Primaquine-Induced Hemolytic Anemia: Susceptibility of Normal versus Glutathione-Depleted Rat Erythrocytes to 5-Hydroxyprimaquine

Article abstract of DOI:10.1124/jpet.103.062984

Primaquine is an important antimalarial agent because of its activity against exoerythrocytic forms of Plasmodium spp. Methemoglobinemia and hemolytic anemia, however, are dose-limiting side effects of primaquine therapy. These hemotoxic effects are believed to be mediated by metabolites, although the identity of the toxic specie(s) and the mechanism underlying hemotoxicity have remained unclear. Previous studies showed that an N-hydroxylated metabolite of primaquine, 6-methoxy-8-hydroxylaminoquinoline, was capable of mediating primaquine-induced hemotoxicity. The present studies were undertaken to investigate the hemolytic potential of 5-hydroxyprimaquine (5-HPQ), a phenolic metabolite that has been detected in experimental animals. 5-HPQ was synthesized, isolated by flash chromatography, and characterized by NMR spectroscopy and mass spectrometry. In vitro exposure of ⁵¹Cr-labeled erythrocytes to 5-HPQ induced a concentration-dependent decrease in erythrocyte survival (TC₅₀ of ca. 40 μM) when the exposed cells were returned to the circulation of isologous rats. 5-HPQ also induced methemoglobin formation and depletion of glutathione (GSH) when incubated with suspensions of rat erythrocytes. Furthermore, when red cell GSH was depleted (>95%) by titration with diethyl maleate to mimic GSH instability in human glucose-6-phosphate dehydrogenase deficiency, a 5-fold enhancement of hemolytic activity was observed. These data indicate that 5-HPQ also has the requisite properties to contribute to the hemotoxicity of primaquine. The relative contribution of N-hydroxy versus phenolic metabolites to the overall hemotoxicity of primaquine remains to be assessed.

Full text of DOI:10.1124/jpet.103.062984

0022-3565/04/3091-79–85$20.00
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
Copyright © 2004 by The American Society for Pharmacology and Experimental Therapeutics
JPET 309:79–85, 2004
Vol. 309, No. 1
62984/1135046
Printed in U.S.A.
Primaquine-Induced Hemolytic Anemia: Susceptibility of
Normal versus Glutathione-Depleted Rat Erythrocytes to
5-Hydroxyprimaquine
Zachary S. Bowman, John E. Oatis, Jr., Jennifer L. Whelan, David J. Jollow, and
David C. McMillan
Department of Cell and Molecular Pharmacology, Medical University of South Carolina, Charleston, South Carolina
Received November 12, 2003; accepted December 16, 2003
ABSTRACT
Primaquine is an important antimalarial agent because of its ac- spectrometry. In vitro exposure of 51Cr-labeled erythrocytes to
tivity against exoerythrocytic forms of Plasmodium spp. Methe- 5-HPQ induced a concentration-dependent decrease in eryth-
moglobinemia and hemolytic anemia, however, are dose-limiting rocyte survival (TC50 of ca. 40 ␮M) when the exposed cells were
side effects of primaquine therapy. These hemotoxic effects are returned to the circulation of isologous rats. 5-HPQ also in-
believed to be mediated by metabolites, although the identity of duced methemoglobin formation and depletion of glutathione
the toxic specie(s) and the mechanism underlying hemotoxicity (GSH) when incubated with suspensions of rat erythrocytes.
have remained unclear. Previous studies showed that an N-hy- Furthermore, when red cell GSH was depleted (Ͼ95%) by
droxylated metabolite of primaquine, 6-methoxy-8-hydroxy- titration with diethyl maleate to mimic GSH instability in human
laminoquinoline, was capable of mediating primaquine-induced glucose-6-phosphate dehydrogenase deficiency, a 5-fold en-
hemotoxicity. The present studies were undertaken to investi- hancement of hemolytic activity was observed. These data
gate the hemolytic potential of 5-hydroxyprimaquine (5-HPQ), a indicate that 5-HPQ also has the requisite properties to con-
phenolic metabolite that has been detected in experimental tribute to the hemotoxicity of primaquine. The relative contri-
animals. 5-HPQ was synthesized, isolated by flash chromatog- bution of N-hydroxy versus phenolic metabolites to the overall
raphy, and characterized by NMR spectroscopy and mass hemotoxicity of primaquine remains to be assessed.
Malaria is a widespread, life-threatening parasitic disease capacity to induce hemolytic anemia, particularly in individ-
that is responsible for 300 to 500 million acute illnesses and
an estimated 1.5 to 2.7 million deaths worldwide each year
(Kain and Keystone, 1998). Primaquine, an 8-aminoquino-
line antimalarial drug, is effective against the exoerythro-
cytic forms of all four of the malarial species that infect
humans and is the only radically curative drug for the latent
tissue forms of Plasmodium vivax and Plasmodium ovale
(Tracy and Webster, 2001). Primaquine is also used in com-
bination with chloroquine to combat the problem of multiple
drug resistance in Plasmodium falciparum (Shanks et al.,
2001). Despite its clinical importance and effectiveness, use
of primaquine has long been known to be limited by its
uals with a hereditary deficiency in erythrocytic glucose-6-
phosphate dehydrogenase (G6PD) activity (Dern et al., 1955;
Degowin et al., 1966). Because G6PD deficiency is prevalent
in malarial areas, this dose-limiting toxicity can have a major
impact on the usefulness of this drug in these populations.
