1136 The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
Effect of fetal hemoglobin on microvascular
regulation in sickle transgenic-knockout mice
Dhananjay K. Kaul, Xiao-du Liu, Hee-Yoon Chang, Ronald L. Nagel, and Mary E. Fabry
Division of Hematology, Department of Medicine, Albert Einstein College of Medicine, Bronx, New York, USA.
In sickle cell disease, intravascular sickling and attendant flow abnormalities underlie the chronic inflamma-
tion and vascular endothelial abnormalities. However, the relationship between sickling and vascular tone is not
well understood. We hypothesized that sickling-induced vaso-occlusive events and attendant oxidative stress will
affect microvascular regulatory mechanisms. In the present studies, we have examined whether microvascular
abnormalities expressed in sickle transgenic-knockout Berkeley (BERK) mice (which express exclusively human
α- and βS-globins with <1% γ-globin levels) are amenable to correction with increased levels of antisickling fetal
hemoglobin (HbF). In BERK mice, sickling, increased oxidative stress, and hemolytic anemia are accompanied
by vasodilation, compensatory increases in eNOS and COX-2, and attenuated vascular responses to NO-medi-
ated vasoactive stimuli and norepinephrine. The hypotension and vasodilation (required for adequate oxygen
delivery in the face of chronic anemia) are mediated by non-NO vasodilators (i.e., prostacyclin) as evidenced
by induction of COX-2. In BERK mice, the resistance to NO-mediated vasodilators is associated with increased
oxidative stress and hemolytic rate, and in BERK + γ mice (expressing 20% HbF), an improved response to these
stimuli is associated with reduced oxidative stress and hemolytic rate. Furthermore, BERK + γ mice show nor-
malization of vessel diameters, and eNOS and COX-2 expression. These results demonstrate a strong relation-
ship between sickling and microvascular function in sickle cell disease.
Sickle cell disease is characterized by recurring episodes of pain-
ful vaso-occlusive crisis, acute chest syndrome, and multiple organ
damage. Although hemoglobin S (HbS) polymerization and red
cell sickling under deoxygenated conditions are central to the
pathophysiology of this disease, emerging evidence indicates that
vascular endothelial activation and blood cell–endothelium inter-
actions may contribute significantly to the onset of vaso-occlu-
sive episodes. Vascular endothelial abnormalities may result from
intravascular sickling and red cell–endothelium interaction, as
well as from reperfusion injury involving transient occlusive events
(1–3). Human sickle cell disease patients show chronic inflamma-
tion as evidenced by elevated cytokines, endothelial cell damage,
endothelial cell detachment, and increased leukocyte counts (4–9).
Recent studies have revealed that intravascular sickling is the
causative factor in hypoxia/reoxygenation–induced endothelial
oxidant generation in transgenic sickle mice (3). In fact, both
sickle cell disease patients and transgenic-knockout sickle mice
show increased oxidant production under steady-state condi-
tions (10). The rheological abnormalities of sickle red cells and the
associated transient occlusive events may not only contribute to
inflammation but also affect vasoregulatory function of vascular
endothelium as in other inflammatory diseases (11).
Vascular tone alterations in sickle cell disease are indicated by
reports of lower peripheral resistance (12), intermittent periodic
microcirculatory flow (13), depressed vasomotor response follow-
ing postocclusive hyperemia (14), impaired NO bioavailability (15),
and attenuated vascular response to oxygen (16, 17). Impaired NO
bioavailability in this disorder is indicated by blunted response
to endothelium-dependent vasodilators in sickle mouse models
(10, 18, 19), as well as by a reduced flow-mediated vasodilation in
human sickle cell disease patients (20, 21). In human sickle cell dis-
ease patients, however, plethysmographic measurements of fore-
arm blood flow show a greater vasodilatory response (i.e., increase
in blood flow) to acetylcholine (ACh), an endothelium-depen-
dent vasodilator, as compared with normal subjects (16, 20). The
increased blood flow response to ACh may involve upregulation
of non-NO vasodilators (20), although this aspect has not been
explored. On the other hand, sodium nitroprusside (SNP), an NO
donor, causes comparable increases in forearm blood flow in sickle
cell disease patients and normal individuals. However, the respons-
es to both ACh and SNP are reduced in male sickle cell disease
patients (20). Also, patients with increased hemolytic rate show a
diminished response to SNP (15), which suggests that hemolysis
in this disease would limit NO bioavailability, since NO is con-
sumed by ferrous hemoglobin.
The present intravital studies were designed to examine the
potential role of NO and non-NO vasodilatory mechanisms in
the regulation of microvascular flow in this disease. Because vas-
cular resistance to blood flow is mainly determined by the vascu-
lar tone of arterioles (resistance vessels) in the microcirculation,
it is important to determine the mechanisms affecting the
arteriolar tone in the above context. In contrast to large con-
duit arteries, arterioles are uniquely endowed with the ability
to undergo vasomotion (i.e., rhythmic contraction and dilation
of vessel diameter). The abnormalities of arteriolar tone may
contribute significantly to vaso-occlusive events. In the present
studies, we address the relationship between sickling, vascular
tone, and microvascular flow.
Nonstandard abbreviations used: ACh, acetylcholine; BERK, Berkeley; BERK-trait,
BERK trait L; D, vessel luminal diameter; DAB, 3,3′-diaminobenzidine; HbF, fetal
hemoglobin; HbS, hemoglobin S; L-NAME, nitro-L-arginine methylester; MAP, mean
arterial pressure; MCH, mean corpuscular hemoglobin; NE, norepinephrine; NOS,
NO synthase; Q, volumetric flow rate; SNP, sodium nitroprusside; β-thal, β-thalasse-
mic; Vmean, mean Vrbc; Vrbc, red cell velocity.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J. Clin. Invest. 114:1136–1145 (2004).
The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
Therapeutic strategies to prevent sickling include increasing
the levels of antisickling fetal hemoglobin (HbF) by hydroxyurea
therapy (22), as well as reducing the mean corpuscular hemoglobin
(MCH) concentration (23). In the presence of HbF (α2γ2), polymer
formation is efficiently prevented (24). HbF exerts an ameliorating
effect in sickle cell disease patients both on red cells and in preven-
tion of multiple organ damage, and a recent study has shown that
hydroxyurea therapy increases responsiveness to SNP (20). How-
ever, the impact of HbF on microvascular flow regulation has not
been explored. The availability of transgenic-knockout sickle mice
expressing human α- and βS-globins with varying levels of HbF (24)
provides an opportunity to determine the relationship between
intravascular sickling and microvascular abnormalities.
We hypothesize that sickling-induced transient occlusive episodes
and increased oxidative stress will result in microvascular regula-
tory abnormalities, and that inhibition of sickling by an increase in
HbF will tend to ameliorate these abnormalities. Here, we show that
microcirculatory abnormalities and greatly impaired vascular reac-
tivity in transgenic-knockout Berkeley (BERK) mice (which express
exclusively human α- and βS-globins with very low levels of γ-glo-
bin [<1%]) are associated with increased oxidative stress, sickling,
hemolytic anemia, and a reduced NO bioavailability. The compen-
satory responses in this mouse model include induction of COX-2
and upregulation of eNOS, both vasodilator species. Furthermore,
increasing the level of antisickling HbF in this mouse model results
in significant alleviation of hemolytic anemia, oxidative stress, and
vascular tone abnormalities.
Table 1 presents hematological fea-
tures of mice used in these studies. As
is evident, BERK mice showed thalas-
semic characteristics, i.e., globin chain
imbalance (β-globin:α-globin ratio,
0.79) that is similar to that reported
in β-thalassemic (β-thal) mice, while
in C57BL/6J mice (also called C57BL
mice), the β:α ratio approached 0.93
(24). Consistent with thalassemic
features, BERK mice showed a low
MCH compared with control C57BL
mice (9.3 ± 0.8 and 14.5 ± 1.0 pg/cell,
respectively). BERK + γ mice expressing
20% γ-globin levels showed increased MCH
(10.8 ± 1.3 pg/cell) compared with BERK
mice. As previously described, BERK + γ
mice have heterogeneous expression of
HbF. Some red cells have high levels of
HbF (F cells) and some have none (24).
