Bisphosphonate complexation and calcium doping in silica xerogels as a combined strategy for local and controlled release of active platinum antitumor compounds

Article abstract of DOI:10.1039/b705239a

The production of bone substitute biomimetic materials which could also act as antitumoral drug release agents is of enormous interest. We report in this paper the synthesis and characterization of a novel platinum dinuclear complex containing a geminal bisphosphonate and its embodiment into xerogels prepared by the sol-gel method. Our goal was to obtain a hybrid inorganic matrix that could release a platinum species active against bone tumors or metastases, upon local implant. Two silica xerogels were considered: one was composed of pure silica, while the other contained also some calcium as potential release-modulating agent thanks to its high affinity for bisphophonates. The platinum-complex loading capacity of the inorganic matrices, the release kinetics in buffer simulating physiological conditions, and the stability upon storage were investigated as a function of Pt-complex concentration and calcium addition. We found that the presence of calcium in the composites deeply influences not only the stability of the formulations but also the nature of the platinum complex liberated in solution. The Royal Society of Chemistry.

Full text of DOI:10.1039/b705239a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Bisphosphonate complexation and calcium doping in silica xerogels as a
combined strategy for local and controlled release of active platinum
antitumor compounds
Nicola Margiotta,a Rosa Ostuni,a Deborah Teoli,b Margherita Morpurgo,b Nicola Realdon,b Barbara Palazzoc
and Giovanni Natile*a
Received 5th April 2007, Accepted 31st May 2007
First published as an Advance Article on the web 19th June 2007
DOI: 10.1039/b705239a
The production of bone substitute biomimetic materials which could also act as antitumoral drug
release agents is of enormous interest. We report in this paper the synthesis and characterization of a
novel platinum dinuclear complex containing a geminal bisphosphonate and its embodiment into
xerogels prepared by the sol–gel method. Our goal was to obtain a hybrid inorganic matrix that could
release a platinum species active against bone tumors or metastases, upon local implant. Two silica
xerogels were considered: one was composed of pure silica, while the other contained also some calcium
as potential release-modulating agent thanks to its high affinity for bisphophonates. The
platinum-complex loading capacity of the inorganic matrices, the release kinetics in buffer simulating
physiological conditions, and the stability upon storage were investigated as a function of Pt-complex
concentration and calcium addition. We found that the presence of calcium in the composites deeply
influences not only the stability of the formulations but also the nature of the platinum complex
liberated in solution.
Introduction
Nowadays cisplatin is one of the five most used agents against solid
tumors such as testicular, ovarian, and bladder carcinomas.1 How-
ever, the use of cisplatin is limited by some serious drawbacks such
as nausea, vomiting, ototoxicity, myelotoxicity and concentration-
dependent nephrotoxicity.2,3 Moreover it shows good activity
towards a limited spectrum of cancerogenic cells.
New platinum-based antitumor agents are being developed
in order to overcome the severe side-effects, to improve clinical
effectiveness, and to extend the spectrum of activity. A promising
strategy makes use of carrier ligands able to promote the specific
accumulation of the drug in target organs or cells.4 In this context,
platinum complexes with (aminoalkyl)phosphonic acids, firstly
synthesized and characterized by Appleton and colleagues,5,6
Chart
1
Sketches of platinum complexes with aminotris(methyl-
were exploited by the group of Keppler for their potential use
as anticancer drugs selective for bone tumors.7,8 This prospect
was justified by the clinical use of geminal bisphosphonates
(such as Etidronate, Clodronate, Pamidronate, and Alendronate)
in the treatment of hypercalcaemia and other bone related
malignancies.9–12 Pharmacological studies performed on platinum
complexes with amino-bis/tris(methylenephosphonate) ligands
(ATMP and BPMAA in Chart 1) proved that these multifunc-
tional molecules have a therapeutic activity superior to that of
enephosphonate) (ATMP), bis(phosphonomethyl)aminoacetate (BP-
MAA), and diethyl [(methylsulfinyl)methyl]phosphonate (SMP). A2 = two
amines or a chelated diamine. L2 = two chlorido, a dicarboxylate, or a
chelated diamine ligand(s).
cisplatininanorthotopicallytransplantedratosteosarcomamodel
which closely resembles the human osteosarcoma.13,14
In test animals the compounds containing bis(phos-
phonomethyl)aminoacetate (BPMAA) appear to have a lower
spectrum of toxicity and an higher therapeutic index and to
be better tolerated even in high doses than those contain-
ing aminotris(methylenephosphonate) (ATMP).15 These com-
pounds react with (di)nucleotides releasing the aminophosphonic
moiety.16
In order to limit the reactivity of Keppler’s compounds towards
(di)nucleotides we have recently proposed the use of a chelated
sulfinyl-phosphonate ligand (SMP in Chart 1), which is capable
aDipartimento Farmaco-Chimico, Universita` degli Studi di Bari. Via
E. Orabona 4, 70125, Bari, Italy. E-mail: natile@farmchim.uniba.it;
Fax: +39-080-5442230
bDipartimento di Scienze Farmaceutiche, Universita` degli Studi di Padova,
Via Marzolo 5, 35131, Padova, Italy
cDipartimento di Chimica “G. Ciamician” Alma Mater Studiorum Universita`
di Bologna, Via Selmi 2, 40126, Bologna, Italy
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 3131–3139 | 3131
©

View Article Online
of binding rather strongly to platinum through the sulfur atom
of the sulfoxide and to give chelation through the non-alkylated
phosphorus-bound oxygen atom.17,18
Commercial reagent grade chemicals and solvents (Sigma-
Aldrich) were used without further purification.
We also reasoned that the affinity of bisphosphonate moi-
eties for calcium ions may provide another means for bone
localization of platinum-bisphosphonate complexes. For exam-
ple, platinum-bisphosphonate complexes could be loaded onto
calcium-containing matrices to be used as bone fillers. Such
composites could be administered locally at the site of an
osteosarcoma and act both as bone substitutes and as local drug-
releasing systems with the final goal of inhibiting locally the tumor
re-growth and reducing the systemic toxicity.
