Journal of Inorganic Biochemistry 83 (2001) 5765

www.elsevier.nl / locate / jinorgbio

Metal-ion speciation in blood plasma incorporating the tetraphosphonate,

N,N-dimethylenephosphonate-1-hydroxy-4-aminopropilydenediphosphonate
(APDDMP), in therapeutic radiopharmaceuticals
a,
a
a
b
Jan Rijn Zeevaart *, Neil V. Jarvis , Werner K.A. Louw , Graham E. Jackson
a

Radiochemistry, South African Nuclear Energy Corporation ( Ltd.), P.O. Box 582, Pretoria, 0001, South Africa
b
Department of Chemistry, University of Cape Town, Private Bag, Rondebosch 7700, South Africa
Received 26 April 2000; received in revised form 27 June 2000; accepted 3 July 2000

Abstract
In a quest for more effective radiopharmaceuticals for pain palliation of metastatic bone cancer, this paper relates results obtained with
Ho and 153 Sm complexed to the bone seeking phosphonate, N,N-dimethylenephosphonate-1-hydroxy-4-aminopropylidenediphosphonate (APDDMP). APDDMP is synthesised from the known bone cancer pain palliation agent 1-hydroxy-3-aminopropylidenediphosphonate (APD) and was complexed to lanthanide trivalent metal ions. This work is performed to utilise the idea that the energetic b-particle
emitter, 166 Ho, coupled with phosphonate ligands such as APD and APDDMP could afford a highly effective radiopharmaceutical in the
treatment of bone cancer. Complex-formation constants of APDDMP with the important blood plasma metal-ions, Ca 21 , Mg 21 , and Zn 21
and the trivalent lanthanides Ho 31 and Sm 31 were measured by glass electrode potentiometry at 378C and I5150 mM. Blood plasma
models were constructed using the computer code ECCLES and the results compared with those gathered from animal tests. The
166
Ho-APDDMP complex was found to have little liver or bone uptake while 153 Sm-APDDMP had a moderate bone uptake. This was
primarily due to the high affinity of APDDMP for Ca(II). Clinical observations could be explained by the blood plasma modelling.
2001 Elsevier Science B.V. All rights reserved.
166

Keywords: Bisphosphonates; Blood plasma; Potentiometry; Speciation; Therapeutic radiopharmaceuticals;

1. Introduction
In previous studies, 153 Sm-EDTMP (QuadrametE) has
been shown to be effective in the palliation of pain
associated with metastatic bone cancer [1,2]. The high
affinity of EDTMP for bone, coupled with the nuclear
properties of 153 Sm (t 1 / 2 546.75 h; maximum b-particle
energy51.3310 213 J) enabled the radiopharmaceutical to
alleviate the pain associated with metastatic bone cancer.
In order to improve the efficacy of the radiopharmaceutical, we have replaced the 153 Sm by 166 Ho which has a
higher energy b-particle (t 1 / 2 526.9 h; maximum b-particle energy53.0310 213 J). Despite the chemical similarity between these two metal ions, however, these studies
were unsuccessful, the Ho-EDTMP complex having a
different biodistribution to Sm-EDTMP. Computer modelling studies showed that this was due to subtle changes in
*Corresponding author. Fax: 127-12-305-5944.
E-mail address: radchem@aec.co.za (J.R. Zeevaart).

166

Ho;

153

Sm

the metal EDTMP formation constants, which changed the

in vivo speciation [3]. In an attempt to modify this
biodistribution, a number of bisphosphonates, methylenediphosphonic
acid
(MDP),
1-hydroxy-ethylenediphosphonic acid (HEDP) [4] and APD [5] were
investigated. Blood plasma speciation studies using the
first two bisphosphonates explained low bone uptake
results for their radioactive Sm(III) and Ho(III) complexes
as found in the literature [6]. For 166 Ho-APD, a high liver
uptake was seen in baboon models. This could be explained using blood plasma speciation modelling, the
results of which, showed that, under physiological conditions, a neutral MLH species was formed. While these
studies were encouraging, none of the complexes were an
improvement on 153 Sm-EDTMP [6].
In order to improve the results obtained with APD, a
new ligand was designed which would form charged
Ho(III) complexes in blood plasma. This could be done by
increasing the stability of the anionic, ML, species so that
it formed at lower pH. Alternatively, negatively charged

0162-0134 / 01 / $ see front matter 2001 Elsevier Science B.V. All rights reserved.
PII: S0162-0134( 00 )00125-2

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

58

Fig. 1. Ligands discussed in this paper.