This heterogeneous distribution of HbF
is reminiscent of that in human sickle cell
disease patients (25).
The low hematocrit level in BERK mice
(28.7% ± 4.0% vs. 48% ± 1.0% for C57BL
mice) was associated with high reticulo-
cyte counts (39.5% ± 4.3% vs. 1.4% ± 0.4%
for C57BL) and increased plasma hemo-
globin levels expressed as micromoles of
heme (7.6 ± 1.2 vs. 3.0 ± 0.5 for C57BL, P < 0.00001); these asso-
ciations indicate increased red cell destruction. Plasma hemoglo-
bin levels in control mice were comparable to those reported previ-
ously (26). While BERK + γ mice showed an increase in hematocrit
(41.6% ± 4.0%) as compared with BERK mice, reticulocyte counts
showed only a slight but significant decrease to 31.9% ± 4.2% (P < 0.01).
While anemia was significantly corrected in BERK + γ mice, plasma
hemoglobin levels (5.6 ± 1.0 μmol), although lower than in BERK
mice (P < 0.034), remained higher than in C57BL mice (P < 0.002),
indicating persistence of hemolysis. The reticulocyte counts com-
pared favorably with those previously reported (24). On the other
hand, in β-thal mice, low hematocrit was associated with increased
reticulocyte counts (23.8% ± 2.6%) but no significant increase in
plasma hemoglobin levels (4.4 ± 1.5 μmol) as compared with C57BL
mice, which is consistent with anemia associated with destruction
of red cells in the reticuloendothelial system and/or inefficient
erythropoiesis in human β-thalassemia.
Hemodynamic parameters in BERK mice
We compared hemodynamic and microvascular parameters in
BERK mice with those in C57BL controls, β-thal mice, and BERK
trait L (BERK-trait) mice. We used β-thal mice as additional con-
trols because BERK mice show thalassemic characteristics.
Mean arterial blood pressure. The averaged mean arterial pressure
(MAP) values in BERK mice (n = 13) showed a pronounced decrease
compared with those in C57BL (n = 17), β-thal (n = 4), and BERK-
Hematological parameters in controls and knockout sickle mice
BERK + γ
(%) (%) (%)
βS γ Hematocrit
48 ± 1
32.3 ± 1.6B
41.0 ± 2.3B
28.7 ± 4.0B
41.6 ± 4.0B,C 10.8 ± 1.3B
14.5 ± 1.0
12.7 ± 0.5
13.1 ± 0.5
9.3 ± 0.8B
1.4 ± 0.4
23.8 ± 2.6B
7.8 ± 1.5B
39.5 ± 4.3B
31.9 ± 4.2B,C
3.0 ± 0.5
4.4 ± 1.5
3.4 ± 1.3
7.6 ± 1.2B
5.6 ± 1.0B,C
αH, human α-globin; βS, human βS-globin; γ, human γ-globin; retics, reticulocytes. ANumber of mice used
in these measurements: C57BL/6J, 20; C57BL β-thal, 6; BERK-trait, 4; BERK, 20; BERK + γ, 5. Reticu-
locytes and plasma hemoglobin determinations were done in 7 β-thal mice and in 5 mice from each of
the other genotypes. BP < 0.05–0.0000001 vs. C57BL controls. CP < 0.034–0.001 vs. BERK mice.
Resting microvascular diameters (μm) in C57BL, β-thal, BERK-trait, BERK, and BERK + γ mice
Vessel order C57BL
β-thal BERK-trait BERK BERK + γ
45.8 ± 2.7 (8)
23.0 ± 0.9 (15)
13.9 ± 1.1 (8)
44.1 ± 2.9 (5)
22.2 ± 1.2 (12)
14.3 ± 0.9 (10)
45.2 ± 2.4 (9)
20.7 ± 1.1 (11)
14.7 ± 0.7 (8)
63.5 ± 2.3 (12)A
32.5 ± 1.4 (23)A
19.1 ± 0.7 (24)A
47.3 ± 1.8 (8)A
22.1 ± 1.9 (7)A
14.7 ± 3.7 (6)A
63.4 ± 2.8 (8)
33.1 ± 1.8 (9)
17.2 ± 0.7 (9)
60.7 ± 4.5 (5)
33.0 ± 2.7 (12)
18.0 ± 1.0 (11)
62.0 ± 1.7 (8)
32.9 ± 1.1 (10)
20.9 ± 1.3 (7)
73.2 ± 2.6 (10)A
39.2 ± 1.8 (19)
23.1 ± 1.0 (16)B
61.0 ± 1.5 (7)A
29.1 ± 1.3 (10)A
17.0 ± 0.3 (6)A
Values are mean ± SEM; the numbers in parentheses represent the number of vessels. Number of mice
used in these measurements: C57BL, 6; β-thal, 3; BERK-trait, 4; BERK, 10; BERK + γ, 4. If the average for
each animal is calculated and these values are averaged, the resulting values are not significantly different
from those presented. AP < 0.05 vs. other groups; BP < 0.05 vs. C57BL and β-thal mice.
1138 The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
trait (n = 4) mice (P < 0.05, multiple comparisons by ANOVA).
The averaged MAP in BERK mice showed an average reduction
of approximately 40 mmHg compared with that in C57BL mice
(65.2 ± 2.4 vs. 104.6 ± 1.7 mmHg), while β-thal and BERK-trait
mice showed intermediate MAP values (91.3 ± 4.3 and 91.3 ± 3.2
Vessel diameters and microvascular flow characteristics. Systemic
hypotension in BERK mice was associated with about 40% dilation
of the resistance vessels (arteriolar branching orders A2, A3, and
A4) and about 15% dilation of V2 venules in the cremaster muscle
microcirculation (P < 0.05) compared with those in C57BL control
mice (Table 2). No significant differences in vessel diameters were
noted among control groups (C57BL, β-thal, and BERK-trait mice).
Direct microscopic observations and videoanalysis of the
microvascular flow under resting conditions in the cremaster
muscle preparation in BERK mice revealed frequent red cell adhe-
sion in postcapillary venules (Figure 1, A–C). Red cells with sickle
morphology (probably irreversibly sickled cells in addition to elon-
gated sickled cells) were also observed (Figure 1B). Intravascular
sickling was observed in postcapillary venules as well as in termi-
nal arterioles, often resulting in stasis. We also found increased
leukocyte-endothelium interactions (Figure 1C), indicating a
proinflammatory condition, in BERK mice. This is in accord with
the observations of Turhan el al. (27), who showed increased leu-
kocyte recruitment (as well as red cell–leukocyte interaction) in the
cremaster venules of C57BL mice transplanted with bone marrow
derived from BERK mice.
Next, we investigated the impact of red cell adhesion and
intravascular sickling on red cell velocity (Vrbc), wall shear rate, and
volumetric flow rate (Q) in similar branching orders of arterioles
(A2–A4) and venules (V2–V4) in BERK mice (see Methods). Arterio-
lar Vrbc (in mm/s, mean ± SE) was somewhat lower in BERK mice
(n = 8), showing significant differences (P < 0.05) for A2 arterioles
compared with those of control C57BL mice (n = 6) (Figure 2A).
Wall shear rates (s–1), calculated from vessel diameter and mean
Vrbc (Vmean), showed marked decreases in BERK mice, i.e., 40–58%
decreases in arterioles and an approximately 35% decrease in small-
diameter postcapillary venules (V4 and V3), as compared with
C57BL controls (P < 0.05). There were no significant differences in
arteriovenous wall shear rates among C57BL, BERK-trait, and β-thal
mice (data not shown). To ascertain the effect of vasodilation on
flow in BERK mice, Q was calculated from Vmean and vessel diameter.
As shown in Figure 2C, arteriovenous Q showed a parabolic pattern
in each group of mice. In each group, the highest Q values (in nl/s,
mean ± SE) were obtained for A2 and V2 vessels. Notably, vasodila-
tion was accompanied by 44–56% greater flow rates in A2 and V2
vessels of BERK mice than in those of C57BL controls (P < 0.05).
