic) acid (13). White solid, 55% yield. 1H NMR (200 MHz,
DMSO-d6, TMS): d = 1.40–2.10 (m, 7H, –CH2CH2CH2CH–), 4.33–
4.36 (m, 2H, –OCH2–), 7.76–8.12 (m, 5H, C6H5), 9.88 ppm (br s,
4H, mobile protons); 13C NMR (50 MHz, DMSO-d6, TMS, a signal
is missing, being covered by the solvent): d = 24.0, 28.8, 74.0,
128.8, 130.3, 136.3, 136.8, 148.8, 161.0 ppm; 31P NMR (120 MHz,
D2O, H3PO4): d = 22.7 ppm. Anal. Calcd for C13H18N2O10P2S: C,
34.22; H, 3.98; N, 6.14. Found: C, 34.17; H, 3.98; N, 6.11.
1
(t, JPC = 133 Hz), 62.4 (t), 70.3, 107.4, 122.4, 126.0, 128.7, 130.3,
162.2 ppm. MS m/z 521 (M+H)+. Anal. Calcd for C21H34N2O9P2: C,
48.46; H, 6.58; N, 5.38. Found: C, 48.53; H, 6.73; N, 5.27.
4.1.1.3. Tetraethyl (5-(4-phenylsulfonylfurazan-3-yloxy)pentyl-
idene) bisphosphonate (11). Eluent: CH2Cl2/MeOH 98:2 v/v;
pale yellow oil; 58% yield. 1H NMR (200 MHz, CDCl3, TMS):
d = 1.31–1.38 (m, 12H, –PO(OCH2CH3)2), 1.67–2.13 (m, 6H, –
3
2
CH2CH2CH2CH–), 2.28 (tt, 1H, JHH = 5.8 Hz, JPH = 24.0 Hz,
–
CH2CH[PO(OCH2CH3)2]2), 4.11–4.27 (m, 8H, –PO(OCH2CH3)2), 4.37
4.1.2.4. 5-(3-Phenylsulfonylfuroxan-4-yloxy)pentylidenebis(phos-
phonic) acid (13a). White solid; 77% yield. 1H NMR (200 MHz,
DMSO-d6, TMS): d = 1.50–2.20 (m, 7H, –CH2CH2CH2CH–), 4.38 (m,
2H, –OCH2–), 7.74–8.06 (m, 5H, C6H5), 8.31 ppm (br s, 4H, mobile
protons); 13C NMR (50 MHz, DMSO-d6, TMS, a signal is missing,
being covered by the solvent): d = 24.7, 27.9, 71.3, 110.6, 128.5,
130.2, 136.2, 137.2, 158.9 ppm; 31P NMR (120 MHz, D2O, H3PO4):
d = 23.4 ppm. Anal. Calcd for C13H18N2O11P2S: C, 33.06; H, 3.84;
N, 5.93. Found: C, 32.79; H, 4.01; N, 5.80.
3
(t, 2H, JHH = 6.2 Hz, –OCH2–), 7.60–8.12 ppm (m, 5H, C6H5); 13C
NMR (50 MHz, CDCl3, TMS): d = 16.3 (d), 25.0 (t), 28.2, 36.5 (t,
1JPC = 143 Hz), 62.4 (t), 73.3, 128.8, 129.5, 135.3, 137.7, 148.0,
161.1 ppm. MS m/z 569 (M+H)+. Anal. Calcd for C21H34N2O10P2S:
C, 44.37; H, 6.03; N, 4.93. Found: C, 44.47; H, 6.11; N, 4.87.
4.1.1.4. Tetraethyl (5-(3-phenylsulfonylfuroxan-4-yloxy)pentyl-
idene) bisphosphonate (11a). Eluent: CH2Cl2/MeOH 98:2 v/v;
pale yellow oil; 59% yield. 1H NMR (200 MHz, CDCl3, TMS):
d = 1.35–1.45 (m, 12H, –PO(OCH2CH3)2), 1.80–2.00 (m, 6H, –
4.1.3. General procedure for the preparation of alcohols15, 15a,
16, 16a
3
2
CH2CH2CH2CH–), 2.30 (tt, 1H, JHH = 5.8 Hz, JPH = 24.0 Hz,
–
CH2CH[PO(OCH2CH3)2]2), 4.12–4.28 (m, 8H, –PO(OCH2CH3)2), 4.43
NaOH (50% w/w; 1.5 equiv) was added over a period of 30 min
to a stirred solution of the appropriate phenylsulfonyl derivative
(8, 8a, 14, 14a) (8.0 g) and 1,5-pentanediol (10 equiv) in distilled
THF (100 mL). Acetic acid was added to the mixture that was con-
centrated under reduced pressure. The residue was dissolved in
EtOAc (100 mL) and washed twice with H2O (100 mL). The organic
layer was dried and concentrated under reduced pressure. The
crude product was purified by flash chromatography to give the
pure product as a white solid.
3
(t, 2H, JHH = 6.0 Hz, –OCH2–), 7.60–8.09 ppm (m, 5H, C6H5); 13C
NMR (50 MHz, CDCl3, TMS): d = 16.3 (d), 25.0 (t), 26.1, 36.5 (t,
1JPC = 132 Hz), 62.4 (t), 71.0, 110.3, 137.9, 128.4, 129.5, 135.4,
158.8 ppm. MS m/z 585 (M+H)+. Anal. Calcd for C21H34N2O11P2S:
C, 43.15; H, 5.86; N, 4.79. Found: C, 43.36; H, 6.03; N, 4.68.
4.1.2. General procedure for the preparation of acids 12, 12a,
13, 13a
Bromotrimethylsilane (2.30 mL; 17.6 mmol; 10 equiv) was
added over 15 min to a solution of the appropriate ester (10, 10a,
11, 11a) (1.76 mmol) in dry CH2Cl2 (10 mL) under inert atmo-
sphere. The mixture was stirred at room temperature for 3.5 h,
then ice cooled and slowly diluted with MeOH (15 mL). The result-
ing mixture was stirred at room temperature for 30 min, then con-
centrated under reduced pressure. After dissolving the crude
product in MeOH (15 mL), the resulting solution was concentrated
under reduced pressure three times. The crude product was then
dissolved in CH3CN (2 mL); the resulting solution was diluted with
water (8 mL) and the mixture was purified by flash RP-18 chroma-
tography (eluent CH3CN/H2O 2:8 v/v) obtaining a foamy solid. The
desired compound was obtained as a white solid triturating with
dry Et2O.
