Evaluation of 1,2-dimethyl-3-hydroxy-4-pyridinecarboxylic acid and of other 3-hydroxy-4-pyridinecarboxylic acid derivatives for possible application in iron and aluminium chelation therapy

Article abstract of DOI:10.1016/j.poly.2013.10.007

Four new possible chelating agents for iron and aluminium, 1,2-dimethyl-3-hydroxy-4-pyridinecarboxylic acid (DT712), 3-hydroxy-1,2,6- trimethyl-4-pyridinecarboxylic acid, 2,6-dimethyl-3-hydroxy-4-pyridinecarboxylic acid, and 2-ethyl-3-hydroxy-1-methyl-4-pyridinecarboxylic acid, were synthesized, and their complex formation with Fe(III) and Al(III) was studied by potentiometry, UV-Vis, ¹H NMR, and electrospray mass spectrometry (ESI-MS). Number, stoichiometry, and stability constants of metal-ligand complexes were obtained at 25 C in aqueous (Na)Cl 0.6 m. DT712 is the most promising hydroxypyridinecarboxylic acid considered so far for iron chelation therapy, as it forms the strongest Fe(III) complexes. This compound was further investigated to better clarify its possible behaviour in vivo with particular respect to iron chelation therapy. UV-Vis measurements were performed to determine the kinetics by which DT712 extracts Fe(III) from transferrin. DT712 resulted to have better kinetic properties than existing iron chelators. Ternary metal/DT712/citric acid complexes were studied by ESI-MS to check the competition with a typical low molecular weight ligand in the blood. The formation of only binary Fe(III)/DT712 and Al(III)/DT712 complexes (and ternary complexes in aged solutions), suggests that DT712 effectively compete with citric acid in the metal complexation. Standard reduction potentials of Fe(III)/DT712 complexes, and the kinetic constants of complex formation, were obtained by cyclic voltammetry. Accordingly, no redox cycling is expected to occur at in vivo conditions, and Fe(III)/DT712 complex formation should not be kinetically limited. On the basis of the present results, DT712 is proposed as candidate for iron chelation therapy.

Full text of DOI:10.1016/j.poly.2013.10.007

Polyhedron 67 (2014) 520–528
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Evaluation of 1,2-dimethyl-3-hydroxy-4-pyridinecarboxylic acid and of
other 3-hydroxy-4-pyridinecarboxylic acid derivatives for possible
application in iron and aluminium chelation therapy
a
b
c
d
a
Annalisa Dean , Maria Grazia Ferlin , Mirjana Cvijovic , Predrag Djurdjevic , Francesco Dotto ,
a
a
e
a,
⇑
Denis Badocco , Paolo Pastore , Alfonso Venzo , Valerio B. Di Marco
a
Department of Chemical Sciences, University of Padova, via Marzolo 1, 35131 Padova, Italy
Department of Pharmaceutical and Pharmacological Sciences, University of Padova, via Marzolo 5, 35131 Padova, Italy
The Copper Mill Sevojno, 31000 Uzice, Serbia
Faculty of Science, Chemistry Department, 3400 Kragujevac, Serbia
CNR, Institute of Sciences and Molecular Technologies, via Marzolo 1, 35131 Padova, Italy
b
c
d
e
a r t i c l e i n f o
a b s t r a c t
Article history:
Four new possible chelating agents for iron and aluminium, 1,2-dimethyl-3-hydroxy-4-pyridinecarboxy-
lic acid (DT712), 3-hydroxy-1,2,6-trimethyl-4-pyridinecarboxylic acid, 2,6-dimethyl-3-hydroxy-4-pyri-
dinecarboxylic acid, and 2-ethyl-3-hydroxy-1-methyl-4-pyridinecarboxylic acid, were synthesized, and
Received 8 April 2013
Accepted 2 October 2013
Available online 12 October 2013
1
their complex formation with Fe(III) and Al(III) was studied by potentiometry, UV–Vis, H NMR, and elec-
trospray mass spectrometry (ESI-MS). Number, stoichiometry, and stability constants of metal–ligand
complexes were obtained at 25 °C in aqueous (Na)Cl 0.6 m. DT712 is the most promising hydroxypyridin-
ecarboxylic acid considered so far for iron chelation therapy, as it forms the strongest Fe(III) complexes.
This compound was further investigated to better clarify its possible behaviour in vivo with particular
respect to iron chelation therapy. UV–Vis measurements were performed to determine the kinetics by
which DT712 extracts Fe(III) from transferrin. DT712 resulted to have better kinetic properties than exist-
ing iron chelators. Ternary metal/DT712/citric acid complexes were studied by ESI-MS to check the com-
petition with a typical low molecular weight ligand in the blood. The formation of only binary Fe(III)/
DT712 and Al(III)/DT712 complexes (and ternary complexes in aged solutions), suggests that DT712
effectively compete with citric acid in the metal complexation. Standard reduction potentials of Fe(III)/
DT712 complexes, and the kinetic constants of complex formation, were obtained by cyclic voltammetry.
Accordingly, no redox cycling is expected to occur at in vivo conditions, and Fe(III)/DT712 complex for-
mation should not be kinetically limited. On the basis of the present results, DT712 is proposed as can-
didate for iron chelation therapy.
Keywords:
Iron
Aluminium
Chelation therapy
Potentiometry
Electrospray mass spectrometry
Hydroxypyridinecarboxylic acids
Ó 2013 Elsevier Ltd. All rights reserved.
1
. Introduction
The overload of iron (Fe) and aluminium (Al) produces several
⇑
Corresponding author. Tel.: +39 0498275219; fax: +39 0498275175.
E-mail address: valerio.dimarco@unipd.it (V.B. Di Marco).
URLs: http://www.chimica.unipd.it/valerio.dimarco, http://www.chimica.unip-
pathologies. Their most efficient therapeutic approach is repre-
sented by chelation therapy. The well established chelators for Fe
and Al overload therapies are Desferal (DFO) and Deferiprone
d.it/analitica/english (V.B. Di Marco).
1
1
CA = citric acid, DFO = Desferal, DQ5 = 4-hydroxy-5-methyl-3-pyridinecarboxylic
(L1) . DFO and L1 have several drawbacks, and also the recently
acid, DQ6 = 4-hydroxy-6-methyl-3-pyridinecarboxylic acid, DQ715 = 1,5-dimethyl-4-
hydroxy-3-pyridinecarboxylic acid, DQ716 = 1,6-dimethyl-4-hydroxy-3-pyridine-
carboxylic acid, DQs = 4-hydroxy-3-pyridinecarboxylic acids, DT0 = 3-hydroxy-4-
pyridinecarboxylic acid, DT1 = 3-hydroxy-1-methyl-4-pyridinecarboxylic acid,
DT2 = 3-hydroxy-2-methyl-4-pyridinecarboxylic acid, DT712 = 1,2-dimethyl-3-
hydroxy-4-pyridinecarboxylic acid, DT71201 = 2-ethyl-3-hydroxy-1-methyl-4-pyri-
dinecarboxylic acid, DT726 = 2,6-dimethyl-3-hydroxy-4-pyridinecarboxylic acid,
DT8126 = 3-hydroxy-1,2,6-trimethyl-4-pyridinecarboxylic acid, DTs = 3-hydroxy-
developed Fe chelator Deferasirox (ICL670) has controversial effi-
ciency and side effects. Therefore, the search for alternative mole-
cules is strongly and continuously requested [1–3].
Hydroxypyridinecarboxylic acids (HPs) are currently under
investigation as possible chelating agents for Fe and Al because
they display a number of favourable properties [4]. HPs form
strong complexes with both Fe(III) and Al(III), and have very low
affinity towards Zn(II), which suggests absence of essential metal
decorporation in vivo. Their low molecular weight is prerequisite
4
-pyridinecarboxylic acids, ESI-MS = electrospray mass spectrometry, ICL670 =
Deferasirox, HP = hydroxypyridinecarboxylic acids, = kinetic constant for the
reaction of complex formation, L1 = Deferiprone, Tf = transferrin.
k
f
0
277-5387/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.poly.2013.10.007

A. Dean et al. / Polyhedron 67 (2014) 520–528
521
for oral activity [5]. Toxic side effects induced by redox activity are
COOH
OH
COOH
OH
COOH
OH
COOH
OH
unlikely for both free ligands and Fe(III)/ligand complexes. HPs
investigated so far display negligible toxic effects (IC50 > 0.1 mM)
to cancer cell lines and primary human cells, following a 3 days
exposure.
