Iron and transition metal transport into erythrocytes mediated by nifedipine degradation products and related compounds

Article abstract of DOI:10.1016/S0006-2952(03)00045-5

The aim of this investigation was to determine the mechanism of action of the nitrosophenylpyridine derivative of nifedipine ("nitrosonifedipine", NN) on Fe(II) transport into erythrocytes. Nifedipine is rapidly degraded to NN by daylight. We used rabbit erythrocytes, NN, and several chemically related substances, and examined their effects on the transfer of iron and other transition metals (cadmium, cobalt, manganese, nickel, zinc) into and out of the cells. NN mediated the transfer of iron and zinc but not the other metals into the cell cytosol. The transfer of Fe(II) was not affected by changes in cell membrane potential and could not be ascribed to free radical production. Two hydroxamic acid compounds chemically related to NN also stimulated iron and zinc uptake, but no evidence was obtained for cell-induced transformation of NN to them. In vivo, NN is probably converted to a lactam derivative. This compound was found to have no effect on iron uptake by the cells. It is concluded that NN has a relatively high specificity for the transport of iron compared with other transition metals, and small changes in its structure markedly affect this action. Also, because the lactam to which NN is converted in vivo is inactive, it is unlikely that nifedipine will affect iron metabolism after therapeutic administration.

Full text of DOI:10.1016/S0006-2952(03)00045-5

Biochemical Pharmacology 65 (2003) 1215–1226
Iron and transition metal transport into erythrocytes mediated by
nifedipine degradation products and related compounds
Donna L. Savignia, Dieter Wegeb, Garth S. Cliffb,
Mark L.H. Meestersb, Evan H. Morgana,*
aDepartment of Physiology, The University of Western Australia, 35 Stirling Highway, Crawley, Western Australia 6009, Australia
bDepartment of Chemistry, The University of Western Australia, 35 Stirling Highway, Crawley, Western Australia 6009, Australia
Received 25 February 2002; accepted 19 August 2002
Abstract
The aim of this investigation was to determine the mechanism of action of the nitrosophenylpyridine derivative of nifedipine
(‘‘nitrosonifedipine’’, NN) on Fe(II) transport into erythrocytes. Nifedipine is rapidly degraded to NN by daylight. We used rabbit
erythrocytes, NN, and several chemically related substances, and examined their effects on the transfer of iron and other transition metals
(cadmium, cobalt, manganese, nickel, zinc) into and out of the cells. NN mediated the transfer of iron and zinc but not the other metals into
the cell cytosol. The transfer of Fe(II) was not affected by changes in cell membrane potential and could not be ascribed to free radical
production. Two hydroxamic acid compounds chemically related to NN also stimulated iron and zinc uptake, but no evidence was
obtained for cell-induced transformation of NN to them. In vivo, NN is probably converted to a lactam derivative. This compound was
found to have no effect on iron uptake by the cells. It is concluded that NN has a relatively high specificity for the transport of iron
compared with other transition metals, and small changes in its structure markedly affect this action. Also, because the lactam to which NN
is converted in vivo is inactive, it is unlikely that nifedipine will affect iron metabolism after therapeutic administration.
# 2003 Elsevier Science Inc. All rights reserved.
Keywords: Iron transport; Erythrocytes; Nitrosophenylpyridine derivative of nifedipine; Transition metals
1. Introduction
(‘‘nitrosonifedipine’’). It was suggested that this substance
is able to act as an iron ionophore [1].
Nifedipine (1) is a Ca2þ channel blocking drug that is
widely used in clinical medicine. It is light-sensitive and in
the presence of daylight is readily converted into a series of
products that are no longer active against Ca2þ channels. In
a previous study, we investigated whether iron can enter
erythroid cells via Ca2þ channels by the use of channel
antagonists, including nifedipine [1]. They had little effect
on iron transport. However, after exposure to light, the
photodegradation products of nifedipine were found to
strongly stimulate ferrous iron (Fe(II)) transfer into and
out of the cells. The active constituent of these products
was isolated by chromatographic techniques and shown to
be the nitrosophenylpyridine derivative of nifedipine
The high capacity of nitrosonifedipine (2) to mediate
Fe(II) transport across the cell membrane of erythroid cells
is quite unusual and in magnitude greatly exceeds that of
known iron transport processes [1]. Under normal circum-
stances, mature erythroid cells (erythrocytes) have no capa-
city to transport iron, and their cell membrane is virtually
impermeable to the metal. It is important to determine the
mechanism of action of nitrosonifedipine since this infor-
mation will help in the elucidation of membrane perme-
ability and the properties of substances that can mediate iron
transport across cell membranes. It may also help (a) to
explain the side-effects of nifedipine that are unrelated to its
action as a Ca2þ channel antagonist, and (b) in the design of
substances that can be used to release iron from cells in
clinical conditions associated with iron overload, such as
thalassaemia and genetic hemochromatosis.
* Corresponding author. Tel.: þ61-8-9380-3320; fax: þ61-8-9380-1025.
E-mail address: ehmorgan@cyllene.uwa.edu.au (E.H. Morgan).
Abbreviations: BHA, 2[3]-t-butyl-4-hydroxyanisole; CP20, 1,2-di-
methyl-3-hydroxypyridin-4-one; DFO, desferrioxamine; PCV, packed cell
volume; PIH, pyridoxal isonicotinoyl hydrazone.
The present investigation was undertaken to investigate
the mechanism of action of nitrosonifedipine on iron
0006-2952/03/$ – see front matter # 2003 Elsevier Science Inc. All rights reserved.
doi:10.1016/S0006-2952(03)00045-5

1216
D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
Fig. 1. Structures of nifedipine and other substances used in the investigation. The figure also shows the reactions used to synthesize some of the compounds
and the numbers used to designate them in the text.
transport into and out of erythrocytes. We used rabbit cells
and several substances chemically related to nitrosonifedi-
pine as well as nitrosonifedipine, and examined the action of
these substances on the transfer of Fe(II), Fe(III), and other
transition metals into and out of the cells. In addition, some
comparisons were made with two iron chelators, PIH and
CP20, which have been studied widely and have been used
in the treatment of iron overload disorders. The structures of
the compounds used in this investigation and their numer-
ical designations (compounds 1 to 13) are given in Fig. 1.
2. Materials and methods
2.1. Materials
The radioisotopes 57CoCl2, 63NiCl2, 65ZnCl2, and
109CdCl2 were purchased from Amersham Australia, and
59FeCl3 and 54MnCl2 were purchased from Dupont. PIH
was synthesised as described by Ponka et al. [2], nitroso-
benzene was synthesised as described by Coleman et al.
[3], and CP20 was provided by Dr. E. Baker. The other