groups, could be added to the ligand giving the MLH

species a formal charge. In this paper, the latter option was
taken. APD was used as the molecular frame-work and the
bone localising phosphonate groups left intact. Instead, two
more phosphonate groups were added to the amino tail of
APD. Phosphonates were chosen as they are known to
have a high bone uptake and they are strong coordinators
of hard lanthanide ions. In this way it was hoped that the
stability of the metal complex would be increased and the
resulting complex would have an improved affinity for
bone. A schematic diagram of the final ligand synthesised
as well as other ligands mentioned are shown in Fig. 1.
This paper reports on results obtained with the ligand,
N,N-dimethylenephosphonate-1-hydroxy-4-aminopropilydenediphosphonic acid (APDDMP) which was synthesised from APD by a Mannich type reaction. Screening
of the potential of this ligand to form radiopharmaceuticals
was achieved using the ECCLES computer blood plasma
model after formation constants for APDDMP were measured potentiometrically. Animal tests were performed
using 166 Ho-APDDMP, 153 Sm-APDDMP and 99m TcAPDDMP and are reported here.

2. Experimental

2.1. Reagents
The ligand APDDMP was synthesised from APD by a
Mannich type reaction based on the method reported by
Shcherbakov et al. [7]. 4.7 g (0.02 mol) APD and 4 g
(0.048 mol) H 3 PO 3 were dissolved in 10 ml H 2 O and 10
ml HCl(c). The APD only dissolves after heating the
mixture in a three-necked flask. The solution was brought
to boiling point under reflux conditions. Over a period of
30 min, 7 ml of a 37% CH 2 O (formaldehyde) solution was
added dropwise. The solution was refluxed for a further 2.5
h after which it was allowed to cool. Excess HCl and
formaldehyde were removed by evaporation at 1008C and

reduced pressure. The product was re-dissolved in a small

amount of water and filtered. Upon treating with ethanol an
oil was released. The oil was separated and re-dissolved in
water. This process was repeated three times until a
transparent oil was obtained. Finally, the oil was dried at
1008C over P2 O 5 . Yield, 5 g (60%) of hygroscopic
APDDMP which was stored under dry N 2 . (Found: C,
12.71; H, 4.87; N, 2.95%. Calc. for C 5 H 17 NP4 O 13 .2.8H 2 O:
C, 12.68; H, 4.87; N, 2.96%) The purity of the ligand was
checked by potentiometric titrations and found to be
greater than 99% and was therefore used without further
purification.
Fresh metal-ion solutions were employed in the titrations. These were made by dissolving reagent grade
chloride salts of the metal-ions in distilled water and
standardised by ICP. Where necessary, solutions were
acidified to prevent hydrolysis.

2.2. Potentiometry
Potentiometric titrations were performed using a Metrohm Titroprocessor 670 with a Metrohm 665 dosimat and
a combination glass electrode (Ag /AgCl reference). The
electrode was calibrated regularly using strong acidbase
titration data [8]. To correct the deviation of the electrode
at high pH, the Nernst equation was adjusted to the
following:
E 5 E 0 1 60.4 3 log[H 1 ] 1 Ej
where Ej (51.354310 211 / [H 1 ]) is the liquid junction
potential.
All titrations were performed under inert atmosphere
and solutions were held at a constant ionic strength of 0.15
M NaCl and a temperature of 37.060.18C. The titrations
were performed beginning at low and ending at high pH,
adding 0.10 ml aliquots of 0.050 M NaOH (carbonate-free)
in 0.10 M NaCl. Protonation constants were calculated
from data obtained from titrations of the ligand alone. The
concentration of the ligand was refined together with the
protonation constants in an iterative process. MetalAPDDMP formation constants were calculated from titration data at three different metal:ligand ratios varying from
1:1 to 1:3. Data were analysed using the ESTA [8] library
of programs. During the analysis, the previously determined protonation constants, were held constant. Metal
hydrolysis constants and pK w were taken from the literature [9] and held constant during optimisation procedures.
The models were tested for plausibility by comparing
experimental and calculated formation and deprotonation
curves.

2.3. Blood plasma modelling

All constants measured in this study were added to the
ECCLES database [10]. The APDDMP concentration used
in modelling was 8.5310 25 M. This value is representa-

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

59

tive of actual clinical amounts, and is similar to that

previously employed during EDTMP work [3].

activity on the solvent front indicating 1% colloidal or

unbound 99m Tc.