Importantly, primaquine is not directly toxic to erythro-
cytes at clinically relevant concentrations. Although the he-
motoxicity of primaquine has long been considered to be
dependent on metabolism, the metabolite(s) responsible and
the underlying mechanism(s) have remained unclear. We
have reported recently that 6-methoxy-8-hydroxylamin-
oquinoline, an N-hydroxylated metabolite of primaquine, is a
direct-acting hemolytic and methemoglobinemic agent in
rats, and therefore may be a contributor to the hemotoxicity
observed in primaquine-treated humans (Bolchoz et al., 2001).
Metabolism of primaquine, however, is relatively complex,
and a variety of known and putative phenolic metabolites
have also been considered to be capable of mediating prima-
This study was supported by National Institutes of Health Grant AI46424
(to D.C.M.). The studies reported in this article were presented in part at the
54th Southeast Regional Meeting of the American Chemical Society, Nov.
13–16, 2002, Charleston, SC (Z.S.B.).
Article, publication date, and citation information can be found at
http://jpet.aspetjournals.org.
DOI: 10.1124/jpet.103.062984.
ABBREVIATIONS: G6PD, glucose-6-phosphate dehydrogenase; 5-HPQ, 5-hydroxyprimaquine; GSH, reduced glutathione; HPLC-EC, high-
performance liquid chromatography with electrochemical detection; ESI, electrospray ionization; PBSG, isotonic phosphate-buffered saline with
glucose; GSSG, oxidized glutathione (glutathione disulfide); PSSG, protein-glutathione mixed disulfides; DEM, diethyl maleate.
79

80
Bowman et al.
quine hemotoxicity. In particular, hydroxylation of prima-
quine at the 5-position of the quinoline ring (Fig. 1) is known
to yield redox-active derivatives that are capable of inducing
oxidative stress within normal and G6PD-deficient human
erythrocytes. Several of these compounds, including 5-hy-
droxyprimaquine (5-HPQ), 5-hydroxy-6-desmethylprima-
quine, and their N-dealkylated derivatives, were synthesized
in the 1960s and made available to investigators by the
World Health Organization. Studies with these compounds
in isolated suspensions of red cells have shown that they can
induce methemoglobin formation, glutathione (GSH) deple-
tion, and stimulation of hexose monophosphate shunt activ-
ity (Allahyari et al., 1984; Link et al., 1985; Baird et al., 1986;
Agarwal et al., 1988; Fletcher et al., 1988; Vasquez-Vivar and
Augusto, 1994). However, there is a notable lack of evidence
for their hemolytic activity in vivo.
Progress toward understanding the role of phenolic metab-
olites in primaquine-induced hemolytic anemia has been
hampered because they are no longer available, the synthetic
methods to prepare them are relatively difficult, and the
products are highly unstable. As a first step in our investi-
gation of the potential contribution of phenolic metabolites to
primaquine-induced hemolytic anemia, we have resynthe-
sized 5-HPQ and examined its stability and redox behavior.
In addition, we have assessed the hemolytic potential of
5-HPQ in GSH-normal and GSH-depleted rat red cells. In
view of the critical role proposed for oxidative stress in the
mechanism underlying primaquine-induced hemolytic ane-
mia, we measured the formation of methemoglobin and mon-
itored red cell sulfhydryl status under hemolytic conditions
to correlate the hemolytic response with these indicators of
intracellular oxidative damage. We report that 5-HPQ is an
extremely potent direct-acting hemolytic agent in rats and
that hemolytic activity is associated with methemoglobin
formation and a marked depletion of erythrocytic GSH.
When GSH was depleted from rat red cells to mimic GSH
instability of human G6PD-deficient red cells (Gaetani et al.,
1979), the hemolytic activity of 5-HPQ was markedly en-
hanced. The significance of the data with regard to the over-
all contribution of metabolites to primaquine-induced hemo-
lytic anemia is discussed.
5-Methoxyprimaquine
(5,6-dimethoxy-8-[4-amino-1-methylbu-
tylamino]quinoline) was prepared from 6-methoxy-8-nitroquinoline
as described previously (Elderfield et al., 1955). 5-HPQ (5-hydroxy-
6-methoxy-8-[4-amino-1-methylbutylamino]quinoline) was synthe-
sized from 5-methoxyprimaquine by HBr-catalyzed hydrolysis using
a modification of an established method (Allahyari et al., 1984). The
composition of the reaction mixture was monitored as a function of
time via liquid chromatography-mass spectrometry. The reaction
mixture contained four major components: 5-HPQ (m/z 275.5–276.5;
4.60 min, 21.0%), 5-HPQ quinoneimine (m/z 273.5–274.5; 4.41 min,
42.8%), 5-methoxyprimaquine (m/z 289.5–290.5; 9.58 min, 21.1%),
and 5-hydroxy-6-desmethylprimaquine (m/z 259.5–260.5; 4.09 min,
15.1%). The yield of 5-HPQ was optimized by adjusting the reaction
temperature to 120°C and the reaction time to 20 min. The yield was
increased further by reducing the quinoneimine to the hydroquinone
using sodium dithionite and maintaining the reaction mixture under
argon to minimize oxidation of the hydroquinone. The reaction mix-
ture was then purified in two steps using SepPak Plus C18 car-
tridges (Waters Corporation, Milford, MA). 5-HPQ was eluted from
the first cartridge with 5% acetonitrile/0.05% aqueous trifluoroacetic
acid, lyophilized, and then applied to a second cartridge. 5-HPQ was
eluted from the second cartridge with 5% acetonitrile in water con-
taining 5 mM HBr. After removal of the solvent by lyophilization,
elemental analysis confirmed the presence of the trihydrobromide
salt of 5-HPQ (purity Ͼ99% as judged by HPLC and NMR analysis).