D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
1217
biochemical reagents were purchased from the Sigma
Chemical Co.
oxide to give the title lactam 3 as colourless needles (16 mg,
24%), m.p. 275–2768 (lit. [6] 257–2638 (dec), lit. [7]
270–2728). H NMR (200 MHz, CDCl3) d 2.70, s, 3H,
1
2.2. Synthesis and characterisation of nifedipine (1),
‘‘nitrosonifedipine’’ (2), and compounds 3, 4, 5, 9, and 10
CH3; 3.25, s, 3H, CH3; 4.02, s, 3H, OCH3; 7.16–7.38,
m, 2H, aromatic; 7.58, dd, J 8.2, 8.2 Hz, 1H, aromatic;
7.85, d, J8.4 Hz, 1H, aromatic. Thesedata areverysimilarto
those reported [6,7] for 3 prepared by other methods.
2.2.1. Dimethyl 1,4-dihydro-2,6-dimethyl-4-(2-
nitrophenyl)pyridine-3,5-dicarboxylate (nifedipine) (1)
The following procedure is based on that described by
Phillips [4]. A solution of 2-nitrobenzaldehyde (10.2 g,
68 mmol), methyl acetoacetate (21.1 g, 181 mmol), and
concentrated ammonium hydroxide solution (10 mL of
25%, w/w) in ethanol (40 mL) was heated under reflux
in a steam bath for 2 hr. The solution was diluted with
water (50 mL) and extracted with dichloromethane
(2 Â 50 mL). The extract was washed with water, dried
(MgSO4), and evaporated, and the residue was recrystal-
lized from dichloromethane/petrol to give the title com-
pound as yellow crystals (18.5 g, 79%), m.p. 173–1748 (lit.
[5] 172–1738). 1H NMR (200 MHz, CDCl3) d 2.28, s, 6H,
2.2.4. Dimethyl 2,6-dimethyl-4-(2-nitrophenyl)pyridine-
3,5-dicarboxylate (4)
This was prepared by the general procedure of Phillips
[4]. A mixture of nifedipine (21.3 g) and 4 M nitric acid
(150 mL) was heated on a steam bath for 2 hr. The cooled
mixture was made alkaline with aqueous sodium hydro-
xide, extracted with ether (3 Â 150 mL), and the extract
washed with water, dried (MgSO4), and evaporated. The
solid residue (17.8 g) was recrystallized from dichloro-
methane/petrol to give the title compound as very pale
yellow rhombs (9.97 g, 47%), m.p. 103–1048 (lit. [8] 103–
1058). 1H NMR (200 MHz, CDCl3) d 2.63, s, 6H, 2 Â CH3;
0
0
0
0
0 0
3.48, s, 6H, 2 Â OCH3; 7.17, ddd, J6 5 6.6, J6 4 2.3, J6 3
2 Â CH3; 3.54, s, 6H, 2 Â OCH3; 5.68, s, 1H, H4; 6.17, s,
0.1 Hz, 1H, H60; 7.53, ddd, J4 5 9.4, J4 3 7.5, J4 6 2.3 Hz,
0
0
0
0
0
0 0
0
0
0
0
0 0
1H, NH; 7.21, ddd, J4 5 7.1, J4 3 8.2, J4 6 1.9 Hz, 1H, H4 ;
0 0
H30.
1H, H40; 7.60, ddd, J5 4 9.4, J5 3 1.5, J5 6 6.6 Hz, 1H, H5 ;
0
7.44, m, 2H, H50, H60; 7.64, dd, J3 4 8.2, J3 5 1.2 Hz, 1H,
0
0
0
0
0 0
0
0
8.18, ddd, J3 4 7.5, J3 5 1.5, J3 6 2.3 Hz, 1H, H30.
0
0
0
0
0 0
2.2.5. Methyl 6-hydroxy-2,4-dimethyl-5-oxo-5,6-
2.2.2. Dimethyl 2,6-dimethyl-4-(2-nitrosophenyl)-
dihydrobenzo[c][2,7]naphthyridine-1-carboxylate (5)
The following procedure is based on the general method
of Kim [9]. A solution of dimethyl 2,6-dimethyl-4-
(2-nitrophenyl)pyridine-3,5-dicarboxylate (4) (222 mg,
0.67 mmol) in ethyl acetate (15 mL) was stirred in an
atmosphere of hydrogen over 5% palladium on carbon
catalyst (50 mg) until 26 mL (1.2 mmol) of hydrogen had
been absorbed. The precipitated product was dissolved by
the addition of ethyl acetate, and the catalyst was removed
by filtration. The filtrate was evaporated, and the residue
was subjected to vacuum liquid chromatography on silica
gel. Elution with 40% ethyl acetate in light petroleum
followed by recrystallization from dichloromethane/petrol
gave the title compound as colourless needles (116 mg,
60%), m.p. 190–1958 (lit. [7] double m.p. 2108, 2158).
Mass spectrum m/z 298 (Mþ, 64%), 281 (41), 256 (100),
213 (40), 185 (33), 137 (33), 129 (76), 111 (37), 110 (42).
1H NMR (300 MHz, DMSO-D6) d 2.60, s, 3H, CH3; 3.04,
s, 3H, CH3; 3.96, s, 3H, OCH3; 7.31–7.80, m, 4H, aro-
matic; 11.42, s, 1H, OH. This spectrum is similar to that
reported previously [6]. 13C NMR (75.5 MHz, DMSO-D6)
d 22.6, CH3; 27.2, CH3; 53.1, OCH3; 113.2, CH; 113.6, C;
117.0, C; 120.0, C; 122.5, CH; 125.2, CH; 136.9, C; 138.4,
C; 155.4, C; 155.3, C; 161.8, CO; 169.9, CO.
pyridine-3,5-dicarboxylate (‘‘nitrosonifedipine’’) (2)
A solution of nifedipine (1) (4.40 g) in dichloromethane
(800 mL) was purged with argon and then irradiated for
4 hr in an Oliphant photochemical chamber reactor with a
bank of 16 fluorescent tubes emitting at 350 nm. The
solvent was evaporated and the residual solid was recrys-
tallized from dichloromethane/petrol to give the title nitroso
compound as green needles (2.61 g, 63%), m.p. 948 (lit. [6]
938). Mass spectrum m/z 328 (Mþ, 18%), 284 (22), 269
(100), 267 (33), 253 (40), 252 (28), 237 (24), 193 (28), 152
(28). 1H NMR (300 MHz, CDCl3) d 2.60, s, 6H, 2 Â CH3;
0
0
0
0
0 0
3.32, s, 6H, 2 Â OCH3; 6.47, ddd, J6 5 8.0, J6 4 1.1, J6 3
0.4 Hz, 1H, H60; 7.37, ddd, J5 6 8.0, J5 4 7.3, J5 3 1.3 Hz,
0
0
0
0
0 0
1H, H50; 7.46, ddd, J4 3 7.6, J4 5 7.3, J4 6 1.1 Hz, 1H, H40;
0
0
0
0
0 0
0
7.67, ddd, J3 4 7.6, J3 5 1.3, J3 6 0.4 Hz, 1H, H3 . 13C NMR
(75.5 MHz, CDCl3) d 23.3, 2 Â CH3; 51.9, 2 Â OCH3;
107.6, CH; 127.1, C; 128.8, CH; 130.5, CH; 135.0, CH;
139.9, C; 144.3, C; 155.3, C; 161.5, C; 167.5, 2 Â CO.
Electronic spectrum (dichloromethane) lmax (log e) 230
(4.20), 283 (3.95), 312 nm (3.82).
0
0
0
0
0 0
2.2.3. Methyl 2,4-dimethyl-5-oxo-5,6-
dihydrobenzo[c][2,7]naphthyridine-1-carboxylate (3)
Solutions of nitrosonifedipine (2) (77 mg, 0.24 mmol) in
ethanol (70 mL) and glutathione (700 mg, 2.4 mmol) in
water (100 mL) were mixed, and the resulting solution was
heated at 608 for 30 min. The solution was concentrated on
a rotary evaporator to about half its original volume and
cooled in ice. The resulting crystals were collected by
filtration and dried in a vacuum desiccator over phosphoric
2.2.6. Stability of the hydroxamic acid 5 towards
glutathione
Solutions of hydroxamic acid 5 (30 mg, 0.100 mmol) in
ethanol (30 mL) and glutathione (316 mg, 1.03 mmol) in
water (40 mL) were mixed, and the resulting solution was