2.4. Preparation of 166 Ho-APDDMP and

APDDMP for animal tests

2.6. Animal tests

153

Sm-

12.5 mg Ho 2 O 3 was irradiated for 24 h in the

SAFARI-1 reactor at a neutron flux of 4310 13 neutron
cm 22 s 21 and allowed to cool for 24 h. Using the
necessary shielding, the irradiated oxide was dissolved in
100 ml of 2 M HCl solution at 608C over a period of 1 h 2
ml water was added to form a stock solution of HoCl 3 . An
aliquot (calculated accordingly to the activity needed on
the day of the animal experiment) of this solution was
transferred to a 10 ml vial containing 700 ml of 1 M NaOH
in which 50 mg of APDDMP was dissolved. A 1 M HCl
solution was then used to reduce the pH of the solution to
7.8. The resulting solution was used as a stock solution of
the complex. A similar procedure was used for Sm 2 O 3 .
Complexation of the Ho and Sm was checked by using a
small carboxymethyl sephadex column [11]. The column
was prepared by placing 200 mg CMC-Sephadex in a
column, water (20 ml) added, and the resin allowed to
swell overnight. The suspension was stirred to obtain
homogeneity and a 5 ml sample taken, placed on a column,
and sucked dry. Twenty microlitres of the stock solution
was loaded onto the column and the activity (background
corrected) measured using an ionisation chamber. The
column was then washed with 20 ml physiological saline
solution to elute the complex and the activity measured
again. The difference in the two measurements gave the
percentage formation of the complex. For each complex,
the yield was .96%.

2.5. Preparation of

99 m

Tc-APDDMP for animal tests

As is the procedure for 99m Tc-radiopharmaceuticals,

lyophilised, labelling kits of the ligand were prepared in
advance and stored in a freezer until the day of use. The
kits were prepared by mixing 5 ml of a solution containing
250 mg of SnCl 2 .2H 2 O crystals in 0.5 ml HCl(c) with an
aqueous APDDMP solution (25 mg / ml) where after the
pH was adjusted to 6. This mixture was dispensed into 5
vials and freeze dried. The evacuated vials were stored
below 08C until used. One vial was used for quality
control.
The final solutions for injection were prepared by adding
0.3 ml 99m Tc (1.4 GBq), obtained from a 99 Mo / 99m Tc
generator, to the above vials. Radiochemical purity was
checked chromatographically. Five microlitres of the final
solution was spotted on ITLC-SG and ITLC-SA chromatograms and developed in acetone and physiological saline
respectively. The first showed that 99.9% of the activity
remained at the origin indicating less than 0.1% free
pertechnetate was present. The latter showed 99% of the

Animal experimentation was done according to the

National Code for the Handling and Use of Animals in
Research, Education, Diagnosis and Testing of Drugs and
Related Substances in South Africa, and the protocols
approved by the authorised Ethics Committee of the
University of Pretoria. Because of the preliminary nature
of this study, only a single baboon at the Pretoria Biomedical Research Centre was injected with each isotope.
For 153 Sm and 99m Tc, a dose of 5 and 6 MBq kg 21
respectively was injected. However for 166 Ho, a dose of 3
MBq kg 21 was used to ensure that the animals did not
receive too high a dose from this nuclide. In the supine
position the radiation-sensitive oesophagus is in close
proximity to the dorsal artery and hence is more affected
by a strong b-emitting nuclide. Blood and urine samples
were taken periodically for 3 h. A Siemens Orbiter gcamera was set at 80.5 keV, 20% window to record the
strongest g-line of 166 Ho. For 153 Sm and 99m Tc these
values were 103 keV, 20% and 140.5 keV, 20% respectively. The animal was positioned in the supine position for
planar imaging. Images were recorded continuously for the
first 30 min (dynamic study) and at one, two, three and
four h after administration, static g-scans were taken of the
head, thorax and pelvis. Computer enhanced images, as
well as analyses of blood and urine results obtained by
counting on a multichannel analyser, were used to estimate
the biodistribution of the particular nuclide.

3. Results and discussion

3.1. Potentiometry
APDDMP has nine potential sites for protonation, one
tertiary amine and four phosphonato groups. The pH range
used in this study precluded the determination of the amine
protonation constant. However, Shcherbakov et al. [7]
have determined this to be 10 11.87 at 258C and I5
1.0 M KNO 3 . The basicity of this group is very high but
not unusual when one considers that the corresponding pKa
of APD is 10.95 [5]. Although a phosphonate is electron
withdrawing, deprotonation of the amine group of
APDDMP relative to APD is more difficult because in the
former a proton is removed from LH 72 while in the latter
it is removed from LH 32 . The charge on the ligand is
likely to affect both the entropy and enthalpy of deprotonation. The same reasoning explains the increase in pKa in
the sequence: NH 3 (pKa 59.24); H 2 NCH 2 PO 22
(pKa 5
3
42
10.05); [HN(CH 2 PO 3 ) 2 ]
(pKa 510.79). Increasing the
alkyl chain separating the amine and phosphonate also