4.1.3.1. 5-(4-Phenylfurazan-3-yloxy)pentan-1-ol (15). Eluent:
PE/EtOAc from 80:20 to 60:40 v/v, white solid, 86% yield; mp
59–60 °C (iPr2O). 1H NMR (300 MHz, CDCl3, TMS): d = 1.34 (s, 1H,
OH), 1.50–1.70 (m, 4H), 1.88–1.97 (m, 2H) (–OCH2CH2CH2CH2-
3
CH2OH), 3.66 (t, 2H, JHH = 6.0 Hz, –CH2OH), 4.44 (t, 2H,
3JHH = 6.6 Hz, –FzOCH2–), 7.44–7.48 (m, 3H, C6H5), 7.93–7.97 ppm
(m, 2H, C6H5); 13C NMR (75 MHz, CDCl3, TMS): d = 22.1, 28.6,
32.2, 62.6, 72.8, 125.3, 127.4, 129.0, 130.6, 145.2, 163.6 ppm. MS
m/z 249 (M+H)+. Anal. Calcd for C13H16N2O3: C, 62.89; H, 6.50; N,
11.28. Found: C, 62.50; H, 6.47; N, 11.19.
4.1.3.2. 5-(3-Phenylfuroxan-4-yloxy)pentan-1-ol (15a). Eluent:
PE/EtOAc 80:20 v/v, white solid, 98% yield; mp 91–92 °C (iPr2O).
1H NMR (300 MHz, CDCl3, TMS): d = 1.31 (s, 1H; –OH), 1.54–1.73
(m, 4H), 1.93–2.02 (m, 2H) (–OCH2CH2CH2CH2CH2OH), 3.70 (t,
4.1.2.1. 5-(4-Phenylfurazan-3-yloxy)pentylidenebis(phosphonic)-
acid (12). White solid; 82% yield. 1H NMR (200 MHz, DMSO-d6,
TMS): d = 1.60–2.20 (m, 7H, –CH2CH2CH2CH–), 4.42–4.44 (m, 2H, –
OCH2–), 7.58–7.98 (m, 5H, C6H5), 8.57 ppm (br s, 4H, mobile pro-
tons); 13C NMR (50 MHz, DMSO-d6, TMS; a signal is missing, being
covered by the solvent): d = 24.9, 28.2, 72.8, 124.5, 127.4, 129.5,
131.1, 145.3, 163.4 ppm; 31P NMR (120 MHz, D2O, H3PO4):
d = 23.4 ppm. Anal. Calcd for C13H18N2O8P2ꢂ0.15H2O: C, 39.53; H,
4.67; N, 7.09. Found: C, 39.17; H, 4.7; N, 7.03.
3
3
2H, JHH = 6.0 Hz, –CH2OH), 4.52 (t, 2H, JHH = 6.6 Hz, –FxOCH2–),
7.44–7.54 (m, 3H, C6H5), 8.12–8.15 ppm (m, 2H, C6H5); 13C NMR
(75 MHz, CDCl3, TMS): d = 22.2, 28.5, 32.1, 62.6, 70.8, 107.6,
122.5, 126.2, 128.9, 130.4, 162.4 ppm. MS m/z 265 (M+H)+. Anal.
Calcd for C13H16N2O4: C, 59.08; H, 6.10; N, 10.60. Found: C,
58.98; H, 6.07; N, 10.57.
4.1.3.3. 5-(4-Phenylsulfonylfurazan-3-yloxy)pentan-1-ol (16).
Eluent: PE/EtOAc 60:40 v/v, colourless oil, 57% yield. 1H NMR
(300 MHz, CDCl3, TMS): d = 1.39–1.61 (m, 5H), 1.76–1.85 (m, 2H)
4.1.2.2. 5-(3-Phenylfuroxan-4-yloxy)pentylidenebis(phosphonic)-
acid (12a). White solid; 67% yield. 1H NMR (200 MHz, DMSO-d6,
TMS): d = 1.60–2.20 (m, 7H, CH2CH2CH2CH–), 4.48 (m, 2H, –OCH2–),
7.56–8.08 (m, 5H, C6H5), 9.48 ppm (br s, 4H, mobile protons); 13C
NMR (50 MHz, DMSO-d6, TMS; a signal is missing, being covered by
the solvent): d = 24.9, 28.0, 71.7, 107.6, 122.1, 123.3, 129.2, 130.7,
162.4 ppm; 31P NMR (120 MHz, D2O, H3PO4): d = 23.8 ppm. Anal.
3
(–OCH2CH2CH2CH2CH2OH), 3.61 (t, 2H, JHH = 6.2 Hz, –CH2OH),
3
4.31 (t, 2H, JHH = 6.4 Hz, –FzOCH2–), 7.54–7.59 (t, 2H, C6H5),
7.67–7.72 (t, 1H, C6H5), 8.00–8.03 ppm (d, 2H, C6H5); 13C NMR
(75 MHz, CDCl3, TMS): d = 21.7, 28.0, 31.8, 62.2, 73.6, 128.7,
129.4, 135.2, 137.5, 148.6, 161.1 ppm. MS m/z 313 (M+H)+. Anal.

M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
2435
Calcd for C13H16N2O5S: C, 49.99; H, 5.16; N, 8.97. Found: C, 49.72;
H, 5.25; N, 9.00.
Calcd for C13H14N2O7S: C, 45.61; H, 4.12; N, 8.18. Found: C, 45.96;
H, 4.14; N, 8.13.
4.1.5. General procedure for the preparation of bisphosphonic
acids 19, 19a, 20, 20a
4.1.3.4. 5-(3-Phenylsulfonylfuroxan-4-yloxy)pentan-1-ol
(16a). Eluent: PE/EtOAc 50:50 v/v, white solid, 59% yield; mp
72 °C (iPr2O). 1H NMR (200 MHz, CDCl3): d = 1.42 (br s, 1H, –OH),
1.49–2.00 (m, 6H, –OCH2CH2CH2CH2CH2OH), 3.69–3.74 (m, 2H, –
A solution of the appropriate carboxylic acid (17, 17a, 18, 18a)
(1.0 g) in thionyl chloride (10 mL, fresh distilled) was stirred at
room temperature for 3 h under N2. Thionyl chloride was then dis-
tilled under reduced pressure, and the system was refilled with N2.
Thereafter, dry THF (10 mL) was added to the residue and then dis-
tilled under reduced pressure three times. The residue was dis-
solved in dry THF (10 mL) and to the resulting solution, stirred
under N2 at 0 °C, was added trimethylphosphite (1 equiv). The mix-
ture was then allowed to reach room temperature and stirred for
18 h. The reaction mixture containing the acylphosphonate (B)
was slowly added to a stirred solution of dimethylphosphite
(2 equiv) and diethylamine (0.4 equiv) in dry THF (20 mL) under
N2 and kept at 0 °C. The mixture was stirred at 0 °C for 3 h, allowed
to reach room temperature, and then poured into 0.5 M HCl
(20 mL) and Et2O (30 mL). The ethereal fraction was extracted with
0.5 M HCl (25 ꢁ 20 mL). The collected aqueous layers were ex-
tracted with CH2Cl2 (10 ꢁ 50 mL). The organic layers were dried
and concentrated under reduced pressure. The crude product was
purified by flash chromatography (CH2Cl2/MeOH 98:2 v/v to 9:1
v/v) to yield the unstable tetramethylester (C) as pale yellow oil.