Towards this end, we have recently investigated the adsorp-
tion and desorption of cisplatin, alendronate, and a platinum
complex containing deprotonated methylenebis(phosphonic acid)
(medronate) towards two bio-mimetic synthetic hydroxyapatite
nanocrystal materials having different crystal shape and chemico-
physical properties.19 The results showed a high loading capacity
of the synthetic hydroxyapatite for the bisphosphonate platinum
complex and the ability of the composite material to release a
bisphosphonate free Pt compound to the surrounding medium in
a slow and controlled manner.
In addition to hydroxyapatite, also nanoporous xerogels can
be used as implantable biodegradable and bioerodable delivery
systems. In particular, silica based bioactive xerogels are known to
be biocompatible and slowly bioerodable after implantation20–22
and have already been exploited for the entrapment and slow
release of cisplatin.23,24 Sol–gel silica is also known to be potentially
bioactive in the sense that surface hydroxyapatite deposition can
occur upon its immersion into a simulating physiological buffer.
The bioactivity of sol–gel silica polymers depends upon the gel
properties but is independent from the presence of calcium in the
polymer composition, this was not the case for fused bioactive
glasses originally developed by Hench.25
In this paper we describe the synthesis of a novel platinum com-
plex specifically designed to contain a geminal bisphosphonate
(Chart 2) and its loading into a silica-xerogel with the final aim
of obtaining a hybrid inorganic matrix that might be used for
treatment of bone tumors upon local implant.26
Instrumental measurements
UV spectra were measured using a Cary m50-Scan UV-Vis
spectrophotometer. Inductively Coupled Plasma Optical Emission
Spectrometry (ICP OES) measurements were performed using
a Perkin Elmer Optima 4200 DV spectrometer. NMR spectra
were recorded on a Bruker Avance DPX instrument operating
at 300 MHz (1H). Standard pulse sequences were used for H,
1
1
1
31P{ H} (121.5 MHz), and 195Pt{ H} (64.5 MHz) 1D spectra.
Chemical shifts are given in ppm. 1H chemical shifts were
referenced to internal sodium 3-(trimethylsilyl)propionate (TSP).
31P chemical shifts were referenced to external H3PO4 (85% w/w).
195Pt chemical shifts were referenced to external K2[PtCl4] in D2O
fixed at −1628 ppm. Elemental analyses were carried out with
a Hewlett Packard 185 C, H, and N analyzer. A Crison Micro-
pH meter Model 2002 equipped with Crison micro-combination
electrodes (5 and 3 mm diameter) and calibrated with Crison
standard buffer solutions at pH 2.00, 4.01, 7.02, and 9.26 was
used for pH measurements. The pH readings from the pH meter
for D2O solutions are indicated as pH* values and are uncorrected
for the effect of deuterium on glass electrodes.27
Synthesis of [PtCl2(en)] and [Pt(en)(H2O)(OSO3)] (en =
ethylenediamine). These complexes were prepared as already
reported in the literature.28
Synthesis of 1-hydroxy-2-(N-phthaloylamino)ethane-1,1-diyl-
bisphosphonic acid (1). N-Phthaloylglycine (1.5 g, 7.31 mmol)
was dissolved in SOCl2 (24 mL) and the resulting solution was kept
under reflux for 2 h. The solvent was then evaporated to dryness
under reduced pressure and the obtained pale yellow solid was
suspended in THF (12 mL). This suspension was treated at 0 ◦C
with tris(trimethylsilyl)phosphite (7.33 mL, 21.93 mmol) for 5 min,
then the mixture was stirred at room temperature for 10 min and,
subsequently, methanol was added to the solution. The resulting
mixture was kept under stirring for 1 h at room temperature
and then the volatile components were evaporated under vacuum
affording a yellow oil. This latter was treated several times with
diethyl ether until obtainment of a white solid, which was isolated
by filtration of the solution, abundantly washed with diethyl ether,
and finally dried under vacuum, affording 2.34 g of compound 1
1
(ca. 91% yield). H-NMR (D2O, pH* = 2.5): 7.92–7.82 (4H, m,
phenyl protons) and 4.36–4.30 (2H, m, CH2) ppm. 31P{ H}-NMR
1
(D2O, pH* = 2.5): 17.08 ppm.
Synthesis of 2-amino-1-hydroxyethane-1,1-diyl-bisphosphonic
acid (AHBP-H4, 2). Compound 1 (2.34 g, 6.65 mmol) was
dissolved in concentrated HCl (24 mL) and the resulting mixture
was kept under reflux overnight. The solvent was then removed
under reduced pressure and the brown solid suspended in ethanol
(40 mL) and kept at 70 ◦C (water bath) for 1 h. The solid was then
isolated by filtration of the solution, washed with warm ethanol
and with diethyl ether, and finally dried under vacuum. The pure
product was obtained by crystallization from water. Obtained
1.29 g (88% yield). Anal.: Calculated for C2H9NO7P2: C, 10.87;
Chart 2 Sketch of the novel compound with some atom numbering.
Material and methods
Chemicals
Ethylenediamine, N-phthaloylglycine, tris(trimethylsilyl)phos-
phite, tetraethoxysilane (TEOS), tetrahydrofuran (THF), and
Ba(OH)2 were purchased from Sigma-Aldrich (Milan, Italy).