1218
D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
1
heated at 608 for 30 min. TLC analysis showed the presence
of starting material; the characteristic blue fluorescence
observed for the lactam 3 under 254 nm radiation was not
detected. The solution was concentrated on a rotary eva-
porator to about half its original volume and cooled in ice.
The resulting crystals were collected by filtration and dried
in a vacuum desiccator over phosphoric oxide to give
unchanged 5 as colourless needles (27 mg), m.p. 192–1948.
needles, m.p. 145–1468. H NMR (300 MHz, CDCl3) d
5.30, s, 2H, OCH2; 7.29–7.81, m, 10H, aromatic; 8.25, d, J
8.0 Hz, 1H, aromatic; 8.27, d, J 8.0 Hz, 1H, aromatic; 8.60,
dd, J 8.0, 1.2 Hz, 1H, aromatic. 13C NMR (75.5 MHz,
CDCl3) d 77.0, OCH2; 112.9, CH; 118.3, C; 121.9, CH;
123.0, CH; 123.1, CH; 126.3, C; 128.0, CH; 128.6, CH;
129.1, CH; 129.8, CH; 132.5, CH; 133.0, C; 133.9, C; 136.3,
C; 157.5, CO.
2.2.7. 5-Benzyloxy-6(5H)-phenanthridone (9)
2.2.8. 5-Hydroxy-6(5H)-phenanthridone (10)
solution of 5-benzyloxy-6(5H)-phenanthridone
The following is based on the general procedure of
Glover et al. [10]. A solution of biphenylene-2-carboxylic
acid (3.00 g) in anhydrous benzene (5 mL) was added
dropwise to gently refluxing freshly distilled thionyl chlor-
ide (5.00 g). After the addition was complete, the mixture
was heated under reflux for 30 min before the benzene and
excess thionyl chloride were evaporated. The residue was
distilled (Kugelrohr) at ꢀ1508/0.5 Torr to give biphenylene-
2-carbonyl chloride as a colourless oil (2.80 g, 79%). The
foregoing acid chloride (2.33 g, 10 mmol) was added drop-
wise to an ice-cold stirred suspension of O-benzylhydrox-
ylamine hydrochloride [11] (1.60 g, 10 mmol) in a mixture
of anhydrous dichloromethane (25 mL) and triethylamine
(2.00 g, 20 mmol). The mixture was stirred at room tem-
perature overnight, diluted with dichloromethane (75 mL),
and then washed successively with dilute hydrochloric acid,
10% sodium carbonate solution, and brine. The dichloro-
methane layer was dried (MgSO4) and evaporated. The
residue was subjected to radial chromatography eluting with
petrol containing increasing amounts of ethyl acetate. Eva-
poration of the main band afforded N-benzyloxybiphenyl-
2-carboxamide as a solid (1.60 g, 53%). 1H NMR
(300 MHz, CDCl3) d 4.72, br s, 2H, OCH2; 7.10–7.26, br
m, 2H, aromatic; 7.25–7.32, m, 3H, aromatic; 7.34–7.45, m,
7H, aromatic; 7.50, ddd, J 7.5, 7.5, 1.4 Hz, 1H, aromatic;
7.60, br d, J 7.5 Hz, 1H, aromatic. 13C NMR (75.5 MHz,
CDCl3) d 77.9, OCH2; 127.5, CH; 128.0, CH; 128.5, CH;
128.6, CH; 128.7, CH; 129.0, CH; 130.1, CH; 130.6, CH;
132.3, C; 135.0, C; 139.6, C; 139.9, C; 167.3, CO.
A solution of N-benzyloxybiphenyl-2-carboxamide
(1.0 g, 3.0 mmol) and t-butyl hypochlorite (1.0 g,
10 mmol) in anhydrous benzene (15 mL) was stirred over-
night in the dark. The benzene and excess t-butyl hypo-
chlorite were evaporated under vacuum to give N-chloro-
N-benzyloxybiphenyl-2-carboxamide as a colourless gum
(0.91 g). A portion of the foregoing compound (0.35 g,
1.0 mmol) in anhydrous benzene (25 mL) was treated with
silver tetrafluoroborate (0.25 g, 1.3 mmol), and the mixture
was stirred in the dark at room temperature overnight. The
precipitated silver chloride was removed by filtration and
was washed with chloroform. The combined organic filtrate
was washed with water, dried (MgSO4), evaporated, and the
residue submitted to radial chromatography. Elution with
10% ethyl acetate in petroleum gave 5-benzyloxy-6(5H)-
phenanthridone (9) as a crystalline solid (0.094 g, 30%),
which crystallized from dichloromethane/petrol as colorless
A
(54 mg) in methanol (10 mL) was stirred in an atmosphere
of hydrogen over 10% palladium on carbon (25 mg) until
hydrogen uptake ceased. The methanol was evaporated, the
residue was extracted with hot chloroform and filtered, and
the filtrate was concentrated until crystallization began. 5-
Hydroxy-6(5H)-phenanthridone was obtained as colorless
rhombs (35 mg, 92%), m.p. 256–2578 (lit. [12] 257–2588).
1H NMR (500 MHz, DMSO-D6) d 7.36, ddd, J 8.0, 8.0,
0.9 Hz, 1H, aromatic; 7.63, ddd, J 7.8, 7.8, 1.0 Hz, 1H,
aromatic; 7.67, dd, J 7.5, 7.5 Hz, 1H, aromatic; 7.77, d, J
8.0 Hz, 1H, aromatic; 7.85, ddd, J 7.7, 7.7, 1.2 Hz, 1H,
aromatic; 8.38, dd, J 8.0, 1.3 Hz, 1H, aromatic; 8.49, d, J
7.3 Hz; 8.55, d, J 8.2 Hz, 1H, aromatic; 11.37, br s, 1H,
OH. 13C NMR (125.75 MHz, DMSO-D6) d 113.1, CH;
117.3, C; 122.7, CH; 122.8, CH; 123.4, CH; 125.4, C;
127.5, CH; 128.2, CH; 130.0, CH; 132.3, C; 132.5, CH;
136.9, C; 156.7, CO.
2.3. Cells
Blood was obtained from adult rabbits by bleeding from
a marginal ear vein, using heparin as an anticoagulant. The
blood cells were washed four times with 10–15 vol. of ice-
cold 0.15 M NaCl and centrifuged at 2000 g for 10 min at
48. The supernatant and buffy coat (leucocytes and plate-
lets) were removed after each centrifugation. For metal
uptake experiments the cells were then suspended in
0.15 M NaCl at a PCV of 25–30%. For metal efflux
experiments, the cells were washed one more time in
0.15 M KCl and then suspended in 0.15 M KCl.
2.4. Metal solutions
The solutions of the metal salts were prepared imme-
diately before use by mixing samples of the radioactive
isotopes with divalent non-radioactive chloride salts of Fe,
Cd, Co, Mn, or Zn, or with FeCl3, and diluting with 0.27 M
sucrose to give solutions with 20 times the desired final
concentrations of the metals.
2.5. Incubation procedures
2.5.1. Metal uptake
Samples of the cell suspensions (usually 60 mL) were
incubated at 378 in an oscillating, temperature-controlled