The compounds were immediately used for the next synthetic step.
Bromotrimethylsilane (2.5 equiv) was slowly added to a solution of
the appropriate tetramethylester (C) (0.70 g) in dry CH2Cl2 (20 mL),
then stirred under inert atmosphere for 48 h. The reaction was fol-
lowed via RP-18 TLC (CH3CN/H2O 5:5 v/v). The solution was cooled
at 0 °C, then dry MeOH (10 mL) was added and the mixture was al-
lowed to reach room temperature. The solution was then concen-
trated under reduced pressure. The residue was dissolved in dry
MeOH (10 mL) and concentrated under reduced pressure twice.
The crude material was purified by preparative HPLC (Lichrospher
250 ꢁ 25 C18, MeOH/H2O 3:7 v/v, flow rate 39 mL minꢀ1, k
224 nm, injection volume 2 mL) to give the desired product as a
white solid.
3
CH2OH), 4.45 (t, 2H, JHH = 6.3 Hz, –FxOCH2–), 7.60–7.81 (m, 3H,
C6H5), 8.07 ppm (d, 2H, C6H5); 13C NMR (50 MHz, CDCl3):
d = 21.7, 27.9, 31.8, 62.3, 71.2, 110.2, 128.3, 129.4, 135.4, 137.8,
158.8 ppm. MS m/z 329 (M+H)+. Anal. Calcd for C13H16N2-
O6Sꢂ0.25H2O: C, 46.91; H, 5.00; N, 8.42. Found: C, 47.03; H, 4.80;
N, 8.41.
4.1.4. General procedure for the preparation of carboxylic acids
17, 17a, 18, 18a
A solution of Jones reagent (2.5 M, 2.5 equiv) was added to a
stirred solution of the appropriate alcohol (15, 15a, 16, 16a)
(5.0 g) in acetone (100 mL), cooled at 0 °C. The mixture was al-
lowed to reach room temperature and then stirred for 18 h. iPrOH
(20 mL) was added, then the mixture was concentrated under re-
duced pressure. The residue was dissolved in Et2O (100 mL),
washed with H2O (3 ꢁ 100 mL) and then extracted with a satu-
rated solution of NaHCO3 (4 ꢁ 30 mL). The aqueous layers were
acidified with 6 M HCl and extracted with Et2O (3 ꢁ 100 mL). The
combined organic layers were dried, filtered and concentrated un-
der reduced pressure to give the pure compound as a white solid.
4.1.4.1. 5-(4-Phenylfurazan-3-yloxy)pentanoic acid (17). 72%
yield; mp 85.5–87.5 °C (iPr2O). 1H NMR (300 MHz, CDCl3, TMS):
d = 1.86–2.00 (m, 4H, –OCH2CH2CH2CH2COOH), 2.49 (t, 2H,
3
3JHH = 7.2 Hz, –OCH2CH2CH2CH2COOH), 4.48 (t, 2H, JHH = 6.0 Hz,
–FzOCH2–), 7.47–7.50 (m, 3H, C6H5), 7.97–8.00 (m, 2H, C6H5),
11.55 ppm (br s, 1H, COOH); 13C NMR (75 MHz, CDCl3, TMS):
d = 21.0, 28.2, 33.4, 72.2, 125.2, 127.4, 129.0, 130.7, 145.2, 163.6,
179.3 ppm. MS m/z 263 (M+H)+. Anal. Calcd for C13H14N2O4: C,
59.54; H, 5.38; N, 10.68. Found: C, 59.31; H, 5.37; N, 10.58.
4.1.4.2. 5-(3-Phenylfuroxan-4-yloxy)pentanoic acid (17a). 76%
yield; mp 113.5–114.5 °C (iPr2O). 1H NMR (300 MHz, CDCl3,
TMS): d = 1.84–2.05 (m, 4H, –OCH2CH2CH2CH2COOH), 2.49 (t, 2H,
4.1.5.1. (5-(4-Phenylfurazan-3-yloxy)-1-hydroxypentylidene)-
bis(phosphonic)acid (19). 15% yield. 1H NMR (300 MHz, D2O):
d = 1.60 (br s, 4H, –OCH2CH2CH2CH2–), 1.80–2.00 (m, 2H,-
OCH2CH2CH2CH2–), 4.05 (br s, 2H, –FzOCH2–), 7.07–7.15 (m, 3H,
C6H5), 7.46–7.48 ppm (m, 2H, C6H5); 13C NMR (75 MHz, D2O):
d = 19.8 (t, JPC = 6.0 Hz), 29.0, 33.6, 72.9, 73.6 (t, JPC = 146 Hz),
124.3, 127.1, 129.0, 130.9, 145.3, 163.4 ppm; 31P NMR (120 MHz,
D2O, H3PO4): d = 20.3 ppm. Anal. Calcd for C13H18N2O9P2ꢂ1.5H2O:
C, 35.87; H, 4.86; N, 6.44. Found: C, 35.99; H, 4.73; N, 6.44.
3
3JHH = 7.2 Hz, –OCH2CH2CH2CH2COOH), 4.52 (t, 2H, JHH = 6.3 Hz,
–FxOCH2–), 7.44–7.54 (m, 3H, C6H5), 8.10–8.17 (m, 2H, C6H5),
11.48 ppm (br s, 1H, COOH); 13C NMR (75 MHz, CDCl3, TMS):
d = 21.0, 28.0, 33.3, 70.3, 107.6, 122.4, 126.2, 128.9, 130.5, 162.3,
179.2 ppm. MS m/z 279 (M+H)+. Anal. Calcd for C13H14N2O5: C,
56.11; H, 5.07; N, 10.07. Found: C, 56.10; H, 5.11; N, 10.03.