+
N
+
N
+
+
N
CH3 H C
CH3 H C
N
CH3
CH CH3
2
3
3
H
CH3
CH3
CH3
DT712
DT8126
DT726
DT71201
Among the possible HP isomers, 4-hydroxy-3-pyridinecarboxy-
lic acids (DQs) substituted at the pyridinic ring have been mostly
considered up to now [4,6]. Although their affinity towards Fe(III)
and Al(III) is very high, it is lower than that of chelators available
presently, such as L1. This is the only disadvantage highlighted
up to now for all DQs, and therefore the major efforts have been di-
rected to the synthesis of derivatives displaying higher metal affin-
ity, and to the rationalization of their metal complexation affinity
in relation to their ring substitution. To this connection, the effect
of the electrondonating methyl substitution on the Fe(III) and
Al(III) complex stability was experimentally determined [6]. The
Fig. 1. 1,2-Dimethyl-3-hydroxy-4-pyridinecarboxylic acid (DT712), 3-hydroxy-
1,2,6-trimethyl-4-pyridinecarboxylic acid (DT8126), 2,6-dimethyl-3-hydroxy-4-
pyridinecarboxylic acid (DT726), and 2-ethyl-3-hydroxy-1-methyl-4-pyridinecarboxylic
acid (DT71201). All ligands are shown in their most protonated form.
low molecular weight ligand in the blood. Fe(III)/DT712 and
Fe(III)/DT8126 complexes were studied by cyclic voltammetry to
determine the standard reduction potentials of the complexes
and the kinetics of complex formation.
1
-, 5- and 6-methyl substitutions enhance the complex stability
for both metal ions, i.e. the coordinating oxygens increase more
their Lewis than their Brønsted basicity. The 5-methyl substitution
is slightly more efficient than the 6-methyl one because the former
is closer to the coordinating oxygens than the latter. The 1-methyl
substitution has a very strong effect on the Fe(III) complex stabil-
ity, and a much lower one on the Al(III) complex stability. On the
other hand, the 2-methyl substitution should be avoided because
the ortho effect caused on the carboxylic group inhibits the com-
plex formation. Finally, the methyl effects appear to be addictive.
Among the DQs studied up to now, 1,5-dimethyl-4-hydroxy-3-
pyridinecarboxylic acid (DQ715) [6] and 1,6-dimethyl-4-hydroxy-
2
. Experimental
2
2
.1. Synthesis of 1,2-dimethyl (DT712), 1,2,6-trimethyl (DT8126),
,6-dimethyl (DT726) and 2-ethyl-1-methyl (DT71201) HP derivatives
Melting points were determined on a Gallenkamp MFB 595
10M/B capillary melting point apparatus, and are uncorrected.
0
1
H NMR spectra were recorded on a Bruker 400 MHz spectrometer,
with the solvents indicated. Chemical shifts are reported in d (ppm)
values downfield from tetramethylsilane which is taken as internal
reference. Coupling constants are given in Hertz. In the case of
multiplets, chemical shifts were measured at the approximate cen-
ter. Integrals were satisfactorily in line with those expected on the
basis of compound structure. Elemental analyses were performed
in the Microanalytical Laboratory, Department of Pharmaceutical
Sciences, University of Padova, on a Perkin-Elmer C, H, N elemental
analyzer model 240B, and analyses indicated by the symbols of the
elements were within ±0.4ꢀ of the theoretical values. Mass spectra
were obtained on a Mat 112 Varian Mat Bremen (70 eV) mass spec-
trometer and Applied Biosystems Mariner System 5220 LC/Ms
3
-pyridinecarboxylic acid (DQ716) [7] are the most promising
derivatives for Fe chelation therapy, whereas 4-hydroxy-5-
methyl-3-pyridinecarboxylic acid (DQ5) [6] and 4-hydroxy-6-
methyl-3-pyridinecarboxylic acid (DQ6) [7] are the most promis-
ing DQs for Al chelation therapy.
The 3-hydroxy-4-pyridinecarboxylic acids (DTs) isomers are
now considered, where the hydroxy and carboxylic groups are ex-
changed with respect to the DQs. Three simple DTs have been stud-
ied previously: 3-hydroxy-4-pyridinecarboxylic acid (DT0) [8,9],
3
3
-hydroxy-1-methyl-4-pyridinecarboxylic acid (DT1) [10,11], and
-hydroxy-2-methyl-4-pyridinecarboxylic acid (DT2) [12]. The
(
nozzle potential 250.00). Chemical reactions were monitored by
analytical thin-layer chromatography (TLC) on Merck silica gel 60
F-254 glass plates. Solutions were concentrated on a rotary evapo-
rator under reduced pressure. Starting materials were purchased
from Aldrich Chimica and Acros, solvents from Carlo Erba, Fluka
and Lab-Scan.
The synthetic procedure leading to DT726 (5a), DT8126 (6a),
and DT71201 (6b) is reviewed in Scheme 1 and it is described in
the following subsections. DT712 (5b) has been synthesized as de-
scribed [13].
DTs expected to display the highest metal affinity can now be
selected on the basis of the methylation results obtained for the
DQs, assuming the same effects of the different substituent posi-
tions. It follows that 5-methyl derivatives should be avoided, and
that 1-, 2-, and 6-methylated DTs should display the highest metal
ion affinity, with the 1-derivatives forming stronger complexes
with Fe(III) than with Al(III). Also, results on DQs suggest that
polymethylation should lead to stronger complexes than
monomethylation.
The dimethylated derivatives 1,2-dimethyl-3-hydroxy-4-pyri-
dinecarboxylic acid (DT712) and 2,6-dimethyl-3-hydroxy-4-pyri-
dinecarboxylic acid (DT726) have been considered in this paper.
Measurements for Fe(III) only were performed also on the ethyl
analogue of DT712, 2-ethyl-3-hydroxy-1-methyl-4-pyridinecarb-
oxylic acid (DT71201). A trimethylated DT was considered as well,
2
.1.1. Synthesis of
A solution of commercial
50 mL absolute ethanol and 5 mL HCl 37ꢀ was refluxed for 5 h.
D
,
L
-alanine ethyl ester hydrochloride (2a)
D,L-alanine (5.00 g, 56.12 mmol) in
1
The solvent was then evaporated and the uncoloured oily residue
1
was dried under vacuum; Yield 92ꢀ; H NMR (CDCl
H, J = 7.07 Hz, –OCH CH ), 1.45 (d, 3H, J = 7.25 Hz, –CHCH
q, 2H, J = 7.07 Hz, –OCH CH ), 4.02 ppm (q, 1H, J = 7.25 Hz, –CHCH
3
): d 1.05 (t,
3
(
2
3
3
), 3.48
3
-hydroxy-1,2,6-trimethyl-4-pyridinecarboxylic acid (DT8126).
2
3
3
);
The compounds studied in this work are shown in Fig. 1.
+
HRMS (ESI, 140 eV): m/z [M+H ] calcd for C
found: 118.0794.
5
H12NO
2
: 118.0868
The synthesis of DT712 was recently performed in our laborato-
ries, and it is described [13], whereas that of DT8126, DT726 and
DT71201 is new. The Fe(III) and Al(III) coordination properties of
these DTs were studied by potentiometric, UV–Vis, and (in the case
2.1.2. Synthesis of ethyl 2-acetamidopropanoate (3a)
1
of Al(III)) H NMR measurements. The ability of DT712 and of
In a 250 mL round bottom flask, to 5 g (32.54 mmol) of ethylala-
nine hydrochloride 2a were added 25 mL of triethylorthoacetate,
and the reaction mixture was refluxed for 2 h (oil bath). In the
meantime, the colourless solution became yellow–brown. At the
end, the solution was concentrated in a rotary evaporator giving
a light yellow oily product. Yield 80–90ꢀ; H NMR (CDCl
DQ715 to extract Fe(III) from transferrin was studied by UV–Vis,
in order to determine the kinetic constant of the metal extraction.
Binary metal/DT712 and ternary metal/DT712/citric acid solutions
were studied by electrospray mass spectrometry to evaluate the
Fe(III) and Al(III) competition between the chelator and a typical
1
3
): d

5
22
A. Dean et al. / Polyhedron 67 (2014) 520–528
Scheme 1. Pathway for the syntheses of DT726 (5a), DT8126 (6a), and DT71201 (6b).
1
.30 (t, 3H, J = 7.12 Hz, –OCH
CHCH ), 2.11 (s, 3H, COCH ), 4.89 (q, 1H, J = 7.12 Hz, –CHCH
.31 ppm (q, 1H, J = 7.14 Hz, –CHCH ); HRMS (ESI, 140 eV): m/z
: 160.0929, found: 160.0794.