1H NMR (in D2O): ␦ 8.62 (dd, J ϭ 1.4, 4.4 Hz, 1, H-2), 8.43 (dd, J ϭ
8.5, 1.4 Hz, 1, H-4), 7.46 (s, 1, H-7), 4.40 (dd, 4.3, 8.5 Hz, 1, H-3), 3.76
(m, 2, H-1Ј), 3.82 (s, 3, OCH3), 2.77 (m, 2, H-4Ј), 1.65 (m, 1, H-2Ј), 1.65
(m, 1, H ϭ 3Ј), 1.56 (m, 1, H-2Ј), 1.52 (m, 1, H-3Ј), 1.19 (d, J ϭ 6.6 Hz,
3, H-5Ј). 13C NMR (in D2O): ␦ 148.6 (C-2), 142.7 (C-6), 139.7 (C-5),
134.9 (C-9), 132.4 (C-4), 122.4 (C-8), 121.8 (C-3), 120.4 (C-10), 112.8
(C-7), 57.9 (C-1Ј), 57.6 (OCH3), 38.9(C-4Ј), 29.7 (C-2Ј), 23.1 (C-3Ј),
15.9 (C-5Ј). Because 5-HPQ is unstable, even when stored in the dark
under argon at –80°C, it was routinely prepared for immediate use
(i.e., within 24–48 h) from its more stable precursor, 5-methoxypri-
maquine, as described above.
HPLC Analysis. Chromatography was performed on a Waters
HPLC system consisting of a model 510 pump, a Rheodyne injector
(5-ml loop), and a 250-mm Alltech Platinum EPS C18 reverse phase
column. 5-HPQ was eluted with 10% acetonitrile in water containing
0.05% trifluoroacetic acid at a flow rate of 1.1 ml/min, and was
detected on a Waters model 481 UV-Vis variable wavelength detec-
tor set at 254 nm. For stability studies, an HPLC system (BAS
Bioanalytical Systems, West Lafayette, IN) consisting of a model
PM-80 pump, a Rheodyne 7125 injector (20-␮l loop), and a 150-mm
Alltech Platinum EPS C18 reverse phase column was used. 5-HPQ
was eluted with 7% acetonitrile in water containing 0.05% trifluoro-
acetic acid and 50 mM potassium trifluoroacetate at a flow rate of 1.0
ml/min and was detected using an Epsilon electrochemical detector
(BAS Bioanalytical Systems) equipped with a glassy carbon working
electrode (oxidation mode, ϩ0.35 V) and a Ag/AgCl reference elec-
trode.
Materials and Methods
Chemicals and Materials. 6-Methoxy-8-nitroquinoline, ferrous
bromide, sodium metal, stannous chloride, potassium trifluoroac-
etate, and GSH were obtained from Sigma-Aldrich (St. Louis, MO).
Na251CrO4 in sterile saline (1 mCi/ml, pH 8) was obtained from New
England Nuclear (Billerica, MA). All other chemicals and reagents
were of the best grade commercially available.
NMR Spectroscopy and Mass Spectrometry. Proton and car-
bon NMR spectra were obtained on a Varian Inova spectrometer
operating at 400 and 100 MHz, respectively. Proton assignments
were made by using the double quantum filtered COSY experiment
acquired in the phase sensitive mode; 2 ϫ 256 fids were acquired.
Digital resolution in F1 was increased by linear prediction to 1024
points, processed using the Gaussian weighting function, and then
Fourier transformed. The chemical shifts of unresolved multiplets
were based on the chemical shifts of the cross peaks. Carbon reso-
nances were assigned using gradient versions of the heteronuclear
single quantum coherence (HSQC) and heteronuclear multi-bond
correlation experiments (gHMBC). In the HSQC, 128 fids were ac-
quired. Linear prediction increased the points in F1 to 512, Gaussian
weighted, and then Fourier transformed. In the HMBC, 400 fids
were acquired, linear prediction increased the points in F1 to 1200,
sinebell weighted, and then Fourier transformed. The nuclear Over-
Fig. 1. Putative metabolism of primaquine to 5-HPQ.