D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
1219
water bath in 2 mL solution containing the radioactive
2.7. Statistics
metals. Most incubations were performed for 10 min in
0.15 M NaCl–0.01 M Tris–HEPES, pH 7.4. Except
where indicated to the contrary, the metal concentration
was 20 mM. Solutions of the nifedipine derivatives and
the other compounds tested were prepared at 200 times
the desired final concentrations in DMSO. Aliquots of
these solutions (10 mL) or DMSO for controls were
added to the cell suspension and incubated with them
for 10 min at room temperature (except where indicated
below) before addition of the radioactive metals
and incubation at 378. After incubation the cells were
washed three times with ice-cold 0.15 M NaCl, hemo-
lysed with 1.5 mL of 0.01 M Tris–HEPES, pH 7.4,
centrifuged to separate cytosolic and membrane frac-
tions, and counted for radioactivity as described pre-
viously [13]. These experiments were performed with
mature erythrocytes, which do not contain nuclei or
other intracellular organelles. Hence, the centrifugation
step separates the cells into two fractions that contain
only cytosol and cell plasma membranes (referred to
as membranes).
The experiments were performed three or four times.
Results are presented in the figures as means Æ SEM.
Significance of differences between the means (P < 0:05)
was assessed using Student’s t-test.
3. Results
3.1. Fe(II) uptake
As found earlier using nitrosonifedipine (2) obtained by
photodegradation of nifedipine [1], de novo synthesized 2
greatly enhanced Fe(II) uptake by erythrocytes. The major-
ity of the Fe(II) taken up by the cells was transferred
through the membrane to the cytosolic fraction of the cells
and accumulated there in a linear manner with respect to
time of incubation for at least 20 min (results not shown).
Hence, in the subsequent experiments, a 10-min uptake
time was used so that the cells would be sampled during the
linear phase of uptake and the results would represent a rate
of uptake.
2.5.2. Metal efflux
The amount of Fe(II) taken up by the cells was depen-
dent upon both the concentration of iron and of 2 (Fig. 2).
As little as 0.5 mM 2 led to markedly increased Fe(II)
uptake, while the effect on uptake to the cytosol was
maximal at approximately 10 mM, as was confirmed in
later experiments (Fig. 3). However, uptake to the mem-
brane increased further as the concentration of 2 was raised
above 10 mM (Fig. 3). In most of these experiments, the
total uptake of Fe(II) by the cells (cytosol plus membrane)
considerably exceeded the amount of 2 present in the
incubation solution when compared on a molar basis.
For instance, from the data in Fig. 2 it was calculated that
the ratio of Fe(II) uptake to 2 present varied from 1 to 5 and
2 to 10 when the Fe(II) concentration was 10 or 40 mM,
respectively, as the 2 concentration was changed from 0.2
to 10 mM.
Samples of the cell suspensions were first labelled with
the radioactive metals by incubation for 30 min at 378 with
20 mM solutions of the metals in 0.15 M KCl–0.01 M Tris–
HEPES, pH 7.0. (This leads to efficient uptake of all of the
metals tested [14].) The cells were then washed four times
with ice-cold 0.15 M NaCl, suspended in 2 mL of 0.15 M
NaCl–0.01 M Tris–HEPES, pH 7.4, containing 100 mM
DFO, and 10 mL of DMSO or solutions of the nifedipine
derivatives were added. Incubation was then performed for
periods of up to 30 min at 378, followed by centrifugation
to obtain the efflux solution, and hemolysis of the cells and
recentrifugation as for the uptake experiments. Radioac-
tivity was counted in the efflux solution and in the two
fractions of the cells.
2.6. Analytical methods
The mechanism by which 2 stimulates Fe(II) uptake was
investigated by examining the possible effects of trans-
membrane potential difference and free radical-induced
changes in the cell membrane on the uptake process. We
also compared the effects of 2 on Fe(II) uptake with those
of certain chemically related compounds (3, 4, 5, 10, 11)
and of two iron chelators that have been studied in some
detail previously, PIH (12) and CP20 (13).
Membrane potential was altered by the addition of
valinomycin (0.5 mM) to the incubation medium. This
increased membrane permeability for Kþ and changed
the potential from À9:0 Æ 3:0 to À71 Æ 8:9 mV
(mean Æ SEM, N ¼ 3). However, this was not accompa-
nied by a significant change in iron uptake to the cytosolic
or membrane fractions of the cells incubated with 0.2, 2,
or 20 mM Fe(II) in the presence of 10 mM 2 (results not
shown).
The PCV was determined by the microhematocrit
method. Partition of 59Fe from an aqueous solution to
an organic solvent was performed by shaking 1.0 mL
of the 59Fe solution in 0.15 M NaCl, 0.01 M Tris–HEPES,
pH 7.4, with 1.0 mL of organic solvent followed
by centrifugation and sampling the organic solvent
for measurement of radioactivity. The nifedipine deri-
vatives were added to the aqueous phase before the
addition of the solvent. The radioactivity of 63Ni was
measured in a Beckman LS6500 liquid scintillation coun-
ter (Beckman Instruments Inc.) and the other radioiso-
topes in a three-channel g-scintillation counter (1282
Compugamma; LKB Wallac). Transmembrane potential
differences were measured by the method of Macey et al.
[15].