2
CH
3
), 1.46 (d, 3H, J = 7.14 Hz, –
325–330 °C (dec); 1H NMR (D
3.95 (s, 3H, 2-C–CH ), 7.15 ppm (s, 3H, 5-H); C NMR (D
d 19.91 (6-C–CH ), 40.83 (2-C–CH ), 122.67 (5-C), 137.25 (4-C),
142.12 (6-C), 147.78 (2-C), 158.957 (3-C), 175.50 ppm (COOH);
2
O/NaOD): d 2.54 (s, 3H, 6-C–CH
3
),
1
3
3
3
3
),
3
2
O/NaOD):
4
3
3
3
+
[
7 3
M+H ] calcd for C H13NO
+
8 3
HRMS (ESI, 140 eV): m/z [M+H ] calcd for C H10NO : 168.169416,
2
.1.3. Synthesis of 5-ethoxy-2,4-dimethyloxazole (4a)
In a three necked oven dried 500 mL round bottom flask, a sus-
pension of P (12 g), Celite (6 g) and MgO (6 g) in 150 mL CHCl
was vigorously stirred at room temperature under inert atmo-
sphere (N ), until it became homogeneous. A solution of ethyl 2-
acetamidopropanoate (3a) (5.01 g, 31.41 mmol) dissolved in
0 mL CHCl was slowly added drop wise and the stirring was con-
tinued for 30 min. The mixture was heated to reflux for 20 h. At the
end, a saturated solution (300 mL) of NaHCO was slowly added to
found: 168.0611; Anal. Calc. for C NO (167.16): C, 57.48; H,
5.43; N, 8.38. Found: C 57.08, H, 5.99; N, 8.23ꢀ.
8
H
9
3
2
O
5
3
2.1.5. Synthesis of 3-hydroxy-1-methyl-4-pyridinecarboxylic acid
derivatives (6a and 6b)
As a typical procedure, in a 100 mL round bottom flask, 2 g
(11.96 mmol) of 5a or of 5b (prepared according to Brun et al.
[13]) were suspended in 20 mL of THF. The mixture was added
2
5
3
3
with 0.634 g (5.98 mmol) of Na
2 3
CO . 3.72 mL (59.80 mmol) of
the cooled reaction mixture and then filtered to remove the solids.
The filtrate was extracted with chloroform, the organic phase was
commercial CH I were then added, and the reaction mixture was
3
stirred at room temperature for about 30 min, refluxing for 1 h
and again at room temperature for 4 h. The solvent was moved
off in a rotary evaporator and the residue was taken up in a small
volume of distilled water (30 mL), acidified with 2 M HCl and stir-
red with slight heating for about 30 min in the presence of 2–3
2 4
washed with water, and dried with anhydrous Na SO . On remov-
ing the solvent in a rotary evaporator, a slight yellow oil was ob-
1
tained. Yield 40–60ꢀ; H NMR (CDCl
3
): d 1.383 (t, 3H, J = 7.09 Hz,
), 2.20 (s, 3H, 4-C–CH ),
): d
), 68.74
3
), 126.28 (4-C), 140.78 (5-C), 168.21 ppm (2-C); HRMS
–
4
1
OCH
.25 ppm (q, 2H, J = 7.09 Hz, –OCH
1.20 (4-C–CH ), 14.53 (2-C–CH ), 14.84 (–OCH
CH
ESI, 140 eV): m/z [M+H ] calcd for C
2
CH
3
), 2.16 (s, 3H, 2-C–CH
3
3
13
2
CH
3
); C NMR (CDCl
CH
3
2 2
drops of H O to remove iodine. After concentration of the solution
3
3
2
3
to about a half volume in a rotary evaporator, the formed solid
product was collected by filtration, and a white residue was ob-
tained. To remove the NaCl present as an impurity, the residue
was treated with a small volume of absolute ethanol (15 mL), the
suspension was filtered, and the filtrate was added with an excess
of diethyl ether, obtaining the precipitation of the desired com-
pound. This was collected and dried under vacuum.
(
(
–OCH
2
+
7
H12NO
2
: 142.1751, found:
1
42.0863.
2.1.4. Synthesis of 2,6-dimethyl-3-hydroxy-4-pyridinecarboxylic acid
(
5a)
In a 50 mL round bottom flask, 4.00 g (28.33 mmol) of oxazole
a were reacted with 4.00 g (3.81 mL, 55.5 mmol) of acrylic acid
4
2.1.6. 3-hydroxy-1,2,6-trimethyl-4-pyridinecarboxylic acid (6a)
1
at room temperature. After 30 min the formation of a precipitate
was observed which was collected after some hours at room tem-
perature. The thin white crystalline product was re-crystallized
from water and dried under vacuum at 40 °C. Yield 40ꢀ; mp
Yield 90ꢀ; mp 310–315 °C (dec); H NMR (D
2
O, NaOD): d 2.27
), 3.35 (s, 3H, 1-N–CH ),
O, NaOD): d 17.49 (2-C–
), 127.32 (5-C), 139.12 (4-C),
(s, 3H, 6-C–CH
6.79 ppm (s,1H, 5-C–H); C NMR (D
CH ), 25.96 (6-C–CH ), 38.65 (N–CH
3
), 2.29 (s, 3H, 2-C–CH
3
3
1
3
2
3
3
3

A. Dean et al. / Polyhedron 67 (2014) 520–528
523
1
46.04 (6-C), 148.89 (2-C), 153.51 (3-C), 175.15 ppm (COOH);
[18]. Al(III)/DT712 equilibration times were significant longer than
usual, so that batch titration had to be performed for this system.
Thirty-five solutions containing Al(III) and DT712 at metal-to-li-
gand ratios of 1:1, 1:3, and 1:5, were prepared, and they were
added with different amounts of NaOH. The e.m.f. measurement
was performed after equilibration (ca. 24 h) by the calibrated
VWR glass electrode + Ag/AgCl/0.6 m NaCl reference electrode.
All stability constants were calculated using computer program
PITMAP [19]. PITMAP minimizes the sum of the squares of the dif-
ferences between experimental and calculated e.m.f. values. Opti-
mization is performed using pitmapping [20] or simplex [21] as
nonlinear least squares algorithms. Mass balance equations were
solved, i.e. species concentrations at equilibrium were obtained,
by means of the Newton–Raphson method [21]. Stability constants
for metal/hydroxo complexes have been taken from the literature:
+
HRMS (ESI, 140 eV): m/z [M+H ] calcd for C
found: 182.0812; Anal. Calc. for C Cl (217.6489): C, 49.67;
H, 5.56; N, 6.44. Found: C, 49.22; H, 5.27; N, 6.56ꢀ.
9 3
H12NO : 182.1985,
9 3
H12NO
2.1.7. 2-ethyl-3-hydroxy-1-methyl-4-pyridinecarboxylic acid (6b)
1
Yield 80–90ꢀ; mp 225–230 °C (dec); H NMR (D
.20 (t, 3H, J = 7.63 Hz, 2-C–CH CH ), 3.05 (q, 2H, J = 7.63 Hz, 2-
CH ), 4.15 (s, 3H, NCH ), 7.25 (d, 1H, J5,6 = 6.01 Hz, 5-H),
.42 ppm (d, 1H, J6,5 = 6.01 Hz, 6-H); C NMR (D
0.34 (–CH CH ), 29.62 (–CH CH ), 45.62 (N–CH ), 121.80 (5-C),
2 ,
O NaOD): d
1
2
3
C–CH
2
3
3
13
7
2
1
1
C
2 ,
O NaOD): d
2
3
2
3
3
27.89 (4-C), 141.06 (6-C), 153.19 (2-C), 161.49 (3-C),
+
75.28 ppm (COOH); HRMS (ESI, 140 eV): m/z [M+H ] calcd for
9 3 9 3
H12NO : 183.2039, found: 183.1897; Anal. Calc. for C H12NO Cl
(
217.6489): C, 49.67; H, 5.56; N, 6.44. Found: C, 49.52; H, 5.69;
log bFeOH = ꢁ2.87, logb
= ꢁ6.16 [15], logb
= ꢁ12.16,
N, 6.23ꢀ.
FeðOHÞ
= ꢁ22.16, logbFe2
2
FeðOHÞ
3
logb
= ꢁ2.9, logbFe3
= ꢁ6.3, log
FeðOHÞ4
ðOHÞ
ðOHÞ4
2
1
b
Fe12ðOHÞ34 = ꢁ48.9 [22], logbAlOH = ꢁ5.52, logbAlðOHÞ2 = ꢁ11.3, logbAlðOHÞ3
2.2. Potentiometric, UV–Vis, and H NMR study
=
ꢁ
ꢁ17.3, logb
= ꢁ23.46, logb
= ꢁ13.57, logb
=
Al13ðOHÞ32
AlðOHÞ
4
Al
3
ðOHÞ
4
109.2 [23].