Hemotoxicity of 5-HPQ in Rat Erythrocytes
81
hauser effect spectroscopy (NOESY) experiment was acquired in the acetone was added to packed red cells. After a 15-min incubation at
phase sensitive mode by collecting 2 ϫ 256 fids. Digital resolution in 37°C, the red cells were analyzed for GSH content by HPLC-EC as
F1 was increased by linear prediction to 1024 points, processed using described above. Under these conditions, GSH was reduced to about
the Gaussian weighting function, and then Fourier transformed. 5% of initial levels. The cells were resuspended to a 40% suspension
Presence of a methoxy group in the 6-position of 5-HPQ was verified in PBSG and used on the same day that they were collected.
by the NOESY experiment.
Mass spectra were obtained using a Finnigan LCQ ion trap mass
spectrometer (Thermo Finnigan, San Jose, CA). A 150-mm Alltech
Results
Stability and Electrochemistry of 5-HPQ. NMR stud-
ies undertaken as part of the characterization of the newly
synthesized 5-HPQ indicated that it was stable for over 24 h
when maintained at low pH under strictly anaerobic condi-
tions. This indicated that it could be prepared and kept as a
solution without significant degradation before its experi-
mental use in erythrocyte suspensions. On the other hand,
previous work had shown 5-HPQ to be unstable in the pres-
ence of oxygen (at pH 8.5) due to its facile conversion into its
quinoneimine form (Vasquez-Vivar and Augusto, 1990).
Therefore, to determine the stability of the 5-HPQ hydro-
quinone/quinoneimine redox pair under our experimental
conditions, 5-HPQ (500 ␮M) was added to aerobic PBSG (pH
7.4) in the absence and presence of red cells. Aliquots were
withdrawn at intervals, treated with an excess of sodium
dithionite, and then assayed for 5-HPQ by HPLC-EC (Fig. 2).
Rapid loss of 5-HPQ occurred in both situations with a half-
life of about 45 s in the absence of red cells and about 30 s in
their presence. Because the hydroquinone and quinoneimine
forms of 5-HPQ were not well separated on the HPLC-EC
column, LC-MS analysis (in which both halves of the redox
pair could be detected independently by selected ion moni-
toring) confirmed that the disappearance of 5-HPQ was not
due simply to its oxidation to the quinoneimine during chro-
matographic analysis, but instead was due to the complete
degradation of the redox pair (data not shown).
Platinum EPS C18 reverse phase column was used. The sample was
eluted with 10% acetonitrile in water containing 0.05% trifluoroace-
tic acid at a flow rate of 0.5 ml/min. The column effluent was split
and 10% was directed to the ESI source. Instrument parameters
were as follows: ESI needle voltage, 4.5 kV; ESI capillary tempera-
ture, 200°C; ion energy, 45%; isolation window, 1 amu; and scan
range, 150.0 to 1000.0 amu. Mass spectrometry and tandem mass
spectrometry data were acquired automatically using Xcalibur soft-
ware (version 1.2).
Electrochemical Activity of 5-HPQ. Cyclic voltammetry was
performed using a CV-27 voltammograph (BAS Bioanalytical Sys-
tems), C-1A/B cell stand, and a model RXY recorder. Stock solutions
of 5-HPQ (245 ␮M) were prepared in argon-purged isotonic phos-
phate-buffered saline (pH 7.4) supplemented with 10 mM D-glucose
(PBSG). Samples were scanned at a rate of 150 mV/s at room tem-
perature under an argon atmosphere using a carbon-paste working
electrode, a platinum auxiliary electrode and a Ag/AgCl reference
electrode.
Animals and Erythrocyte Incubation Conditions. Male
Sprague-Dawley rats (75–100 g) were purchased from Harlan (Indi-
anapolis, IN) and maintained on food and water ad libitum. Animals
were acclimated for 1 week to a 12-h light/dark cycle before their use.
Blood from the descending aorta of anesthetized rats was collected
into heparinized tubes and washed three times with PBSG to remove
the plasma and buffy coat. The cells were resuspended to a 40%
hematocrit in PBSG and used the same day they were collected.
Stock solutions of 5-HPQ in argon-purged water were prepared to
deliver the appropriate concentration of 5-HPQ in 10 ␮l to erythro-
cyte suspensions (1–3 ml).
Previous studies have shown that in the presence of an
excess of NADPH and a catalyst (ferredoxin:NADPϩ oxi-
doreductase), 5-HPQ can generate greater than stoichiomet-
ric amounts of hydrogen peroxide (Vasquez-Vivar and Au-
Measurement of Hemolytic Activity. The survival of rat 51Cr-
labeled red cells was determined in vivo after in vitro incubation
with various concentrations of 5-HPQ (25–300 ␮M). After incubation
for 2 h at 37°C, the erythrocytes were washed once and resuspended
in PBSG (40% hematocrit). Aliquots (0.5 ml) were administered
intravenously to isologous rats. T0 blood samples were taken from
the orbital sinus 30 min after administration of labeled red cells.
Additional blood samples were taken every 48 h for 14 days. At the
end of the experiment, the samples were counted in a well type
gamma counter, and the data were expressed as a percentage of the
T0 blood sample. The hemolytic response was quantified by calculat-
ing the fraction of radiolabeled red cells that were removed from the
circulation within the first 48 h for each animal by linear regression
as described previously (McMillan et al., 2001). Statistical signifi-
cance was determined with the use of Student’s t test.