1
H, 4.10; N, 6.34. Found: C, 10.71; H, 4.06; N, 6.19%. H-NMR
3132 | Dalton Trans., 2007, 3131–3139
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
3
(D2O, pH* = 3.5): 3.45 (2H, t, JH-P = 11.81 Hz, CH2) ppm.
tetraethoxysilane (TEOS, 1 equiv), ethanol (0.5 equiv), HCl
(0.01 equiv), and water (2 equiv) which could be either pure or
containing Ca(NO3)2 (0.33 equiv). Pre-hydrolysis was catalyzed
by 2 h sonication in an ice bath in the case of the sol containing
calcium and by 2 h heating at reflux in the case of the calcium-free
sample. In a second phase, the pre-activated sols were mixed with
different amounts of doubly distilled H2O, an aqueous solution of
the Pt(II) complex (45 mg mL−1), and a solution of HF (2.5 M) as a
catalyst for gelation. The final H2O : Si ratio (R) in all samples was
4. Gel formation occurred within 3.5 h independently from the
presence of the platinum complex. Wet gels were aged in a closed
vial for three days and, after removal of the liquid expelled upon
syneresis, they were dried at 50 ◦C until constant weight. Xerogels
were manually crushed in a mortar and sieved to obtain granules
with homogeneous size distribution (ø = 370 180 microns).
31P{ H}-NMR (D2O, pH* = 3.5): 16.46 ppm.
1
Synthesis of [Pt2(en)2(AHBP-H)](HSO4) (3). Compound 2
(124 mg, 0.561 mmol) was dissolved in H2O (10 mL) and the
resulting solution was treated with Ba(OH)2·8H2O (177 mg,
0.561 mmol). The obtained white suspension was treated with
a solution of [Pt(en)(H2O)(OSO3)] (423 mg, 1.146 mmol in 10 mL
of water) and the resulting mixture kept under stirring at 0 ◦C
for 3 h and then left standing at 4 ◦C overnight. After removal
of the white solid (BaSO4), the filtered solution (pH ca. 3.0) was
treated with additional Ba(OH)2·8H2O until the pH reached a
value of 5.0 (required 30.3 mg, 0.096 mmol). The newly formed
white solid was removed by filtration of the solution and this latter
taken to dryness by evaporation of the solvent under vacuum.
The yellow crystalline residue was dissolved in H2O (1.5 mL) and
the pH was lowered to 1.0 by addition of H2SO4 (0.5 M). The
resulting acidic solution was treated with methanol which caused
the precipitation of the desired product having HSO4− as counter-
ion. The precipitate was isolated by filtration of the solution,
washed with methanol and diethyl ether, and finally dried under
vacuum. Obtained 329 mg (80.6% yield). Anal.: Calculated for
[Pt2(en)2(AHBP-H)](HSO4)·5H2O (C6H32N5O16P2Pt2S): C, 7.88;
Determination of platinum content
To determine the amount of complex embedded in the final
formulations, weighted amounts of each xerogel were digested
by treatment with 10% HF and the Pt concentration in the
digested mixtures was determined by ICP-OES analysis. The
mole composition of the final xerogels is reported in Table 1.
For ICP analysis, the HF digested samples were diluted up to
a final 0.25% acid concentration and then nebulized using a
compact nebulizer having a small diameter capillary (Teflon) and
a polyvinylidine difluoride (PVDF) body to minimize undesired
large drop formation and improve HF tolerance. Four emission
lines were used for the determination of platinum. Solution
concentrations were obtained as an average of two records.
1
H, 3.53; N, 7.66. Found: C, 7.94; H, 3.12; N, 7.35%. H-NMR
(D2O, pH* = 3.5): 5.93 (4H, br, ethylenediamine NH2), 5.43 (4H,
3
br, ethylenediamine NH2), 3.38-3.31 (2H, dd, JH–P = 10.07 and
10.30 Hz, C(2)H2), 2.62 (8H, br, C(3)H2) ppm. 31P{ H}-NMR
1
(D2O, pH* = 3.5): 37.60 ppm.
NMR experiments at different pH and calculation of pKa values
Samples containing ca. 0.0016 mmol of the compound under
investigation (AHBP-H4 or [Pt2(en)2(AHBP-H)](HSO4)) dissolved
in 0.7 mL of D2O were transferred into an NMR tube. The pH*
of the samples was adjusted to the required values by addition
of 0.9 M D2SO4 and 2.0 or 0.5 M NaOD solutions and the pH*
value was measured by using a 3 mm diameter electrode for NMR
tube. No control of the ionic strength was performed. Portions of
the pH titration curve were fitted separately to the Henderson–
Hasselbalch equation using the program KaleidaGraph.29
Complex release, erosion, and stability of the formulations
Release of [Pt2(en)2(AHBP-H)]+ was investigated by immersing
the xerogel granules (200 mg) into Tris buffer (100 mL of 10 mM
trishydroxymethylaminomethane, 150 mM NaCl, pH 7.4) at 37 ◦C,
under constant stirring. At scheduled times, small fractions (2 mL)
of the solution were removed for Pt quantification and replaced
by fresh buffer. Pt concentration was determined by ICP OES
analysis. For this purpose, the samples were treated with ultrapure
nitric acid in order to obtain a metal concentration between 1 and
8 ppm in 1% nitric acid, and analysed as described above. The
reported concentrations are the average of two determinations. A
separate set of experiments was carried out to determine whether
different Pt species were released in the different phases of the
process. For this purpose, the xerogels were immersed in the buffer
solution as described above and, at scheduled times (1.5 and 20 h),
the receiving medium was completely removed (and replaced by
fresh buffer) and analyzed by UV-Vis spectrophotometry. The
Preparation of Pt(II) complex-loaded silica xerogels
Two xerogel formulations, differing in the presence or absence
of calcium (Ca(+) and Ca(−), respectively), were considered.
Each formulation was loaded with two different amounts of Pt(II)
complex (Table 1). Complex-free formulations were also used for
control.