2
1
4.1.4.3. 5-(4-Phenylsulfonylfurazan-3-yloxy)pentanoic acid
(18). 65% yield; mp 96–97 °C (iPr2O). 1H NMR (200 MHz, CDCl3):
d = 1.74–2.00 (m, 4H, –OCH2CH2CH2CH2COOH), 2.47 (t, 2H,
4.1.5.2. (5-(3-Phenylfuroxan-4-yloxy)-1-hydroxypentylidene)-
bis(phosphonic)acid (19a). 31% yield. 1H NMR (300 MHz, D2O):
d = 1.78–2.04 (m, 6H, –OCH2CH2CH2CH2–), 4.49 (t, 2H, 3JHH = 6.9 Hz,
–FxOCH2–), 7.54–7.56 (m, 3H, C6H5), 7.99–8.00 ppm (m, 2H, C6H5);
3
3JHH = 7.1 Hz, –OCH2CH2CH2CH2COOH), 4.41 (t, 2H, JHH = 5.8 Hz,
–FzOCH2–), 7.60–7.81 (m, 3H, C6H5), 8.11 ppm (d, 2H, C6H5); 13C
NMR (50 MHz, CDCl3): d = 20.7, 27.7, 33.1, 73.2, 128.6, 129.5,
135.3, 137.6, 146.6, 161.1, 177.0 ppm. MS m/z 327 (M+H)+. Anal.
Calcd for C13H14N2O6S: C, 47.85; H, 4.32; N, 8.58. Found: C,
47.63; H, 4.29; N, 8.50.
2
13C NMR (75 MHz, D2O): d = 20.7 (t, JPC = 5.5 Hz), 29.6, 36.1, 72.2,
1
76.8 (t, JPC = 134.0 Hz), 110.3, 121.7, 126.7, 129.2, 131.4,
163.0 ppm; 31P NMR (120 MHz, D2O, H3PO4): d = 20.4 ppm. Anal.
Calcd for C13H18N2O10P2ꢂ2.0H2O: C, 33.92; H, 4.82; N, 6.09. Found:
C, 33.93; H, 4.56; N, 6.04.
4.1.4.4. 5-(3-Phenylsulfonylfuroxan-4-yloxy)pentanoic acid
(18a). 70% yield; mp 117–120 °C (iPr2O). 1H NMR (200 MHz,
CDCl3): d = 1.79–2.04 (m, 4H, –OCH2CH2CH2CH2COOH), 2.50 (t, 2H,
4.1.5.3. (5-(4-Phenylsulfonylfurazan-3-yloxy)-1-hydroxypenty-
lidene)bis(phosphonic)acid (20). 25% yield. 1H NMR (200 MHz,
D2O): d = 1.70–1.88 (m, 4H), 1.98–2.12 (m, 2H, –OCH2CH2CH2CH2–
3
3JHH = 7.1 Hz, –OCH2CH2CH2CH2COOH), 4.45 (t, 2H, JHH = 5.8 Hz, –
3
), 4.35 (t, 2H, JHH = 6.0 Hz, –FzOCH2–), 7.70 (t, 2H, C6H5), 7.83 (t,
FxOCH2–), 7.58–7.81 (m, 3H, C6H5), 8.06 ppm (d, 2H, C6H5); 13C
NMR (50 MHz, CDCl3): d = 20.6, 27.5, 33.1, 70.7, 110.2, 128.3,
129.5, 135.4, 137.8, 158.7, 179.2 ppm. MS m/z 343 (M+H)+. Anal.
1H, C6H5), 8.06 ppm (d, 2H, C6H5); 13C NMR (50 MHz, D2O):
1
d = 17.0, 26.0, 30.5, 70.6 (t, JPC = 143.0 Hz), 71.4, 126.1, 127.4,
133.1, 133.7, 145.7, 158.6 ppm; 31P NMR (80 MHz, D2O, H3PO4):

2436
M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
1
d = 20.3 ppm. Anal. Calcd for C13H18N2O11P2Sꢂ0.75H2O: C, 32.14; H,
2JPC = 6.7 Hz), 27.4, 38.2, 47.7, 55.1, 70.2 (t, JPC = 140.0 Hz), 108.5,
4.04; N, 5.77. Found: C, 31.78; H, 4.09; N, 5.66.
147.4, 155.1 ppm. Anal. Calcd for C9H18N4O10P2ꢂ1.5 H20: C, 25.07;
H, 4.91; N, 12.99. Found: C, 25.27; H, 4.82; N, 12.63.
4.1.5.4. (5-(3-Phenylsulfonylfuroxan-4-yloxy)-1-hydroxypenty-
lidene)bis(phosphonic)acid (20a). 30% yield. 1H NMR (200 MHz,
D2O): d = 1.71–2.05 (m, 6H, –OCH2CH2CH2CH2–), 4.34 (t, 2H,
3JHH = 5.6 Hz, –FxOCH2–), 7.55–7.78 (m, 3H, C6H5), 7.95 ppm (d,
2H, C6H5); 13C NMR (50 MHz, D2O): d = 17.2, 25.9, 30.4, 69.1, 70.7
4.1.7. Determination of pKa values
Potentiometric titrations of compounds were performed with
the GlpKa apparatus (Sirius Analytical Instruments Ltd, Forest
Row, East Sussex, UK). Ionisation constants were determined as
outlined in literature: the aqueous titrations were carried out un-
der N2 at 25.0 0.1 °C, and final data were obtained with the Mul-
tiset approach.35 Ionisation constant measurement of compounds
24, 24a, 25, 25a was repeated in different methanol-water mix-
tures (18–40 wt %); aqueous pKa values were obtained by extrapo-
lation at 0% methanol using the Yasuda-Shedlovsky procedure.22
1
(t, JPC = 141.0 Hz), 108.5, 125.8, 127.3, 133.5, 133.8, 156.7 ppm;
31P NMR (80 MHz, D2O, H3PO4): d = 20.2 ppm. Anal. Calcd for
C13H18N2O12P2Sꢂ2.5H2O: C, 29.28; H, 4.35; N, 5.25. Found: C,
29.18; H, 4.09; N, 5.07.
4.1.6. General procedure for the preparation of bisphosphonic
acids 24, 24a, 25, 25a
The pH of a solution of 23 (0.36 g; 1.38 mmol) in water (4 mL)
was adjusted to 12 by slowly adding 2 M NaOH, then a solution
of the appropriate bromomethyl-derivative (21, 21a, 22, 22a)
(0.9 equiv) in CH3CN (4 mL) was added. The mixture was stirred
at room temperature maintaining the pH at 12 by addition of
2 M NaOH. Subsequently, phenylsulfonyl chloride (1 equiv) was
added, the mixture was stirred at pH 12 until a clear solution
was obtained, then pH was adjusted to 8 with 1 M HCl. The mix-
ture was washed with CH2Cl2 (3 ꢁ 10 mL), then concentrated un-
der reduced pressure. The residue (10 mL) was percolated
through Dowex 50W-X8 (200–400 mesh hydrogen form; 20 mL re-
sin bed, 34 mequiv) eluting with water until neutrality of the elu-
ate. The eluate was concentrated under reduced pressure to obtain
the desired compound as a white solid.