Determination of Methemoglobin Formation and Sulfhy-
dryl Status. Methemoglobin levels in erythrocyte suspensions
treated with 5-HPQ (25–1000 ␮M) were measured using a modifica-
tion of the spectrophotometric technique of Evelyn and Malloy (1938)
as described previously (Harrison and Jollow, 1987).
For determination of sulfhydryl status, aliquots (200 ␮l) of the
erythrocyte suspensions were removed at various intervals after
addition of 5-HPQ and assayed for GSH, oxidized glutathione
(GSSG), and glutathione-protein mixed disulfides (PSSG) by HPLC
with electrochemical detection (HPLC-EC) as described previously
(Grossman et al., 1992). The amount of sulfhydryl present in the
samples was determined by comparison of peak areas to prepared
standards.
Fig. 2. Stability of 5-HPQ in blood versus buffer. 5-HPQ (500 ␮M) was
added to buffered saline in the absence and presence of erythrocytes (40%
hematocrit) and allowed to incubate aerobically at 37°C. Aliquots were
removed at designated intervals, treated with dithionite, and assayed for
5-HPQ concentration using HPLC with electrochemical detection. Values
are expressed as a percentage of the concentration at T0 and are means Ϯ
S.D. (n ϭ 3).
GSH Depletion of Erythrocyte Suspensions. Diethyl maleate
(DEM) was used to deplete GSH in red cell suspensions as described
previously (Bolchoz et al., 2002). Briefly, DEM (750 ␮M) dissolved in

82
Bowman et al.
gusto, 1992), which suggests that this compound has the
ability to redox cycle. To determine directly whether oxida-
tion of 5-HPQ to its quinoneimine form is reversible, we
examined its electrochemical activity in argon-purged PBSG
(pH 7.4) using cyclic voltammetry. As shown in Fig. 3, when
an excitation potential scan was initiated in the positive
direction, two oxidation products (peaks A and B) were ob-
served. When the potential scan was reversed (in the nega-
tive direction), a reduction product (peak C) was observed.
These data are consistent with a concerted oxidation of
5-HPQ hydroquinone to 5-HPQ quinoneimine (peak B) via a
semiquinone radical intermediate (peak A), as suggested by
previous work (Vasquez-Vivar and Augusto, 1992). Peak C
corresponds to the reduction of the oxidized products formed
on the forward scan back to 5-HPQ hydroquinone. These
electrochemical data support the concept that 5-HPQ can
undergo cycling as a fully reversible redox couple at physio-
logical pH.
Fig. 4. A, survival of rat 51Cr-labeled erythrocytes in vivo after in vitro
exposure to 5-HPQ. Radiolabeled erythrocytes were incubated for 2 h at
37°C with the indicated concentrations of 5-HPQ; control cells were
incubated with vehicle (10 ␮l of H2O). The erythrocytes were then washed
and readministered intravenously to isologous rats. T0 blood samples
were taken 30 min after administration of labeled cells. Data points are
means Ϯ S.D. (n ϭ 4). B, concentration dependence for the hemolytic
response after 5-HPQ exposure. Values are means Ϯ S.D. (n ϭ 4).
Direct Hemolytic Activity of 5-HPQ. Although a vari-
ety of studies on the oxidative activity of 5-HPQ in red cells
have been published, its direct hemolytic activity has not
been established. To investigate the hemolytic potential of
5-HPQ, rat 51Cr-labeled erythrocytes were incubated with
various concentrations of 5-HPQ for 2 h at 37°C. The cells
Methemoglobin Formation by 5-HPQ. 5-HPQ has been
were then washed and returned to the circulation of isologous previously shown to deplete red cell GSH and induce methe-
rats. A T0 blood sample was taken from the orbital sinus 30 moglobin formation (Allahyari et al., 1984; Link et al., 1985;
min after administration of the labeled red cells and then Baird et al., 1986; Agarwal et al., 1988; Fletcher et al., 1988;
serial blood samples were taken at 48-h intervals for 14 days. Vasquez-Vivar and Augusto, 1994). To determine the rela-
As shown in Fig. 4A, exposure of the labeled cells to 5-HPQ tionship between these endpoints and the hemolytic re-
caused a concentration-dependent increase in the rate of sponse, we examined the time and concentration dependence
removal of radioactivity from the circulation compared with of methemoglobin formation in rat erythrocyte suspensions
controls. Figure 4B shows the concentration response curve exposed to 5-HPQ. As shown in Fig. 5A, incubation of a rat
for the hemolytic activity of 5-HPQ. The concentration-re- red cell suspension with a maximal hemolytic concentration
sponse curve for 5-HPQ was extremely sharp, with an appar- of 5-HPQ (100 ␮M) resulted in the rapid formation of methe-
ent threshold concentration of about 25 ␮M, a TC50 of ap- moglobin. This concentration produced a peak methemoglo-
proximately 40 ␮M, and a maximal response at about 100 bin level of only about 20%, which nevertheless remained
␮M.
constant over the 2-h incubation period. Figure 5B depicts
the concentration dependence of the methemoglobinemic re-
sponse to 5-HPQ at 30 min post-exposure. Methemoglobin
levels ranged from approximately 3.5% at 25 ␮M 5-HPQ to a
maximum of about 40% at 300 ␮M (TC50 of ca. 100 ␮M).