All potentiometric measurements were performed using Radi-
UV–Vis measurements were performed at 25.0 ± 0.1 °C in 0.1,
ometer ABU93 Tri-burette and Metrohm 765 Dosimat apparatus.
UV–Vis and H NMR spectra were recorded using a Perkin-Elmer
Lambda 20 spectrophotometer and a Bruker DRX-400 spectrome-
ter operating at 400.13 MHz, respectively. All analyte concentra-
tions were expressed in molality scale (mol/kg of water). For
potentiometric and UV–Vis measurements, working solutions of
HCl (0.1 m), NaOH (0.1–0.2 m, freshly prepared), FeCl
containing HCl 0.3 m), and AlCl (0.04 m, containing HCl 0.3 m),
were prepared and standardised as described previously [14,15].
The ligands were used as synthesized to prepare 0.0067 m
1
0.5 or 1 cm quartz cuvettes for solutions containing the free ligands
(
DT712, DT8126, DT726, DT71201), for solutions containing each
ligand + Fe(III), and for solutions containing each ligand + Al(III),
at various pH values. The pH was measured with the same elec-
trodes and procedures as for potentiometric titrations. pH values
below 2 were computed from the stoichiometric concentration of
3
(0.05 m,
+
HCl, because [H ] modifications produced by all other species were
3
negligible under these conditions. The pKa1 for both ligands
(
(
(
Table 1) and the log b values for some metal/ligand complexes
Table 2) were computed by the program PITMAP. The values
molal absorbivity coefficient) of the various species at the
e
(
(
DT712), 0.0066 m (DT726), 0.0041 m (DT8126), and 0.0042 m
DT71201) working solutions. Ionic strength of all solutions was
1
given wavelengths were computed as well by the program. For
adjusted to 0.6 m (0.594 M) (Na)Cl [28]. Solutions for H NMR mea-
surements were prepared by dissolving weighed amounts of ligand
and AlCl in D O, and the pH was set by NaOD or DCl additions.
3 2
Internal reference for H NMR measurements was Me
COOH (TSP, Aldrich 99ꢀ+).
Potentiometric measurements were carried out at 25.0 ± 0.1 °C;
duplicate potentiometric measurements were performed using
two glass electrodes (VWR 662-1792) and an Ag/AgCl/0.6 m NaCl
reference electrode [16,17] with a J-shaped junction. The following
titrations were performed: glass electrode calibration, base stan-
dardisation (which gave the water ionization constant,
pK
gave the pK
metal/ligand mixtures. Metal ion to ligand ratios were from 1:1
3
example, for Fe(III)/DT712 at 425 nm,
e
FeL = (1.65 ± 0.04) ꢀ 10
ꢁ1
ꢁ1
mol kg cm
.
1
1
H NMR spectra were obtained at room temperature. Chemical
shift values are given in d units with reference to internal TSP. Suit-
able integral values for the proton signals were obtained by a pre-
scan delay of 10 s. The assignment of the proton resonances was
performed by standard chemical shift correlations and NOESY
3 2
SiCH CH2-
2
measurements when necessary. Spectra were recorded in D O
solutions containing each free ligand, or the ligand + Al(III). pH
1
for H NMR solutions was set by DCl/NaOD (Aldrich) additions.
pH was measured with a Crison 5014 combined glass electrode
previously calibrated in buffered aqueous solutions at pH = 4 and
pH = 7. In heavy water the values of pD were computed by adding
0.41 pH units to the pH meter readings [24] in order to correct for
w
= 13.67 ± 0.01), ligand standardisation experiments (which
a
of the ligands, Table 1) [14], and titrations of binary
ꢁ4
to 1:6, Fe(III) and Al(III) concentrations ranged from 3 ꢀ 10 to
ꢁ3
isotopic and solvent effects due to the use of D
.3. Kinetics of Fe(III) removal from transferrin
The kinetic UV–Vis measurements were performed using a Var-
2 2
O instead of H O.
1
ꢀ 10 m. Experimental points were regarded as non-equilib-
rium points and rejected in subsequent data analysis if the poten-
tial did not stabilize within 6 min after each addition of titrant
2
(
maximum allowed rate of e.m.f. change: 0.05 mV/min). In the
acidic pH range of Al(III) titrations, measured e.m.f. drifted and
reached a constant value only after few hours, suggesting a low
complex formation rate. This finding is in agreement with previous
results of other DTs [8,10,12]. Experimental details regarding the
handling of the slow kinetics during the titrations are reported
ian-Cary 100 Bio with a Varian-Cary temperature controller. The
temperature was set at 37.0 ± 0.1 °C. The following solutions were
prepared: FeCl
3
200
lM (dilution of the solution used in potentio-
metric titrations), olo-transferrin 100
lM (transferrin saturated
with Fe(III), Sigma, 98ꢀ) DT712 0.050 M, DQ715 0.050 M, and Def-
eriprone 0.050 M (Sigma–Aldrich, 98ꢀ). The Fe(III) solution was
added with NaCl 0.15 M. A buffer solution containing NaCl
Table 1
pK
a
values of DT712, DT8126, DT726, and DT71201, at 25 °C in aqueous (Na)Cl 0.6 m.
0
3
.15 M, NaHCO 0.05 M (J.T. Baker, 99+ꢀ) and Hepes 0.05 M (Sigma,
Standard deviations are given in brackets.
9
9.5+ꢀ) was set at pH = 7.4, and it was used as ‘‘solvent’’ to prepare
pK
a
DT712
DT8126
DT726
DT71201
all other solutions and as blank in the UV–Vis measurements.
Preliminary UV/Vis spectra were collected for all ligands, for
olo-transferrin, and for 1:1 Fe(III)-ligand mixtures at room temper-
ature, in order to individuate the most selective wavelength for
the kinetic measurements. The chosen wavelengths are reported
pKa1
pKa2
pKa3
0.4 (0.1)a
7.70 (0.01)
–
0.6 (0.1)a
8.06 (0.01)
–
0.6 (0.1)a
5.32 (0.01)
11.09 (0.03)
0.5 (0.1)a
7.82 (0.01)
–
a
Values obtained by UV–Vis.

5
24
A. Dean et al. / Polyhedron 67 (2014) 520–528
Table 2
Stability constants (log b) for metal/ligand complexes, at 25 °C in aqueous (Na)Cl
.6 m (reactions: p M + q L + r H = M ). Standard deviations are given in brackets.
ues) are reported in the captions of Tables 4 and 5. The solution
pH was adjusted by the addition of trifluoroacetic acid (Fluka, Buc-
hs, Switzerland, >99ꢀ) or triethylamine (Fluka, Buchs, Switzerland,
0
p q r
L H
L is the completely deprotonated form, which has charge ꢁ1 for DT712, DT8126, and
>99.5ꢀ), and it was measured by glass combined-electrode mea-
DT71201, and has charge ꢁ2 for DT726.
surement. All solutions were aged for two days before recording
the spectra. Spectra for solutions containing metal, DT712 and CA
were recorded also at larger aging times (15, 30, 40, 60 days).
p, q, r
DT712
DT8126
DT726
DT71201
Fe3+
1
1
1
1
1
1
1
1
1
1
1
1
, 1, 1
, 1, 0
, 1, –1
, 2, 2
, 2, 1
, 2, 0
, 2, –1
, 2, –2
, 3, 3
, 3, 2
, 3, 1
, 3, 0
20.21 (0.05)a
17.22 (0.06)
11.14 (0.01)a
8.63 (0.02)
10.54 (0.03)a
7.46 (0.01)
10.84 (0.04)a
7.70 (0.04)
2.5. Voltammetric study
37.52 (0.03)
32.82 (0.02)
Cyclic voltammetry experiments were carried out with a home-
made three-electrode cell having a 3 mm glassy carbon working
electrode, a saturated Ag/AgCl/KCl, and a Pt counter electrode.
The potential wave-form applied to the working electrode was
generated by an AUTOLAB PGSTAT100 (Quebec, Canada) potentio-
stat piloted with a GPES software (General Purpose Electrochemical
System, version 4.9). The working electrode was mechanically pol-
ished with SiC 4000 grits, rinsed with water, and conditioned at
0.8 V for 20 s. Solutions were degassed with nitrogen for 5 min
prior each experiment. Potential was spanned from 0.80 to
19.54 (0.01)
15.35 (0.02)
7.74 (0.09)
19.29 (0.04)
14.62 (0.04)
19.60 (0.07)
15.32 (0.09)
6.54 (0.08)
53.61 (0.04)
46.30 (0.09)
40.05 (0.04)
27.03 (0.09)
8.51 (0.05)
24.91 (0.02)
26.30 (0.09)
Al3+
1
1
, 1, 1
, 1, 0
18.68 (0.08)a
9.08 (0.03)a
8
.64 (0.02)a
ꢁ1
1
1
1
, 2, 2
, 2, 1
, 2, 0
32.82 (0.06)
29.44 (0.09)
ꢁ0.30 V at 10 mV s . A Crison GLP 21 pH-meter and an Ingold
combined glass electrode were used for pH measurements. The
DIGISIM software [25] versus 3.03b (February 2004) was used for
mechanism simulation.