4.1.8. Adsorption of BPs to HAP
HAP (150 mg, Aldrich) was equilibrated in 50 mL of 0.05 M Tris–
HCl pH 7.4 for 24 h at 37 °C according to literature.17,19 The BPs
(1 mM) were incubated in HAP suspension at 37 °C and magneti-
cally stirred. After 0 h, 6 h and 24 h, 1 mL of suspension was centri-
fuged (10,000 rpm, 10 min), and the concentration of the BPs in the
supernatant was determined by RP-HPLC. The HPLC system con-
sisted of an Agilent 1200 chromatograph system equipped with a
multiple wavelength detector SL (Agilent) and a ZORBAX Eclipse
XDB C18 column (150 ꢁ 4 mm, 5
lm, Agilent). The mobile phase
consisted of methanol/10 mM phosphate buffer pH 7.4 containing
1 mM tetrabutylammonium hydrogen sulfate (60:40 v/v for com-
pounds 12, 12a, 13, 13a, 19, 19a, 20, 20a; 40:60 v/v for compounds
24, 24a, 25, 25a). The flow-rate was 1.2 mL minꢀ1, and the injec-
tion volume was 20
lL (Rheodyne, Cotati, CA). The peaks were
4.1.6.1. [4-(N-Methyl-N-[(4-methylfurazan-3-yl)methyl]-1-hydro-
xybutylidene]bis(phosphonic)acid (24). 36% yield; mp 138 °C. 1H
NMR (200 MHz, D2O): d = 1.74–2.14 (m, 4H, –CH2CH2C(OH)[-
PO(OH)2]2), 2.30 (s, 3H, FzCH3), 2.87 (s, 3H, –CH2N(CH3)CH2CH2–),
monitored at 226 nm.
4.2. Biological methods
3.09–3.37 (m, 2H, –CH2N(CH3)CH2CH2–), 4.52 ppm (s, 2H,
–
Recombinant human M-CSF was purchased from R&D Systems
(Abingdon, UK) and recombinant human RANKL was from Alexis
(San Diego, CA). Samples from peripheral blood (PB) were obtained
from healthy controls, showing normal bone metabolism (evalu-
ated by bone densitometry). In order to enrol a homogeneous
group of healthy controls we adopted the following exclusion cri-
teria: chronic diseases; therapies with medications which could in-
crease osteoclastogenesis (steroids, glucocorticoids); therapies
with calcium, vitamin D, phosphorus. Informed consent was ob-
tained to comply with institutional policies.
CH2N(CH3)CH2CH2–); 13C NMR (50 MHz, D2O): d = 4.54, 16.5 (t,
1
2JPC = 6.3 Hz), 27.5, 37.8, 44.9, 54.4, 70.1 (t, JPC = 140.0 Hz), 144.6,
148.5 ppm. Anal. Calcd for C9H19N3O8P2ꢂ1.1H20: C, 28.52; H, 5.64;
N, 11.07. Found: C, 28.17; H, 5.40; N, 10.88.
4.1.6.2. [4-(N-Methyl-N-[(3-methylfuroxan-4-yl)methyl]-1-hydro-
xybutylidene]bis(phosphonic)acid (24a). 71% yield; mp 214 °C
dec. 1H NMR (200 MHz, D2O): d = 1.70–2.07 (m, 4H, –CH2CH2C(OH)-
[PO(OH)2]2), 2.07 (s, 3H, FxCH3), 2.87 (s, 3H, –CH2N(CH3)CH2CH2–),
3.06–3.41 (m, 2H, –CH2N(CH3)CH2CH2–), 4.45 ppm (s, 2H, –CH2-
N(CH3)CH2CH2–); 13C NMR (50 MHz, D2O): d = 4.39, 16.5 (t,
4.2.1. Cell cultures
1
2JPC = 6.8), 27.5, 37.9, 46.2, 54.4, 70.2 (t, JPC = 140.0 Hz), 113.3,
PBMCs, collected from healthy controls, were isolated after cen-
trifugation over a density gradient (Lymphoprep, Nycomed Phar-
ma, Norway) and cultured in a 24-well plate at a density of
148.0 ppm. Anal. Calcd for C9H19N3O9P2ꢂ1.5H20: C, 26.87; H, 5.51;
N, 10.45. Found: C, 26.54; H, 5.30; N, 10.17.
2 ꢁ 106 cells mLꢀ1 in
a-MEM (Invitrogen, UK) supplemented with
4.1.6.3. [4-(N-Methyl-N-[(4-carbamoylfurazan-3-yl)methyl]-1-
hydroxybutylidene]bis(phosphonic)acid (25). 34% yield. 1H
NMR (200 MHz, D2O): d = 1.76–2.18 (m, 4H, –CH2CH2C(OH)-
[PO(OH)2]2), 2.89 (s, 3H, –CH2N(CH3)CH2CH2–), 3.18–3.41 (m, 2H,
–CH2N(CH3)CH2CH2–), 4.74–4.87 ppm (m, 2H, –CH2N(CH3)-
10% Fetal Bovine Serum (FBS, BioWhittaker, Valkersville, MD), pen-
icillin (100 U mLꢀ1) and streptomycin (100
fully differentiated human OCs, PBMCs were then cultured with/
without M-CSF (25 ng mLꢀ1) from day 0 and RANKL (30 ng mLꢀ1
l
g mLꢀ1). To obtain
)
from day 10. Culture supernatants were changed on days 5 and
10, when medium was either refreshed and supplemented or not
with the agents described above. They were all prepared as stock
solutions at 10ꢀ3 M concentration in distilled water, aliquoted
and stored at ꢀ20 °C. Cultures were stopped after 15 days, mature
OCs were identified as multinucleated cells containing three or
2
CH2CH2–); 13C NMR (50 MHz, D2O): d = 16.4 (t, JPC = 7.3 Hz),
1
27.4, 38.2, 45.7, 54.9, 70.2 (t, JPC = 140 Hz), 144.9, 145.9,
157.3 ppm. Anal. Calcd for C9H17NaN4O9P2ꢂ1.14 H20: C, 25.10; H,
4.51; N, 13.01. Found: C, 25.50; H, 4.65; N, 12.60.