Effect of 5-HPQ on Rat Erythrocyte Sulfhydryl Sta-
tus. To examine the fate of red cell GSH after treatment with
hemolytic concentrations of 5-HPQ, aliquots were taken at
various intervals and analyzed for GSH, GSSG, and PSSG
levels by HPLC-EC. As shown in Fig. 6A, addition of 100 ␮M
5-HPQ to rat red cells resulted in a complete depletion of
GSH within 15 min. The loss of GSH was matched by an
increase in PSSG; GSSG remained low throughout the incu-
bation period. The concentration dependence of the 5-HPQ-
induced depletion of GSH is shown in Fig. 6B. As with the
hemolytic response (Fig. 4B), a sharp concentration-response
curve was observed, with a TC50 of approximately 40 ␮M.
Hemolytic Activity of 5-HPQ in GSH-Depleted Eryth-
rocytes. The enhanced susceptibility displayed by G6PD-defi-
cient individuals to primaquine-induced hemolytic anemia is
thought to be due to an inability to maintain sufficient levels of
NADPH, and thus reduced glutathione, in response to the oxi-
dative stress. To reproduce in rat erythrocytes the instability of
GSH known to occur in human G6PD-deficient erythrocytes,
51Cr-labeled red cells were titrated with DEM to deplete GSH
by Ͼ95%. The GSH-depleted red cells were then exposed to
Fig. 3. Cyclic voltammogram of 5-HPQ (245 ␮M) in argon-purged PBSG
(pH 7.4) at room temperature. Working electrode, carbon paste; reference
electrode, Ag/AgCl; auxiliary electrode, platinum. Scan rate, 150 mV/s. A
and B, anodic (oxidation) peaks. C, cathodic (reduction) peak.

Hemotoxicity of 5-HPQ in Rat Erythrocytes
83
Fig. 5. Effect of 5-HPQ on methemoglobin for-
mation in rat erythrocytes. A, rat erythrocytes
were treated with 5-HPQ (100 ␮M) and assayed
for methemoglobin levels over time; control cells
were incubated with vehicle (10 ␮l of H2O). Data
points are means Ϯ S.D. (n ϭ 3). B, concentration
dependence for methemoglobin formation. Ali-
quots of the incubation mixture were assayed for
methemoglobin 30 min after exposure to 5-HPQ
(25–1000 ␮M). Data points are means Ϯ S.D.
(n ϭ 3).
various concentrations of 5-HPQ in vitro for 2 h at 37°C, and likely candidates for mediating both the hemolytic and met-
their survival was determined in vivo. As shown in Fig. 7A, the
survival of untreated GSH-depleted red cells (T50 ϭ 11.0 Ϯ 1.9
days) was not significantly different from the survival of GSH-
normal red cells (Fig. 4A; T50 ϭ 9.8 Ϯ 0.8 days). As expected
from the previous experiment, the rate of removal of GSH-
normal red cells exposed to a subhemolytic concentration of
5-HPQ (10 ␮M) was also not significantly different from the
controls (Fig. 7A). In contrast, exposure of GSH-depleted red
cells to a 10 ␮M concentration of 5-HPQ provoked a dramatic
increase in their rate of removal. Quantitation of the hemolytic
response for GSH-depleted red cells (Fig. 7B) revealed the con-
centration-response curve to be shifted significantly to the left
of the response curve for GSH-normal cells (Fig. 4B), with a
TC50 under these conditions of about 7.5 ␮M.
hemoglobinemic responses that have been observed during
the course of therapy with this antimalarial drug. Consider-
able attention has been given to the 5-hydroxy- and 5,6-
dihydroxy metabolites of primaquine because they have the
potential to redox cycle (via quinoneimine and 5,6-quinone
formation, respectively) and generate reactive oxygen spe-
cies. Support for the importance of phenolic metabolites has
come from a variety of in vitro studies that showed that these
compounds were able to induce oxidative changes within red
cells, such as stimulation of hexose monophosphate shunt
activity, GSH depletion, and hemoglobin oxidation. What has
been missing from these efforts, however, is evidence that
links these biochemical changes observed in vitro to loss of
erythrocyte viability in vivo.
The present results demonstrate that a redox active phe-
nolic metabolite of primaquine, 5-HPQ, is a direct-acting
hemolytic agent in the rat (Fig. 4). This loss of erythrocyte
viability in vivo was correlated with a rapid and extensive
Discussion
Oxidative metabolism has long been known to play a crit-
ical role in the onset of primaquine-induced hemotoxicity,
and phenolic metabolites have been considered the most depletion of GSH (Fig. 6A), which exhibited a concentration
Fig. 6. Effect of 5-HPQ on rat erythrocyte sulf-
hydryl status. A, rat red cells were incubated at
37°C in PBSG containing 5-HPQ (100 ␮M). At
the indicated time points, aliquots were with-
drawn and assayed for GSH, GSSG, and PSSG.
Data points are means Ϯ S.D. (n ϭ 3). B, concen-
tration dependence for GSH depletion by 5-HPQ
(5–100 ␮M) in rat erythrocytes. GSH concentra-
tion was determined before addition of 5-HPQ
and again at 15 min postexposure. Values are
expressed as a percentage of the initial level and
are means Ϯ S.D. (n ϭ 3).