1220
D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
Fig. 2. Effect of the concentrations of Fe(II) and 2 on iron uptake by rabbit erythrocytes. The cells were incubated for 10 min at 378 with 2–50 mM Fe(II) in
the presence of 0–10 mM 2; then they were washed and separated into cytosolic and membrane fractions. The results are the means Æ SEM of four
experiments.
An increase in membrane permeability to iron due to
free radical production and lipid peroxidation in the cell
membrane has been reported for iron uptake by other cell
types [16,17]. The possibility that a similar effect was
responsible for the results of the present investigation was
therefore investigated by the addition of reagents that act as
free radical producers or scavengers to the incubation
medium. The free radical producers tested were t-butyl
hydroperoxide (20 mM), cumene hydroperoxide (10 mM),
H2O2 (100 mM), and 3-methylindole (100 mM). The sca-
vengers were 2[3]-t-butyl-4-hydroxyanisole (50 mM),
N,N0-diphenyl-1,4-phenylenediamine (20 mM), melatonin
(1 mM), a-tocopherol (100 mM), and superoxide dismutase
(200 IU/mL). None of these reagents, with the exception of
superoxide dismutase, had a significant effect on Fe(II)
uptake in the presence of 10 mM 2. Superoxide dismutase
produced a 12% reduction in Fe(II) uptake (results not
shown).
Two compounds chemically related to 2, compounds 5
and 10, stimulated Fe(II) uptake to the cytosol while others,
3, 4 and 11, as well as 12 and 13, had no effect (Fig. 3). Iron
uptake to the membrane was also enhanced by 5 and 10
Fig. 3. Effect of various concentrations (0–20 mM) of 2, 5, 10, 11, 12, 3, 4, and 13 on Fe(II) uptake by rabbit erythrocytes. The cells were incubated for
10 min at 378 with 20 mM Fe(II) in the presence of the reagents. Values are the means Æ SEM of three experiments.

D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
1221
and, to a lesser extent, by 11 and 12. The effects of 5 were
very similar to those of 2, while 10 was only about half as
effective in stimulating Fe(II) uptake to the cytosol,
although it was nearly as effective with respect to iron
uptake to the membrane.
The similarity of the effects of 2 and the hydroxamic
acid 5 also raised the possibility that the action of 2 is the
result of its cell-induced transformation to 5 and that this is
the active substance. To test this, we measured the time
course of Fe(II) uptake mediated by the two substances
when they were added to the cells either 10 min before or
simultaneously with the radiolabelled Fe(II). The rate of
iron uptake was almost the same in the presence of each of
the two substances and was not influenced by the time of
their addition to the cells (results not shown).
It has been demonstrated that, in vivo, 2 is converted
rapidly to the lactam, 3 [18,19]. It was considered possible
that this conversion could occur during incubation of 2
with erythrocytes and that the observed effects could be
attributed to the action of 3. However, as demonstrated
above, 3 lacks the capacity of 2 to mediate Fe(II) uptake by
erythrocytes. The ability of the cells to convert 2 to 3 under
the incubation conditions used in these experiments was
examined as follows. Erythrocytes were incubated with
10 mM 2 for various times up to 60 min at 378, centrifuged,
and the supernatant used in a second incubation with 59Fe
using fresh samples of cells. Compared with 2 which had
been incubated in the absence of cells, the cell supernatants
produced approximately 80% as much Fe(II) uptake even
after a 60-min preincubation with 5 times as many cells as
used during iron uptake measurements (Fig. 4).
with Cd2þ, Co2þ, Mn2þ, Zn2þ, Ni2þ, and Fe(III) under the
same conditions as were used to increase Fe(II) uptake.
Only Zn2þ and Fe(III) were transported into the cytosol but
at a much lower rate than Fe(II). Co2þ and Ni2þ as well as
Zn2þ and Fe(III) were taken up by the membrane fraction
of the cells (Fig. 5). However, the uptake of the other
metals was lower than that of Fe(II) or Fe(III), especially at
lower concentrations of 2. Although the membrane uptake
of Fe(III) was greater than that of Fe(II) at all concentra-
tions of 2 except 20 mM, this could be attributed to a much
larger uptake of Fe(III) than Fe(II) in the absence of 2. If
these values are subtracted from the uptake found in the
presence of 2, the uptake of Fe(II) mediated by 2 is found to
be greater than that of Fe(III).
Of the substances chemically related to 2 that were
tested, only the hydroxamic acids 5 and 10 had any effect
on divalent metal uptake, and their actions were confined to
Co2þ, Zn2þ, and Ni2þ (Fig. 6). None of the substances
influenced Mn2þ or Cd2þ uptake by the cells (results not
shown). Compound 10 strongly stimulated Co2þ, Zn2þ
,
and Ni2þ uptake to the cytosol and, to a lesser degree, to the
membrane. Compound 5 mimicked the action of 2, stimu-
lating Zn2þ uptake into the cytosol and Co2þ, Zn2þ, and
Ni2þ uptake to the membrane. Compounds 5 and 10 at
10 mM also stimulated Fe(III) uptake to the cytosol, to
levels approximately 30% as great as Fe(II) uptake (results
not shown).
3.3. Metal efflux from erythrocytes
To study the effects of 2 and related compounds on metal
efflux from erythrocytes, the cells were first incubated with
the radiolabelled metals under conditions that allow effi-
cient uptake of all of the metals under investigation. They
were then washed and reincubated in the presence of DFO
plus the compounds, since it was shown previously that 2
3.2. Uptake of other divalent metals and Fe(III)
The specificity of the action of 2 on other divalent metals
as well as iron was assessed by performing incubations
Fig. 4. Effect of preincubation of 2 with erythrocytes on its subsequent ability to mediate Fe(II) uptake by fresh samples of cells. Compound 2 (10 mM) was
incubated at 378 with no cells (*), 60 mL cells (*), or 300 mL cells (D) of PCV of 25% in 2 mL NaCl, pH 7.4, for 0–60 min. The cell suspension was then
centrifuged at 2000 g for 10 min at 48, and the supernatant was used to measure Fe(II) uptake by a fresh sample of cells in the standard manner. The figure
shows the means Æ SEM of three experiments.