There were no significant differences in Q among C57BL, β-thal,
and BERK-trait mice (data not shown).
Hemodynamic parameters: the effect of HbF
BERK + γ mice, which express 80% βS-globin and 20% γ-globin,
were examined to determine the antisickling effect of HbF and its
consequences for hemodynamic and microvascular flow param-
eters. In contrast to BERK mice (<1% γ-globin), BERK + γ mice
(n = 9) showed a markedly increased MAP (90.3 ± 1.9 vs. 65.2 ± 2.4
mmHg; P < 0.00001), accompanied by a significant decrease in
microvascular diameters (P < 0.05); the resulting vessel diameters
were not significantly different from those in control groups of
mice (Table 2). Furthermore, both arteriovenous wall shear rates
and volumetric flow showed significant reductions in BERK + γ
Videomicrographs showing in vivo adhesion of red cells to endothelium
of postcapillary venules in the cremaster muscle microcirculation of the
BERK mouse. (A) Adherent red cells and a leukocyte (L) during flow
(large arrow). The small arrow indicates an irreversibly sickled cell.
Scale bar: 15 μm. (B) Adherent red cells (arrowheads) in a postcapillary
venule during flow (large arrow). Scale bar: 15 μm. (C) Adherent red
cells (arrowheads) and large-diameter leukocytes (L) at the confluence of
postcapillary venules during flow (large arrow). Scale bar: 10 μm.
Arteriovenous Vrbc (A), wall shear rate (B), and Q (C) profiles in the
resting cremaster muscle microcirculation of C57BL, BERK, and BERK
+ γ mice. Microvascular blood flow in the BERK mice is characterized
by a pronounced decline in arteriolar wall shear rates, and by a greater
Q in A2 and V2 vessels. Note the normalization of wall shear rates and
Q in BERK + γ mice to control values. *P < 0.05 vs. C57BL and BERK
mice (Kruskal-Wallis test for ANOVA); †P < 0.05 vs. C57BL mice.
The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
mice compared with BERK mice (P < 0.05; Figure 2, B and C) and
were not significantly different from those in control mice; this
confirmed a trend toward normalization of microhemodynamic
parameters in BERK + γ mice.
Vasoactive molecules: the effect of HbF
To determine whether the pronounced vasodilation and sys-
temic hypotension in BERK mice are secondary to upregulation
and/or induction of vasodilator species, we explored the status of
NO synthase (NOS) and COX-2 in BERK, BERK + γ, and control
mice. While chronic hypoxia associated with low hematocrit and
intravascular sickling may cause upregulation of NOS, COX-2 is
induced during hypoxic and inflammatory conditions.
Upregulation of eNOS protein expression. Western blot analysis
using a mAb directed against eNOS showed that expression of
eNOS (140 kDa) was distinctly increased in BERK mice (Figure
3A). Densitometric analysis of Western blots confirmed an aver-
age 2-fold increase (range, 1.4- to 3.3-fold) in eNOS in BERK mice
compared with controls (n = 6 each) (Figure 3B). A comparison
of eNOS expression among various groups of mice confirmed
higher expression in BERK mice (P < 0.05, multiple comparisons
by ANOVA), and no appreciable differences in eNOS expression
were noted between BERK + γ (n = 3) and other control groups
(C57BL, β-thal [n = 3], and BERK-trait [n = 3]); this indicated an
ameliorating effect of HbF (Figure 3B).
Lack of induction of iNOS. Using a mAb specific to iNOS, we did
not find any evidence of iNOS expression using Western blots
of the cremaster and kidney lysates in 3 C57BL control mice
and 3 BERK mice.
COX-2 protein expression. Western blot analysis of cremaster
muscle lysates for COX-2 was carried out using murine polyclonal
antibodies against COX-2. As shown in Figure 3A, C57BL con-
trol mice showed a low level of COX-2 expression. A low level of
COX-2 expression has been previously reported in control brain
tissue and unperturbed endothelial cells (28, 29). In contrast,
BERK mice showed a significantly increased expression of COX-2.
Densitometric analysis of Western blots showed an average 3.7-
fold increase (range, 1.6- to 6.2-fold) in COX-2 in BERK mice
compared with controls (n = 8 each) (Figure 3C). A comparison
of COX-2 expression among various groups of mice confirmed a
higher expression in BERK mice (P < 0.05), while no appreciable
differences were observed between BERK + γ mice (n = 3) and con-
trol groups (C57BL, β-thal [n = 3], and BERK-trait [n = 3]) (Figure
3C); this further suggested an anti-inflammatory effect of antis-
ickling HbF in BERK + γ mice.
Western blot analysis of cremaster muscle lysates for eNOS and COX-2
in BERK, BERK + γ, and control groups (C57BL, β-thal, and BERK-
trait) of mice. (A) Note the higher expression of eNOS and COX-2
proteins in BERK mice. (B) Densitometric analysis of Western blot
confirmed an average 2-fold increase in eNOS in BERK mice as com-
pared with control C57BL mice, but no appreciable differences in the
eNOS expression were observed between BERK + γ and control mice.
(C) Densitometric analysis of COX-2 expression showed an average
3.7-fold increase compared with control C57BL mice. No appreciable
differences were noted between BERK + γ mice and control groups.
*P < 0.05 (multiple comparisons by ANOVA).
Immunoperoxidase staining for eNOS in the cremaster muscle
microvasculature of C57BL, BERK, and BERK + γ mice. (A) Negative
control. Arrowheads indicate blood vessels. (B) The same vessels in
an adjacent section from control C57BL cremaster muscle show posi-
tive reaction for eNOS in the vessel wall (arrowheads). (C) BERK mice
show a strongly positive reaction in the vessel wall (arrowheads). (D)
In contrast to BERK mice, BERK + γ mice show a distinct decrease in
the intensity of staining for eNOS in vessels (arrowheads).
1140 The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
Immunohistochemistry. Microvascular expression of eNOS and
COX-2 was determined in the cremaster muscle preparation of
C57BL, BERK, and BERK + γ mice. As shown in Figure 4B, vas-
cular endothelium of C57BL mice was distinctly positive for
eNOS. In contrast, the staining for eNOS in the cremaster vessels
of BERK mice was greatly enhanced (Figure 4C) compared with
that in the control tissue, while BERK + γ mice showed a marked
decrease in the staining for eNOS in blood vessels compared
with BERK mice (Figure 4D). Immunohistochemistry for COX-2
revealed negative to weakly positive expression of the enzyme in
the cremaster blood vessels of control C57BL mice (Figure 5B). In
contrast, BERK mice showed a marked enhancement of COX-2
expression in the endothelium of most blood vessels (Figure 5, C
and D). On the other hand, COX-2 expression in BERK + γ mice
was distinctly low compared with that in BERK mice (Figure 5E)
and similar to that observed in control C57BL mice. The specific-
ity of immunohistochemical reactions was confirmed by negative
controls for eNOS and COX-2 (Figures 4A and 5A).
Vascular reactivity: the effect of HbF
Response to ACh, an endothelium-dependent vasodilator. As shown in
Figure 6A, topical ACh (10–6 M) caused 87.7% and 73.6% increases,
respectively, in arteriolar (A3) diameters (mean ± SE) in C57BL con-
trols (22.6 ± 2.9 to 41.5 ± 4.7 μm, P < 0.0003, n = 7) and BERK-trait
mice (21.6 ± 2.5 to 36.5 ± 3.0 μm, P < 0.0001, n = 10). In contrast,
BERK mice showed an attenuated response as evidenced by an only
7% increase (P < 0.00001 vs. C57BL and BERK-trait mice) in arterio-
lar diameters from 27.8 ± 2.3 to 29.8 ± 2.8 μm (n = 11). In BERK + γ
mice, ACh resulted in an approximately 33% increase in the diameter
(25.8 ± 2.3 to 33.9 ± 2.3 μm, P < 0.002, n = 7), and this response was
significantly greater than that in BERK mice (P < 0.001), although
the percentage increase in the diameter was significantly less pro-
nounced than in C57BL and BERK-trait mice (P < 0.01).