60

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

increases the basicity of the nitrogen [H 2 NCH 2 PO 22

3 ,
22
10.05; H 2 NCH 2 CH 2 PO 22
,
11.05;
H
NCH
CH
CH
PO
3
2
2
2
2
3 ,
11.07]-electronwithdrawing group moves further away
from the amine.
The results of an ESTA [8] analysis of the protonation
titrations in the pH range 210 are given in Table 1. The
low Hamilton R-factor and standard deviation in the
equilibrium constants lends confidence to the results. A
graphical display of the data is given in Fig. 2 in which the
amine proton is assumed not to dissociate. At high pH, the
curve levels off at 0, indicating that no further proton is
lost. Below pH54, the curves level off at 4 indicating that
4 protons have been added to APDDMP. It is these four
protonation constants that have been determined in this
study. H 5 APDDMP (H 6 L) is only 30% formed at pH52.5

(Fig. 3) and this is reflected in the higher standard

deviation in the equilibrium constant for this species. The
final three protonation constants are all ,2.0 and could not
be determined in this study. The agreement between the
calculated function (represented by a solid line) and the
experimental data is good and lends confidence to the
model.
The protonation constants reported here for APDDMP
are approximately 0.5 log units higher than those reported
in the literature [7]. This may be due to the different
experimental conditions of the two measurements but may
also be experimental error. The constants determined in
this study are in agreement with the corresponding constants for APD. A comparative value for the diphosphonate
centre is the second protonation constant for APD which is

Table 1
Protonation and formation constants for APDDMP determined in this study at 378C and I5150 mM NaCl. All constants calculated using potentiometry
Equilibrium a

log K
[this work]

H1L;HL
H1HL;H 2 L
H 1H 2 L;H 3 L
H 1H 3 L;H 4 L
H 1H 4 L;H 5 L
H 1H 5 L;H 6 L
Ca1L1H;Ca(HL)
2Ca1L;Ca 2 L
Ca 2 L1H;Ca 2 (HL)
Ca 2 (HL)1H;Ca 2 (H 2 L)
Ca1L1OH;CaL(OH)
Ca1L12OH;CaL(OH) 2
2Mg1L;Mg 2 L
Mg 2 L1H;Mg 2 (HL)
Mg 2 (HL)1H;Mg 2 (H 2 L)
Mg 2 L1OH;Mg 2 L(OH)
Mg1L12OH;MgL(OH) 2
2Sr1L;Sr 2 L
Sr 2 L1H;Sr 2 (HL)
Sr 2 (HL)1H;Sr 2 (H 2 L)
Sr 2 L1OH;Sr 2 L(OH)
Zn1L;ZnL
ZnL1H;Zn(HL)
Zn(HL)1H;Zn(H 2 L)
Zn(H 2 L)1H;Zn(H 3 L)
Zn1L1OH;ZnL(OH)
Zn1L12OH;ZnL(OH) 2
Ho1L;HoL
HoL1H;Ho(HL)
2Ho1L12H;Ho 2 (H 2 L)
Ho1L1OH;HoL(OH)
Ho1L12OH;HoL(OH) 2
Sm1L;SmL
SmL1H;Sm(HL)
2Sm1L12H;Sm 2 (H 2 L)
Sm1L1OH;SmL(OH)
Sm1L12OH;SmL(OH) 2

Not Found c
9.82(,1)
7.21(,1)
6.09(1)
4.89(1)
2.04(2)
13.87(3)
10.99(3)
7.30(4)
5.91(4)
8.60(3)
11.61(2)
11.10(3)
7.34(4)
5.72(4)
4.47(4)
12.19(2)
10.17(2)
7.32(3)
5.71(4)
3.25(4)
9.53(2)
7.19(3)
6.03(3)
4.26(4)
13.46(3)
15.49(7)
12.01
7.81(3)
32.32(4)
19.14(4)
22.72(5)
13.10(3)
6.80(3)
31.50(4)
18.93(4)
22.96(4)

Number of
data points

Hamilton
R factor

643

0.00509

298

0.0128

290

0.0122

600

0.0104

267

0.0134

364

0.0262

342

0.0232

log K b
Shcherbakov [7]
11.87
10.30
6.82
5.70
4.57
,2
12.83

a
Charges on metal-ions, APDDMP and complexes have been omitted for simplicity. L represents H-APDDMP except for the protonation constants
where L represents APDDMP for comparison.
b
I51 M KNO 3 and 258C. The metal to ligand ratio studied is 1:1.
c
The value is probably to high to be determined by the experimental setup used in this work.

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

Fig. 2. Experimental (points) and modelled (lines) protonation formation

curves for APDDMP. The three separate titrations are represented by (s)
0.00102 M APDDMP and 0.00662 M HCl; (h) 0.00118 M APDDMP
and 0.00861 M HCl and (n) 0.00148 M APDDMP and 0.01099 M HCl
vs. 0.0500 M NaOH in 0.010 M NaCl. All solutions were at 378C and
0.15 M NaCl or 0.15 M total ionic strength.

also 9.80 [5]. pKa5 (fourth protonation constant of

APDDMP) compares favourably with the third protonation
constant for APD (6.01) which is also at a diphosphonate
centre. The third protonation constant for APDDMP
(which is in between the two above) would thus be
expected to represent protonation at a methylene-phosphonate centre. This constant compares well with the first
protonation constant found for the methylene-phosphonate
centre of EDTMP [3] viz. 7.63. However, for other
methylenephosphonates, lower values have been reported.
For example, the protonation constant for N-isopentyliminodi(methylenephosphonic) acid [12] is 6.26.
This discrepancy can once again be explained by the
decreased difficulty in removing a proton from the latter
5(LH 22
2 ) than from APDDMP and EDTMP (LH 3 ). Fig. 4
gives a graphical representation of the proposed protonation scheme.