15.40 (0.05)
1
16.99 (0.02)
16.99 (0.08)
5.33 (0.01)a
a
1
1
1
1
, 3, 3
, 3, 2
, 3, 1
, 3, 0
51.10 (0.06)
44.85 (0.07)
37.41 (0.09)
29.45 (0.09)
20.98 (0.06)
22.87 (0.03)
Table 4
a
2
2.9 (0.5)
Experimental positive mass measurements in ESI-MS analysis of free DT712
(
C
3
pH = 5.16, 6.93, CDT712 = 1.0 ꢀ 10ꢁ M), Fe(III)/DT712 (pH = 2.05, 3.14, 4.99, 5.97,
4
a
Values obtained by UV–Vis.
ꢁ4
ꢁ4
Fe = 0.5 ꢀ 10 M, CDT712 = 2.0 ꢀ 10 M) and of Al(III)/DT712 solutions (pH = 2.95,
ꢁ4 ꢁ4
.85, 5.15, 6.96, CAl = 1.0 ꢀ 10 M, CDT712 = 4.0 ꢀ 10 M). DT712 is symbolized as
HL, which has charge 0 and M
m
= 167. If not differently specified, detected ions have
in Table 3. The kinetic UV–Vis measurements were performed at
7 °C in quartz cells with optical length = 1 cm, on mixtures con-
taining each ligand 0.020 M and olo-transferrin 50 M. Each solu-
tion was warmed at ca. 40 °C before mixing and starting the kinetic
plot. Absorbance values were collected every 15 s for a total time of
charge +1.
3
Experimental m/z
Identified species/ions
l
+
+
+
1
1
68, 150, 122,
06, 94
HL (+H ), (+H –H
2
O), (+H –HCOOH),
+
+
(+H –H
2
O–CO
2
), (+H –HCOOH–CO)
b
a
3+
+
+
110.5 , 190 ,
214, 239,
3
Fe + HL (–2H ), (–CH OH–H ),
(+K –3H –C
(+K +2H O–3H –CO)
2
Fe + 2HL (–2H –C
+
+
+
1
h. Data were fitted to Eq. (1) using a commercial non-linear least
2
H
7
N), (+H
2
O–2H ),
+
+
2
3
3
5
2
2
2
67
squares algorithm (SigmaPlot 2004 for Windows version 9.01).
3+
+
17,
2
H
7
N–C ),
2 2
H
a
+
+
81 , 388
55, 577
03, 446, 504
14, 266
69,
(+2H
2
O–H –CO–CH
4
), (ꢁ2H )
3
+
+
+ +
Fe + 3HL (ꢁ2H ), (+Na –3H )
2.4. Mass spectrometric study
unassigned
3+
+
+
+
+
Al + HL (+Na –3H ), (+2H
2
O + K –3H )
Al + 2HL (–2H –HCOOH–CO ),
(–2H –2CO ), (–2H –CH –C H ), (–2H )
Electrospray mass spectrometry (ESI-MS) and ESI-MS/MS
3+
+
2
+
+
+
experiments were performed on a Waters UPLC System (with ESI
and APCI source, triple quad analyser and TQ detector). Data acqui-
sition and evaluation was made with MASSLYNX software. The instru-
ment operated in positive ion mode, capillary voltage 3.50 kV, cone
voltage 25 V, source temperature 150 °C, desolvation temperature
271, 317, 359
526, 548
2
4
+
2
+
2
3+
+
Al + 3HL (–2H ), (+Na –3H )
a
Fe2+ species (ESI-MS reduction from Fe3+).
Ion with charge +2.
b
3
50 °C. Solutions were directly infused in the ESI source using a
Harvard Apparatus Model 11 syringe pump (Harvard Apparatus,
Holliston, MA, USA) at a flow rate of 10 L/min. For MS/MS mea-
Table 5
Experimental positive mass measurements in ESI-MS analysis of Fe(III)/DT712/CA
l
ꢁ
4
ꢁ4
ꢁ4
surements, the collision energy was set between 20ꢀ and 40ꢀ of
the maximum Vrf. The precursor ions for the MS experiments were
selected with an isolation width of 1.0 m/z. The spectra of the fol-
lowing aqueous solutions were obtained: free DT712, Fe(III)/
DT712, Al(III)/DT712, Fe(III)/DT712/CA (citric acid, p.a., Merck);
Al(III)/DT712/CA. Experimental details (concentration and pH val-
(pH = 3.01, 5.17, CFe = 0.5 ꢀ 10 M, CDT712 = 2 ꢀ 10 M, CCA = 0.5 ꢀ 10 M) and
2
ꢁ4 ꢁ4
Al(III)/DT712/CA solutions (pH = 3.16, 5.05, CAl = 1.0 ꢀ 10 M, CDT712 = 4.0 ꢀ 10 M,
ꢁ4
C
CA = 1 ꢀ 10 M). DT712 is symbolized as HL, which has charge 0 and M
m
= 167, CA is
symbolized as H X, which has charge 0 and M = 192. Only peaks of the ternary
3
m
species are highlighted in this Table.
Experimental m/z
Identified species/ions
+
+
3
5
51
53
Fe3 + HL + H
3
X (ꢁ2H –CO
2
2
–H O)
a
3+
+
Fe + 2HL + H
Fe + 3HL + H
Fe + 4HL + H
3
3
3
X (ꢁH –CO)
3+
+
Table 3
765
968
661
511
5
7
8
4
X (+H
2
O–2H )
3+
+
Kinetic constants for the removal of Fe(III) from Tf by some chelators (T = 37 °C; data
were collected at the given wavelengths). The standard deviations obtained from the
fitting procedure are given in brackets.
X (+3H
2
O–2H )
unassigned
3
+
+
+
Al + 2HL + H
Al + 2HL + H
Al + 3HL + H
Al + 4HL + H
3
3
3
3
X (+Na –3H –CO
2
–H
2
O)
3
+
+
53
20
83
X (+H
X (+H
X (+H
2
2
2
O–2H –CH
O–2H –CH
4
)
)
ꢁ
Chelator
k (nm)
k (min 1
)
3+
+
4
3
+
+
O–5H –CO
2
)
DT712
DQ715
L1
420
464, 334
350
0.277 (0.002)
0.164 (0.005)
0.113 (0.001)
95, 564
unassigned
a
2
+
3
+
)
.
F
e
s
p
e
c
i
e
s
(
E
S
I
-
M
S
r
e
d
u
c
t
i
o
n
f
r
o
m
F
e

A. Dean et al. / Polyhedron 67 (2014) 520–528
525
3
. Results and discussion
.1. Acidity constants
The pK values for each ligand are reported in Table 1. They
0
0
.8
.6
pH = 1.30
3
A
pH = 0.29
a
were determined by potentiometric titrations, except for all pKa1
which were computed from UV–Vis spectra. The assignment of
0.4
0.2
0.0
the pK
ities: for DT726, pKa1 is assigned to –COOH, pKa2 to –NH, pKa3 to
OH; for DT712, DT8126, and DT71201, in which the pyridinic
nitrogen cannot deprotonate, pKa1 is assigned to –COOH and pKa2
to –OH [4]. All pK values are ca. 0.5–1 log units larger than the
a
values is straightforward on the basis of the intrinsic acid-
–
300
400
500
600
700
λ (nm)
a
corresponding ones obtained for the less or non-methylated DT0,
DT1, and DT2 [8–12], because of the inductive effect of the methyl
groups.
Fig. 3. UV–Vis spectra at various pH values for Fe(III)/DT726 solutions: aqueous
ꢁ3
ꢁ3
(
Na)Cl 0.6 m, T = 25 °C;
C
M
= 1.32 ꢀ 10 m,
C
L
= 1.30 ꢀ 10 m, optical
length = 0.1 cm.