Response to SNP, an endothelium-independent vasodilator. Figure 6B
shows the effect of SNP, an NO donor, on the arteriolar (A3) diame-
ter. Topical SNP (10–6 M) caused 78% and 64% increases, respectively,
in arteriolar diameters (mean ± SE) in C57BL controls (22.3 ± 2.9 to
40.5 ± 4.5 μm, P < 0.00001, n = 7) and BERK-trait mice (19.9 ± 2.1
to 32.4 ± 2.9 μm, P < 0.001, n = 6). On the other hand, BERK mice
showed an attenuated response as evidenced by an only 13% increase
in the diameter (27.4 ± 2.5 to 30.7 ± 2.5 μm, n = 11, P < 0.001 vs. C57BL
and BERK-trait mice). In BERK + γ mice, SNP caused an almost 50%
increase in the diameter (23.9 ± 2.8 to 33.9 ± 2.1 μm, P < 0.001).
The percentage increase in the diameter in BERK + γ mice was signifi-
cantly greater than that in BERK mice (P < 0.002), but significantly
less than that in C57BL mice (P < 0.01).
Response to nitro-L-arginine methylester. Next, we evaluated NOS
activity by evaluating the response of MAP to intravenous infusion
of nitro-L-arginine methylester (L-NAME) (20 mg/kg for 10 min-
utes), a potent nonselective inhibitor of NOS. As is evident from
Figure 7A, L-NAME caused significant blood pressure increases
(∼25–28%) at 30 minutes in male C57BL (n = 5) and BERK + γ (n = 3)
mice compared with the baseline (pre–L-NAME) values (P < 0.001
and 0.044, respectively). In contrast, male BERK mice (n = 7)
Immunoperoxidase staining for COX-2 in the cremaster muscle
microvasculature of C57BL, BERK, and BERK + γ mice. (A) Nega-
tive control. Arrowheads indicate blood vessels. (B) The same vessels
in an adjacent section from control C57BL cremaster muscle show
negative to weakly positive reaction (arrowheads). (C and D) Strongly
positive reaction for COX-2 in vascular endothelium of blood vessels in
BERK mice (arrowheads). (E) BERK + γ mice show negative or weakly
positive reaction for COX-2 in vessel walls (arrowheads).
Arteriolar diameter responses (percent increase) to topical application of
ACh (10–6 M) and SNP (10–6 M) in C57BL, BERK-trait, BERK, and BERK
+ γ mice. Note the attenuated response of arterioles in BERK mice to
ACh (A) and SNP (B). ACh and SNP caused significant increases in
arteriolar diameters of BERK + γ mice as compared with those in BERK
mice (∼33% and ∼50% increases, respectively). *P < 0.005–0.000001
vs. C57BL and BERK-trait mice. †P < 0.00–0.002 vs. BERK mice.
The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
showed an attenuated response as indicated by a less than 7%
increase in MAP after L-NAME infusion (P = 0.29).
In separate experiments, we evaluated the effect of L-NAME
infusion on MAP in female C57BL (n = 5) and BERK (n = 6) mice
to ascertain whether the MAP response showed a sexual dimor-
phism. In female C57BL mice, L-NAME caused an approximately
23% increase in MAP (98.6 ± 1.8 to 120.8 ± 2.3 mmHg, P < 0.003),
which showed no significant difference from the values depicted
for the males in Figure 7A. On the other hand, in female BERK
mice, L-NAME caused a greater than 13% increase (69.7 ± 6.1 to
79.0 ± 5.8 mmHg, P > 0.004), which showed a still muted but great-
er response to L-NAME than in male BERK mice (Figure 7A).
The effect of norepinephrine. As depicted in Figure 7B, intravenous
infusion of norepinephrine (NE) (1–4 μg/kg, n = 4 each) produced
a significant progressive increase in MAP in both control C57BL
and BERK + γ mice; in each group NE at 4 μg/kg resulted in a 30%
increase in MAP (P < 0.05). On the other hand, NE caused an attenu-
ated response in BERK mice, and there was no progressive rise in
MAP with the increasing dose of NE. Moreover, at any given dose of
NE, MAP remained significantly lower in BERK mice (n = 4) than in
C57BL and BERK + γ mice (P < 0.05, multiple comparisons).
In the above experiments wherein we used cremaster preparation,
we cannot discount some inflammatory effect of tissue exterior-
ization, which has been observed by others (27, 30). Nevertheless,
our results demonstrate distinct differences in hemodynamic pro-
files and vascular reactivity among control, BERK, and BERK + γ
mice subjected to the same exteriorization procedure. We also
observed distinct differences in eNOS and COX-2 expression in
rapidly excised cremaster tissue among these mice.
Nitrotyrosine: the effect of HbF
Since BERK mice exhibited attenuated responses to both NO-
mediated vasodilators (ACh and SNP) and a NOS inhibitor
(L-NAME), we investigated whether these effects were related
to increased oxidative stress. Western blot analysis of cremaster
muscle extracts (Figure 8) showed 2 prominent bands of nitrated
proteins, corresponding approximately to 66 and 26 kDa. In 2
BERK mice, tyrosine nitration of the 66-kDa protein was increased
6.2- and 4.8-fold compared with that in the C57BL control, while
tyrosine nitration of the 26-kDa protein was increased 2- and 1.7-
fold. In the BERK + γ mouse, tyrosine nitration of the 66- and
26-kDa proteins showed smaller increases (i.e., 2.3- and 1.3-fold,
respectively) compared with that in the C57BL control, indicating
a protective effect of HbF against oxidative stress. The BERK-trait
mouse showed smaller increases of 1.1- and 1.4-fold, respectively,
for nitration of the 66- and 26-kDa proteins, while the respective
ratios for the β-thal mouse were 0.1 and 1.0.
We show that in the BERK mouse model of sickle cell disease,
intravascular sickling, increased oxidative stress, and hemolytic
anemia are accompanied by vasodilation, compensatory increases
in eNOS and COX-2, and attenuated vascular responses to NO-
mediated vasoactive stimuli and NE. The hypotension and vasodi-
lation (required for adequate oxygen delivery in the face of chronic
anemia) in this mouse model are mediated by non-NO vasodilators
(i.e., prostacyclin) as evidenced by induction of COX-2. The resis-
tance to NO-mediated vasodilators in BERK mice is associated with
an increased oxidative stress and hemolytic rate, and an improved
response to these stimuli in BERK + γ mice (expressing 20% HbF)
is associated with a decreased oxidative stress and hemolytic rate.
Also, in the presence of antisickling HbF, BERK + γ mice show nor-
malization of vessel diameters, and eNOS and COX-2 expression.
(A) The effect of L-NAME (20 mg/kg) on MAP in C57BL, BERK, BERK-
trait, and BERK + γ mice. L-NAME caused significant increases in MAP
in C57BL and BERK-trait mice as compared with pre–L-NAME values.
In contrast, BERK mice showed an attenuated response. (B) The effect
of norepinephrine (NE; 1, 2, and 4 μg) on MAP in C57BL, BERK, and
BERK + γ mice. NE caused dose-dependent increases in MAP in C57BL
and BERK + γ mice compared with pre-NE values, but an attenuated
response in BERK mice. *P < 0.05 vs. C57BL and BERK + γ mice (mul-
tiple comparisons by ANOVA). †P < 0.05 vs. pre-NE values (t test).
Western blot analysis of cremaster muscle lysates for the expression of
nitrotyrosine. Two prominent bands of nitrated proteins were detected
by the antibody to nitrotyrosine, corresponding to approximately 66 and
26 kDa. BERK mice showed increased tyrosine nitration of both 66- and
26-kDa proteins (average increase, 5-fold and about 2-fold, respective-
ly), while the BERK + γ mouse showed smaller increases as compared
with C57BL controls. The nitrotyrosine levels in BERK-trait and β-thal
mice showed no increase as compared with C57BL controls. Equal
loading of the samples was ascertained using anti–actin antibody.