Fig. 3. Species distribution curves for the protonation of APDDMP at

378C and 0.15 M as calculated from the protonation constants in Table 1.
The initial APDDMP concentration was 0.00148 M.

61

For metalligand systems, a typical example of the

approach used in model validation is shown for the Ca(II)APDDMP system in Fig. 5. In this diagram, the calculated
and experimental formation curves, (the average number of
metal ions bound per ligand) follow each other closely. At
low concentrations of free ligand (high pH) the formation
curves suddenly rise and turn in the opposite direction.
This phenomenon is referred to as backfanning [13] and
indicates that hydrolysis becomes important. From the
deprotonation curves,, (the average number of protons per
metal ion released upon complexation) for a similar
system, viz. Mg(II)-APDDMP (Fig. 6), hydroxy-species
can also be deduced. At high pH these curves rise above
the dashed curve (the protonation state of the ligand in the
absence of the metal ion), which is interpreted as more
protons being released from the complex than are available
to dissociate from the ligand-implying hydroxy species
formation. Using this information the potentiometric data
were analysed and the results presented in Table 1.
From the results it can be noted that for Ca(II), Mg(II)
and Sr(II) binuclear species are reported. The formation of
binuclear species has been found for other diphosphonate
ligands such as MDP, HEDP and APD [4,5] and is a well
established phenomenon. Experimental evidence for these
species can be found in Fig. 6. In the pH region 5 to 9, a
value of 0.5 is recorded. If the particular pH region of 89
is considered this means that half a proton is released per
metal ion on complexation. In the absence of the metal, the
predominant ligand species would be HL. Combining this
information suggests that one proton is released from HL
per two metal ions resulting in an M 2 L species. The same
is true for the other binuclear species at lower pH. As was
the case for the complexation of APD and Ca(II) [5], a
mixed system is reported for Ca(II)-APDDMP where an
MLH species is included together with the binuclear
species. However from species distribution diagrams and
ECCLES modelling (Table 2) it is clear that, at the ligand
to metal ion ratios employed in this study, MLH is a minor
species and is included only to improve the fit. As the
values reported in the literature [7] do not include binuclear species, the only comparison with this work is the
MLH species for Ca(II) which differs somewhat from the
value reported here.
The constants found for the M 2 L species of Ca(II),
Mg(II) and Sr(II) show the same trend as that found for
APD. The values for the complexation of APD with these
metal ions are 10.79, 10.85 and 9.80 respectively. The
higher M 2 L constant for Ca(II)-APDDMP than for APD
would indicate a higher affinity of APDDMP for Ca(II).
For the equilibrium M 2 L1OHM 2 L(OH), the log K
value is 4.47. This is significantly higher than the first
hydrolysis constant of MgOH. (2.58) and implies that on
formation of Mg 2 L(OH) the proton is not released from
coordinated water but rather, the amine (pKa 511.87 [7]) is
de-protonated. Thus it would be more correct to represent
the complex as M 2 LH 21 rather than M 2 L(OH). If this

62

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

Fig. 4. Proposed protonation scheme for APDDMP.

argument is correct, the drop in pKa of the amine from

11.87 to 4.47 must imply coordination of the metal-ion to
the amine.

For Zn(II) no evidence of binuclear species (Q51

for
pH559) was found which confirms earlier findings from
polarography that for Zn(II)-diphosphonate ligand systems,
no binuclear species are present [4,5]. The log K values for
the formation of ML(OH) and ML(OH) 2 are calculated as
3.93 and 5.96 respectively. Both these values are respectively lower than the first and second hydrolysis constants
for Zn(II) viz. 5.0 and 10.2. This indicates that both
protons are released from coordinated water molecules.
For Sm(III) only one hydroxy species, SmL(OH), is
formed. Log K for this complex is 5.83 which is similar to

the first hydrolysis constant for Sm(III) viz. 5.39. For

Ho(III) and Sm(III) precipitation occurred up to pH54.
This pH region could thus not be interpreted. For Ho(III)
the value for the ML constant has a high standard
deviation of 0.18 and was not included in the final model.
The value is nonetheless reported.