1H NMR spectra were obtained to confirm the Al(III)-DT specia-
tion picture and to get some structural information on the metal
complexes. Spectra of each free ligand show the typical signals
which appear in the aliphatic and in the aromatic d range (see
Experimental). The aromatic signals of free DT712 are shown in
Fig. 4 (label ‘‘0’’). In the presence of Al(III), the resonances of the
free ligands were still observed, together with new peaks due to
3
.2. Metal/ligand complexes
Potentiometric data were elaborated in a sequential manner, as
described previously [8], to avoid the difficulties produced by the
strong correlation between parameters to be optimised in the pres-
ence of a large number of metal/ligand species. Stoichiometry and
stability constants of metal/ligand complexes identified in solution
are reported in Table 2. Distribution diagrams for metal/DT712
solutions, computed at 5 ꢀ 10 m metal ion and 2 ꢀ 10 m li-
gand, are shown in Fig. 2. The observed speciation, though rather
complicated, resembles the general characteristics observed for
the other metal/HPs solutions studied so far. Through discussions
are reported [4,8,9].
1
the complex formation (Fig. 4, labels ‘‘1’’, ‘‘2’’, ‘‘3’’). H NMR Spectra
ꢁ4
ꢁ3
can be interpreted with the help of a distribution diagram, drawn
at the same metal and ligand concentration. This diagram is not re-
ported here, but a similar one can be referred to in Fig. 2b. Accord-
ingly, the minor and narrow doublets at pD = 2.5 can be attributed
2
+
+
2
to AlL . The broad signals at pD = 4.3 can be attributed to AlL . All
All UV–Vis spectra of metal/DT solutions display the typical
absorption of substituted pyridinic rings. One or two charge-trans-
fer peaks appear between 400 and 500 nm when Fe(III) coordinates
to each ligand. As a typical example, UV–Vis spectra for Fe(III)/
DT726 solutions at various pH values are shown in Fig. 3. Al(III)
complexation produces modest changes of the absorption coeffi-
cients in the wavelengths range 250–350 nm.
the 1:2 Al(III)/HPs complexes give broad signals [6–12], due to the
formation of several isomers (up to 4 diastereoisomers can exist in
solution for this complex) and to a slow kinetics of interconversion
3
between them. Pattern of AlL signals (pD = 7.4) is peculiar of
1
Al(III)/HPs H NMR spectra, and it is due to the presence of two in-
ert isomers in solution. As result, for each signal of the free ligand
3
four signals are obtained when AlL is formed (for a complete dis-
The pKa1 of each DT, and the log b values of the ‘‘first’’ 1:1
Fe(III)-to-DT complex (FeLH for DT726, FeL for the other DTs),
could not be obtained by potentiometric titrations, because the
cussion see Refs. [8,10]). Ten peaks are observed between d 7.8 and
8.0 in this case (four of them are partially merged): two belongs to
the free ligands, and the other eight are due to AlL . Part of this pat-
3
2+
2+
corresponding species are already formed at very acidic pH values
(
tern can be observed also in the spectrum at pD = 4.3.
see e.g. the distribution of FeL2 in Fig. 2a). Therefore, these values
+
1
Semiquantitative data obtained from H NMR integrals give the
were computed by UV–Vis (Tables 1 and 2). In the case of Al(III)/
DTs, UV–Vis measurements were performed at acidic and neutral
pH, to check whether the slow kinetics of complex formation
following percentage values of free ligand-to-total ligand: 74ꢀ,
46ꢀ, and 21ꢀ at pD 2.5, 4.3, and 7.4, respectively. These values
can be compared with those calculated on the basis of the specia-
tion data reported in Tables 1 and 2, computed at the same Al(III)
and DT712 concentrations and at the same pH values. Values are,
(
see Experimental) affected potentiometric results. The agreement
between the stability constants determined by the two techniques
Table 2) is good, thus suggesting the reliability of our speciation
data.
respectively, 73ꢀ, 45ꢀ, and 15ꢀ. The good agreement between 1
H
(
a 1.0
1.0
b
FeL
AlL3
Al
FeL3
Al(OH)4
0
0
0
0
0
.8
.6
.4
.2
.0
0.8
0.6
0.4
0.2
0.0
AlL
AlL2
α
α
FeL2
FeLH-1
FeL H
2
Fe
-1
0
2
4
6
8
10
0
2
4
6
8
10
pH
pH
= 5 ꢀ 10ꢁ m,
4
Fig. 2. Distribution diagrams of the most important Fe(III) (a) and Al(III) (b) species in the presence of DT712 in aqueous (Na)Cl 0.6 m, T = 25 °C; C
M
= 2 ꢀ 10ꢁ m. Charges of the species are omitted for simplicity. The dashed line shows the theoretical starting pH for Fe(III) hydroxide precipitation.
3
C
L

5
26
A. Dean et al. / Polyhedron 67 (2014) 520–528
0
apy. This should not be a problem if the very low in vitro toxicity of
0
all HPs [4] will be confirmed in vivo. It is also important to note that
DT712 (and its analogue DT71201) forms Fe(III) complexes which
are more stable by 0.5–1 log units at physiological pH (Fig. 5a) with
respect to the complexes formed by all other HPs investigated so
far. The high stability of Fe(III)/DT712 complexes was expected
0
0
1
1
1
1
(
see Introduction), whereas its higher stability with respect to
Fe(III)/DQ715 and Fe(III)/DQ716 complexes was not. So, methyl-
ated DTs appear to be better Fe chelator candidates than DQs.
The importance of the 1-methyl substitution for the enhancement
of the Fe(III) complexation strength, highlighted for DQs, is con-
firmed for DTs, too, as DT712 forms much stronger Fe(III) com-
plexes than DT726. On the other hand, the trimethylated DT8126
forms weaker complexes than DT712, thus indicating that the
overmethylation of the pyridinic ring is not beneficial to increase
the complex stability. The picture for Al(III) (Fig. 5b) is similar, as
DT726 and DT8126 form weaker Al(III) complexes than DT712.
However, DQ5 and DQ6 still form stronger Al(III) complexes. For
the Al chelation therapy, therefore, DQs appear to be better candi-
dates than DTs.
pD = 2.5
0
0
0
0
2
2
3
3
3
pD = 4.3
3
3
3
3
3
0
0
0
0
3
The above results show that, among all methylated HPs, DT712
is the strongest Fe(III) chelator. This compound was therefore
investigated to better clarify its possible behaviour in vivo with
particular respect to Fe chelation therapy. The following thermody-
namic and kinetic properties were evaluated, and results are re-
ported in the subsequent sections: the ability of DT712 to
remove the metal ion from transferrin, the formation of ternary
complexes with citric acid, the redox potentials of some Fe(III)/
DT712 complexes, and their kinetic constants. Other useful data
for DT712, i.e. its in vitro toxicity and anti-oxidant properties, have
been obtained previously [13]. It was shown that DT712 has negli-
gible citotoxicity (IC50 < 0.1 mM) and displays significant antioxi-
dant activity.
pD = 7.4
8
.2 8.1 8.0 7.9 7.8 7.7
ppm
Fig. 4. 1H NMR spectra (aromatic d range) of D
O solutions containing Al(III) and
2
ꢁ3
ꢁ
3
DT712 (CAl = 1.30 ꢀ 10 m, CDT712 = 5.05 ꢀ 10 m) at various pD values. Signal
2
+
+
2 3
labels: ‘‘0’’ = free ligand, ‘‘1’’ = AlL , ‘‘2’’ = AlL , ‘‘3’’ = AlL .
NMR and potentiometric – UV–Vis data further suggests reliability
of speciation results.
3.3. Stability of the complexes: discussion
3.4. Kinetics of Fe(III) removal from transferrin
The strength of Fe(III) and Al(III) complexes formed by the
Transferrin (Tf), the major Fe transporter protein of the blood, is
completely saturated when Fe overload occurs [1]. An ideal chelat-
ing agent should be able to extract Fe(III) from the two Tf binding
sites. For the existing chelating agents DFO, L1, and ICL670, the
most limiting property in this process is the kinetics, rather than
the thermodynamics, because the time required to remove the me-
tal from Tf is larger than the half-life of the drug in the blood. For
example, DFO extracts Fe(III) from Tf within ca. 24 h, but its blood
half-life is ca. 10 min. [29]. Also ICL670 has no favourable kinetic
properties [30]. Among the chelators tested by Turcot et al. [31],
L1 showed the best extraction kinetics, and to the best of our
knowledge it is still the best chelator as regards this property.
However, even for L1 the time required for Fe(III) extraction
(1–2 h) [31] is comparable with that required for drug metabolization
investigated DTs can be compared with those formed by the DQs
displaying the strongest metal affinity, i.e. DQ715 and DQ716 for
Fe(III), DQ5 and DQ6 for Al(III) [6,7]. To this aim, pM versus pH
graphs (pM = ꢁlog[M]) were computed (Fig. 5) for the selected li-
gands: higher pM values indicate stronger metal/ligand complexes.