1142 The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
Oxidative stress in BERK mice. The increased oxidative stress in this
mouse model is likely the result of extensive intravascular sickling,
the presence of β-thalassemia features, blood cell–endothelium
interactions, and recurring transient occlusive events. Sickling
and transient vaso-occlusive events can result in endothelial acti-
vation, as shown in a transgenic mouse model in which hypoxia/
reoxygenation resulted in an exaggerated inflammatory response
(i.e., endothelial oxidant generation and leukocyte recruitment)
compared with that in normal mice (2, 3). These results imply
that intravascular sickling and associated hypoxia/reoxygenation
events in sickle cell disease will result in chronic oxidative stress.
Consistent with these observations, Aslan and coworkers (10) have
reported that both sickle cell disease patients and BERK mice show
elevated plasma xanthine oxidase activity accompanied by increased
vascular superoxide (O2•–) production.
The present studies show that the oxidative stress in BERK
mice is associated with increased production of nitrotyrosine.
Tyrosine nitration can occur when O2•– reacts with NO to form
peroxynitrite (ONOO–), which can nitrate tyrosine residues (31). A
major mechanism contributing to this process is hemin-mediated
nitrated-tyrosine formation in the presence of nitrite and hydro-
gen peroxide (H2O2) (32). Furthermore, the increased oxidative
stress and chronic hypoxia (intravascular sickling and low hema-
tocrit) in the BERK model are associated with the induction of
COX-2. Unlike COX-1, which is constitutively expressed, COX-2 is
an inducible enzyme with low or undetectable levels under normal
conditions. COX-2 is induced under conditions of chronic hypoxia
and inflammation, resulting in a greater production of PGE2, a
vasodilator prostanoid (33, 34).
In the BERK mouse, COX-2 expression is confirmed by both
Western blotting and immunohistochemistry, with the latter con-
firming its localization in microvascular endothelium. We propose
that the induction of COX-2 and associated vasodilation is a com-
pensatory response to accomplish adequate oxygen delivery under
conditions of anemia. Also, another vasodilator enzyme, heme
oxygenase-1, with antioxidant properties is induced in both trans-
genic sickle mice and human sickle cell disease patients (35, 36).
During ischemic/reperfusion events, endothelial generation of
H2O2, a vasodilator and hyperpolarizing factor (37, 38), may also
contribute to vasodilation. We also find increased expression of
eNOS in the BERK model and confirm its upregulation in vascular
endothelium. However, the bioavailability of NO may be impaired
by increased oxidant generation, as well as by cell-free hemoglobin
(15). On the other hand, the cremaster tissue, while positive for
eNOS, was negative for iNOS, which is in agreement with previous
reports in transgenic sickle mice (18, 19).
Microvascular and hemodynamic abnormalities in BERK mice. The
results show that microcirculatory blood flow in BERK mice is
markedly affected by intravascular sickling, hemolysis, and atten-
dant anemia. This is evidenced by pronounced 40% dilation of the
resistance vessels (i.e., arterioles), average reduction of 40 mmHg
in MAP, and an almost 50% increase in Q as compared with those
in C57BL mice. No significant differences in the vessel diameters
and flow were noted among C57BL, BERK-trait, and β-thal mice.
Notably, β-thal mice have anemia but no intravascular sickling or
significant hemolysis. In contrast, in BERK mice, anemia com-
bined with sickling and hemolysis results in vasodilation to facili-
tate optimal oxygen delivery. In addition, in BERK mice, sickling,
increased blood cell–endothelium interactions, and vasodilation
contribute to a marked decrease in wall shear rates in both arte-
rioles (40–58% decrease) and venules (V4 and V3, approximately a
35% decrease) compared with those in C57BL controls.
Systemic hypotension coupled with arteriolar dilation indicates
decreased vascular resistance in BERK mice. Although no compa-
rable microcirculatory measurements exist for sickle cell disease
patients, several studies have demonstrated that low blood pres-
sure in these patients is associated with almost 50% reduction in
the peripheral resistance (12, 39). The reduced peripheral resis-
tance in sickle cell disease patients is also associated with a sig-
nificant increase in Q in large arteries (16, 20). Such hemodynamic
alterations in sickle cell disease patients may be due to altered
microvascular control (vasodilation) as observed in BERK mice
and may reflect contribution of non–NO-mediated vasodilators,
Vascular reactivity. Further insights into microvascular control
mechanisms in BERK mice come from the characteristics of vessel
responses to the NO-mediated vasodilators ACh and SNP. In con-
trast to 65–80% arteriolar dilation induced by ACh (an endothelium-
dependent vasodilator) and SNP (an NO donor) in control mice
(C57BL and BERK-trait mice), BERK mice showed minimal effects
of these NO-mediated vasodilators on arteriolar diameters (7% and
13%, respectively). Furthermore, L-NAME, an inhibitor of NOS
activity, caused a significant increase in blood pressure in both
male and female C57BL mice. Among BERK mice, compared with
the attenuated response in males, the females showed a statistically
significant increase in blood pressure that, however, did not reach
the level of C57BL controls. Sex differences in L-NAME responses
have been reported in humans with sickle cell disease (20) and may
indicate sexual dimorphism in NO bioactivity.
In the BERK model, almost complete resistance to exogenous
NO (i.e., SNP) demonstrates inactivation and/or destruction of
NO. The BERK model differs markedly from ischemic coronary
artery disease, which is characterized by blunted responses to L-
NMMA but normal responses to SNP (11). In sickle cell disease,
increased oxidative stress (10) and systemic effects of chronic
hemolysis (15) both may contribute to NO inactivation. NO reacts
at least 1,000 times more rapidly with cell-free ferrous hemoglo-
bin than with red cells. Also, an increase in endothelium-bound
xanthine oxidase, which has been reported in BERK mice (10),
could catalyze the increased generation of O2•– and H2O2, thereby
interfering with vessel responses to NO-mediated vasodilators.
While increased O2•– could potentially react with NO, as suggested
by Aslan et al. (10), H2O2 is required in hemin-mediated nitrated-
tyrosine formation (32). Thus, in BERK mice, NO bioavailability is
likely impaired by increased oxidative stress and hemolytic rate.
Next, increased oxidative stress and formation of nitrated pro-
teins (via ONOO– and/or hemin-mediated pathways) may itself
cause chronic vascular injury and impaired vascular reactivity.
Previous studies have shown that nitrotyrosine (3-nitro-L-tyrosine)
infusion in rats leads to attenuated hemodynamic responses to
catecholamines (epinephrine and NE) and angiotensin II (40, 41),
which is comparable to the attenuated blood pressure response to
NE observed in the present studies. Previously reported attenu-
ated hemodynamic response to angiotensin II in sickle cell disease
patients (42) may also involve a similar mechanism.
The effect of HbF. BERK + γ mice expressing 80% HbS and 20% γ-glo-
bin levels show improvement in both β-globin:α-globin chain ratio
and MCH, which indicates a significant alleviation of thalassemic
features of BERK mice. We also find a distinct increase in hemato-
crit levels accompanied by reduced hemolytic rate as compared with
The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
BERK mice. However, higher-than-normal levels of plasma hemoglo-
bin in BERK + γ mice indicate persistent hemolysis, a condition com-
parable to that observed in sickle cell disease patients on hydroxyurea
therapy (36). Nevertheless, the inhibitory effect of HbF on red cell
sickling in BERK + γ mice and the likely decrease in sickling-induced
transient vaso-occlusive events result in normalization of several
microvascular and hemodynamic parameters. First, in these mice,
microvascular diameters, wall shear rates, and Q are not significantly
different from those in control C57BL mice. Second, in BERK + γ
mice, MAP shows a significant increase as compared with that in
BERK mice, further indicating normalization of vascular tone. Third,
the normalization of vascular tone (i.e., arteriolar diameter) is asso-
ciated with a marked decrease in COX-2 and eNOS expression in
microvascular endothelium, almost to the control levels. Finally, in
contrast to BERK mice, BERK + γ mice exhibit an improved response
to NO-mediated vasoactive stimuli (ACh, SNP, and L-NAME), as well
as to a vasoconstrictor (NE), indicating a trend toward normalization
of vascular reactivity. It is, however, worthwhile to note that, while
some pathology is completely corrected by the introduction of HbF
(e.g., vessel diameters, COX-2 and eNOS expression), the response to
vasoactive stimuli such as Ach and SNP is only partially corrected.