1222
D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
Fig. 5. Concentration-dependent effects of 2 on the uptake of divalent metals (Fe, Co, Zn, Ni, Cd, and Mn) and Fe(III) by rabbit erythrocytes. The
incubations were performed for 10 min at 378 in the presence of 0–20 mM 2 and a 20 mM concentration of the metals. There was virtually no uptake of Co,
Cd, Mn, or Ni in the cytosolic fraction of the cells. The results are the means Æ SEM of three experiments.
Fig. 6. Effects of 2 and related compounds on Co, Zn, and Ni uptake by rabbit erythrocytes. The cells were incubated for 10 min with 20 mM concentrations
of the metals in the presence of 0–20 mM 2 (*), 5 (*), 10 (&), 3, 4 (D), or 11 (~). Values are the means Æ SEM of 3–4 measurements.

D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
1223
Fig. 7. Efflux of iron from rabbit erythrocytes in the presence of various concentrations of 2. The cells were prelabelled with 59Fe as described in the text,
washed, and reincubated at 378 in the presence of 100 mM DFO and 0 (*), 0.5 (*), 1 (D), 2 (~), 5 (&), or 10 (&) mM 2. The figure shows the percentage of
the 59Fe present in the efflux solution, cytosol, and membrane of the cells after 0-, 10-, 20-, and 30-min reincubation at 378. Each value is the mean Æ SEM of
four experiments.
stimulates iron efflux only if DFO, an iron chelator that
does not enter erythrocytes, is present in the extracellular
fluid [1]. Under these conditions, 2 was found to mediate
iron efflux from the cells in a concentration-dependent
manner (Fig. 7). Most of the iron lost from the cells
was derived from the cytosol, although there was some
decrease in membrane 59Fe (Fig. 7).
Compound 5 stimulated iron efflux to a degree similar to
that of 2 (Fig. 8), while 3, 4, 10, 11, 12, and 13 had no
significant effect (results not shown). Compound 10 caused
a shift of iron from the cytosolic to the membrane fraction
of the cells but had little effect on efflux to the incubation
solution. DFO by itself produced no significant efflux of
iron (or other metals) or redistribution between cytosol and
membrane (Fig. 8).
When the other divalent metals were examined, an effect
was found only with Co2þ (Fig. 8). None of the reagents
affected efflux of Mn2þ, Cd2þ, Zn2þ, or Ni2þ (results not
shown). With Co2þ, 10 was found to stimulate efflux. This
was associated with a drop in cytosolic Co2þ but no change
in membrane Co2þ (Fig. 8).
3.4. Two-phase partition of iron
The effect of 2 and related compounds on the partition of
iron from an aqueous to an organic phase was tested
initially using 2 and a variety of organic solvents. The
highest level of partitioning was found with 1-octanol,
which was used in subsequent studies, while lower levels
of partitioning were found with cyclohexane, petroleum
Fig. 8. Iron and cobalt efflux from erythrocytes mediated by DFO, 2, 5,
and 10. The cells were prelabelled with 59Fe or 57Co, washed, and
reincubated at 378 for 30 min in the presence of no reagents (Nil), 100 mM
DFO, or 100 mM DFO plus 10 mM 2, 5, or 10. The figure shows the
percentage of the 59Fe or 57Co that was present in the efflux solution,
cytosol, and membrane at the end of the reincubation period. Each value is
the mean of four experiments.

1224
D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
Fig. 9. Partitioning of Fe(II) and Fe(III) from the aqueous phase to 1-octanol in the presence of various concentrations of 2, 5, and 10. Compounds 4 and 11
produced no measurable partitioning of iron. Values are the means Æ SEM of three experiments.
ether, n-hexane, and carbon tetrachloride. Compounds 2, 5,
and 10 were approximately equally effective in mediating
iron partitioning Fe(II) into 1-octanol, while the nitro
compound 4 and nitrosobenzene (11) were without effect
(Fig. 9). Compounds 5 and 10 also partitioned Fe(III) into
the octanol phase to a degree similar to that for Fe(II), but 2
was less effective with Fe(III) than with Fe(II) (Fig. 9).
taken up by the cells exceeded the amount of 2 initially
present in the incubation solution. To achieve this, 2 could
be acting as a stimulator of an Fe(II) transport process, a
channel former, or an ionophore. An effect on known Fe(II)
transport processes was ruled out by earlier experiments
[1]. Moreover, such processes either disappear or diminish
considerably when immature erythroid cells such as reti-
culocytes mature into mature erythrocytes [14], yet Fe(II)
uptake mediated by 2 tends to be greater in mature cells
than in reticulocytes [1]. Channel formation as a means of
action of 2 is unlikely since Fe(II) uptake mediated by 2
was not affected by a change of membrane potential.
Hence, it is probable that 2, 5, and 10 act as lipophilic
ionophores, having the ability to transport the Fe(II) into
and across the cell membrane, deposit it at these sites, and
return to the incubation solution. Presumably there are
binding sites in the cytosol and cell membrane that can
compete with these compounds for the Fe(II). Only when
there is a membrane impermeable chelator such as des-
ferrioxamine outside the cell is the direction of iron
exchange reversed.
Compounds 2 and 5 produced similar results with
respect to all aspects of Fe(II) transport that were studied
in the present work. They also had a similar capacity to
partition Fe(II) into 1-octanol. Hence, their mode of action
is probably the same, acting as Fe(II) ionophores. Com-
pound 10 was less effective in mediating Fe(II) uptake to
the cytosol than were compounds 2 or 5 even though it
partitioned Fe(II) to 1-octanol as well as they did. Thus, it
also probably acts as a lipophilic ionophore but with
quantitatively different properties from 2 and 5. The ability
of metal complexing agents to mediate metal exchange
across cell membranes depends largely on their binding
affinity for the metals and their lipophilicity [20,21]. In
addition, the presence of binding sites and their affinities
for the metals in the cell membrane and cytosol will
influence the distribution of the metals in the cell. All of
these factors probably came into play in the present studies
4. Discussion
These experiments show that synthetic 2 has similar
effects on iron transport into erythrocytes as reported
previously for the photodegradation product of nifedipine
[1]. This confirms that the active substance is the nitroso
derivative of nifedipine. Using the pure compound, we
have obtained results quantitatively similar to those of the
earlier work, although the pure substance was found to be
slightly more active per unit weight than the substance
derived from photodegradation of nifedipine. Thus, at a
20 mM iron concentration, maximal Fe(II) uptake was
found to occur at a concentration of 2 of 10 mM, whereas
in the earlier work uptake was slightly higher at 20 mM
than at 10 mM. These differences may be due to the
presence of impurities in the earlier preparations.
A striking feature of the results is the very high rate of
Fe(II) transport into the cells in the presence of 2 and 5.
This is illustrated by the following calculation. At 5–
10 mM concentrations of these compounds and 20 mM
Fe(II), the rate of Fe(II) uptake was equivalent to about
5 mmol/L cells/hr. The iron concentration in mature ery-
throcytes is approximately 20 mmol/L. Hence, if the rate of
Fe(II) uptake in the presence of 2 or 5 continued in a linear
manner, it would take only 4 hr to acquire this amount of
iron, whereas the developing red cell in vivo takes 4–5 days
to do this.
Under most concentrations of Fe(II) and 2 used in these
experiments, the quantity of Fe(II) in molar terms that was