Computations were performed at metal and ligand concentrations
ꢁ6
which are representative of physiologic conditions (10 m metal
ꢁ5
ion and 5 ꢀ 10 m ligand [26]). The pFe and pAl plots for L1 at
the same conditions (speciation data from Clarke and Martell
[
27] and from Clevette et al. [28]) are reported for comparison.
Fig. 5 demonstrates that HPs are weaker Fe(III) and Al(III) chelators
than L1. This means that a similar metal depletion efficiency can be
obtained if higher doses of HP are administered in chelation ther-
a 24
b 16
DT712
2
2
0
8
6
4
2
0
DT71201
1
4
L1
DQ5
2
L1
DQ716
DQ715
DQ6
pFe
pAl
1
1
1
1
1
12
10
8
DT8126
DT712
DT726
DT726
DT8126
6
4
5
6
7
8
9
3
4
5
6
7
8
pH
pH
Fig. 5. pFe (a) and pAl (b) plots for several ligands (see text); Cmetal = 10 6 m, Cligand = 5 ꢀ 10ꢁ5 m. pFe plots stop at the theoretical starting pH for hydroxide precipitation.
ꢁ

A. Dean et al. / Polyhedron 67 (2014) 520–528
527
(
2–3 h) [32], so that L1 can only partially remove Fe(III) from Tf
in vivo.
For any chelator, the kinetics of Fe(III) removal from Tf follows a
more DT712 molecules were detected as well (Table 5). The inten-
sity of these peaks increased by aging the solution, reaching an
approximately constant level after 20 days. Ternary peaks are then
stable (spectra were collected for over 60 days), indicating that
their formation is thermodynamically favoured, but kinetically
hindered. If a dynamic solution like blood is considered, the binary
metal/DT712 complexes are expected to be the only significant
species forming in the presence of Fe(III) or Al(III), CA, and DT712.
two-step mechanism, as the two binding sites of the protein act
independently. The kinetic law explaining this process is therefore
a biexponential one [33]. However, the two kinetic constants are
generally very similar, so that the biexponential law can be simpli-
fied to a monoexponential one [34]:
Þeꢁ þ A
kt
ð1Þ
A
t
¼ ðA
0
ꢁ A
1
1
3
.6. Voltammetric study
where A is the measured property at times 0, t, and 1, and k is the
pseudo-first order kinetic constant. The k value is generally ob-
tained by UV–Vis absorbance measurements versus time at a fixed
wavelength, by fitting the experimental values with Eq. (1) [31]. The
same technique was used in this work to determine the k values for
DT712, DQ715, and L1. Experimental results are reported in Table 3.
These k values indicate that the two investigated HPs, and espe-
cially DT712, display significantly better kinetic properties than L1,
as DT712 and DQ715 remove Fe(III) from Tf ca. 2.5 and 1.5 times
faster than L1, respectively. This property appears to be very posi-
tive. However, it cannot be completely judged because the metab-
olization half times of all HPs are presently unknown. The only HP
which deserved biochemical interest up to now is pyridoxic acid
The voltammetric study was performed on DT712, DT8126 and
on their Fe(III) complexes. The cyclic voltammetries for the two
systems were similar. The free ligands showed no redox activity
in the investigated e.m.f. range. Fig. 6 shows the results for
Fe(III)/DT8126 complexes. The voltammetric curves point out the
presence of three redox couples. The redox couple marked I and
0
I refers to the reduction of free Fe(III) to free Fe(II) [6,37]. The
intensity of the cathodic peak decreases by increasing the ligand
concentration. In the same conditions the corresponding anodic peak
remains practically unchanged, indicating that the Fe(II)/ligand
complexes formed upon Fe(III) reduction are labile, and that
the dissociation kinetic constant is fast with respect to the experi-
ment time. The redox couples marked II and II refer to the reduc-
tion of at least two Fe(III)/ligand complexes. Simulations were
performed on the voltammetric curves using the DIGISIM routine
(
5-hydroxymethyl-3-hydroxy-2-methyl-4-pyridinecarboxylic
0
acid), because is a vitamine B6 metabolite. For this HP, no metab-
olization in vivo is reported in the literature to occur. Metaboliza-
tion studies for selected HPs are in progress in our laboratories.
[
25], assuming the speciation indicated in Tables 1 and 2. Accord-
0
ing to the simulation results, the redox couples II and II are
Fe(III)L /Fe(II)L and Fe(III)LHꢁ1 /Fe(II)LHꢁ1. The standard reduc-
tion potentials (E ) obtained by DIGISIM for all redox processes
are shown in Table 6. The simulations gave other parameters in-
3.5. Mass spectrometric study
2+
+
+
0
Besides Tf, there are several low molecular weight metal bind-
ers in blood, such as citric acid (CA), phosphate, acetate, adenosine
triphosphate, etc. These ligands can bind Fe(III) and Al(III) at metal
ion overload conditions, when Tf is completely saturated. It is
desirable that chelators used to achieve negative Fe and Al balance
can remove the metals from these complexes. As regards Fe(III),
May et al. [35] predicted the dominant metal binder in plasma to
be CA, on the basis of the relative concentrations of potential li-
gands. Therefore, CA was considered to study whether low molec-
ular weight metal binders in blood can compete with DT712 in the
metal complexation. The ternary Fe(III)/DT712/CA and Al(III)/
DT712/CA systems were studied by electrospray mass spectrome-
try (ESI-MS). The binary Fe(III)/DT712 and Al(III)/DT712 systems
were studied as well to verify the speciation picture obtained by
volved in the voltammetric study, i.e., the electron transfer coeffi-
cients of the species (
electron transfer (ksh), and the kinetic constants for the reactions
of complex formation (k ). The simulated values for these parame-
a), the heterogeneous rate constants of the
f
ters are very similar to those obtained previously for the other
ꢁ2 ꢁ1
investigated HPs [6,7,12,37], e.g.
a
ꢂ 0.6,
k
sh ꢂ 10
s , and
5
ꢁ1 ꢁ1
k
f
> 10 M
s
.
0
These E results suggest that Fe(III) is not able to generate redox
cycling at physiologic environment in the presence of DT712 or
DT8126, as the reduction standard redox potentials of Fe(III)/ligand
complexes predominating at physiologic pH are lower than the
limit potential of 0.53 V (versus saturated Ag/AgCl/KCl) indicated
f
by Merkofer et al. [38]. As well, the very large values of k indicate
1
potentiometry, UV–Vis and H NMR. ESI-MS represents a powerful
that both ligands react very quickly with Fe(III) to form the com-
plexes. Therefore, no kinetic hindrance is expected to occur at
physiologic conditions during a Fe chelation therapy treatment
by DT712 or DT8126.
technique in speciation analysis, as it is able to give direct evidence
of the stoichiometry of the complexes formed in solution [6,36].
Tables 4 and 5 report the assignments of the peaks detected in
ESI-MS spectra collected for binary and ternary systems, respec-
tively. Molecular mass, isotopic distributions (especially useful
for Fe-containing peaks) and MS/MS spectra, allowed unambigu-
ous peak assignments.
5.0e-7
I'
II'
0.0
The ESI-MS analysis for binary systems confirms the results ob-
tained by equilibrium techniques, because the same Fe(III) and
Al(III) complexes (ML, ML and ML ) were detected in the gas
2 3
i (μA)
-5.0e-7
-1.0e-6
-1.5e-6
phase (ESI-MS is not able to discriminate differently protonated
species [36]). Several fragmentation or adduct products were ob-
C = 0
L
3
+
served, as usual. In some cases, the ESI-MS reduction of Fe to
C = 1.22 mM
L
II
I
2
+
Fe was detected, too [36].
-
0.2
0.0
0.2
0.4
0.6
0.8
The ESI-MS analysis for ternary systems showed mainly the
peaks of the corresponding binary metal/DT712 complexes. No
peaks related to metal/CA complexes were observed. This is the
most important result, as it stems that DT712 can successfully
compete with CA in the metal complexation. Minor peaks related
to mixed complexes containing one metal ion, one CA and one or
E (V) (vs. saturated Ag/AgCl/KCl)
Fig. 6. Cyclic voltammetries of a 0.32 mM Fe(III) solution added with aliquots of
DT8126 (in 3.3ꢀ of ligand in alumina). Ligand concentration: 0, 0.04, 0.09, 0.14,
.19, 0.24, 0.32, 0.45, 0.70, and 1.22 mM. pH = 2.42 with HCl. Redox peaks I and II,
and corresponding back scan peaks I and II , are highlighted (see text).
0
0
0

5
28
A. Dean et al. / Polyhedron 67 (2014) 520–528
Table 6
References
0
E
values (V) versus saturated Ag/AgCl/KCl obtained for Fe(III)/DT712 and Fe/III)/
DT8126 complexes. The uncertainty of the values is 0.01 V.