The gels were synthesized by a two-step catalyzed proce-
dure. Initially, two pre-hydrolyzed sols were prepared by mixing
Table 1 Mole ratios of silica sol formulations and amount of loaded platinum complex
Formulation
Wet gel mole compositiona
Matrix mole compositionb
Pt(II) complex/mg g−1 xerogel
Ca(−)1
Ca(−)2
Ca(+)1
Ca(+)2
1 : 0 : 4 : 0.025 : 1.5 × 10−4
1 : 0 : 4 : 0.025 : 2.9 × 10−4
1 : 0.33 : 4 : 0.025 : 1.6 × 10−4
1 : 0.33 : 4 : 0.025 : 3.14 × 10−4
1 : 0 : 0.025
1 : 0 : 0.025
1 : 0.33 : 0.025
1 : 0.33 : 0.025
1.9
2.98
1.24
1.57
a Si: Ca: H2O: F−: Pt(II)-complex mole ratio. b SiO2: CaO: F− mole ratio.
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 3131–3139 | 3133
©

View Article Online
spectra of the release samples were compared with those of
[Pt2(en)2(AHBP-H)](HSO4) and of [PtCl2(en)].
product corresponding to a mononuclear species having one
bisphosphonate molecule per Pt(en) moiety. Unfortunately the two
products are extremely difficult to separate. Therefore we repeated
the reaction using two equivalents of [Pt(en)(H2O)(OSO3)] for
one equivalent of barium bisphosphonate so as to favor the
formation of the dinuclear species.32,33 The elemental analysis of
the resulting compound was in accordance with the presence of
one aminobisphosphonate and two Pt(en) residues (3, Chart 2).
The characterization of 3 has been performed via 1H, 31P, and 195Pt-
NMR in D2O. The 1H-NMR spectrum (Fig. 1) shows two broad
singlets at 5.93 and 5.43 ppm, assigned to the aminic protons
of coordinated en (geminal protons of each aminic group are
made inequivalent by the lack of symmetry with respect to the
coordination plane).34,35 The acidity conditions slow down the
exchange of the aminic protons with the deuterium of the solvent,
allowing detection of the aminic protons in water solution. The
doublet of doublets centered at 3.34 ppm was assigned to the
protons of the methylene group of the bisphosphonate (position
2 in Chart 2). These protons are shifted at higher field (Dd =
−0.11 ppm) as compared to the free ligand. The two methylenic
protons are magnetically equivalent, however each of them has
Matrix erosion was analyzed by measuring the amount of
silicic acid released in the medium using the molybdenum-blue
colorimetric test.26 Release and erosion tests were performed one
and thirty days after xerogel preparation to verify the stability
of the formulations upon storage (in closed vials and at room
temperature). Release and erosion data were mathematically
elaborated and were plotted as (Mt/Mtot)% vs time, where Mt
is the amount of total compound released at time t, and Mtot is the
total compound contained in the sample.
Results and discussion
Synthesis and characterization of the platinum complex
The synthesis of the ligand (2-amino-1-hydroxyethane-1,1-
diyl)bisphosphonic acid (AHBP-H4, 2) has been carried out by
a mixture of two different published procedures (Scheme 1).30,31
Initially a NH2-protected glycine (N-phthaloylglycine) was con-
verted into the corresponding acyl chloride and used as substrate in
a modified Arbuzov reaction in which tris(trimethylsilyl)phosphite
was used as phosphorylating agent. The reaction led to the silicic
ester of the geminal bisphosphonic acid, which was hydrolyzed by
treatment with methanol to give compound 1. After evaporation
of the volatile fractions, deprotection of the aminic group was
achieved by treatment with concentrated HCl. This treatment led
to formation of 2 and phthalic acid which was removed by washing
with hot absolute ethanol.
The characterization of 2 (present in the zwitterionic form)
has been performed via elemental analysis and 1H and 31P-NMR
spectroscopy in D2O at pH* = 3.5. The triplet at 3.45 ppm in the
1H-NMR spectrum was assigned to the methylene protons. The
splitting of the methylene signal in a triplet is caused by coupling
with two equivalent phosphorus atoms (3JH,P = 11.81 Hz). The
1
31P{ H} NMR (pH* = 3.5) showed a singlet at 16.46 ppm assigned
to the two P atoms.
The synthesis of [Pt2(en)2(AHBP-H)](HSO4) (3) was per-
formed starting from the barium salt of 2 (BaAHBP-H2) and
[Pt(en)(H2O)(OSO3)]. Initially, the reaction was carried on by mix-
ing the reactants in a 1 : 1 molar ratio (the aim was to prepare the
1 : 1 adduct), however, under these conditions, we obtained
a mixture of two products: a major product corresponding
to the dinuclear species described in this work and a minor
Fig. 1 1H (A) and 31P (B) NMR spectra of compound 3 in D2O.
Scheme 1
3134 | Dalton Trans., 2007, 3131–3139
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
different coupling constants with the two phosphorus atoms
(3JH–P = 10.07 and 10.30 Hz) resulting in a doublet of doublets.
Finally, the broad signal at 2.62 ppm was assigned to the methylene
protons of the Pt(en) residues.
chemical shift as a function of pH in the case of free AHBP is
given as Fig. 3. The titration curve shows changes in the slope
(inflexion points) corresponding to the various deprotonation
steps (Scheme 2). We did not determine the constant related
to the first deprotonation step (pKa1 in Scheme 2) because this
requires ionic hydrogen concentrations beyond the reach of our
experimental conditions.
The 31P-NMR spectrum (Fig. 1) of [Pt2(en)2(AHBP-H)]+ in D2O
at pH* = 3.5 shows a singlet with unresolved platinum satellites
(platinum satellites usually broaden with an increase in molecular
size and chemical shift anisotropy)35 falling at 37.60 ppm and
assigned to the two phosphorus atoms of the phosphonic groups,
which are magnetically equivalent and shifted at lower field (Dd =
21 ppm) with respect to the free AHBP ligand at the same pH*.
Such a lower field shift of the phosphorus nuclei is characteristic
of a phosphorus atom included in a 5-membered ring,6,36 however
in 3 the P atoms are included in 6-membered rings, therefore
the downfield shift likely depends upon electronic rather than
geometric effects.