D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
1225
to produce the observed results. As judged from the
partition of Fe(II) into 1-octanol, 2, 5, and 10 have similar
lipophilicity, but this does not rule out the possibility that
10 is more soluble in the cell membranes than are the other
two compounds. Alternatively, or additionally, its affinity
for Fe(II) may be higher so that it exchanges the Fe(II) with
binding sites in the cell less effectively than 2 or 5. Either
property could lead to a relatively greater accumulation of
Fe(II) in the membranes compared with cytosol than with 2
or 5, and movement of iron from the cytosol to membranes
during the efflux experiments.
Fe(II) uptake by erythrocytes. However, several scavengers
of free radicals were found to have little effect on Fe(II)
uptake in the presence or absence of 2, while free radical
producers that may be expected to lead to increased uptake
were also without effect. Hence, there is no evidence that
cell membrane changes due to iron-catalysed lipid perox-
idation are involved. More likely, as concluded above, 2
forms a neutral complex with Fe(II) and acts as an iono-
phore which transfers the iron across the cell membrane.
Compound 2 is rapidly reduced to the lactam 3 when
incubated with glutathione [18,19], and this procedure was
used to produce 3 in the present study. The same change
was observed when 2 was incubated in plasma or blood,
and after intravenous injection into rats [18,19]. Hence, it
appears that 2 is rapidly reduced to a stable lactam 3 in
vivo, and any 2 produced by photodegradation of nifedi-
pine in vivo would be converted to the lactam and have no
effect on cellular iron metabolism. However, related com-
pounds that are stable in vivo may have some effects and
would be worthy of investigation. Compounds 5 and 10 are
possible examples of such compounds.
Erythrocytes are known to possess cell membrane reduc-
tase activity [22], and this may contribute to the production
of the lactam that occurs in blood and in vivo. Indeed, some
decrease in Fe(II) transporting activity was found after
incubation of 2 with erythrocytes. However, this effect was
small, probably due to the small quantities of erythrocytes
used in the present investigation, so that Fe(II) uptake in
the presence of 2 continued in a linear manner during a 20-
min incubation and was not affected significantly by the
addition of 2 to the cells 10 min prior to adding the iron.
Compound 2 has been shown to promote iron uptake by
human epidermal keratinocytes [23] and gallium uptake by
cultured tumour cells [24]. On the basis of these investiga-
tions it was proposed that the nifedipine derivative may be
of value clinically, by mediating iron excretion via the skin
in patients with iron overload [23] and to improve the
imaging of tumours by the use of 67Ga [24]. However, the
observations that 2 is rapidly converted to a stable lactam
compound 3 in vivo [19] and that this product does not
mediate Fe(II) transport in erythrocytes (present study)
suggest that it does not have the potential for use in vivo.
Moreover, this conversion is probably the reason why the
effects of alterations of iron metabolism have not been
reported for patients undergoing treatment with nifedipine.
In terms of three-dimensional structure and polarity, the
nitro compound 4 and nitrosonifedipine 2 are very similar,
yet only the latter is active in Fe(II) transport. The nitroso
group of 2 is unlikely to be solely responsible for this
property as the parent nitrosobenzene (11) is inactive.
Nitrosobenzene (PhNO) is reduced rapidly to phenylhy-
droxylamine (PhNOH) in a number of cell systems,
[25,26], and it is possible that in the current system
an analogous reduction of the nitroso group of 2 could
generate a hydroxylamine group that cyclises onto the
neighbouring ester group to deliver the hydroxamic acid
A remarkable property of 2 and 5 was their high degree of
selectivity for the transport of Fe(II) into the cytosol when
compared with the other metals examined in this study.
Only Zn2þ was transported to the cytosol by these com-
pounds, but to a lesser extent than Fe(II). However, Co2þ
and Ni2þ as well as Zn2þ were taken up by the membrane
fraction of the cells. By contrast, 10 carried all of these
metals including Fe(II) into the cytosol. As discussed above,
these results are probably dependent upon the solubility
of the metal complexes in the cell membrane, the affinity
of the compounds for the metals, and the number and
affinity of binding sites for the metals within the cells.
Compounds 5 and 10 are hydroxamic acids, which are
known to be strong chelators of Fe(III) [21]. However, the
partition studies show that they can complex Fe(II),
although probably much less strongly than Fe(III). The
lower affinity for Fe(II) would aid dissociation of the iron
within the cell and its accumulation there, thus accounting
for the greater cellular uptake of Fe(II) than Fe(III) pro-
duced by 5 and 10. Probably 2 has an affinity for Fe(II)
similar to that of 5, as well as lipophilicity, resulting in the
observed similarity in iron transporting properties. Com-
pound 2 also had some affinity for Fe(III) and some ability
to mediate Fe(III) uptake by the cells, although lower than
for Fe(II). Compounds 12 and 13 are lipophilic chelators of
Fe(III), which is probably why they showed little activity in
this study which used Fe(II). Overall, the results discussed
above are compatible with the conclusion that 2, 5, and 10
can function as lipophilic ionophores for iron, cobalt, zinc,
and nickel.
Fe(II) uptake by the cells was not affected by the
addition of valinomycin. This Kþ ionophore increases
the membrane conductance of Kþ and changes the mem-
brane potential of erythrocytes from one dependent upon
the [Cl1/Cl0] distribution ratio to one dependent upon Kþ.
In the present study this was accompanied by a change in
membrane potential from À9.0 to À71 mV (inside relative
to outside). The fact that such a large change in potential
difference did not affect Fe(II) uptake indicates that the
iron enters the cells in the form of an electrically neutral
complex with 2.
Iron-induced free radical production can lead to lipid
peroxidation and damage to cell membranes. This may
result in increased permeability to Fe(II) [16,17] and could
provide an explanation for nitrosonifedipine-stimulated