4.1.6.4. [4-(N-Methyl-N-[(3-carbamoylfuroxan-4-yl)methyl]-1-
hydroxybutylidene]bis(phosphonic)acid (25a). 40% yield. 1H
NMR (200 MHz, D2O): d = 1.81–2.04 (m, 4H, –CH2CH2C(OH)-
[PO(OH)2]2), 2.89 (s, 3H, –CH2N(CH3)CH2CH2–), 3.29 (br t, 2H, –
CH2N(CH3)CH2CH2–), 4.70 ppm (covered by the solvent peak, 2H, –
CH2N(CH3)CH2CH2–); 13C NMR (50 MHz, D2O): d = 16.4 (t,
more nuclei and positive for the expression of TRAP and
(vitronectin receptor).
aVb3
4.2.2. Culture of RAW 246.7 cells
As a model system of osteoclastogenesis we have used a clone
of RAW 246.7 murine monocyte/macrophagic cell line (TIB-71),

M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
2437
purchased from American Type Culture Collection (ATCC, Rockville,
4.2.5. Inhibition of squalene synthase activity in rat liver
microsomes
MD). RAW 246.7 cells were cultured in a 24-well plate at a density
of 1 ꢁ 104 cell mLꢀ1 in
a
-MEM (Invitrogen, UK) supplemented with
The
substrate
[1-3H]-FPP
(0.02
l
Ci,
26.2 Ci mmolꢀ1
,
10% Fetal Bovine Serum (FBS, BioWhittaker, Walkersville, MD),
969.4 GBq mmolꢀ1
) (PerkinElmer Life and Analytical Science),
penicillin (100 U mLꢀ1) and streptomycin (100
l
g mLꢀ1). After
0.2 mM U14266A (oxidosqualene cyclase inhibitor) and different
concentrations of bisphosphonates, dissolved in methanol, were
added to test tubes containing Tween-80 (final concentration
0.2 mg mLꢀ1). To the mixture freed from solvent and redissolved
24 h of culture, media were changed and replaced by one contain-
ing RANKL (30 ng mLꢀ1) and tested compounds. At day 3 media
were removed and cells were washed with PBS and subsequently
fixed for 30 min in a solution of 4% paraformaldehyde and stained
for tartrate-resistant acid phosphatase solution (TRAP, Sigma)
according to the manufacturer.
in 50
l
L of PBS containing 5 mM MgCl2, 30 mM nicotinamide,
L of microsomes (3 mg protein)
20 mM NaF and 2 mM NADPH, 50
l
were added. Sodium ibandronate, dissolved in PBS, was used as ref-
erence inhibitor. Assay tubes were flushed with N2, capped, and
incubated at 30 °C for 60 min. The enzymatic reaction was stopped
by adding methanolic KOH (1 mL, 10% w/v) and heating at 80 °C for
30 min in a water bath. The reaction mixture was extracted twice
with petroleum ether (1.5 and 1 mL) and the synthesised [3H]-
squalene was counted in Ready GelTM (Beckman) using a Packard
liquid scintillation counter. Control incubations were carried out in
the same conditions, in the absence of bisphosphonate inhibitors.
The efficiency of the petroleum ether extraction was about 90%,
as indicated by the recovery of [14C]-squalene added at the end
of the enzymatic reaction as a tracer. Extracts from both control
and inhibition tests were analysed by TLC, using cyclohexane/ethyl
acetate (85:15, v/v) as a developing solvent and authentic stan-
dards of cholesterol, lanosterol, oxidosqualene, and squalene as
reference compounds. Labelled products were determined by scan-
ning with a System 200 Imaging Scanner (Hewlett–Packard, Palo
Alto, CA, USA). The radiochromatographic analysis showed that,
in the control test, 95% of radioactivity accumulated into a single
peak co-chromatographing with squalene. In order to calculate
the IC50 values, the results were expressed as % of radioactivity ex-
tracted with respect to the control incubations in the absence of
inhibitors.
4.2.3. Microsomal enzyme preparation
Cellular fractions (10,000g, 100,000g supernatants and micro-
somes, respectively) of rat liver were prepared as described by
Popjak.36 All procedures were performed at 4 °C. Briefly, livers
were homogenised in 2.5 volumes of 0.1 M potassium phosphate
buffer, pH 7.4, containing 5 mM MgCl2 and 30 mM nicotinamide.
Homogenates were centrifuged at 10,000g for 30 min and the
10,000g supernatant (S10) was aliquoted and stored at ꢀ80 °C.
To obtain the microsomes, the S10 was further centrifuged at
100,000g for 60 min. The pelleted microsomes were resuspended
in one-tenth volume of the original homogenate. Microsomes
and 100,000g supernatant (S100) were aliquoted and stored at
ꢀ80 °C. Proteins were assayed by the BCA method (Pierce) using
bovine serum albumin as a standard.
4.2.4. Inhibition of the section of sterol biosynthesis ranging
from mevalonic acid to squalene in rat liver S10 or S100 cellular
fractions
For theinhibitiontest, bisphosphonates 12, 12a, 20, 20a were dis-
solved in methanol and sodium ibandronate, tested as a reference
inhibitor, was dissolved in 0.1 M sodium phosphate buffer, pH 7.4
(PBS). Rat liver S10 or S100 were incubated with (R,S)-[2-14C]meva-
4.2.6. Statistical analyses
lonic acid (0.05 l
Ci, 55 mCi mmolꢀ1, 2.04 GBq mmolꢀ1) (Amersham
Statistical analyses were performed with Prism 5.0. We com-
pared the results by means of Student’s paired t-test. The results
were considered statistically significant for p <0.05. IC50 values
(inhibitor concentrations that decreased enzymatic conversion by
50%) relative to squalene synthase activity were calculated by non-
linear regression analysis of the residual activity versus the log of
inhibitor concentration using statistical software XLfit 5.1 from
IDBS (Guildford, Surrey, UK).