3.2. In vivo speciation calculated by ECCLES

The speciation of Ho(III) and Sm(III) in the presence of
APDDMP in blood plasma is presented in Fig. 7. Transferrin was not included as kinetics of complexation for
lanthanides is slow and would probably not occur before
bone localisation. From the results it is clear that the
lanthanide-APDDMP complexes that form in blood plasma
viz. ML, ML(OH) and M(HL) are charged and not neutral
species that might cause liver uptake. The major species
for the Ho(III)-APDDMP systems proves to be the
ML(OH) species, in contrast to the neutral M(HL) species
of APD [5]. This was because of the stronger complexation
of APDDMP to Ho(III) (which was one of the design
features set out to be achieved). However, Ho(III) is still
complexed to the extent of 61.3% by citrate in blood
plasma. For Sm(III) the picture is totally different with

Table 2
Speciation of APDDMP in normal blood plasma as calculated by
ECCLES. L5APDDMP 27
Species
Fig. 5. Experimental (points) and modelled (lines) formation curves for
Ca(II) complexation by APDDMP. pA is the negative logarithm of the
free ligand concentration. The three separate titrations are represented by
(s) 0.00105 M Ca(II), 0.00122 M APDDMP and 0.00583 M HCl; (h)
0.00100 M Ca(II), 0.00162 M APDDMP and 0.00562 M HCl and (n)
0.00096 M Ca(II), 0.00310 M APDDMP and 0.01063 M HCl vs. 0.0500
M NaOH in 0.010 M NaCl. All solutions were at 378C and 0.15 M NaCl
or 0.15 M total ionic strength.

Mol%
32

[Ca 2 (L)]
[Ca 2 (HL)] 22
[Mg 2 (L)] 32
[Mg 2 (HL)] 22
[Ca(HL)] 42
[Ca 2 (H 2 L)] 2
[Mg 2 (LOH)] 42

42.4
33.0
11.6
10.1
1.1
1.1
0.3

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

Fig. 6. Experimental (points) and modelled (lines) deprotonation curves

for Mg(II) complexation by APDDMP. The dashed line is the curve and
represents the protonation state of the ligand in the absence of the
metal-ion. The three separate titrations are represented by (s) 0.00113 M
Mg(II), 0.00122 M APDDMP and 0.00590 M HCl; (h) 0.00119 M
Mg(II), 0.00205 M APDDMP and 0.01241 M HCl and (^) 0.00145 M
Mg(II), 0.00310 M APDDMP and 0.01083 M HCl vs. 0.0500 M NaOH
in 0.010 M NaCl. All solutions were at 378C and 0.15 M NaCl or 0.15 M
total ionic strength.

63.1% remaining bound to APDDMP while only 35.4% of

the Sm(III) is bound to citrate. If this is compared to the
plasma results of EDTMP [3] where 79% of the metal
remains bound to the ligand we would predict that 153 SmAPDDMP would not be as good a radiopharmaceutical as
153
Sm-EDTMP. The main reason for this discrepancy
between the two ligands lies in their affinity for Ca(II). In
the case of APDDMP in blood plasma, 77.6% of the ligand
is bound to Ca(II) rather than Ln(III). This means that
more of the lanthanide ion is available to complex to
citrate. It is noted that, once again, only a small difference
in formation constants for Ho(III) and Sm(III) with a
ligand (in this case APDDMP) has profound effects on the

Fig. 8. Scintigrams of the baboon head indicating skeletal localisation of (left)

63

Fig. 7. Speciation of Ho(III) and Sm(III) in blood plasma with APDDMP

added. H 3 cit5citric acid and H 2 sal5salicylic acid.

in vivo speciation of the metal-ion in blood plasma. The

final speciation of APDDMP in blood plasma is presented
in Table 2. 77.6% of the APDDMP is complexed to Ca(II)
(even higher than APD) confirming its affinity for Ca(II).
This explains the poor performance of APDDMP at
complexing / retaining Ho(III) in blood plasma. APDDMP
prefers to be complexed to Ca(II) rather than to Ho(III)
which is preferentially complexed to citrate.
Using the Sr(II) database of Williams and co-workers
added to the ECCLES database [14], a model of blood
plasma shows that 98.6% of Sr(II) is complexed by
phosphate and the remaining 1.4% is complexed by
ascorbate. As was the case for APD [5] addition of
APDDMP to the model had no influence on phosphate
complexation of Sr(II) in blood plasma and would therefore not improve the bone uptake of 89 Sr which is found in
the radiotherapeutic drug MetastronE.

166

Ho-APDDMP, (middle)

153

Sm-APDDMP and (right)

99m

Tc-APDDMP.

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

64

Fig. 9. Scintigrams of the baboon thorax in the same order as Fig. 8.