[
1] G. Crisponi, M. Remelli, Coord. Chem. Rev. 252 (2008) 1225.
E0 (V)
L = DT712
E0 (V)
L = DT8126
[2] G. Crisponi, V.M. Nurchi, V. Bertolasi, M. Remelli, G. Faa, Coord. Chem. Rev. 256
Reactions
(
2012) 89.
[
3] R. Galanello, S. Campus, R. Origa, Expert Opin. Drug Metab. Toxicol. 8 (2012)
123.
ꢁ
Fe(III) + e = Fe(II)
0.52
0.19
<0.1
<0.1
0.52
0.20
<0.15
<ꢁ0.3
ꢁ
[4] G. Crisponi, A. Dean, V. Di Marco, J.I. Lachowicz, V.M. Nurchi, M. Remelli, A.
Tapparo, Anal. Bioanal. Chem. 405 (2013) 585.
[5] Z.D. Liu, R.C. Hider, Med. Res. Rev. 22 (2002) 26.
Fe(III)L + e = Fe(II)L
Fe(III)LHꢁ1 + e = Fe(II)LHꢁ1
ꢁ
ꢁ
Fe(III)L
2
+ e = Fe(II)L
2
[
6] A. Dean, É. Sija, É. Zsigó, M.G. Ferlin, D. Marton, V. Gandin, C. Marzano, D.
Badocco, P. Pastore, A. Venzo, R. Bertani, T. Kiss, V. Di Marco, Eur. J. Inorg.
Chem. (2013) 1310.
4
. Conclusions
DT712, DT726, DT71201, and DT8126 were synthesized, and
[
7] A. Dean, M.G. Ferlin, P. Brun, I. Castagliuolo, R.A. Yokel, D. Badocco, P. Pastore,
A. Venzo, G.G. Bombi, V.B. Di Marco, Dalton Trans. (2009) 1815.
[8] V.B. Di Marco, R.A. Yokel, M.G. Ferlin, A. Tapparo, G.G. Bombi, Eur. J. Inorg.
Chem. (2002) 2648.
their complex formation with Fe(III) and Al(III) was studied by
means of a multimethodological approach. Number, stoichiometry,
and stability constants of metal–ligand complexes were obtained.
The role of the methyl substitution in the Fe(III)/DTs and Al(III)/
DTs complex formation strength appears very similar to that high-
lighted for the DQs. The methyl groups in 2 and 6 increase Fe(III)
and Al(III) complex stability, and the 1-methyl substitution in-
creases the Fe(III) complex stability much more than the Al(III)
one. An unexpected result was that the trimethylated DT8126
was less effective than the other dimethylated DTs, indicating that
the over-methylation of the HP pyridinic ring is not a successful
strategy for the synthesis of stronger metal chelators. The thermo-
dynamic results suggest that DTs are better Fe chelators than DQs,
and that the latter are better Al chelators than the former.
DT712 is the most promising HP considered so far for Fe chela-
tion therapy, as it forms the strongest Fe(III) complexes. This com-
pound was therefore further investigated. DT712 resulted to have
the best kinetic properties in the extraction of Fe(III) from Tf, being
ca. 2.5 times faster than L1. DT712 is able to effectively compete
with CA in the metal complexation. The standard reduction poten-
tials of Fe(III)/DT712 complexes suggest the absence of redox cy-
cling at in vivo conditions, which is a prerequisite to ensure
absence of toxic effects induced by the chelator. Finally, kinetic
constants indicate that Fe(III) complex formation should not be
kinetically limited. Also other data reported in the literature for
DT712 (no cytotoxicity, antioxidant activity) [13] are very positive
for the employment of this compound as drug. On the basis of the
present results, DT712 is a promising candidate for Fe chelation
therapy. It can and should therefore be submitted to in vitro and
in vivo experiments in order to evaluate its behaviour in the rele-
vant physiological environment.
[
9] V.B. Di Marco, R.A. Yokel, H. Li, A. Tapparo, G.G. Bombi, Inorg. Chim. Acta 357
2004) 3753.
(
[10] V.B. Di Marco, A. Dean, M.G. Ferlin, R.A. Yokel, H. Li, A. Venzo, G.G. Bombi, Eur.
J. Inorg. Chem. (2006) 1284.
11] V.B. Di Marco, A. Dean, R.A. Yokel, H. Li, G.G. Bombi, Polyhedron 26 (2007)
227.
[12] A. Dean, M.G. Ferlin, P. Brun, I. Castagliuolo, D. Badocco, P. Pastore, A. Venzo,
G.G. Bombi, V.B. Di Marco, Dalton Trans. (2008) 1689.
13] P. Brun, A. Dean, V. Di Marco, P. Surajit, I. Castagliuolo, D. Carta, M.G. Ferlin,
Eur. J. Med. Chem. 62 (2013) 486.
14] V.B. Di Marco, A. Tapparo, G.G. Bombi, Anal. Chim. 89 (1999) 397.
[15] V.B. Di Marco, A. Tapparo, G.G. Bombi, Anal. Chim. 91 (2001) 595.
16] G.J. Janz, Silver – Silver Halide Electrodes, in Reference Electrodes: Theory and
Practice, in: D.J.G. Ives, G.J. Janz (Eds.), Academic Press, New York – London,
[
3
[
[
[
1961.
[17] G. Biedermann, L.G. Sillén, Ark. Kemi 5 (1953) 425.
[
[
[
18] V.B. Di Marco, A. Tapparo, G.G. Bombi, Anal. Chim. 89 (1999) 535.
19] V.B. Di Marco, (Ph.D. thesis), University of Padova, 1998.
20] L.G. Sillén, Acta Chem. Scand. 18 (1964) 1085.
[21] W.H. Press, B.P. Flannery, S.A. Teukolsky, W.T. Vetterling, Numerical Recipes:
The Art of Scientific Computing, Cambridge University Press, Cambridge, 1986.
22] C.M. Flynn Jr., Chem. Rev. 84 (1984) 31.
23] L.O. Öhman, Inorg. Chem. 27 (1988) 2565.
[24] A.K. Covington, M. Paabo, R.A. Robinson, R.G. Bates, Anal. Chem. 40 (1968) 700.
25] M. Rudolph, D.P. Reddy, S.W. Feldberg, Anal. Chem. 66 (1994) 589A.
26] T. Kiss, E. Farkas, Perspect. Bioinorg. Chem. 3 (1996) 199.
27] E.T. Clarke, A.E. Martell, Inorg. Chim. Acta 191 (1992) 57.
[28] D.J. Clevette, W.O. Nelson, A. Nordin, C. Orvig, S. Sjöberg, Inorg. Chem. 28
(1989) 2079.
29] G. Crisponi, V.M. Nurchi, M. Crespo-Alonso, L. Toso, Curr. Med. Chem. 19
2012) 2794.
30] G.J. Kontoghiorghes, Hemoglobin 30 (2006) 183.
[
[
[
[
[
[
(
[
[31] I. Turcot, A. Stintzi, J. Xu, K.N. Raymond, J. Biol. Inorg. Chem. 5 (2000) 634.
[
[
[
32] Australian Government Department of Health and Ageing, Therapeutic Goods
Administration, Australian Public Assessment Report for Deferiprone, 2011.
33] I. Bertini, J. Hirose, C. Luchinat, L. Messori, M. Piccioli, A. Scozzafava, Inorg.
Chem. 27 (1988) 2405.
34] W.R. Harris, C.E. Brook, C.D. Spilling, S. Elleppan, W. Peng, M. Xin, J. Van Wyk, J.
Inorg. Biochem. 98 (2004) 1824.
35] P.M. May, P.W. Linder, D.R. Williams, J. Chem. Soc., Dalton Trans. (1977) 588.
36] V.B. Di Marco, G.G. Bombi, Mass Spectrom. Rev. 25 (2006) 347.
37] D. Badocco, M. Marcon, A. Mondin, A. Dean, V.B. Di Marco, P. Pastore, Dalton
Trans. (2009) 2415.
[
[
[
Acknowledgements
[
38] M. Merkofer, R. Kissner, R.C. Hider, U.T. Brunk, W.H. Koppenol, Chem. Res.
Toxicol. 19 (2006) 1263.
This work was supported by University of Padova, ‘‘Progetto
Ateneo 2008, CPDA083904/08, 54000 €’’. We thank Daniele Lena,
Fabio Todesco, Marta Da Pian, and Elisa Agnoletto (University of
Padova) for their excellent work, and Dr. Cristina Tubaro for the ki-
netic UV–Vis measurements.