The 195Pt-NMR spectrum (Fig. 2) shows two broad signals at
−1691 and −1707 ppm. These signals are at lower field with respect
to the starting [Pt(en)(H2O)2]2+ substrate (−1922 ppm). The same
trend has already been observed in platinum complexes with a
diamine and a dicarboxylate ligand with respect to the species
with a diamine and two aqua ligands.37–39
Fig. 3 Plot of 31P chemical shift vs pH* for 2.
Fig. 2 195Pt NMR of 3 in D2O.
The presence of only one signal in the 31P spectrum and of
two signals in the 195Pt spectrum strongly supports a structure
in which one bisphosphonate unit bridges two platinum atoms
(W conformation).33 The two phosphorus atoms are chemically
equivalent because of the presence of a plane of symmetry
passing through the two platinum atoms and perpendicular to
the coordination plane. In contrast, the absence of a plane of
symmetry passing through the P–C(1)–P atoms and perpendicular
to the coordination plane (presence of two different substituents,
Scheme 2
+
–OH and –CH2–NH3 , on the C(1) carbon atom), renders the
two platinum atoms inequivalent. The protonated amino group
is expected to exert a deshielding effect on the nearby Pt atom,
therefore the less shielded platinum signal (−1691 ppm) is assigned
to the platinum atom facing the protonated aminomethyl group.
Our data exclude the involvement of a deprotonated hydroxylic
group in coordination to platinum as observed in the complexes
of many other metals with tricoordinate bisphosphonates.40
As the pH* of the sample was increased from ca. 0 to 3
we observed a shift to higher field of the 31P chemical shift
with an inflexion point (calculated by fitting the points between
pH* 0 and 3 in Fig. 3 to the Henderson–Hasselbalch equation)
falling at pH* 1.12. This first constant corresponds to the
second deprotonation step (pKa2, Scheme 2) of the AHBP ligand.
Deprotonation of a phosphonate OH is known to shield the
phosphorus nucleus directly attached to it and to deshield the
distant second phosphonate group present in the molecule. The
former effect is called the “direct effect” while the latter effect
is called the “remote effect,” both effects were described very
accurately by Appleton in the pH titrations performed on the
BPMAA ligand (Chart 1).6 We propose that these effects apply
also to the present case and that for the second deprotonation
(pKa2) the direct effect prevails over the remote effect. The global
result is a shielding of the phosphorus nuclei.
Determination of pKa’s via 31P-NMR pH titrations
For a full characterization of the new compounds we performed
31P-NMR experiments at different pH on the free aminobisphos-
phonate ligand and on the corresponding platinum complex.
Acid dissociation constants have been previously reported
for alendronate (4-amino-1-hydroxybutane-1,1-diyl-bisphos-
phonic acid),33 which is a commercial drug having a structure
very similar to that of the aminobisphosphonate used in this
study. Alendronate exhibits five protonation constants spread
over the whole range of pH. The plot showing the change in 31P
By further increasing the pH* from 3 to 7 we observed a
deshielding of the phosphorus nuclei with an inflexion point falling
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 3131–3139 | 3135
©

View Article Online
at pH* 5.61. This deprotonation step corresponds to bisdeproto-
nation of only one phosphonate group (pKa3 in Scheme 2). The
deshielding of the 31P nuclei indicates that in this deprotonation
step the remote effect prevails over the direct effect.
The four samples were analyzed for platinum release the day
after xerogel preparation and one month later (Fig. 4). Moreover,
simultaneously with the platinum release, was also investigated
the matrix erosion (Fig. 5). The results thus provide information
regarding the effect of calcium and of Pt complex concentration
on (i) matrix loading capacity, (ii) rate of release, and (iii) stability
upon long term storage.
By further increasing the pH* from 7 to 14 we observed a
further deshielding of the phosphorus signal with two inflexion
points: one falling at pH* 9.23 (pKa4) and the other at pH*
13.14 (pKa5). The former deprotonation step (pKa4) most likely
corresponds to the bisdeprotonation of both phosphonate groups.
Once again the remote effect is higher than the direct effect, but
the difference is not large so that only a small global deshielding of
the phosphorus atoms is observed. The last inflexion point (pKa5)
is characterized by a huge deshielding of the phosphorus nuclei
and is typical for the deprotonation of the ammonium group in an
aminoalkylphosphonate species.41 The great deshielding is caused
by the loss of the zwitterionic form in which the positively charged
ammonium moiety is electrostatically attracted by the (partially or
fully) deprotonated phosphonate groups and in this conformation
causes a shielding of the phosphorus atoms.42 Therefore our
data are in full agreement with the data obtained for analogous
aminobisphosphonates.43
The pH titration was also performed on a sample of the platinum
complex 3. Differently from the free aminobisphosphonate ligand,
a constant 31P chemical shift was observed indicating that this is
not influenced by pH* (pH* range 0–8). This result supports the
involvement of all phosphonate hydroxyl oxygens in coordination
to platinum. The platinum-coordinated oxygen atoms cannot
undergo either protonation or deprotonation steps as the pH* is
changed and therefore there are no breaks in the curve of dP against
pH. Unfortunately, compound 3 undergoes structural changes at
pH* higher than 8.0 with the formation of new product(s) having
inequivalent phosphorus atoms. Most probably the deprotonated
aminic group coordinates to platinum displacing one oxygen atom
of a bis-coordinated phosphonate. The resulting product with
a bis-coordinated and a monocoordinated phosphonate group
would have inequivalent phosphorus atoms. No release of free
AHBP was observed, indicating that a rearrangement and not
a decomposition of the dinuclear platinum complex occurs at
pH* > 8.
Fig. 4 Kinetics of Pt complex release from (A) Ca(−) and (B) Ca(+)
formulations containing Pt : Si ratio of ca. 1.5 × 10−4 (1, full symbols)
and 3.0 × 10−4 (2, open symbols) as measured one day after xerogel
preparation (d1, circles) and 30 days after preparation (d30, squares).