1226
D.L. Savigni et al. / Biochemical Pharmacology 65 (2003) 1215–1226
[11] Nicolaus BJR, Pagani G, Testa E. Synthese und Eigenschaften der O-
(phenylethyl)-hydroxylamine. Helv Chim Acta 1962;45:1381–95.
[12] Moriconi EJ, Spano FA. Heteropolar ozonization of aza-aromatics and
their N-oxides. J Am Chem Soc 1964;86:38–46.
5. Indeed, a synthetic sample of 5 was found to be almost as
active as 2. In addition, the simpler hydroxamic acid 10
was also observed to possess activity. However, the effect
of 2 on Fe(II) uptake was observed immediately after its
addition to the cells, and with very short incubation times it
was just as great as that of 5. Hence, it is unlikely that the
action of 2 is due to conversion to 5. Probably the compar-
able effects of these two substances result from their
similar lipid solubility and ability to complex the divalent
metals. Indeed, their abilities to mediate the partitioning of
Fe(II) into 1-octanol were almost identical.
[13] Morgan EH. Membrane transport of non-transferrin-bound iron by
reticulocytes. Biochim Biophys Acta 1988;943:428–39.
[14] Savigni DL, Morgan EH. Transport mechanisms for iron and other
transition metals in rat and rabbit erythroid cells. J Physiol (Lond)
1998;508:837–50.
[15] Macey RI, Adorante JS, Orme FW. Erythrocyte membrane potentials
determined by hydrogen ion distribution. Biochim Biophys Acta
1978;512:284–95.
[16] Fodor I, Marx JJM. Lipid peroxidation of rabbit small intestinal
microvillus membrane vesicles by iron complexes. Biochim Biophys
Acta 1988;961:96–102.
[17] Richardson DR, Ponka P. Identification of a mechanism of iron uptake
by cells which is stimulated by hydroxyl radicals generated via the
iron-catalysed Haber–Weiss reaction. Biochim Biophys Acta 1995;
1269:105–14.
References
[1] Savigni DL, Morgan EH. Mediation of iron uptake and release in
erythroid cells by photoreduction products of nifedipine. Biochem
Pharmacol 1996;51:1701–9.
[18] De Vries H, Beijersbergen van Henegouwen GMJ. Photodegradation
of nifedipine under in vivo-related circumstances. Photochem Photo-
biol 1995;62:959–63.
[2] Ponka P, Borova J, Neuwirt J, Fuchs O. Mobilization of iron from
reticulocytes. Identification of pyridoxal isonicotinoyl hydrazone as a
new chelating agent. FEBS Lett 1979;97:317–21.
[3] Coleman GH, McClosky CM, Stuart FA. Nitrosobenzene. Org Synth
1945;25:80–3.
[19] De Vries H, Beijersbergen van Henegouwen GMJ. Photoreactivity
of nifedipine in vitro and in vivo. J Photochem Photobiol B 1998;43:
217–21.
[20] Ponka P, Richardson DR, Edward JT, Chubb FL. Iron chelators of the
pyridoxal isonicotinoyl hydrazone class. Relationship of the lipophi-
licity of the apochelator to its ability to mobilise iron from reticulo-
cytes in vitro. Can J Physiol Pharmacol 1994;72:659–66.
[21] Zanninelli H, Glickstein H, Breuer W, Milgram P, Brissot P, Hider RC,
Konijn AM, Libman J, Shanzer A, Cabantchik I. Chelation and
mobilization of iron by different classes of chelators. Mol Pharmacol
1997;51:842–85.
[4] Phillips AP. Hantzsch’s pyridine synthesis. J Am Chem Soc 1949;71:
4003–7.
[5] Loev B, Goodman MM, Snader MK, Tedeschi R, Macko E. ‘‘Hantzsch-
type’’ dihydropyridine hypotensive agents. 3. J Med Chem 1974;17:
956–65.
[6] Hayase N, Itagaki Y, Ogawa S, Akutsu S, Inagaki S, Abiko Y. Newly
discovered photodegradation products of nifedipine in hospital pre-
scriptions. J Pharm Sci 1994;83:532–8.
[22] Zamudio I, Cellino M, Canessa-Fischer M. The relation between
structure and NADH (acceptor) oxidoreductase activity of erythrocyte
ghosts. Arch Biochem Biophys 1969;129:336–45.
¨
[7] Gorlitzer K, Bub D. 3,6-Diazaphenanthrene aus nifedipin. Arch
Pharm (Weinheim) 1985;318:97–105.
[8] Kurfu¨rst A, Trska P, Goljer I. Interpretation of 13C NMR chemical
shifts of Hantzsch’s pyridines and dihydropyridines. Coll Czech Chem
Commun 1984;49:2393–9.
[23] Gruen AB, Zhou J, Morton KA, Milstone LM. Photodegraded nife-
dipine stimulates uptake and retention of iron in human epidermal
keratinocytes. J Invest Dermatol 2001;116:774–7.
[9] Kim DH. Rearrangement of 4-(2-aminophenyl)-1,4-dihydro-2,6-di-
methyl-3,5-pyridinedicarboxylic acid diethyl ester. J Heterocyclic
Chem 1986;23:1471–4.
[24] Luttropp CA, Vu C, Morton KA. Photodegraded nifedipine promotes
transferrin-independent gallium uptake by cultured tumor cells. J Nucl
Med 1999;40:159–65.
¨
[25] Eyer P, Kampffmeyer H, Maister H, Rosch-Oehme E. Biotransforma-
[10] Glover SA, Goosen A, McCleland CW. N-Alkoxy-N-acylnitrenium
ions as possible intermediates in intramolecular substitution: novel
formation of N-acyl-3,4-dihydro-1H-2,1-benzoxazines and N-acyl-
4,5-dihydro-1H,3H-2,1-benzoxazepine. J Chem Soc [Perkin 1] 1984;
1:2255–60.
tion of nitrosobenzene, phenylhydroxylamine, and aniline in the
isolated perfused rat liver. Xenobiotica 1980;10:499–516.
[26] Eyer P, Lierheimer E. Biotransformation of nitrosobenzene in the red
cell and the role of glutathione. Xenobiotica 1980;10:517–26.