Pharmacia Biotech, UK), in the presence of 0.2 mM U14266A, an
oxidosqualene cyclase inhibitor.37 The labelled substrate,
U14266A, bisphosphonates 12, 12a, 20, 20a were added as organic
solution to test tubes containing Tween-80 (final concentration
0.2 mg mLꢀ1) and the solvent was evaporated under a stream of
nitrogen. Residues were redissolved in 50
5 mM MgCl2, 4 mM MnCl2, 30 mM nicotinamide, 2 mM ATP. 50
lL of PBS, containing
lL
of S10 or S100 (2 mg protein) were then added to test tubes. Assay
tubes were flushed with N2, capped, and incubated at 30 °C for
60 min. The inhibition by sodium ibandronate was tested under
the same conditions, except for the addition of the inhibitor, which
was directly transferred as a PBS solution into the test tubes contain-
ing the reaction mixture freed from solvent. The enzymatic reaction
4.2.7. Vasodilator activities
Thoracic aortas were isolated from male Wistar rats weighing
180–200 g. As few animals as possible were used. The purposes
and the protocols of our studies were approved by the Ministero
della Salute, Rome, Italy. The endothelium was removed and the
vessels were helically cut: three strips were obtained from each
aorta. The tissues were mounted under 1.0 g tension in organ baths
containing 30 mL of Krebs-bicarbonate buffer with the following
composition (mM): NaCl 111.2, KCl 5.0, CaCl2 2.5, MgSO4 1.2,
KH2PO4 1.0, NaHCO3 12.0, glucose 11.1, maintained at 37 °C and
gassed with 95% O2–5% CO2 (pH 7.4). The aortic strips were al-
was stopped by adding 20 lL of 5 M HCl and the mixture was kept at
37 °C for 30 min to complete the hydrolysis of prenyl pyrophos-
phate. Hydrolysis products were extracted twice with petroleum
ether (1.5 and 1 mL). Extracts were spotted on TLC plates (Alufolien
Kieselgel 60F254, Merck, Darmstadt, Germany) using benzene/ace-
tone (90:10 v/v) as the developing solvent, and authentic standards
of geraniol, farnesol, squalene, oxidosqualene, lanosterol and cho-
lesterol as reference compounds. Labelled metabolites were deter-
mined by scanning TLC plates with a System 200 Imaging Scanner
(Hewlett–Packard, Palo Alto, CA, USA). Prenol references were
detected by iodine vapour. Under our experimental conditions
(plugged test tubes with no NADPH added, in the presence of an
oxidosqualene cyclase inhibitor) only one major radioactive product
(about 80% of total extract radioactivity) co-chromatographing with
squalene was detected in control tests (no bisphosphonate inhibi-
tors) incubated with the S10 cellular fraction.
lowed to equilibrate for 120 min and then contracted with 1
-phenylephrine. When the response to the agonist reached a pla-
teau, cumulative concentrations of the vasodilating agent were
added. Results are expressed as EC50 SE ( M). The effects of
M ODQ on relaxation were evaluated in separate series of
lM
L
l
1
l
experiments. ODQ was added to the organ bath 5 minutes before
the contraction. Responses were recorded by an isometric trans-
ducer connected to the MacLab System PowerLab (ADInstruments
Ltd, Oxfordshire, UK). Addition of the drug vehicle (DMSO) had no
appreciable effect on contraction level.

2438
M. L. Lolli et al. / Bioorg. Med. Chem. 18 (2010) 2428–2438
4. Dunford, J. E.; Kwaasi, A. A.; Rogers, M. J.; Barnett, B. L.; Ebetino, F. H.; Russell, R.
4.3. Molecular modelling
G. G.; Oppermann, U.; Kavanagh, K. L. J. Med. Chem. 2008, 51, 2187.
5. Green, J. R.; Clezardin, P. Am. J. Clin. Oncol. (CCT) 2002, 25, S3.
6. Rogers, M. J.; Frith, J. C.; Luckman, S. P.; Coxon, F. P.; Benford, H. L.; Mönkkönen,
J.; Auriola, S.; Chilton, K. M.; Russell, R. G. G. Bone 1999, 24, 73S.
7. Wildler, L.; Jaeggi, K. A.; Glatt, M.; Müller, K.; Bachmann, R.; Bisping, M.; Born,
A.-R.; Cortesi, R.; Guiglia, G.; Jeker, H.; Klein, R.; Ramseier, U.; Schmid, J.;
Schreiber, G.; Seltenmeyer, Y.; Green, J. R. J. Med. Chem. 2002, 45, 3721.
8. Szabo, C. M.; Matsumura, Y.; Fukura, S.; Martin, M. B.; Sanders, J. M.; Sengupta,
S.; Cieslak, J. A.; Loftus, T. C.; Lea, C. R.; Lee, H.-J.; Koohang, A.; Coates, R. M.;
Sagami, H.; Oldfield, E. J. Med. Chem. 2002, 45, 2185.
The molecular models of ibandronate, 20 and 20a were con-
structed using standard bond lengths and angles with the MOE
software package.38 In accordance with their pKa values (Table
1), all compounds were modelled with three of the four phospho-
nate hydroxyl groups in their ionised state; in the case of ibandro-
nate the basic nitrogen atom was modelled as protonated.
Following truncated Newton–Raphson geometry optimisation
with the MMFF94s force field (MMFF94 charges, GB/SA implicit
solvent model) until the gradient was lower than
0.001 kcal molꢀ1 Åꢀ1, a Monte Carlo conformational search was
carried out with the StochasticCSearch module implemented in
MOE. The most stable conformer was then chosen for further opti-
misation by an ab initio HF/6–31G(d) method using the GAMESS-
US package;39 on the equilibrium geometry, atom-centred point
charges were derived by fitting the quantum-mechanical electro-
static potential with the RESP method.40 The experimental X-ray
structures of human FPPS in complex with ibandronate (PDB ID
2F94), human SQS in complex with CP-320473 (PDB ID 1EZF)
and human GGPPS in complex with geranylgeranyl diphosphate
(PDB ID 2Q80) were retrieved from the Protein Data Bank.41 While
only one monomer is present in the FPPS asymmetric unit, a trimer
and an hexamer are found for SQS and GGPPS, respectively. Since
neither active site involves the interface between monomers, only
one monomer was considered. Two short loops are missing from
the crystal structure of SQS and GGPPS (10-aminoacid and 6-ami-
noacid long, respectively). Since these loops do not contribute to
defining the shape of the active sites they were not modelled;
rather, the C- and N-chain terminals at the boundaries of the miss-
ing segments were simply capped with a N-methylamide and an
acetyl group, respectively. After removing co-crystallised inhibi-
tors, missing hydrogen atoms were added to the three enzymes
in standard positions, then their coordinates were optimised with
the SANDER module of the AMBER 10 software package,42 while
heavy atoms were harmonically restrained to the initial X-ray crys-
tal positions with a force constant of 1000 kcal molꢀ1 Åꢀ2. All
molecular mechanics optimisations on the proteins were accom-
plished using parameters and electrostatic charges of the AM-
BER99 force field. The volume of the active sites was computed
with the VOIDOO program.43 Docking of the inhibitors was carried
out using AUTODOCK 4.0.1.44–46 A grid with a 0.375 Å step size was
centred on the active site and energy maps were pre-calculated
9. Reszka, A.; Rodan, G. A. Curr. Rheumatol. Rep. 2003, 5, 65.
10. Ylitalo, R. Gen. Pharmacol. 2002, 35, 287.
11. McFarlane, S. I.; Muniyappa, R.; Shin, J. J.; Bahtiyar, G.; Sowers, J. R. Endocrine
2004, 23, 1.
12. Hamerman, D. Q. J. Med. 2005, 98, 467.
13. Gasco, A.; Schoenafinger, K. In Nitric Oxide Donors; Wang, P. G., Cai, T. B.,
Taniguchi, N., Eds.; Wiley-VCH: Weinheim, 2005; pp 131–175.