A: (᭹) Ca(−)1-d1; (ꢀ) Ca(−)1-d30; (᭺) Ca(−)2-d1; (ꢁ) Ca(−)2-d30. B:
(᭹) Ca(+)1-d1; (ꢀ) Ca(+)1-d30; (᭺) Ca(+)2-d1; (ꢁ) Ca(+)2-d30.
Pt-complex release from silica xerogels, matrix erosion, and
stability of the formulations
The Pt–bisphosphonate complex was loaded onto two xerogel
formulations differing in the presence or absence of calcium in
their compositions. The rationale for introducing calcium into
the gel was twofold: (i) to improve its bioactivity and (ii) to
increase its affinity for bisphosphonic moieties. In particular, the
presence of calcium in silica polymers is known to improve their
bioactivity44,45,25 by promoting hydroxyapatite (HA) deposition on
their surface, which favors implant grafting to the surrounding
bone tissue. Therefore calcium containing xerogel formulations
known from the literature to be bioactive were selected for
this work.20,22 Moreover, since bisphosphonic moieties have high
affinity for calcium ions, the presence of calcium could influence
significantly also the loading and the release processes. Each
formulation (Ca(+) and Ca(−)) was loaded with two different
amounts of Pt complex so as to measure the loading capacity of
each formulation.
All release profiles were characterized by an initial burst
(occurring within the first 60 min) followed by a slow and long
lasting Pt-release process. The amount of complex released in
the initial burst varied from 20 to 60% of the loaded compound,
depending upon the formulation under investigation and the time
of storage before testing. It is likely that the portion of complex
that is released during the initial burst is located at the granule
surface and is released by a dissolution process. In contrast, the
portion of complex that is released in the slow phase is embedded
in the polymer matrix. The amount of complex embedded in
each formulation was thus calculated by subtracting the fraction
released in the initial burst from the total amount used in the
3136 | Dalton Trans., 2007, 3131–3139
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
capacity (ca. 0.6–0.8 mg complex g−1 xerogel) which does not
depend upon the initial drug load. Finally, the release behavior
of the Ca(+) samples and their erosion profile do not change
upon storage for 30 days, thus indicating that these samples are
stable with time and therefore suitable for practical application.
In contrast, the Ca(−) samples display a different behavior:
the loading capacity appears to be higher and to depend upon
the initial platinum concentration and the aging of the sample.
Moreover the release behaviour and the erosion kinetics also
change upon aging indicating that such formulations are rather
unstable.
In order to better understand the mechanism underlying the
release process, the slow Pt-release was compared with the erosion
process. In order to allow also the Ca(−) sample to reach stability,
the investigation was performed only on samples stored for 30 days
at room temperature. In both cases (Ca(+) and Ca(−) samples) the
platinum release and the matrix erosion data, plotted against the
time square root (data not shown), were linear, indicating that
both phenomena depend upon a diffusion process, in agreement
with previous observations.46 The times required for total matrix
erosion and Pt release (time of exhaustion, t-ex) were estimated
from the slopes of the linear plots and are reported in Table 3. The
data indicate that platinum release is faster than matrix erosion
for both Ca(+) and Ca(−) formulations and that both erosion and
diffusion were significantly faster in the presence of calcium.
The different Pt-complex loading capability demonstrated by
the Ca(+) and Ca(−) formulations probably reflects structural
differences between the two xerogels. Indeed, the calcium content
in the Ca(+) formulation was quite high (Ca : Si = 0.33 : 1 mol :
mol, Table 1) and could have a significant impact on the inner
structure of the polymer network resulting in a reduced polymer
“solubility” of the platinum compound. The latter structural
feature could be totally unrelated to the chemical affinity of
calcium ions for the bisphosphonic moiety of the platinum
complex, which, instead, can deeply affect the release behaviour.
Fig. 5 Silica erosion from (A) Ca(−) and (B) Ca(+) xerogels containing
Pt : Si ratio of ca. 1.5 × 10−4 (1, full symbols) and 3.0 × 10−4 (2, open
symbols) analyzed one day after their preparation (d1, circles) and 30 days
later (d30, squares). A: (᭹) Ca(−)1-d1; (ꢀ) Ca(−)1-d30; (᭺) Ca(−)2-d1;
(ꢁ) Ca(−)2-d30. B) (᭹) Ca(+)1-d1; (ꢀ) Ca(+)1-d30; (᭺) Ca(+)2-d1; (ꢁ)
Ca(+)2-d30.
Table 3 Estimated time required for total Pt-complex release (t-ex Pt-
release) and silica erosion (t-ex Silica erosion) performed on aged samples
(30 days storage)
xerogel preparation (Table 2). The concentration of complex in
each matrix (Pt mg g−1 xerogel) was thus calculated and was
defined as the ‘loading capacity’ of that formulation.
Formulation
t-ex silica erosion/days
t-ex Pt release/days
The results indicate that the presence of calcium reduces the
matrix loading capacity while increasing the amount of platinum
compound released in the initial burst. Moreover the calcium
containing formulations appear to have a maximum loading
Ca(−)1-d30
Ca(−)2-d30
Ca(+)1-d30
Ca(+)2-d30
222
270
36
57
54
6
35
3
Table 2 Percentages of complex 3 released in the initial burst or left embedded in the xerogels (loading capacity). The determinations were performed 1
(d1) and 30 (d30) days after preparation
Pt(II) complex released in
the burst/mg g−1 xerogel
Pt(II) complex left embedded in the
matrix after the burst/mg g−1 xerogel
Formulation
% Burst
Ca(−)1-d1
Ca(−)1-d30
Ca(−)2-d1
Ca(−)2-d30
Ca(+)1-d1
Ca(+)1-d30
Ca(+)2-d1
Ca(+)2-d30
31
18
41
21
36
30
60
60.5
0.59
0.35
1.22
0.63
0.45
0.38
0.94
0.95
1.31
1.55
1.76
2.35
0.79
0.86
0.63
0.62
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 3131–3139 | 3137
©

View Article Online
The nature of the released species was investigated by UV-Vis
spectroscopy. A highly significant difference was observed between
the Ca(−) and Ca(+) formulations. The Pt species released from
the Ca(−) gel always displayed a UV spectrum similar to that
of the initially entrapped dinuclear complex 3 (data not shown),
demonstrating that the species released in the initial burst and in
the subsequent slow process is the same as the one embedded. On
the contrary, the Pt species released from the Ca(+) xerogel had
UV spectra depending upon which phase of the release process is
analyzed. In particular, the compound released in the initial burst
had UV spectrum similar to that of the initially entrapped complex
3, whereas the compound released in the second phase had UV
spectrum with a maximum of absorbance at 300 nm, which is
similar to that of [PtCl2(en)] under similar conditions. These results
clearly indicate that calcium acts as an anchoring moiety for the
bisphosphonic ligand while the released Pt compound is either
[PtCl2(en)] (formed by reaction with chloride ions present in the
buffer) or related aqua species. The pharmacological relevance of
this finding is related to the different therapeutic potency of the
two Pt species involved ([Pt2(en)2(AHBP-H)]+ and [PtCl2(en)]) and
is currently under investigation.