This suggests that, even when the vessel diameter is normal, the sig-
naling pathways for the maintenance of vascular tone are still dis-
rupted in BERK + γ mice, perhaps because of higher-than-normal
oxidant generation and hemolytic rates in these mice.
In conclusion, the present studies demonstrate that HbS polymer-
ization, intravascular sickling, and attendant vaso-occlusive events
in the BERK mouse are associated with significant microvascular
and hemodynamic adaptations. The arteriolar dilation coupled
with increased blood flow is a compensatory response to provide
adequate oxygen delivery in the face of chronic anemia. The vasodi-
lation is associated with induction of COX-2 and increased expres-
sion of eNOS. However, resistance to NO-mediated vasodilators
and L-NAME suggests inactivation of NO secondary to increased
oxidative stress and hemolytic rate. The attenuated response to NE
indicates that oxidative damage and tyrosine nitration may affect
vascular smooth muscle function. In contrast, increasing HbF levels
to 20% results not only in normalization of microvascular parame-
ters in BERK + γ mice, but also in decreased expression of eNOS and
COX-2 almost to the control levels. Importantly, in BERK + γ mice,
the decrease in oxidative stress and hemolytic rate is accompanied
by improved responses to NO-mediated vasoactive stimuli and NE.
These results demonstrate a strong relationship between sickling
and microvascular function in sickle cell disease.
BERK mice expressing cointegrated 6.4-kb miniLCR, a 1.5-kb PstI fragment
of human α1 gene, and a 39-kb KpnI fragment containing human GγAγδ and
βS-globin genes were generated as previously described (43). These mice are
homozygous for the mouse α-knockout (44) and homozygous for the mouse
β-knockout (45) and express exclusively human hemoglobin via the hemi-
zygous copy of the BERK transgene (43). We generated BERK + γ mice by
breeding in the γ-transgene generated by Gilman (46), as described by Fabry
et al. (24). These mice are also homozygous for both the α- and the β-knock-
outs and express exclusively human hemoglobin via a hemizygous copy of the
BERK transgene and a hemizygous copy of the γ-transgene. The BERK-trait
mouse was used as one of the controls. This mouse is homozygous for the
α-knockout, hemizygous for the β-knockout, and hemizygous for the BERK
transgene. One copy of the BERK transgene results in expression of 15%
βS-globin and 85% mouse β-globin. The globin composition in adult trans-
genic-knockout mice was determined by HPLC as previously described (24).
The controls included C57BL/6J (C57BL) mice and C57BL β-thal (β-thal) mice.
The control C57BL and β-thal mice were maintained on a standard diet
and water ad libitum. Sickle mice were maintained on “sickle chow” devel-
oped by Paszty et al. (43) (diet no. 5740C; Purina Mills Inc.), without added
arginine, and had access to Nestlets (Ancare) nesting material. All experimen-
tal protocols were approved by the Animal Institute Committee of the Albert
Einstein College of Medicine.
Blood samples obtained from the tail were analyzed for the MCH using
the Sysmex SE-9000 system (TOA Medical Electronics Co.). Percentage
reticulocytes was determined in smears stained by New methylene blue
(Sigma-Aldrich). Hematocrit in plasma was directly measured using a
microcapillary centrifuge (model MB; International Equipment Com-
pany). For plasma-free hemoglobin, blood samples were drawn from the
mouse left ventricle, and determinations were made using a benzidine-
based assay kit (Catachem Inc.).
C57BL control (n = 51), BERK (n = 51), BERK + γ (n = 31), β-thal (n = 17),
and BERK-trait (n = 19) mice weighing approximately 25–30 g (4–6 months
old) were used. Mice were anesthetized intraperitoneally with 10% urethane
and 2% α-chloralose in saline (5 ml/kg). The animals were tracheostomized.
The right jugular vein was cannulated for infusion of vasoactive substanc-
es. To monitor arterial pressure, the left carotid artery was cannulated
using PE10 polyethylene tubing. In vivo microcirculatory observations
were made in the open cremaster muscle preparation, prepared according
to the method of Baez (47). The suffusion and maintenance of the mouse
cremaster preparation were done as previously described (17). Briefly, the
open cremaster preparation was suffused with bicarbonate Ringer’s solu-
tion (in mmol: NaCl 135.0, KCl 5.0, NaHCO3 27.0, MgCl2 0.64, and glucose
11.6; pH adjusted to 7.35–7.4 by bubbling with 94.6% N2 and 5.6% CO2).
The osmolarity of Ringer’s solution was adjusted to 330 mOsm, which
is similar to that of the mouse plasma, and the temperature of suffusion
solution was maintained at 34.5–35°C. Microscopic observations were car-
ried out using a Nikon microscope (model E400; Morrell Instrument Co.)
equipped with a CCD television camera (model CCD-300T-RC; Dage-MTI
Inc.) and a Sony U-matic video recorder (model VO5800; Sony Corp.).
Diameter and Vrbc measurements were made in A2, A3, and A4 arterioles
and V2, V3, and V4 venules. Vessel luminal diameter (D) was measured on-
line using an image shearing device (model 908; Vista Electronic Co.). Vrbc
was measured along the vessel centerline using a “dual-slit” photodiode and
a velocity cross-correlator (model 102 BF; Vista Electronic Co.) (48, 49). The
centerline Vrbc was converted to the Vmean across the vessel diameter using a
conversion factor of 1.6 (Vrbc/Vmean = 1.6), originally described by Baker and
Wayland (50) and later validated by Seki and Lipowsky (51). Q was deter-
mined from Vmean and the vessel cross-sectional area (πD2/4) as described
elsewhere (50). Shear rates along the wall of a microvessel of a given luminal
diameter were calculated using the relationship 8 × Vmean/D (17).
In a series of experiments, arteriolar responses to topical application of
ACh (10–6 M; Sigma-Aldrich) and SNP (10–6 M), both NO-mediated vaso-
dilators, were compared in control, BERK, and BERK + γ sickle mice. After
base-line measurements, suffusion with bicarbonate Ringer’s solution was
interrupted, and ACh or SNP prepared in the same medium was topically
applied. Vessel diameter and Vrbc were measured after 3 minutes of the
topical application. As reported previously, both ACh and SNP resulted
in maximal vasodilatory effects in control mice when the concentration
1144 The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
of each was increased from 10–8 to 10–6 M, and increasing the concentra-
tion to 10–5 M caused no appreciable further increase in the diameter (18).
Hence, the concentration of 10–6 M was selected in each case. Ten minutes
were allowed to lapse before the next topical application. In general, diam-
eters returned to baseline levels within 10 minutes.
In separate experiments, blood pressure response to the infusion of
L-NAME (Sigma-Aldrich) and NE was evaluated in C57BL, BERK, and
BERK + γ mice. L-NAME (20 mg/kg) was infused intravenously over 10
minutes, and MAP was measured after 30 minutes. NE (Abbott Laborato-
ries) was infused intravenously via the jugular vein at increasing concentra-
tions of 1, 2, and 4 μg/kg, as described elsewhere (40).