14. Kerwin, J. F., Jr; Lancaster, J. R.; Feldman, P. L. J. Med. Chem. 1995, 38, 4343.
15. Nitric Oxide and the Cardiovascular System; Loscalzo, J., Vita, J. A., Eds.; Humana
Press: Totowa, New Jersey, 2000.
16. Wallace, J. L. Br. J. Pharmacol. 2007, 152, 421.
17. Van’t Hof, R. J.; Ralstone, S. H. Immunology 2001, 103, 255.
18. Cummings, S. R.; Schwartz, A. V.; Black, D. M. N. Eng. Med. 2007, 356, 1895.
19. Defilippi, A.; Sorba, G.; Calvino, R.; Garrone, A.; Gasco, A.; Orsetti, M. Arch.
Pharm. (Weinheim) 1988, 321, 77.
20. Nancollas, G. H.; Tang, R.; Phipps, R. J.; Henneman, Z.; Gulde, S.; Wu, W.;
Mangood, A.; Russell, R. G. G.; Ebetino, F. H. Bone 2006, 38, 617.
21. Lazzarato, L.; Rolando, B.; Lolli, M. L.; Tron, G. C.; Fruttero, R.; Gasco, A.; Deleide,
G.; Guenther, H. L. J. Med. Chem. 2005, 48, 1322.
22. Avdeef, A.; Comer, J. E. A.; Thompson, S. J. Anal. Chem. 1993, 65, 42.
23. Cohen, H.; Solomon, V.; Alferiev, I. S.; Breuer, E.; Ornoy, A.; Patlas, N.; Eidelman,
N.; Hägele, G.; Golomb, G. Pharm. Res. 1998, 15, 606.
24. Rondeau, J.-M.; Bitsch, F.; Bourgier, E.; Geiser, M.; Hemmig, R.; Kroemer, M.;
Lehmann, S.; Ramage, P.; Rieffel, S.; Strauss, A.; Green, J. R.; Jahnke, W.
ChemMedChem 2006, 1, 267.
25. Pandit, J.; Danley, D. E.; Schulte, G. K.; Mazzalupo, S.; Pauly, T. A.; Hayward, C.
M.; Hamanaka, E. S.; Thompson, J. F.; Harwood, H. J., Jr. J. Biol. Chem. 2000, 275,
30610.
26. Kavanagh, K. L.; Dunford, J. E.; Bunkoczi, G.; Russell, R. G.; Oppermann, U. J. Biol.
Chem. 2006, 281, 22004.
27. Di Stilo, A.; Chegaev, K.; Lazzarato, L.; Fruttero, R.; Gasco, A.; Rastaldo, R.;
Cappello, S. Arzneim.-Forsch. 2009, 59, 111.
28. Jork, H.; Funk, W.; Fischer, W.; Wimmer, H.. In Thin Layer Chromatography;
Weinheim, VCH: Germany, 1990; Vol. 1a.
29. Calvino, R.; Mortarini, V.; Gasco, A.; Sanfilippo, A.; Ricciardi, M. L. Eur. J. Med.
Chem. 1980, 15, 485.
30. Boschi, D.; Di Stilo, A.; Cena, C.; Lolli, M. L.; Fruttero, R.; Gasco, A. Pharm. Res.
1997, 14, 1750.
31. Sorba, G.; Ermondi, G.; Fruttero, R.; Galli, U.; Gasco, A. J. Heterocycl. Chem. 1996,
33, 327.
32. Bertinaria, M.; Galli, U.; Sorba, G.; Fruttero, R.; Gasco, A.; Brenciaglia, M. I.;
Scaltrito, M. M.; Dubini, F. Drug. Dev. Res. 2003, 60, 225.
33. Kieczykowski, G. R.; Jobson, R. B.; Melillo, D. G.; Reinhold, D. F.; Grenda, V. J.;
Shinkai, I. J. Org. Chem. 1995, 60, 8310.
34. Di Stilo, A.; Visentin, S.; Cena, C.; Gasco, A. M.; Ermondi, G.; Gasco, A. J. Med.
Chem. 1998, 41, 5393.
with AUTOGRID, then flexible docking was carried out with AUTODOCK
.
35. Avdeef, A. J Pharm. Sci. 1993, 82, 183.
The target proteins were kept rigid, while ligands were left free
to explore the conformational space inside the enzyme cavities;
200 separate docking simulations were run on each target using
the Lamarckian genetic algorithm with default parameters. Dock-
ing poses were clustered according to their RMS deviation from
the starting reference structure.
36. Popjak, G. Methods Enzymol. 1969, 15, 393.
37. Field, R. B.; Holmlund, C. E.; Whittaker, N. F. Lipids 1979, 14, 741.
38. MOE Version 2008.10, Chemical Computing Group Inc.: Montreal, Quebec,
Canada, 2008.
39. Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen,
J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis,
M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347.
40. Wang, J.; Cieplak, P.; Kollman, P. A. J. Comput. Chem. 2000, 21, 1049.
41. The RCSB Protein Data Bank, http://www.rcsb.org/ (last accessed Jan 29 2010).
42. Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling, C. L.; Wang, J.; Duke,
R. E.; Luo, R.; Crowley, M.; Walker, R. C.; Zhang, W.; Merz, K. M.; Wang, B.;
Hayik, S.; Roitberg, A.; Seabra, G.; Kolossvary, I.; Wong, K. F.; Paesani, F.;
Vanicek, J.; Wu, X.; Brozell, S. R.; Steinbrecher, T.; Gohlke, H.; Yang, L.; Tan, C.;
Mongan, J.; Hornak, V.; Cui, G.; Mathews, D. H.; Seetin, M. G.; Sagui, C.; Babin,
V.; Kollman, P. A. AMBER 10, University of California: San Francisco (USA),
2008.
Acknowledgement
This work was supported by a grant from Regione Piemonte
(Ricerca sanitaria finalizzata 2008).
References and notes
43. Kleywegt, G. J.; Jones, T. A. Acta Crystallogr., Sect. D 1994, D50, 178.
44. Goodsell, D. S.; Olson, A. J. Proteins 1990, 8, 195.
1. Fleisch, H. Bisphosphonates in Bone Diseases, 3rd ed.; The Parthenon Publishing
Group Inc.: New York, 1997.
45. Morris, G. M.; Goodsell, D. S.; Huey, R.; Olson, A. J. J. Comput. Aided Mol. Des.
1996, 10, 293.
2. Boyle, W. J.; Simonet, W. S.; Lacey, D. L. Nature 2003, 423, 337.
3. Russell, R. G. G.; Croucher, P. I.; Rogers, M. J. Osteoporos. Int. 1999, 9, 66.
46. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.; Belew, R. K.;
Olson, A. J. J. Comput. Chem. 1998, 19, 1639.