hybrid material by changing the inorganic network composition,
either pure silica or Ca++ added silica. The presence of calcium
in the matrix was found to reduce its loading capacity but to
improve its stability upon storage, a property that is fundamental
for practical applications. In addition, the presence of calcium
affected the nature of the Pt complex released in the slow diffusion-
controlled process following the initial burst. When present, Ca++
was able to retain the bisphosphonic ligand so that only the Pt(en)
residue was released from the xerogel. In contrast, in the absence
of calcium, the Pt complex was released in its original dinuclear
form with bridging bisphosphonate. Hopefully, the Pt complex
concentration will be sufficient to exert therapeutic activity only
at the site of the implant while it will be too low to exert undesired
toxic effects on the neighbouring tissues and at a systemic level. We
are currently performing cytotoxicity tests on both the dinuclear
platinum complex and the platinum-embedded silica xerogels to
investigate how the different nature of the released platinum
species affects the biological activity.
Acknowledgements
It is worth noting that the release experiments were carried
out under sink conditions always allowing fresh buffer to be in
contact with the platinum-loaded xerogel. These conditions are
likely to mimic the in vivo situation where constant renovation of
body fluids occurs around the implant. It is important to point
out that the concentration of Pt in the medium never reached
the therapeutic value (even assuming that the released compound
had similar potency as cisplatin). This means that the therapeutic
activity of the matrices here investigated can be exerted only locally
at their surface, while the surrounding tissue will not be affected.
We also want to point out that the trend observed in the
present study (calcium acting as an anchoring moiety for a
bisphosphonic ligand coordinated to a platinum atom) is very
similar to what we have found in a previous study conducted on
a Pt-complex with medronate ([Pt2(en)2(MDP)]; en = ethylenedi-
amine, MDP = methylenebis(phosphonate)) embedded into two
biomimetic synthetic hydroxyapatite nanocrystalline materials.19
We found that the release of platinum from the hydroxyapatite
composite materials occurred through complete breakage of the
platinum–medronate bonds, the process depending upon the
surface stoichiometry (Ca : P ratio).
The authors thank the Universities of Bari, Bologna, and Padova,
the Italian “Ministero dell’Istruzione, Universita` e Ricerca
(MIUR)” (PRIN 2004 no. 2004059078_006), and the EC (COST
Chemistry project D39/0004/06) for support.
References
1 E. R. Jamieson and S. J. Lippard, Chem. Rev., 1999, 99, 2467–2498.
2 D. D. Von Hoff, R. Schilsky, C. M. Reichert, R. L. Reddick, M.
Rozencweig, R. C. Young and F. M. Muggia, Cancer Treat. Rep., 1979,
63, 152–1531.
3 G. Daugaard and U. Abildgaard, Cancer Chemother. Pharmacol., 1989,
25, 1–9.
4 E. Wong and C. M. Giandomenico, Chem. Rev., 1999, 99, 2451–2466.
5 T. G. Appleton, J. R. Hall and I. J. McMahon, Inorg. Chem., 1986, 25,
720–725.
6 T. G. Appleton, J. R. Hall and I. McMahon, Inorg. Chem., 1986, 25,
726–734.
7 T. Klenner, P. Valenzuela-Paz, B. K. Keppler, G. Angres, H. R. Scherf,
F. Wingen, F. Amelung and D. Schma¨hl, Cancer Treatment Rev., 1990,
17, 253–259.
8 B. K. Keppler, M. R. Berger, T. Klenner and M. E. Heim, Adv. Drug.
Res., 1990, 19, 243–310.
9 J. M. Sanders, Y. Song, J. M. W. Chan, Y. Zhang, S. Jennings, T.
Kostzowski, S. Odeh, R. Flessner, C. Schwerdtfeger, E. Kotsikorou,
G. A. Meints, A. O. Gomez, D. Gonza´lez-Pacanowska, A. M. Raker,
H. Wang, E. R. van Beek, S. E. Papapoulos, C. T. Morita and E.
Oldfield, J. Med. Chem., 2005, 48, 2957–2963.
Conclusions
10 E. Kotsikorou and E. Oldfield, J. Med. Chem., 2003, 46, 2932–2944.
11 L. Widler, K. A. Jaeggi, M. Glatt, K. Mu¨ller, R. Bachmann, M. Bisping,
A.-R. Born, R. Cortesi, G. Guiglia, H. Jeker, R. Klein, U. Ramseier, J.
Schmid, G. Schreiber, Y. Seltenmeyer and J. R. Green, J. Med. Chem.,
2002, 45, 3721–3738.
In this work we have synthesized and fully characterized by
multinuclear NMR spectroscopy a geminal aminobisphosphonic
acid (2) and the corresponding platinum complex (3). The complex
was found to have a dinuclear structure with the bisphosphonate
12 J. H. Lin, Bone, 1996, 18, 75–85.
bridging two platinum moieties in a W conformation. 1H and 31
P
13 T. Klenner, P. Valenzuela-Paz, B. K. Keppler and H. R. Scherf, J. Cancer
Res. Clin. Oncol., 1990, 116, 453–458.