84
Bowman et al.
Fig. 7. A, survival of normal versus GSH-de-
pleted rat 51Cr-labeled erythrocytes in vivo after
in vitro exposure to 5-HPQ. Radiolabeled red
cells were treated with DEM to deplete intracel-
lular GSH (Ͼ95%). The cells were incubated for
2 h at 37°C with the indicated concentrations of
5-HPQ; GSH-depleted control cells were incu-
bated with vehicle (10 ␮l of H2O). Data points are
means Ϯ S.D. (n ϭ 4). B, concentration depen-
dence for the hemolytic response in GSH-de-
pleted red cells. Values are means Ϯ S.D. (n ϭ 3).
dependence that coincided with that of the hemolytic re- bility of 5-HPQ in the presence of red cells (Fig. 2). Alterna-
sponse (Fig. 6B). The disappearance of GSH was matched by tively, 5-HPQ may interfere with the normal reduction of
the formation of mixed disulfides between GSH and the sol- methemoglobin, either by depletion of reducing cofactors
uble protein of the red cell. The importance of GSH status in (NADH/NADPH) and/or inhibition of cellular reductases, or
determining the sensitivity of rat red cells to this hemolytic by generating more stable oxidants that continue to generate
agent is illustrated by the data in Fig. 7A, which shows that
depletion of GSH with DEM before 5-HPQ exposure caused a
marked enhancement of the hemolytic response. These data
strongly support the concept that the hemolytic response has
a discrete dose threshold and that this threshold is depen-
dent on the presence GSH in the red cell.
methemoglobin at a rate that exceeds its reduction. In any
case, the concentration-response data suggest that the mech-
anisms underlying methemoglobin formation and hemolytic
activity of 5-HPQ may be unrelated.
Together, these data strongly support a role for 5-HPQ in
primaquine-induced hemolytic anemia, and furthermore,
they may provide an explanation for the dramatic difference
in primaquine sensitivity between G6PD-deficient and
G6PD-normal individuals. Data published by Degowin et al.
(1966) showed that doses of primaquine necessary to provoke
a hemolytic response in G6PD-deficient humans are about
20-fold lower than those required to elicit a similar response
in G6PD-normal individuals, whereas the doses of dapsone
required to induce similar responses in G6PD-deficient ver-
sus normal differed by only a factor of 2. Although the reason
for the difference in susceptibility between dapsone and pri-
maquine is not yet understood, it may be related to the fact
that dapsone hemotoxicity is mediated by a single hydroxyl-
amine metabolite, whereas primaquine hemotoxicity may be
mediated by the synergistic action of multiple toxic metabo-
lites, including N-hydroxy, quinoneimine, and quinone.
In summary, we have demonstrated that a phenolic me-
tabolite of primaquine, 5-HPQ, is directly hemotoxic to the
rat red cell. We have also shown that the hemotoxicity is
highly dependent on the level of GSH in the red cell, which
suggests that GSH status may underlie the apparent thresh-
old for primaquine hemotoxicity in G6PD deficiency. The
actual contribution of this metabolite, however, to prima-
quine hemotoxicity remains to be assessed.
Although the in vitro exposure/in vivo survival data pre-
sented in Fig. 4 do not allow for a direct assessment of the
role of 5-HPQ in primaquine hemotoxicity, this assay does
permit the hemolytic damage observed in vivo to be repro-
duced in vitro under controlled conditions during a 2-h incu-
bation period before the red cells are returned to the circu-
lation of rats, and thus serves as a useful indicator of the
relative potency among direct-acting hemolytic agents. Inter-
estingly, 5-HPQ is the most potent hemolytic agent we have
examined to date. The TC50 of 5-HPQ (ca. 40 ␮M) was about
3.5-fold lower than that of dapsone hydroxylamine (TC50 of
ca. 150 ␮M), an N-hydroxy metabolite known to be the sole
mediator of the hemolytic activity of dapsone, and about
8.5-fold lower than that of 6-methoxy-8-hydroxylaminoquino-
line (TC50 of ca. 350 ␮M), an N-hydroxy metabolite shown
recently by our laboratory to have the requisite properties to
mediate primaquine hemotoxicity. Of interest, the potency of
5-HPQ was increased by more than 5-fold in GSH-depleted
red cells (TC50 of ca. 7.5 ␮M).
As shown in Fig. 5A, hemolytic concentrations of 5-HPQ
were associated with the formation of methemoglobin; how-
ever, the concentration-response curve for methemoglobin
formation (Fig. 5B) was shifted well to the right of the hemo-
lytic concentration-response curve. In addition, the methe-
moglobinemic efficacy of 5-HPQ was limited to about 40% of
the maximum response, even when extremely high (1 mM)
concentrations were used. Although the reason for this lack
of efficacy and low relative potency is unknown and requires
Acknowledgments
We thank Jennifer Schulte for excellent technical support. We
acknowledge the Medical University of South Carolina NMR Re-
further investigation, it may be related to the marked insta- source Facility and the Medical University of South Carolina Mass

Hemotoxicity of 5-HPQ in Rat Erythrocytes
85
Gaetani GF, Mareni C, Salvidio E, Galiano S, Meloni T, and Arese P (1979) Favism:
erythrocyte metabolism during haemolysis and reticulocytosis. Br J Haematol
43:39–48.