3.3. Animal tests

In interpreting the in vivo results, it must be kept in
mind that only the distribution of the radio-nuclide entity
of the complex is observed which is not necessarily the
same as the radio-nuclide-ligand complex that has been
administered, due to possibilities of in vivo transchelators.
In the baboon experiments, static g-camera images of
the head and chest regions were recorded 4 h after
injection of 166 Ho-, 153 Sm- or 99m Tc-APDDMP. These
results are shown in Figs. 8 and 9. Although the soft g-ray
of 166 Ho tends to somewhat diffuse the images it is quite
clear that the lowest bone uptake is achieved with 166 HoAPDDMP. ECCLES modelling suggests that only 34.6%
of the injected Ho(III) stays bound to APDDMP in blood
plasma and the observed low bone localisation could be
attributed to this. The alternative localisation is via the
ability of citrate (which competes for 61.3% of the Ho(III))
to localise Ho(III) on bone. Neves et al. [15] report that
localisation with citrate is slow and incomplete.
The distribution of the radioactive isotope in the various
organs after 4 h is presented in Table 3. The values
reported are comparisons of the activity measured (counted
pixels) in a defined area on the organ as compared to the
activity in the surrounding tissue (background) and are not
normalised between the organs. The bone uptake of 166 HoAPDDMP is 55.4% which is lower than that of 166 Ho-

APD (58.1%). Concomitantly, liver uptake has been

reduced from 68.3% to 57.1%. This phenomenon is
explained by speciation modelling as resulting from the
increased affinity of APDDMP relative to APD for endogenous Ca(II) in blood plasma. In addition, the most
predominant Ho-APDDMP complex under physiological
conditions is charged and this militates against liver
uptake.
Clearly better bone-uptake is achieved with 153 SmAPDDMP and 99m Tc-APDDMP where selective skeletal
and little non-osseous tissue accumulation is visible. For
153
Sm and 99m Tc the bone uptake is 66.7 and 82.2%
respectively. However these values do not compare well
with the recorded uptake of 90% for 153 Sm-EDTMP [16].
Also, the kidney excretion values for 153 Sm are much
higher (20% of the injected dose) than those for 99m Tc
(7%). Blood and urine samples taken during the 99m TcAPDDMP experiment were analysed using CM Sephadex
columns and chromatograms, and showed that no free
99m
TcO 42 was present. This was also concluded from the
scintigraphic images where no thyroid uptake was observed (thyroid uptake would indicate free pertechnetate).
Because of the high selectivity of APDDMP for Ca(II),
this ligand shows some similarities with HEDP. From the
literature it is known that 99m Tc-HEDP gives a better
contrast than 99m Tc-MDP (where the hydroxy group is
absent) between regions of higher and lower calcification

Table 3
Distribution of the various nuclides in different organs of baboons
Area

% Uptake in various organs after 4 h a

166

Bone
Liver
Kidneys
a

Ho-APD b

58.1
68.3
58.4

166

55.4
57.1 c
60 c

Corrections were made for 166 Ho, 153 Sm and 99m Tc decay.
Values for 3 h.
c
Values extrapolated from recordings after 2 h.
b

Ho-APDDMP

153

Sm-APDDMP

66.7
58.8
72

99m

Tc-APDDMP

82.2
52.4
62

J.R. Zeevaart et al. / Journal of Inorganic Biochemistry 83 (2001) 57 65

[17,18]. The same increased lesion to normal (L / N) bone

uptake ratio was reported for 188 Re / 186 Re-HEDP [19]. It
might be speculated that 153 Sm-APDDMP could have an
increased tumour selectivity above its normal bone uptake.
However when the blood plasma model with the total
Ca(II) concentration increased tenfold (to simulate the
high concentration of Ca(II) in calcificated or osteoclast
effected areas) was re-run and the speciation for Sm(III)
calculated, a decrease in Sm(III) involved in APDDMP
complexes was noted. This is in contrast to EDTMP where
an unexpected moderate increase in Sm(III) complexed to
EDTMP was calculated at elevated Ca(II) levels [3]. From
this it was concluded that 153 Sm-EDTMP is absorbed onto
bone followed by dissociation and subsequent Sm(III)
uptake. Although 153 Sm-EDTMP is found to have an
increase in the L / N bone uptake ratio in animal and human
studies, this might also be for other reasons such as the
increased blood flow and metabolism within the osteoclasts. The diphosphonate ligands, Etidronate (HEDP),
Pamindronate (APD) and Alendronate [20], with their
tridentate binding (the hydroxy-group being the third
binding location) display better binding strength. When the
calculated speciation for 153 Sm-HEDP [4] in blood plasma
was remodelled with a Ca(II) concentration raised by an
order of magnitude, the distribution of Sm(III) was reversed. Whereas 62% of Sm(III) was complexed to HEDP
and 37% to citrate it changes to 76% for citrate and only
13% for HEDP. This is opposite to that shown by 153 SmEDTMP. In spite of this 186 Re-HEDP achieves favourable
lesion to normal bone uptake ratios [19]. A recent publication by Claessens and Kolar [21] showed, using absorption calculations, that Sn(II) diphosphonates (like HEDP
and MDP) dissociate before adsorption to hydroxyapatite
and that Sn(II) and the diphosphonates bind at different
binding sites. (Sn(II) is used in 99m Tc-diphosphonate-based
radiopharmaceutical preparations to keep Tc in the reduced
state). Furthermore ligands with high complex stability and
high affinity for hydroxyapatite have higher target-to-nontarget ratios [21].
Therefore APDDMP, a tridentate binding diphosphonate,
might have a high L / N ratio potential-feature which could
be pursued.