14 T. Klenner, F. Wingen, B. K. Keppler, B. Krempien and D. Schma¨hl,
J. Cancer Res. Clin. Oncol., 1990, 116, 341–350.
15 M. Galanski, S. Slaby, M. A. Jakupec and B. K. Keppler, J. Med.
Chem., 2003, 46, 4946–4951.
16 M. J. Boemink, J. P. Dorenbos, R. J. Heeterbrij, B. K. Keppler, J. Reedijk
and H. Zahm, Inorg. Chem., 1994, 33, 1127–1132.
17 M. Laforgia, L. Cerasino, N. Margiotta, C. Cardellicchio, M. A. M.
Capozzi, F. Naso and G. Natile, Eur. J. Inorg. Chem., 2004, 3445–3452.
18 M. Laforgia, N. Margiotta, F. Capitelli, V. Bertolasi and G. Natile,
Eur. J. Inorg. Chem., 2005, 1710–1715.
NMR experiments at different pH values have allowed the deter-
mination of the acidity constants for the free aminobisphosphonic
acid and an estimation of the stability of the Pt-complex. Complex
3 is stable at pH values not exceeding 8.0, which is compatible with
carrying on pharmacological investigations.
The Pt(II) bisphosphonate complex was thus embedded into two
different silica-based polymers to obtain biocompatible hybrid
materials to be used for the local treatment of bone tumors. Our
investigation is aimed at controlling the release properties of the
3138 | Dalton Trans., 2007, 3131–3139
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
19 B. Palazzo, M. Iafisco, M. Laforgia, N. Margiotta, G. Natile, C. L.
Bianchi , D. Walsh, S. Mann, N. Roveri, Adv. Funct. Mater., 2007, in
press.
20 G. Palumbo, L. Avigliano, G. Strukul, F. Pinna, D. Del Principe, I.
D’Angelo, M. Annicchiarico-Petruzzelli, B. Locardi and N. Rosato,
J. Mater. Sci., 1997, 8, 417–21.
21 P. Kortesuo, M. Ahola, S. Karlsson, I. Kangasniemi, A. Yli-Urpo and
J. Kiesvaara, Biomaterials, 2000, 21, 193–198.
22 R. F. S. Lenza, W. L. Vasconcelos, J. R. Jones and L. L. Hench, J. Mater.
Sci., 2002, 13, 837–42.
33 K. Libson, E. Deutsch and B. L. Barnett, J. Am. Chem. Soc., 1980, 102,
2476–2478.
34 S. J. Berners-Price and P. J. Sadler, Coord. Chem. Rev., 1996, 151, 1–40.
35 S. J. Berners-Price, L. Ronconi and P. J. Sadler, Prog. Nucl. Magn.
Reson. Spectrosc., 2006, 49, 65–98.
36 D. G. Gorenstein, Prog. Nucl. Magn. Reson. Spectrosc., 1984, 16, 1–98.
37 Y.-A. Lee, K. C. Young and S. S. Youn, J. Inorg. Biochem., 1997, 68,
289–294.
38 F. D. Rochon and L. M. Gruia, Inorg. Chim. Acta, 2000, 306, 193–
204.
23 K. Czarnobaj and J. Łukasiak, Drug Delivery, 2004, 11, 341–344.
24 K. Czarnobaj, M. Prokopowicz and J. Łukasiak, Drug Delivery, 2006,
13, 339–344.
39 F. D. Rochon and V. Buculei, Inorg. Chim. Acta, 2005, 358, 2040–
2056.
40 E. R. van Beek, C. W. G. M. Lo¨wik, F. H. Ebetino and S. E. Papapoulos,
Bone, 1998, 23, 437–442.
25 L. L. Hench, J. Mater. Sci., 2006, 17, 967–978.
26 M. Morpurgo, D. Teoli, B. Palazzo, E. Bergamin, N. Realdon and M.
Guglielmi, Farmaco, 2005, 60, 675–683.
27 R. D. Feltham and R. G. Hayter, J. Chem. Soc., 1964, 4587–4591.
28 A. Pasini, C. Caldirola, S. Spinelli and M. Valsecchi, Synth. React.
Inorg. Met.-Org. Chem., 1993, 23, 1021–1060.
29 KaleidaGraph 3.5, Synergy Software, Reading, PA, USA, 2000.
30 M. Lecouvey, I. Mallard, T. Bailly, R. Burgada and Y. Leroux,
Tetrahedron Lett., 2001, 42, 8475–8478.
41 E. Matczak-Jon, B. Kurzak, P. Kafarski and A. Woz´na, J. Inorg.
Biochem., 2006, 100, 1155–1166.
42 T. G. Appleton, J. R. Hall, A. D. Harris, H. A. Kimlin and I. J.
McMahon, Aust. J. Chem., 1984, 37, 1833–1840.
43 V. Kub´ıcˇek, J. Kotek, P. Hermann and I. Lukesˇ, Eur. J. Inorg. Chem.,
2007, 333–344 references therein.
44 J. Wilson, G. H. Pigott, F. J. Schoen and L. L. Hench, J. Biomed. Mater.
Res., 1981, 15, 805–17.
31 D. V. Griffiths and J. M. Hughes, Tetrahedron, 1997, 53, 17815–17822.
32 R. Bau, S. K. S. Huang, J.-A. Feng and C. E. McKenna, J. Am. Chem.
Soc., 1988, 110, 7546–7547.
45 L. L. Hench and J. Wilson, MRS Bull., 1991, 16, 62–74.
46 D. Teoli, L. Parisi, N. Realdon, M. Guglielmi, A. Rosato and M.
Morpurgo, J. Controlled Release, 2006, 116, 295–303.
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 3131–3139 | 3139
©