Spectrometry Resource Facility for assistance in NMR and mass
spectral analyses, respectively.
Grossman SJ, Simson J, and Jollow DJ (1992) Dapsone-induced hemolytic anemia:
effect of N-hydroxy dapsone on the sulfhydryl status and membrane proteins of rat
erythrocytes. Toxicol Appl Pharmacol 117:208–217.
Harrison J Jr and Jollow DJ (1987) Contribution of aniline metabolites to aniline-
induced methemoglobinemia. Mol Pharmacol 32:423–431.
Kain KC and Keystone JS (1998) Malaria in travelers. Epidemiology, disease and
prevention. Infect Dis Clin North Am 12:267–284.
References
Agarwal S, Gupta UR, Gupta RC, Anand N, and Agarwal SS (1988) Susceptibility of
glucose-6-phosphate dehydrogenase deficient red cells to primaquine enantiomers
and two putative metabolites. I. Effect on reduced glutathione, methemoglobin
content and release of hemoglobin. Biochem Pharmacol 37:4605–4609.
Allahyari R, Strother A, Fraser IM, and Verbiscar AJ (1984) Synthesis of certain
hydroxy analogues of the antimalarial drug primaquine and their in vitro methe-
moglobin-producing and glutathione-depleting activity in human erythrocytes.
J Med Chem 27:407–410.
Baird JK, McCormick GJ, and Canfield CJ (1986) Effects of nine synthetic putative
metabolites of primaquine on activity of the hexose monophosphate shunt in intact
human red blood cells in vitro. Biochem Pharmacol 35:1099–1106.
Bolchoz LJ, Budinsky RA, McMillan DC, and Jollow DJ (2001) Primaquine-induced
hemolytic anemia: formation and hemotoxicity of the arylhydroxylamine metabo-
lite 6-methoxy-8-hydroxylaminoquinoline. J Pharmacol Exp Ther 297:509–515.
Bolchoz LJ, Morrow JD, McMillan DC, and Jollow DJ (2002) Primaquine-induced
hemolytic anemia: effect of 6-methoxy-8-hydroxylaminoquinoline on rat erythro-
cyte sulfhydryl status, membrane lipids, cytoskeletal proteins and morphology.
J Pharmacol Exp Ther 303:1121–1129.
Degowin RL, Eppes RB, Powell RD, and Carson PE (1966) The haemolytic effects of
diaphenylsulfone (DDS) in normal subjects and in those with glucose-6-phosphate-
dehydrogenase deficiency. Bull World Health Organ 35:165–179.
Dern RJ, Beutler E, and Alving AS (1955) The hemolytic effect of primaquine. V.
Primaquine sensitivity as a manifestation of a multiple drug sensitivity. J Lab
Clin Med 45:30–39.
Elderfield RC, Mertel HE, Mitch RT, Wempen IM, and Werble E (1955) Synthesis of
primaquine and certain of its analogs. J Am Chem Soc 77:4816–4819.
Evelyn KA and Malloy HT (1938) Microdetermination of oxyhemoglobin, methemo-
globin and sulfhemoglobin in a single sample of blood. J Biol Chem 126:655–662.
Fletcher KA, Barton PF, and Kelly JA (1988) Studies on the mechanisms of oxidation
in the erythrocyte by metabolites of primaquine. Biochem Pharmacol 37:2683–
2690.
Link CM, Theoharides AD, Anders JC, Chung H, and Canfield CJ (1985) Structure-
activity relationships of putative primaquine metabolites causing methemoglobin
formation in canine hemolysates. Toxicol Appl Pharmacol 81:192–202.
McMillan DC, Bolchoz LJ, and Jollow DJ (2001) Favism: effect of divicine on rat
erythrocyte sulfhydryl status, hexose monophosphate shunt activity, morphology
and membrane skeletal proteins. Toxicol Sci 62:353–359.
Shanks GD, Kain KC, and Keystone JS (2001) Malaria chemoprophylaxis in the age
of drug resistance. II. Drugs that may be available in the future. Clin Infect Dis
33:381–385.
Tracy JW and Webster LT (2001) Drugs used in the chemotherapy of protozoal
infections: Malaria, in Goodman & Gilman’s The Pharmacological Basis of Ther-
apeutics (Hardman JG and Limbird LE eds) pp 1069–1095, McGraw-Hill Compa-
nies, New York.
Vasquez-Vivar J and Augusto O (1990) ESR detection of free radical intermediates
during autoxidation of 5-hydroxyprimaquine. Free Rad Res Commun 9:383–389.
Vasquez-Vivar J and Augusto O (1992) Hydroxylated metabolites of the antimalarial
drug primaquine. Oxidation and redox cycling. J Biol Chem 267:6848–6854.
Vasquez-Vivar J and Augusto O (1994) Oxidative activity of primaquine metabolites
on rat erythrocytes in vitro and in vivo. Biochem Pharmacol 47:309–316.
Address correspondence to: Dr. David C. McMillan, Department of Phar-
macology, Medical University of South Carolina, 171 Ashley Ave., Charleston,
SC 29425. E-mail: mcmilldc@musc.edu