4. Conclusion
It was once again gratifying to find that ECCLES is able
to give much insight into the in vivo behaviour of
radiopharmaceuticals. The ECCLES output enabled the
authors to accurately predict that 166 Ho-APDDMP would
have little liver and bone uptake. It furthermore predicted
that 153 Sm-APDDMP would have a moderate bone uptake,

65

which was recorded in vivo. Although a high bone uptake

was set out to be achieved for 166 Ho-APDDMP, attention
was drawn to 153 Sm-APDDMP and its possibilities. The
latter has the potential to achieve a high lesion to normal
uptake ratio which will be further investigated.

Acknowledgements
The authors thank the South African Nuclear Energy
Corporation for permission to publish this work. We also
thank Prof. Irene Dormehl for the baboon tests, Dr. A.V.
Kolesnikov and Dr. D Tomasevic for the translation of the
Russian article.

References
[1] A.S. Alberts, B.J. Smit, W.K.A. Louw, A.J. van Rensburg, A. van
Beek, V. Kritzinger, J.S. Nel, Radiother. Oncol. 43 (1997) 175179.
[2] USA Food and Drug Administration, Approval for 153 Sm-EDTMP
(EQuadramet), Granted during March 1997.
[3] N.V. Jarvis, J.M. Wagener, G.E. Jackson, J. Chem. Soc. Dalton
Trans. ((1995)) 14111415.
[4] J.R. Zeevaart, N.V. Jarvis, I. Cukrowski, G.E. Jackson, S. Afr. J.
Chem. 50 (1997) 189194.
[5] J.R. Zeevaart, N.V. Jarvis, W.K.A. Louw, G.E. Jackson, I. Cukrowski, C.J. Mouton, J. Inorg. Biochem. 73 (1999) 265272.
[6] J. Simon, D.A. Wilson, J.R. Garlich, D.E. Troutner, Patent 899734,
Dow Chemical Company, 1989.
[7] B.K. Shcherbakov, F.I. Belskii, M.P. Komarova, Y.M. Polikarpov,
T. Y Medved, M.I. Kabachnik, Izv. Izv. Akad. Nauk. SSSR Ser.
Khim. 3 (1982) 560564.
[8] P.M. May, K. Murray, Talanta 35 (1988) 927941, and references
therein.
[9] A.E. Martell, R.M. Smith, Critical Stability Constants, Vols. 3 and 6,
Plenum, New York, 1977 and 1989, and references therein.
[10] P.A. May, D.R. Williams, P.W. Linder, J. Chem. Soc. Dalton Trans.
((1977)) 588595.
[11] W.F. Goeckeler, B. Edwards, W.A. Volkert, R.A. Holmes, J. Simon,
D. Wilson, J. Nucl. Med. 28 (1987) 495504.
[12] B. Kurzak, A. Kamecka, K. Kurzak, J. Jezierska, P. Kafarski,
Polyhedron 17 (1998) 44034413.
[13] N.V. Jarvis, J.M. Wagener, Talanta 41 (1994) 747754.
[14] G.L. Smith, S.B. Rees, D.R. Williams, Polyhedron 4 (1985) 713
716.
[15] M. Neves, M. Reis, F. Waerenborgh, E. Martinho, L. Patricio, Inorg.
Chim. Acta 140 (1987) 359360.
[16] W.K.A. Louw, I.C. Dormehl, A.J. van Rensburg et al., Nucl. Med.
Biol. 23 (1996) 935941.
[17] J.S. Arnold, Kinetic analysis of bone imaging agents, in: Principles
of Radiopharmacology, CRC Press, Chicago, 1979, pp. 205233.
[18] I. Fogelman, D.L. Citrin, J.H. McKillop, J.G. Turner, R.G. Bessent,
W.R. Greig, J. Nucl. Med. 20 (1979) 98101.
[19] W.-Y. Lin, C. Lin, S. Yen et al., Eur. J. Nucl. Med. 24 (1997)
590595.
[20] H. Fleisch, Endocrine Rev. 19 (1998) 80100.
[21] R.A.M.J. Claessens, Z.I. Kolar, Langmuir 16 (2000) 13601367.