The cremaster tissue was rapidly excised, rinsed in saline, and homogenized
in boiling buffer (1% SDS in 10 mM Tris-HCl, pH 7.4). The homogenate was
boiled for 5 minutes, cooled in ice, and centrifuged at 9,000 g for 5 minutes
to remove insoluble material. After 2 passages of the supernatant through a
26-gauge needle, 50 μl of the supernatant was mixed with 0.45 ml of deion-
ized H2O and 2 ml of biuret reagent, and protein concentration was deter-
mined on a spectrophometer at 320 μm. For Western blotting of eNOS and
iNOS (mol wt, 140 and 130 kDa, respectively), SDS-PAGE was performed
using 7.5% acrylamide gels for separation. Molecular weight standards and
a positive control (human endothelial lysate for eNOS, or mouse macro-
phage lysate for iNOS) and a negative control (lacking specific antibody)
were also run. Monoclonal antibodies directed against human eNOS and
mouse macrophage iNOS (Transduction Laboratories) were used as previ-
ously described (18). Western blotting of COX-2 (mol wt, 70 kDa) was done
using a 10% acrylamide gel. Molecular weight standards and a positive con-
trol (Cos-1 cell lysate) were run as described above. After the transfer, the
nitrocellulose membranes were treated with the blocking buffer and incu-
bated with rabbit anti–mouse COX-2 polyclonal antibody (dilution, 1:1,000;
Cayman Chemical Co.) for 1 hour at room temperature, followed by incu-
bation with anti–rabbit IgG-HRP secondary antibody (1:10,000) for 1 hour
at room temperature. Western blotting of nitrotyrosine-containing proteins
was done using a 4–20% linear-gradient gel (Bio-Rad Laboratories Inc.).
Tyrosine-nitrated proteins with a range of molecular weights served as posi-
tive control (Upstate Biotechnology Inc.). The membranes were incubated
overnight at 4°C with a monoclonal anti–nitrotyrosine antibody (clone 1A5;
Upstate Biotechnology Inc.) diluted 1:800 in the blocking buffer, followed by
goat anti–mouse IgG-HRP (1:2,000) for 1 hour. To ascertain equal loading
of samples, the membranes were stripped off in stripping buffer (Chemicon
International Inc.) for 15 minutes, blocked for 30 minutes in the blocking
buffer, and reincubated for 1 hour with goat anti–actin antibody (Santa Cruz
Biotechnology Inc.) followed by incubation with anti–goat IgG-HRP second-
ary antibody (1:2,000) for 1 hour. In the above Western blotting procedures,
the specific proteins were detected by ECL (DuPont). The protein bands
on the developed film were scanned and quantified by computerized laser
densitometry and ImageQuant software (Molecular Dynamics).
eNOS. The cremaster tissue was obtained from 3 groups of mice: control
C57BL, BERK, and BERK + γ. Frozen cremaster tissue sections (6 μm
thick) were fixed in acetone, dried for 20 minutes, washed with PBS,
and blocked with 2% BSA in PBS containing 5% rabbit serum. Sections
were incubated overnight at 4°C with primary monoclonal anti–human
eNOS (1:1,000; Transductions Laboratories). Omission of primary anti-
body was used as control. After 3 washes with PBS, the sections were
then incubated with biotinylated rabbit anti–mouse antibody (1:1,000;
Chemicon International Inc.). Antibody-binding sites were visualized by
incubation in avidin-biotin peroxidase complex solution (ABC complex,
VECTASTAIN; Vector Laboratories Inc.) and 3,3′-diaminobenzidine
(DAB; DAKO Corp.). Counterstaining was done with hematoxylin (Har-
ris Hematoxylin; Poly Scientific R&D Corp.). Survey and photomicros-
copy were done using a Zeiss Axiophot microscope (Carl Zeiss Inc.) at a
magnification of ×400.
COX-2. Cremaster muscle tissue was obtained from the same 3 groups
of mice as described above. The tissue was fixed in 10% neutral-buff-
ered formalin immediately after exteriorization. Paraffin sections (5
μm thick) were deparaffinized in xylene, rehydrated through graded
ethanols to water, and equilibrated in 1× PBS. Sections were pretreated
for 20 minutes by steaming in 0.01 M Na-citrate buffer (pH 6.0; Vec-
tor Laboratories Inc.). Endogenous peroxidase was quenched with 0.3%
H2O2 for 30 minutes, and 5% BSA was used to block nonspecific bind-
ing. Sections were incubated overnight at 4°C with polyclonal rabbit
anti–mouse COX-2 antiserum (1:50; Cayman Chemical Co.). Omission
of primary antibody was used as negative control. Sections were then
incubated with biotinylated goat anti–rabbit IgG (1:200; Vector Labora-
tories Inc.) for 1 hour at room temperature. Visualization was by DAB
as described above.
Statistical analysis was performed using Student’s t test or 1-way ANOVA
with Newman-Keuls multiple comparisons. Where tests for normality
failed, or Bartlett’s test for homogeneity of variance showed significant dif-
ference in the standard deviations, nonparametric tests such as the Kruskal-
Wallis test for ANOVA or the Wilcoxon 2-sample test were used. Values are
expressed as the mean ± SE. P values less than 0.05 were considered signifi-
cant. The statistical analysis was performed using STATGRAPHICS*Plus
3.0 for Windows (Manugistics Inc.).
The authors express thanks to S. Suzuka for invaluable help
with genotyping of the knockout mice, and M. Cammer (Albert
Einstein College of Medicine, Bronx, New York, USA) for assis-
tance with photomicrography. This work was supported by
NIH grants RO1 HL070047, HL55552, U54 HL38655, and P60
HL55435, and by a Grant-in-Aid from the American Heart Asso-
ciation (Heritage Affiliate).
Received for publication March 17, 2004, and accepted in revised
form August 3, 2004.
Address correspondence to: Dhananjay K. Kaul, Department of Medi-
cine, Room U-917, Albert Einstein College of Medicine, 1300 Morris
Park Avenue, Bronx, New York 10461, USA. Phone: (718) 430-3702;
Fax: (718) 430-8969; E-mail: email@example.com.
1. Kaul, D.K., Fabry, M.E., and Nagel, R.L. 1996. The
pathophysiology of vascular obstruction in the
sickle syndromes. Blood Rev. 10:29–44.
2. Osarogiagbon, U.R., et al. 2000. Reperfusion injury
pathophysiology in sickle transgenic mice. Blood.
3. Kaul, D.K., and Hebbel, R.P. 2000. Hypoxia/reoxy-
genation causes inflammatory response in trans-
genic sickle mice but not in normal mice. J. Clin.
4. Francis, R.B.J., and Haywood, L.J. 1992. Elevat-
ed immunoreactive tumor necrosis factor and
interleukin-1 in sickle cell disease. J. Natl. Med.
5. Oh, S.O., Ibe, B.O., Johnson, C., Kurantsin-Mills,
J., and Raj, J.U. 1997. Platelet-activating factor in
plasma of patients with sickle cell disease in steady
state. J. Lab. Clin. Med. 130:191–196.
6. Klug, P.P., Kaye, N., and Jensen, W.N. 1982.
Endothelial cell and vascular damage in the sickle
cell disorders. Blood Cells. 8:175–184.
7. Sowemimo-Coker, S.O., Meiselman, H.J., and Fran-
cis, R.B., Jr. 1989. Increased circulating endothelial
cells in sickle cell crisis. Am. J. Hematol. 31:263–265.
research article Download full-text
The Journal of Clinical Investigation http://www.jci.org Volume 114 Number 8 October 2004
8. Solovey, A., et al. 1997. Circulating activated
endothelial cells in sickle cell anemia. N. Engl. J.
9. Platt, O.S. 2000. Sickle cell anemia as an inflamma-
tory disease. J. Clin. Invest. 106:337–338.
10. Aslan, M., et al. 2001. Oxygen radical inhibition
of nitric oxide-dependent vascular function in
sickle cell disease. Proc. Natl. Acad. Sci. U. S. A.
11. Conger, J.D., and Weil, J.V. 1995. Abnormal vascu-
lar function following ischemia-reperfusion injury.
J. Investig. Med. 43:431–442.
12. Lonsdorfer, J., et al. 1983. Cardiorespiratory adjust-
ments in chronic sickle cell anemia. Bull. Eur. Phys-
iopathol. Respir. 19:339–344.
13. Rodgers, G.P., et al. 1990. Microcirculatory adap-
tations in sickle cell anemia: reactive hyperemia
response. Am. J. Physiol. 258:H113–H120.
14. Lipowsky, H.H., Sheikh, N.U., and Katz, D.M. 1987.
Intravital microscopy of capillary hemodynamics
in sickle cell disease. J. Clin. Invest. 80:117–127.
15. Reiter, C.D., et al. 2002. Cell-free hemoglobin lim-
its nitric oxide bioavailability in sickle-cell disease.
Nat. Med. 8:1383–1389.
16. Belhassen, L., et al. 2001. Endothelial dysfunction
in patients with sickle cell disease is related to selec-
tive impairment of shear stress-mediated vasodila-
tion. Blood. 97:1584–1589.
17. Kaul, D.K., Fabry, M.E., Costantini, F., Rubin, E.M.,
and Nagel, R.L. 1995. In vivo demonstration of red
cell-endothelial interaction, sickling and altered
microvascular response to oxygen in the sickle
transgenic mouse. J. Clin. Invest. 96:2845–2853.
18. Kaul, D.K., Liu, X.D., Fabry, M.E., and Nagel, R.L.
2000. Impaired nitric oxide-mediated vasodilation
in transgenic sickle mouse. Am. J. Physiol. Heart Circ.
19. Nath, K.A., et al. 2000. Mechanisms of vascular
instability in a transgenic mouse model of sickle
cell disease. Am. J. Physiol. Regul. Integr. Comp. Physiol.
20. Gladwin, M.T., et al. 2003. Divergent nitric oxide
bioavailability in men and women with sickle cell
disease. Circulation. 107:271–278.
21. Eberhardt, R.T., et al. 2003. Sickle cell anemia is
associated with reduced nitric oxide bioactivity in
peripheral conduit and resistance vessels. Am. J.
22. Steinberg, M.H., et al. 1997. Fetal hemoglobin in
sickle cell anemia: determinants of response to
hydroxyurea. Multicenter Study of Hydroxyurea.
23. Brugnara, C., et al. 1996. Therapy with oral clotrim-
azole induces inhibition of the Gardos channel and
reduction of erythrocyte dehydration in patients
with sickle cell disease. J. Clin. Invest. 97:1227–1234.
24. Fabry, M.E., et al. 2001. Second generation knockout
sickle mice: the effect of HbF. Blood. 97:410–418.
25. Horiuchi, K., Osterhout, M.L., Kamma, H., Bekoe,
N.A., and Hirokawa, K.J. 1995. Estimation of fetal
hemoglobin levels in individual red cells via fluo-
rescence image cytometry. Cytometry. 20:261–267.
26. Lim, S.K., et al. 1998. Increased susceptibility in
Hp knockout mice during acute hemolysis. Blood.
27. Turhan, A., Weiss, L.A., Mohandas, N., Coller,
B.S., and Frenette, P.S. 2002. Primary role for
adherent leukocytes in sickle cell vascular occlu-
sion: a new paradigm. Proc. Natl. Acad. Sci. U. S. A.
28. Schmedtje, J.F., Jr., Ji, Y.S., Liu, W.L., DuBois,
R.N., and Runge, M.S. 1997. Hypoxia induces
cyclooxygenase-2 via the NF-kappaB p65 transcrip-
tion factor in human vascular endothelial cells.
J. Biol. Chem. 272:601–608.
29. Okamoto, H., Ito, O., Roman, R.J., and Hudetz,
A.G. 1998. Role of inducible nitric oxide synthase
and cyclooxygenase-2 in endotoxin-induced cere-
bral hyperemia. Stroke. 29:1209–1218.
30. Fiebig, E., Ley, K., and Arfors, K.E. 1991. Rapid leu-
kocyte accumulation by “spontaneous” rolling and
adhesion in the exteriorized rabbit mesentery. Int. J.
Microcirc. Clin. Exp. 10:127–144.
31. Beckman, J.S., and Koppenol, W.H. 1996. Nitric
oxide, superoxide, and peroxynitrite: the good, the
bad, and ugly. Am. J. Physiol. 271:C1424–C1437.
32. Thomas, D.D., Espey, M.G., Vitek, M.P., Miranda,
K.M., and Wink, D.A. 2002. Protein nitration is
mediated by heme and free metals through Fenton-
type chemistry: an alternative to the NO/O2- reac-
tion. Proc. Natl. Acad. Sci. U. S. A. 99:12691–12696.
33. Wu, G., et al. 2003. Hypoxia induces myocyte-
dependent COX-2 regulation in endothelial cells:
role of VEGF. Am. J. Physiol. Heart Circ. Physiol.
34. Oltman, C.L., et al. 2003. Reactive oxygen species
mediate arachidonic acid-induced dilation in por-
cine coronary microvessels. Am. J. Physiol. Heart Circ.
35. Nath, K.A., et al. 2001. Oxidative stress and induc-
tion of heme oxygenase-1 in the kidney in sickle
cell disease. Am. J. Pathol. 158:893–903.
36. Jison, M.L., et al. 2004. Blood mononuclear cell
gene expression profiles characterize the oxidant,
hemolytic, and inflammatory stress of sickle cell
disease. Blood. 104:270–280.
37. Thengchaisri, N., and Kuo, L. 2003. Hydrogen
peroxide induces endothelium-dependent and -
independent coronary arteriolar dilation: role of
cyclooxygenase and potassium channels. Am. J.
Physiol. Heart Circ. Physiol. 285:H2255–H2263.
38. Matoba, T., et al. 2000. Hydrogen peroxide is an
endothelium-derived hyperpolarizing factor in
mice. J. Clin. Invest. 106:1521–1530.
39. Denenberg, B.S., Criner, G., Jones, R., and Spann,
J.F. 1983. Cardiac function in sickle cell anemia.
Am. J. Cardiol. 51:1674–1678.
40. Kooy, N.W., and Lewis, S.J. 1996. Nitrotyro-
sine attenuates the hemodynamic effects of
adrenoceptor agonists in vivo: relevance to the
pathophysiology of peroxynitrite. Eur. J. Pharmacol.
41. Kooy, N.W., and Lewis, S.J. 1996. The peroxynitrite
product 3-nitro-L-tyrosine attenuates the
hemodynamic responses to angiotensin II in vivo.
Eur. J. Pharmacol. 315:165–170.
42. Hatch, F.E., Crowe, L.R., Miles, D.E., Young, J.P.,
and Portner, M.E. 1989. Altered vascular reactivity
in sickle hemoglobinopathy. A possible protective
factor from hypertension. Am. J. Hypertens. 2:2–8.
43. Paszty, C., et al. 1997. Transgenic knockout mice
with exclusively human sickle hemoglobin and
sickle cell disease. Science. 278:876–878.
44. Paszty, C., et al. 1995. Lethal alpha-thalassaemia
created by gene targeting in mice and its genetic
rescue. Nat. Genet. 11:33–39.
45. Shehee, W.R., Oliver, P., and Smithies, O. 1993.
Lethal thalassemia after insertional disruption of
the mouse major adult beta-globin gene. Proc. Natl.
Acad. Sci. U. S. A. 90:3177–3181.
46. Gilman, J. 1995. Developmental changes of human
Gγ and Aγ and mouse embryonic εγ1, εγ2 and βη in
transgenic mice with HS4-Gγ-Aγ [abstract]. Blood.
47. Baez, S. 1973. An open cremaster muscle prepara-
tion for the study of blood vessels by in vivo micros-
copy. Microvasc. Res. 5:384–394.
48. Silva, J., and Intaglietta, M. 1974. The correlation of
photometric signals derived from in vivo red blood
cell flow in microvessels. Microvasc. Res. 7:156–169.
49. Wayland, H., and Johnson, P.C. 1967. Erythrocyte
velocity measurement in microvessels by a two-slit
photometric method. J. Appl. Physiol. 22:333–337.
50. Baker, M., and Wayland, H. 1974. On-line volume
flow rate and velocity profile measurement for
blood in microvessels. Microvasc. Res. 7:131–143.
51. Seki, J., and Lipowsky, H.H. 1989. In vivo and in vitro
measurements of red cell velocity under epifluores-
cence microscopy. Microvasc. Res. 38:110–124.